8624075105
The motivation of this change is simply to reduce test duplication. As can be seen in the (massive) test delta, we have many tests whose output differ only due to the use of addi on rv32 vs addiw on rv64 when the high bits are don't care. As an aside, we don't need to worry about the non-zero immediate restriction on the compressed variants because we're not directly forming the compressed variants. If we happen to get a zero immediate for the ADDI, then either a later optimization will strip the useless instruction or the encoder is responsible for not compressing the instruction.
22199 lines
768 KiB
LLVM
22199 lines
768 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc -mtriple=riscv32 -verify-machineinstrs < %s \
|
|
; RUN: | FileCheck -check-prefix=RV32I %s
|
|
; RUN: llc -mtriple=riscv32 -mattr=+a -verify-machineinstrs < %s \
|
|
; RUN: | FileCheck -check-prefixes=RV32IA,RV32IA-WMO %s
|
|
; RUN: llc -mtriple=riscv32 -mattr=+a,+experimental-ztso -verify-machineinstrs < %s \
|
|
; RUN: | FileCheck -check-prefixes=RV32IA,RV32IA-TSO %s
|
|
; RUN: llc -mtriple=riscv64 -verify-machineinstrs < %s \
|
|
; RUN: | FileCheck -check-prefix=RV64I %s
|
|
; RUN: llc -mtriple=riscv64 -mattr=+a -verify-machineinstrs < %s \
|
|
; RUN: | FileCheck -check-prefixes=RV64IA,RV64IA-WMO %s
|
|
; RUN: llc -mtriple=riscv64 -mattr=+a,+experimental-ztso -verify-machineinstrs < %s \
|
|
; RUN: | FileCheck -check-prefixes=RV64IA,RV64IA-TSO %s
|
|
|
|
define i8 @atomicrmw_xchg_i8_monotonic(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i8_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_exchange_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xchg_i8_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: li a3, 255
|
|
; RV32IA-NEXT: sll a3, a3, a0
|
|
; RV32IA-NEXT: andi a1, a1, 255
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB0_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a4, (a2)
|
|
; RV32IA-NEXT: mv a5, a1
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: and a5, a5, a3
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB0_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a4, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i8_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_exchange_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_xchg_i8_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: li a3, 255
|
|
; RV64IA-NEXT: sllw a3, a3, a0
|
|
; RV64IA-NEXT: andi a1, a1, 255
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB0_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a4, (a2)
|
|
; RV64IA-NEXT: mv a5, a1
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: and a5, a5, a3
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB0_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a4, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i8 %b monotonic
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xchg_i8_acquire(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i8_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_exchange_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_i8_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB1_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV32IA-WMO-NEXT: mv a5, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB1_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_i8_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB1_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: mv a5, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB1_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i8_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_exchange_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_i8_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB1_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV64IA-WMO-NEXT: mv a5, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB1_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_i8_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB1_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: mv a5, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB1_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i8 %b acquire
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xchg_i8_release(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i8_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_exchange_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_i8_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB2_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-WMO-NEXT: mv a5, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB2_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_i8_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB2_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: mv a5, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB2_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i8_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_exchange_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_i8_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB2_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-WMO-NEXT: mv a5, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB2_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_i8_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB2_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: mv a5, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB2_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i8 %b release
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xchg_i8_acq_rel(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i8_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_exchange_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_i8_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB3_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV32IA-WMO-NEXT: mv a5, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB3_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_i8_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB3_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: mv a5, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB3_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i8_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_exchange_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_i8_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB3_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV64IA-WMO-NEXT: mv a5, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB3_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_i8_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB3_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: mv a5, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB3_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i8 %b acq_rel
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xchg_i8_seq_cst(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i8_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_exchange_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xchg_i8_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: li a3, 255
|
|
; RV32IA-NEXT: sll a3, a3, a0
|
|
; RV32IA-NEXT: andi a1, a1, 255
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB4_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a4, (a2)
|
|
; RV32IA-NEXT: mv a5, a1
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: and a5, a5, a3
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB4_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a4, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i8_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_exchange_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_xchg_i8_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: li a3, 255
|
|
; RV64IA-NEXT: sllw a3, a3, a0
|
|
; RV64IA-NEXT: andi a1, a1, 255
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB4_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a4, (a2)
|
|
; RV64IA-NEXT: mv a5, a1
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: and a5, a5, a3
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB4_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a4, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i8 %b seq_cst
|
|
ret i8 %1
|
|
}
|
|
|
|
; Ensure the following 'atomicrmw xchg a, {0,-1}` cases are lowered to an
|
|
; amoand or amoor with appropriate mask.
|
|
|
|
define i8 @atomicrmw_xchg_0_i8_monotonic(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_0_i8_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a1, 0
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_exchange_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xchg_0_i8_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a1, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: li a2, 255
|
|
; RV32IA-NEXT: sll a2, a2, a0
|
|
; RV32IA-NEXT: not a2, a2
|
|
; RV32IA-NEXT: amoand.w a1, a2, (a1)
|
|
; RV32IA-NEXT: srl a0, a1, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_0_i8_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a1, 0
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_exchange_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_xchg_0_i8_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a1, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: li a2, 255
|
|
; RV64IA-NEXT: sllw a2, a2, a0
|
|
; RV64IA-NEXT: not a2, a2
|
|
; RV64IA-NEXT: amoand.w a1, a2, (a1)
|
|
; RV64IA-NEXT: srlw a0, a1, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i8 0 monotonic
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xchg_0_i8_acquire(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_0_i8_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: li a1, 0
|
|
; RV32I-NEXT: call __atomic_exchange_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_0_i8_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a2, 255
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: not a2, a2
|
|
; RV32IA-WMO-NEXT: amoand.w.aq a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_0_i8_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a2, 255
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: not a2, a2
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_0_i8_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: li a1, 0
|
|
; RV64I-NEXT: call __atomic_exchange_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_0_i8_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a2, 255
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: not a2, a2
|
|
; RV64IA-WMO-NEXT: amoand.w.aq a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_0_i8_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a2, 255
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: not a2, a2
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i8 0 acquire
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xchg_0_i8_release(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_0_i8_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: li a1, 0
|
|
; RV32I-NEXT: call __atomic_exchange_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_0_i8_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a2, 255
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: not a2, a2
|
|
; RV32IA-WMO-NEXT: amoand.w.rl a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_0_i8_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a2, 255
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: not a2, a2
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_0_i8_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: li a1, 0
|
|
; RV64I-NEXT: call __atomic_exchange_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_0_i8_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a2, 255
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: not a2, a2
|
|
; RV64IA-WMO-NEXT: amoand.w.rl a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_0_i8_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a2, 255
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: not a2, a2
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i8 0 release
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xchg_0_i8_acq_rel(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_0_i8_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: li a1, 0
|
|
; RV32I-NEXT: call __atomic_exchange_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_0_i8_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a2, 255
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: not a2, a2
|
|
; RV32IA-WMO-NEXT: amoand.w.aqrl a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_0_i8_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a2, 255
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: not a2, a2
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_0_i8_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: li a1, 0
|
|
; RV64I-NEXT: call __atomic_exchange_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_0_i8_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a2, 255
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: not a2, a2
|
|
; RV64IA-WMO-NEXT: amoand.w.aqrl a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_0_i8_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a2, 255
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: not a2, a2
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i8 0 acq_rel
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xchg_0_i8_seq_cst(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_0_i8_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: li a1, 0
|
|
; RV32I-NEXT: call __atomic_exchange_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_0_i8_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a2, 255
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: not a2, a2
|
|
; RV32IA-WMO-NEXT: amoand.w.aqrl a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_0_i8_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a2, 255
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: not a2, a2
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_0_i8_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: li a1, 0
|
|
; RV64I-NEXT: call __atomic_exchange_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_0_i8_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a2, 255
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: not a2, a2
|
|
; RV64IA-WMO-NEXT: amoand.w.aqrl a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_0_i8_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a2, 255
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: not a2, a2
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i8 0 seq_cst
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xchg_minus_1_i8_monotonic(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_minus_1_i8_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a1, 255
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_exchange_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xchg_minus_1_i8_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a1, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: li a2, 255
|
|
; RV32IA-NEXT: sll a2, a2, a0
|
|
; RV32IA-NEXT: amoor.w a1, a2, (a1)
|
|
; RV32IA-NEXT: srl a0, a1, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_minus_1_i8_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a1, 255
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_exchange_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_xchg_minus_1_i8_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a1, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: li a2, 255
|
|
; RV64IA-NEXT: sllw a2, a2, a0
|
|
; RV64IA-NEXT: amoor.w a1, a2, (a1)
|
|
; RV64IA-NEXT: srlw a0, a1, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i8 -1 monotonic
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xchg_minus_1_i8_acquire(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_minus_1_i8_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a1, 255
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_exchange_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_minus_1_i8_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a2, 255
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.aq a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_minus_1_i8_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a2, 255
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_minus_1_i8_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a1, 255
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_exchange_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_minus_1_i8_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a2, 255
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.aq a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_minus_1_i8_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a2, 255
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i8 -1 acquire
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xchg_minus_1_i8_release(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_minus_1_i8_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a1, 255
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_exchange_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_minus_1_i8_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a2, 255
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.rl a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_minus_1_i8_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a2, 255
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_minus_1_i8_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a1, 255
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_exchange_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_minus_1_i8_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a2, 255
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.rl a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_minus_1_i8_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a2, 255
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i8 -1 release
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xchg_minus_1_i8_acq_rel(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_minus_1_i8_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a1, 255
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_exchange_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_minus_1_i8_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a2, 255
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.aqrl a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_minus_1_i8_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a2, 255
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_minus_1_i8_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a1, 255
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_exchange_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_minus_1_i8_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a2, 255
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.aqrl a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_minus_1_i8_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a2, 255
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i8 -1 acq_rel
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xchg_minus_1_i8_seq_cst(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_minus_1_i8_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a1, 255
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_exchange_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_minus_1_i8_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a2, 255
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.aqrl a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_minus_1_i8_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a2, 255
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_minus_1_i8_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a1, 255
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_exchange_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_minus_1_i8_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a2, 255
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.aqrl a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_minus_1_i8_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a2, 255
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i8 -1 seq_cst
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_add_i8_monotonic(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i8_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_add_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_add_i8_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: li a3, 255
|
|
; RV32IA-NEXT: sll a3, a3, a0
|
|
; RV32IA-NEXT: andi a1, a1, 255
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB15_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a4, (a2)
|
|
; RV32IA-NEXT: add a5, a4, a1
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: and a5, a5, a3
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB15_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a4, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i8_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_add_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_add_i8_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: li a3, 255
|
|
; RV64IA-NEXT: sllw a3, a3, a0
|
|
; RV64IA-NEXT: andi a1, a1, 255
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB15_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a4, (a2)
|
|
; RV64IA-NEXT: add a5, a4, a1
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: and a5, a5, a3
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB15_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a4, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i8 %b monotonic
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_add_i8_acquire(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i8_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_add_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_add_i8_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB16_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV32IA-WMO-NEXT: add a5, a4, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB16_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_add_i8_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB16_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: add a5, a4, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB16_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i8_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_add_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_add_i8_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB16_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV64IA-WMO-NEXT: add a5, a4, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB16_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_add_i8_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB16_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: add a5, a4, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB16_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i8 %b acquire
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_add_i8_release(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i8_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_add_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_add_i8_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB17_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-WMO-NEXT: add a5, a4, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB17_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_add_i8_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB17_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: add a5, a4, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB17_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i8_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_add_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_add_i8_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB17_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-WMO-NEXT: add a5, a4, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB17_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_add_i8_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB17_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: add a5, a4, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB17_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i8 %b release
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_add_i8_acq_rel(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i8_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_add_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_add_i8_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB18_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV32IA-WMO-NEXT: add a5, a4, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB18_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_add_i8_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB18_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: add a5, a4, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB18_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i8_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_add_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_add_i8_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB18_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV64IA-WMO-NEXT: add a5, a4, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB18_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_add_i8_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB18_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: add a5, a4, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB18_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i8 %b acq_rel
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_add_i8_seq_cst(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i8_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_add_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_add_i8_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: li a3, 255
|
|
; RV32IA-NEXT: sll a3, a3, a0
|
|
; RV32IA-NEXT: andi a1, a1, 255
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB19_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a4, (a2)
|
|
; RV32IA-NEXT: add a5, a4, a1
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: and a5, a5, a3
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB19_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a4, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i8_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_add_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_add_i8_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: li a3, 255
|
|
; RV64IA-NEXT: sllw a3, a3, a0
|
|
; RV64IA-NEXT: andi a1, a1, 255
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB19_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a4, (a2)
|
|
; RV64IA-NEXT: add a5, a4, a1
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: and a5, a5, a3
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB19_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a4, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i8 %b seq_cst
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_sub_i8_monotonic(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i8_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_sub_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_sub_i8_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: li a3, 255
|
|
; RV32IA-NEXT: sll a3, a3, a0
|
|
; RV32IA-NEXT: andi a1, a1, 255
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB20_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a4, (a2)
|
|
; RV32IA-NEXT: sub a5, a4, a1
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: and a5, a5, a3
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB20_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a4, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i8_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_sub_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_sub_i8_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: li a3, 255
|
|
; RV64IA-NEXT: sllw a3, a3, a0
|
|
; RV64IA-NEXT: andi a1, a1, 255
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB20_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a4, (a2)
|
|
; RV64IA-NEXT: sub a5, a4, a1
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: and a5, a5, a3
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB20_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a4, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i8 %b monotonic
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_sub_i8_acquire(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i8_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_sub_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_sub_i8_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB21_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV32IA-WMO-NEXT: sub a5, a4, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB21_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_sub_i8_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB21_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: sub a5, a4, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB21_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i8_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_sub_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_sub_i8_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB21_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV64IA-WMO-NEXT: sub a5, a4, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB21_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_sub_i8_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB21_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: sub a5, a4, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB21_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i8 %b acquire
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_sub_i8_release(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i8_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_sub_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_sub_i8_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB22_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-WMO-NEXT: sub a5, a4, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB22_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_sub_i8_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB22_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: sub a5, a4, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB22_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i8_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_sub_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_sub_i8_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB22_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-WMO-NEXT: sub a5, a4, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB22_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_sub_i8_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB22_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: sub a5, a4, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB22_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i8 %b release
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_sub_i8_acq_rel(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i8_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_sub_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_sub_i8_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB23_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV32IA-WMO-NEXT: sub a5, a4, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB23_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_sub_i8_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB23_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: sub a5, a4, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB23_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i8_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_sub_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_sub_i8_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB23_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV64IA-WMO-NEXT: sub a5, a4, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB23_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_sub_i8_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB23_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: sub a5, a4, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB23_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i8 %b acq_rel
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_sub_i8_seq_cst(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i8_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_sub_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_sub_i8_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: li a3, 255
|
|
; RV32IA-NEXT: sll a3, a3, a0
|
|
; RV32IA-NEXT: andi a1, a1, 255
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB24_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a4, (a2)
|
|
; RV32IA-NEXT: sub a5, a4, a1
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: and a5, a5, a3
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB24_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a4, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i8_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_sub_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_sub_i8_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: li a3, 255
|
|
; RV64IA-NEXT: sllw a3, a3, a0
|
|
; RV64IA-NEXT: andi a1, a1, 255
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB24_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a4, (a2)
|
|
; RV64IA-NEXT: sub a5, a4, a1
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: and a5, a5, a3
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB24_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a4, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i8 %b seq_cst
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_and_i8_monotonic(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i8_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_and_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_and_i8_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: li a3, 255
|
|
; RV32IA-NEXT: sll a3, a3, a0
|
|
; RV32IA-NEXT: not a3, a3
|
|
; RV32IA-NEXT: andi a1, a1, 255
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: or a1, a3, a1
|
|
; RV32IA-NEXT: amoand.w a1, a1, (a2)
|
|
; RV32IA-NEXT: srl a0, a1, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i8_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_and_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_and_i8_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: li a3, 255
|
|
; RV64IA-NEXT: sllw a3, a3, a0
|
|
; RV64IA-NEXT: not a3, a3
|
|
; RV64IA-NEXT: andi a1, a1, 255
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: or a1, a3, a1
|
|
; RV64IA-NEXT: amoand.w a1, a1, (a2)
|
|
; RV64IA-NEXT: srlw a0, a1, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i8 %b monotonic
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_and_i8_acquire(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i8_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_and_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_and_i8_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: not a3, a3
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: or a1, a3, a1
|
|
; RV32IA-WMO-NEXT: amoand.w.aq a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_and_i8_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: not a3, a3
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: or a1, a3, a1
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i8_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_and_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i8_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: not a3, a3
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: or a1, a3, a1
|
|
; RV64IA-WMO-NEXT: amoand.w.aq a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i8_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: not a3, a3
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: or a1, a3, a1
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i8 %b acquire
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_and_i8_release(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i8_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_and_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_and_i8_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: not a3, a3
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: or a1, a3, a1
|
|
; RV32IA-WMO-NEXT: amoand.w.rl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_and_i8_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: not a3, a3
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: or a1, a3, a1
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i8_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_and_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i8_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: not a3, a3
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: or a1, a3, a1
|
|
; RV64IA-WMO-NEXT: amoand.w.rl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i8_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: not a3, a3
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: or a1, a3, a1
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i8 %b release
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_and_i8_acq_rel(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i8_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_and_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_and_i8_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: not a3, a3
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: or a1, a3, a1
|
|
; RV32IA-WMO-NEXT: amoand.w.aqrl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_and_i8_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: not a3, a3
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: or a1, a3, a1
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i8_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_and_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i8_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: not a3, a3
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: or a1, a3, a1
|
|
; RV64IA-WMO-NEXT: amoand.w.aqrl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i8_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: not a3, a3
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: or a1, a3, a1
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i8 %b acq_rel
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_and_i8_seq_cst(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i8_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_and_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_and_i8_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: not a3, a3
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: or a1, a3, a1
|
|
; RV32IA-WMO-NEXT: amoand.w.aqrl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_and_i8_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: not a3, a3
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: or a1, a3, a1
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i8_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_and_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i8_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: not a3, a3
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: or a1, a3, a1
|
|
; RV64IA-WMO-NEXT: amoand.w.aqrl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i8_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: not a3, a3
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: or a1, a3, a1
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i8 %b seq_cst
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_nand_i8_monotonic(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i8_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_nand_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_nand_i8_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: li a3, 255
|
|
; RV32IA-NEXT: sll a3, a3, a0
|
|
; RV32IA-NEXT: andi a1, a1, 255
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB30_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a4, (a2)
|
|
; RV32IA-NEXT: and a5, a4, a1
|
|
; RV32IA-NEXT: not a5, a5
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: and a5, a5, a3
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB30_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a4, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i8_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_nand_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_nand_i8_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: li a3, 255
|
|
; RV64IA-NEXT: sllw a3, a3, a0
|
|
; RV64IA-NEXT: andi a1, a1, 255
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB30_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a4, (a2)
|
|
; RV64IA-NEXT: and a5, a4, a1
|
|
; RV64IA-NEXT: not a5, a5
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: and a5, a5, a3
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB30_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a4, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i8 %b monotonic
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_nand_i8_acquire(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i8_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_nand_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_nand_i8_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB31_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV32IA-WMO-NEXT: and a5, a4, a1
|
|
; RV32IA-WMO-NEXT: not a5, a5
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB31_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_nand_i8_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB31_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: and a5, a4, a1
|
|
; RV32IA-TSO-NEXT: not a5, a5
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB31_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i8_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_nand_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_nand_i8_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB31_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV64IA-WMO-NEXT: and a5, a4, a1
|
|
; RV64IA-WMO-NEXT: not a5, a5
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB31_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_nand_i8_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB31_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: and a5, a4, a1
|
|
; RV64IA-TSO-NEXT: not a5, a5
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB31_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i8 %b acquire
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_nand_i8_release(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i8_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_nand_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_nand_i8_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB32_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-WMO-NEXT: and a5, a4, a1
|
|
; RV32IA-WMO-NEXT: not a5, a5
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB32_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_nand_i8_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB32_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: and a5, a4, a1
|
|
; RV32IA-TSO-NEXT: not a5, a5
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB32_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i8_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_nand_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_nand_i8_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB32_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-WMO-NEXT: and a5, a4, a1
|
|
; RV64IA-WMO-NEXT: not a5, a5
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB32_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_nand_i8_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB32_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: and a5, a4, a1
|
|
; RV64IA-TSO-NEXT: not a5, a5
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB32_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i8 %b release
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_nand_i8_acq_rel(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i8_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_nand_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_nand_i8_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB33_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV32IA-WMO-NEXT: and a5, a4, a1
|
|
; RV32IA-WMO-NEXT: not a5, a5
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB33_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_nand_i8_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB33_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: and a5, a4, a1
|
|
; RV32IA-TSO-NEXT: not a5, a5
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB33_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i8_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_nand_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_nand_i8_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB33_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV64IA-WMO-NEXT: and a5, a4, a1
|
|
; RV64IA-WMO-NEXT: not a5, a5
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB33_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_nand_i8_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB33_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: and a5, a4, a1
|
|
; RV64IA-TSO-NEXT: not a5, a5
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB33_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i8 %b acq_rel
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_nand_i8_seq_cst(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i8_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_nand_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_nand_i8_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: li a3, 255
|
|
; RV32IA-NEXT: sll a3, a3, a0
|
|
; RV32IA-NEXT: andi a1, a1, 255
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB34_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a4, (a2)
|
|
; RV32IA-NEXT: and a5, a4, a1
|
|
; RV32IA-NEXT: not a5, a5
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: and a5, a5, a3
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB34_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a4, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i8_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_nand_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_nand_i8_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: li a3, 255
|
|
; RV64IA-NEXT: sllw a3, a3, a0
|
|
; RV64IA-NEXT: andi a1, a1, 255
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB34_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a4, (a2)
|
|
; RV64IA-NEXT: and a5, a4, a1
|
|
; RV64IA-NEXT: not a5, a5
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: and a5, a5, a3
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB34_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a4, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i8 %b seq_cst
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_or_i8_monotonic(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i8_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_or_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_or_i8_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: andi a1, a1, 255
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: amoor.w a1, a1, (a2)
|
|
; RV32IA-NEXT: srl a0, a1, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i8_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_or_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_or_i8_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: andi a1, a1, 255
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: amoor.w a1, a1, (a2)
|
|
; RV64IA-NEXT: srlw a0, a1, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i8 %b monotonic
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_or_i8_acquire(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i8_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_or_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_or_i8_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.aq a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_or_i8_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i8_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_or_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i8_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.aq a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i8_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i8 %b acquire
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_or_i8_release(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i8_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_or_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_or_i8_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.rl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_or_i8_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i8_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_or_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i8_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.rl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i8_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i8 %b release
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_or_i8_acq_rel(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i8_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_or_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_or_i8_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.aqrl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_or_i8_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i8_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_or_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i8_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.aqrl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i8_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i8 %b acq_rel
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_or_i8_seq_cst(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i8_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_or_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_or_i8_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.aqrl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_or_i8_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i8_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_or_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i8_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.aqrl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i8_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i8 %b seq_cst
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xor_i8_monotonic(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i8_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_xor_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xor_i8_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: andi a1, a1, 255
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV32IA-NEXT: srl a0, a1, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i8_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_xor_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_xor_i8_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: andi a1, a1, 255
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV64IA-NEXT: srlw a0, a1, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i8 %b monotonic
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xor_i8_acquire(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i8_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_xor_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xor_i8_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoxor.w.aq a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xor_i8_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i8_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_xor_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i8_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoxor.w.aq a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i8_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i8 %b acquire
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xor_i8_release(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i8_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_xor_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xor_i8_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoxor.w.rl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xor_i8_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i8_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_xor_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i8_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoxor.w.rl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i8_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i8 %b release
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xor_i8_acq_rel(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i8_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_xor_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xor_i8_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoxor.w.aqrl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xor_i8_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i8_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_xor_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i8_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoxor.w.aqrl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i8_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i8 %b acq_rel
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_xor_i8_seq_cst(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i8_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_xor_1@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xor_i8_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoxor.w.aqrl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xor_i8_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i8_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_xor_1@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i8_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoxor.w.aqrl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i8_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i8 %b seq_cst
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_max_i8_monotonic(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i8_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 24
|
|
; RV32I-NEXT: srai s2, a0, 24
|
|
; RV32I-NEXT: j .LBB45_2
|
|
; RV32I-NEXT: .LBB45_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB45_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB45_4
|
|
; RV32I-NEXT: .LBB45_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 24
|
|
; RV32I-NEXT: srai a0, a0, 24
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: blt s2, a0, .LBB45_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB45_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB45_1
|
|
; RV32I-NEXT: .LBB45_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_max_i8_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: andi a3, a0, 24
|
|
; RV32IA-NEXT: li a4, 255
|
|
; RV32IA-NEXT: sll a4, a4, a0
|
|
; RV32IA-NEXT: slli a1, a1, 24
|
|
; RV32IA-NEXT: srai a1, a1, 24
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: xori a3, a3, 24
|
|
; RV32IA-NEXT: .LBB45_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a5, (a2)
|
|
; RV32IA-NEXT: and a7, a5, a4
|
|
; RV32IA-NEXT: mv a6, a5
|
|
; RV32IA-NEXT: sll a7, a7, a3
|
|
; RV32IA-NEXT: sra a7, a7, a3
|
|
; RV32IA-NEXT: bge a7, a1, .LBB45_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB45_1 Depth=1
|
|
; RV32IA-NEXT: xor a6, a5, a1
|
|
; RV32IA-NEXT: and a6, a6, a4
|
|
; RV32IA-NEXT: xor a6, a5, a6
|
|
; RV32IA-NEXT: .LBB45_3: # in Loop: Header=BB45_1 Depth=1
|
|
; RV32IA-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-NEXT: bnez a6, .LBB45_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a5, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i8_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 56
|
|
; RV64I-NEXT: srai s2, a0, 56
|
|
; RV64I-NEXT: j .LBB45_2
|
|
; RV64I-NEXT: .LBB45_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB45_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB45_4
|
|
; RV64I-NEXT: .LBB45_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 56
|
|
; RV64I-NEXT: srai a0, a0, 56
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s2, a0, .LBB45_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB45_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB45_1
|
|
; RV64I-NEXT: .LBB45_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_max_i8_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: andi a3, a0, 24
|
|
; RV64IA-NEXT: li a4, 255
|
|
; RV64IA-NEXT: sllw a4, a4, a0
|
|
; RV64IA-NEXT: slli a1, a1, 56
|
|
; RV64IA-NEXT: srai a1, a1, 56
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: xori a3, a3, 56
|
|
; RV64IA-NEXT: .LBB45_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a5, (a2)
|
|
; RV64IA-NEXT: and a7, a5, a4
|
|
; RV64IA-NEXT: mv a6, a5
|
|
; RV64IA-NEXT: sll a7, a7, a3
|
|
; RV64IA-NEXT: sra a7, a7, a3
|
|
; RV64IA-NEXT: bge a7, a1, .LBB45_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB45_1 Depth=1
|
|
; RV64IA-NEXT: xor a6, a5, a1
|
|
; RV64IA-NEXT: and a6, a6, a4
|
|
; RV64IA-NEXT: xor a6, a5, a6
|
|
; RV64IA-NEXT: .LBB45_3: # in Loop: Header=BB45_1 Depth=1
|
|
; RV64IA-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-NEXT: bnez a6, .LBB45_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a5, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i8 %b monotonic
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_max_i8_acquire(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i8_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 24
|
|
; RV32I-NEXT: srai s2, a0, 24
|
|
; RV32I-NEXT: j .LBB46_2
|
|
; RV32I-NEXT: .LBB46_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB46_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB46_4
|
|
; RV32I-NEXT: .LBB46_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 24
|
|
; RV32I-NEXT: srai a0, a0, 24
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: blt s2, a0, .LBB46_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB46_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB46_1
|
|
; RV32I-NEXT: .LBB46_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_max_i8_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV32IA-WMO-NEXT: li a4, 255
|
|
; RV32IA-WMO-NEXT: sll a4, a4, a0
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 24
|
|
; RV32IA-WMO-NEXT: srai a1, a1, 24
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: xori a3, a3, 24
|
|
; RV32IA-WMO-NEXT: .LBB46_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a5, (a2)
|
|
; RV32IA-WMO-NEXT: and a7, a5, a4
|
|
; RV32IA-WMO-NEXT: mv a6, a5
|
|
; RV32IA-WMO-NEXT: sll a7, a7, a3
|
|
; RV32IA-WMO-NEXT: sra a7, a7, a3
|
|
; RV32IA-WMO-NEXT: bge a7, a1, .LBB46_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB46_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a6, a5, a1
|
|
; RV32IA-WMO-NEXT: and a6, a6, a4
|
|
; RV32IA-WMO-NEXT: xor a6, a5, a6
|
|
; RV32IA-WMO-NEXT: .LBB46_3: # in Loop: Header=BB46_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a6, .LBB46_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a5, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_max_i8_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV32IA-TSO-NEXT: li a4, 255
|
|
; RV32IA-TSO-NEXT: sll a4, a4, a0
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 24
|
|
; RV32IA-TSO-NEXT: srai a1, a1, 24
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: xori a3, a3, 24
|
|
; RV32IA-TSO-NEXT: .LBB46_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a5, (a2)
|
|
; RV32IA-TSO-NEXT: and a7, a5, a4
|
|
; RV32IA-TSO-NEXT: mv a6, a5
|
|
; RV32IA-TSO-NEXT: sll a7, a7, a3
|
|
; RV32IA-TSO-NEXT: sra a7, a7, a3
|
|
; RV32IA-TSO-NEXT: bge a7, a1, .LBB46_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB46_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a6, a5, a1
|
|
; RV32IA-TSO-NEXT: and a6, a6, a4
|
|
; RV32IA-TSO-NEXT: xor a6, a5, a6
|
|
; RV32IA-TSO-NEXT: .LBB46_3: # in Loop: Header=BB46_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a6, .LBB46_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a5, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i8_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 56
|
|
; RV64I-NEXT: srai s2, a0, 56
|
|
; RV64I-NEXT: j .LBB46_2
|
|
; RV64I-NEXT: .LBB46_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB46_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB46_4
|
|
; RV64I-NEXT: .LBB46_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 56
|
|
; RV64I-NEXT: srai a0, a0, 56
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s2, a0, .LBB46_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB46_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB46_1
|
|
; RV64I-NEXT: .LBB46_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_max_i8_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV64IA-WMO-NEXT: li a4, 255
|
|
; RV64IA-WMO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 56
|
|
; RV64IA-WMO-NEXT: srai a1, a1, 56
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: xori a3, a3, 56
|
|
; RV64IA-WMO-NEXT: .LBB46_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a5, (a2)
|
|
; RV64IA-WMO-NEXT: and a7, a5, a4
|
|
; RV64IA-WMO-NEXT: mv a6, a5
|
|
; RV64IA-WMO-NEXT: sll a7, a7, a3
|
|
; RV64IA-WMO-NEXT: sra a7, a7, a3
|
|
; RV64IA-WMO-NEXT: bge a7, a1, .LBB46_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB46_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a6, a5, a1
|
|
; RV64IA-WMO-NEXT: and a6, a6, a4
|
|
; RV64IA-WMO-NEXT: xor a6, a5, a6
|
|
; RV64IA-WMO-NEXT: .LBB46_3: # in Loop: Header=BB46_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a6, .LBB46_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a5, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_max_i8_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV64IA-TSO-NEXT: li a4, 255
|
|
; RV64IA-TSO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 56
|
|
; RV64IA-TSO-NEXT: srai a1, a1, 56
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: xori a3, a3, 56
|
|
; RV64IA-TSO-NEXT: .LBB46_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a5, (a2)
|
|
; RV64IA-TSO-NEXT: and a7, a5, a4
|
|
; RV64IA-TSO-NEXT: mv a6, a5
|
|
; RV64IA-TSO-NEXT: sll a7, a7, a3
|
|
; RV64IA-TSO-NEXT: sra a7, a7, a3
|
|
; RV64IA-TSO-NEXT: bge a7, a1, .LBB46_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB46_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a6, a5, a1
|
|
; RV64IA-TSO-NEXT: and a6, a6, a4
|
|
; RV64IA-TSO-NEXT: xor a6, a5, a6
|
|
; RV64IA-TSO-NEXT: .LBB46_3: # in Loop: Header=BB46_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a6, .LBB46_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a5, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i8 %b acquire
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_max_i8_release(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i8_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 24
|
|
; RV32I-NEXT: srai s2, a0, 24
|
|
; RV32I-NEXT: j .LBB47_2
|
|
; RV32I-NEXT: .LBB47_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB47_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB47_4
|
|
; RV32I-NEXT: .LBB47_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 24
|
|
; RV32I-NEXT: srai a0, a0, 24
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: blt s2, a0, .LBB47_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB47_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB47_1
|
|
; RV32I-NEXT: .LBB47_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_max_i8_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV32IA-WMO-NEXT: li a4, 255
|
|
; RV32IA-WMO-NEXT: sll a4, a4, a0
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 24
|
|
; RV32IA-WMO-NEXT: srai a1, a1, 24
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: xori a3, a3, 24
|
|
; RV32IA-WMO-NEXT: .LBB47_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a5, (a2)
|
|
; RV32IA-WMO-NEXT: and a7, a5, a4
|
|
; RV32IA-WMO-NEXT: mv a6, a5
|
|
; RV32IA-WMO-NEXT: sll a7, a7, a3
|
|
; RV32IA-WMO-NEXT: sra a7, a7, a3
|
|
; RV32IA-WMO-NEXT: bge a7, a1, .LBB47_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB47_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a6, a5, a1
|
|
; RV32IA-WMO-NEXT: and a6, a6, a4
|
|
; RV32IA-WMO-NEXT: xor a6, a5, a6
|
|
; RV32IA-WMO-NEXT: .LBB47_3: # in Loop: Header=BB47_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a6, .LBB47_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a5, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_max_i8_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV32IA-TSO-NEXT: li a4, 255
|
|
; RV32IA-TSO-NEXT: sll a4, a4, a0
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 24
|
|
; RV32IA-TSO-NEXT: srai a1, a1, 24
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: xori a3, a3, 24
|
|
; RV32IA-TSO-NEXT: .LBB47_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a5, (a2)
|
|
; RV32IA-TSO-NEXT: and a7, a5, a4
|
|
; RV32IA-TSO-NEXT: mv a6, a5
|
|
; RV32IA-TSO-NEXT: sll a7, a7, a3
|
|
; RV32IA-TSO-NEXT: sra a7, a7, a3
|
|
; RV32IA-TSO-NEXT: bge a7, a1, .LBB47_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB47_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a6, a5, a1
|
|
; RV32IA-TSO-NEXT: and a6, a6, a4
|
|
; RV32IA-TSO-NEXT: xor a6, a5, a6
|
|
; RV32IA-TSO-NEXT: .LBB47_3: # in Loop: Header=BB47_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a6, .LBB47_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a5, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i8_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 56
|
|
; RV64I-NEXT: srai s2, a0, 56
|
|
; RV64I-NEXT: j .LBB47_2
|
|
; RV64I-NEXT: .LBB47_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB47_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB47_4
|
|
; RV64I-NEXT: .LBB47_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 56
|
|
; RV64I-NEXT: srai a0, a0, 56
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s2, a0, .LBB47_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB47_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB47_1
|
|
; RV64I-NEXT: .LBB47_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_max_i8_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV64IA-WMO-NEXT: li a4, 255
|
|
; RV64IA-WMO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 56
|
|
; RV64IA-WMO-NEXT: srai a1, a1, 56
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: xori a3, a3, 56
|
|
; RV64IA-WMO-NEXT: .LBB47_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a5, (a2)
|
|
; RV64IA-WMO-NEXT: and a7, a5, a4
|
|
; RV64IA-WMO-NEXT: mv a6, a5
|
|
; RV64IA-WMO-NEXT: sll a7, a7, a3
|
|
; RV64IA-WMO-NEXT: sra a7, a7, a3
|
|
; RV64IA-WMO-NEXT: bge a7, a1, .LBB47_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB47_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a6, a5, a1
|
|
; RV64IA-WMO-NEXT: and a6, a6, a4
|
|
; RV64IA-WMO-NEXT: xor a6, a5, a6
|
|
; RV64IA-WMO-NEXT: .LBB47_3: # in Loop: Header=BB47_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a6, .LBB47_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a5, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_max_i8_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV64IA-TSO-NEXT: li a4, 255
|
|
; RV64IA-TSO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 56
|
|
; RV64IA-TSO-NEXT: srai a1, a1, 56
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: xori a3, a3, 56
|
|
; RV64IA-TSO-NEXT: .LBB47_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a5, (a2)
|
|
; RV64IA-TSO-NEXT: and a7, a5, a4
|
|
; RV64IA-TSO-NEXT: mv a6, a5
|
|
; RV64IA-TSO-NEXT: sll a7, a7, a3
|
|
; RV64IA-TSO-NEXT: sra a7, a7, a3
|
|
; RV64IA-TSO-NEXT: bge a7, a1, .LBB47_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB47_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a6, a5, a1
|
|
; RV64IA-TSO-NEXT: and a6, a6, a4
|
|
; RV64IA-TSO-NEXT: xor a6, a5, a6
|
|
; RV64IA-TSO-NEXT: .LBB47_3: # in Loop: Header=BB47_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a6, .LBB47_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a5, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i8 %b release
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_max_i8_acq_rel(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i8_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 24
|
|
; RV32I-NEXT: srai s2, a0, 24
|
|
; RV32I-NEXT: j .LBB48_2
|
|
; RV32I-NEXT: .LBB48_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB48_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB48_4
|
|
; RV32I-NEXT: .LBB48_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 24
|
|
; RV32I-NEXT: srai a0, a0, 24
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: blt s2, a0, .LBB48_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB48_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB48_1
|
|
; RV32I-NEXT: .LBB48_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_max_i8_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV32IA-WMO-NEXT: li a4, 255
|
|
; RV32IA-WMO-NEXT: sll a4, a4, a0
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 24
|
|
; RV32IA-WMO-NEXT: srai a1, a1, 24
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: xori a3, a3, 24
|
|
; RV32IA-WMO-NEXT: .LBB48_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a5, (a2)
|
|
; RV32IA-WMO-NEXT: and a7, a5, a4
|
|
; RV32IA-WMO-NEXT: mv a6, a5
|
|
; RV32IA-WMO-NEXT: sll a7, a7, a3
|
|
; RV32IA-WMO-NEXT: sra a7, a7, a3
|
|
; RV32IA-WMO-NEXT: bge a7, a1, .LBB48_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB48_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a6, a5, a1
|
|
; RV32IA-WMO-NEXT: and a6, a6, a4
|
|
; RV32IA-WMO-NEXT: xor a6, a5, a6
|
|
; RV32IA-WMO-NEXT: .LBB48_3: # in Loop: Header=BB48_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a6, .LBB48_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a5, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_max_i8_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV32IA-TSO-NEXT: li a4, 255
|
|
; RV32IA-TSO-NEXT: sll a4, a4, a0
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 24
|
|
; RV32IA-TSO-NEXT: srai a1, a1, 24
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: xori a3, a3, 24
|
|
; RV32IA-TSO-NEXT: .LBB48_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a5, (a2)
|
|
; RV32IA-TSO-NEXT: and a7, a5, a4
|
|
; RV32IA-TSO-NEXT: mv a6, a5
|
|
; RV32IA-TSO-NEXT: sll a7, a7, a3
|
|
; RV32IA-TSO-NEXT: sra a7, a7, a3
|
|
; RV32IA-TSO-NEXT: bge a7, a1, .LBB48_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB48_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a6, a5, a1
|
|
; RV32IA-TSO-NEXT: and a6, a6, a4
|
|
; RV32IA-TSO-NEXT: xor a6, a5, a6
|
|
; RV32IA-TSO-NEXT: .LBB48_3: # in Loop: Header=BB48_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a6, .LBB48_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a5, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i8_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 56
|
|
; RV64I-NEXT: srai s2, a0, 56
|
|
; RV64I-NEXT: j .LBB48_2
|
|
; RV64I-NEXT: .LBB48_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB48_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB48_4
|
|
; RV64I-NEXT: .LBB48_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 56
|
|
; RV64I-NEXT: srai a0, a0, 56
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s2, a0, .LBB48_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB48_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB48_1
|
|
; RV64I-NEXT: .LBB48_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_max_i8_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV64IA-WMO-NEXT: li a4, 255
|
|
; RV64IA-WMO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 56
|
|
; RV64IA-WMO-NEXT: srai a1, a1, 56
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: xori a3, a3, 56
|
|
; RV64IA-WMO-NEXT: .LBB48_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a5, (a2)
|
|
; RV64IA-WMO-NEXT: and a7, a5, a4
|
|
; RV64IA-WMO-NEXT: mv a6, a5
|
|
; RV64IA-WMO-NEXT: sll a7, a7, a3
|
|
; RV64IA-WMO-NEXT: sra a7, a7, a3
|
|
; RV64IA-WMO-NEXT: bge a7, a1, .LBB48_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB48_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a6, a5, a1
|
|
; RV64IA-WMO-NEXT: and a6, a6, a4
|
|
; RV64IA-WMO-NEXT: xor a6, a5, a6
|
|
; RV64IA-WMO-NEXT: .LBB48_3: # in Loop: Header=BB48_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a6, .LBB48_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a5, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_max_i8_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV64IA-TSO-NEXT: li a4, 255
|
|
; RV64IA-TSO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 56
|
|
; RV64IA-TSO-NEXT: srai a1, a1, 56
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: xori a3, a3, 56
|
|
; RV64IA-TSO-NEXT: .LBB48_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a5, (a2)
|
|
; RV64IA-TSO-NEXT: and a7, a5, a4
|
|
; RV64IA-TSO-NEXT: mv a6, a5
|
|
; RV64IA-TSO-NEXT: sll a7, a7, a3
|
|
; RV64IA-TSO-NEXT: sra a7, a7, a3
|
|
; RV64IA-TSO-NEXT: bge a7, a1, .LBB48_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB48_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a6, a5, a1
|
|
; RV64IA-TSO-NEXT: and a6, a6, a4
|
|
; RV64IA-TSO-NEXT: xor a6, a5, a6
|
|
; RV64IA-TSO-NEXT: .LBB48_3: # in Loop: Header=BB48_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a6, .LBB48_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a5, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i8 %b acq_rel
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_max_i8_seq_cst(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i8_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 24
|
|
; RV32I-NEXT: srai s2, a0, 24
|
|
; RV32I-NEXT: j .LBB49_2
|
|
; RV32I-NEXT: .LBB49_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB49_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB49_4
|
|
; RV32I-NEXT: .LBB49_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 24
|
|
; RV32I-NEXT: srai a0, a0, 24
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: blt s2, a0, .LBB49_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB49_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB49_1
|
|
; RV32I-NEXT: .LBB49_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_max_i8_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: andi a3, a0, 24
|
|
; RV32IA-NEXT: li a4, 255
|
|
; RV32IA-NEXT: sll a4, a4, a0
|
|
; RV32IA-NEXT: slli a1, a1, 24
|
|
; RV32IA-NEXT: srai a1, a1, 24
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: xori a3, a3, 24
|
|
; RV32IA-NEXT: .LBB49_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a5, (a2)
|
|
; RV32IA-NEXT: and a7, a5, a4
|
|
; RV32IA-NEXT: mv a6, a5
|
|
; RV32IA-NEXT: sll a7, a7, a3
|
|
; RV32IA-NEXT: sra a7, a7, a3
|
|
; RV32IA-NEXT: bge a7, a1, .LBB49_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB49_1 Depth=1
|
|
; RV32IA-NEXT: xor a6, a5, a1
|
|
; RV32IA-NEXT: and a6, a6, a4
|
|
; RV32IA-NEXT: xor a6, a5, a6
|
|
; RV32IA-NEXT: .LBB49_3: # in Loop: Header=BB49_1 Depth=1
|
|
; RV32IA-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV32IA-NEXT: bnez a6, .LBB49_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a5, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i8_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 56
|
|
; RV64I-NEXT: srai s2, a0, 56
|
|
; RV64I-NEXT: j .LBB49_2
|
|
; RV64I-NEXT: .LBB49_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB49_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB49_4
|
|
; RV64I-NEXT: .LBB49_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 56
|
|
; RV64I-NEXT: srai a0, a0, 56
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s2, a0, .LBB49_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB49_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB49_1
|
|
; RV64I-NEXT: .LBB49_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_max_i8_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: andi a3, a0, 24
|
|
; RV64IA-NEXT: li a4, 255
|
|
; RV64IA-NEXT: sllw a4, a4, a0
|
|
; RV64IA-NEXT: slli a1, a1, 56
|
|
; RV64IA-NEXT: srai a1, a1, 56
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: xori a3, a3, 56
|
|
; RV64IA-NEXT: .LBB49_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a5, (a2)
|
|
; RV64IA-NEXT: and a7, a5, a4
|
|
; RV64IA-NEXT: mv a6, a5
|
|
; RV64IA-NEXT: sll a7, a7, a3
|
|
; RV64IA-NEXT: sra a7, a7, a3
|
|
; RV64IA-NEXT: bge a7, a1, .LBB49_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB49_1 Depth=1
|
|
; RV64IA-NEXT: xor a6, a5, a1
|
|
; RV64IA-NEXT: and a6, a6, a4
|
|
; RV64IA-NEXT: xor a6, a5, a6
|
|
; RV64IA-NEXT: .LBB49_3: # in Loop: Header=BB49_1 Depth=1
|
|
; RV64IA-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV64IA-NEXT: bnez a6, .LBB49_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a5, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i8 %b seq_cst
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_min_i8_monotonic(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i8_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 24
|
|
; RV32I-NEXT: srai s2, a0, 24
|
|
; RV32I-NEXT: j .LBB50_2
|
|
; RV32I-NEXT: .LBB50_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB50_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB50_4
|
|
; RV32I-NEXT: .LBB50_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 24
|
|
; RV32I-NEXT: srai a0, a0, 24
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bge s2, a0, .LBB50_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB50_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB50_1
|
|
; RV32I-NEXT: .LBB50_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_min_i8_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: andi a3, a0, 24
|
|
; RV32IA-NEXT: li a4, 255
|
|
; RV32IA-NEXT: sll a4, a4, a0
|
|
; RV32IA-NEXT: slli a1, a1, 24
|
|
; RV32IA-NEXT: srai a1, a1, 24
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: xori a3, a3, 24
|
|
; RV32IA-NEXT: .LBB50_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a5, (a2)
|
|
; RV32IA-NEXT: and a7, a5, a4
|
|
; RV32IA-NEXT: mv a6, a5
|
|
; RV32IA-NEXT: sll a7, a7, a3
|
|
; RV32IA-NEXT: sra a7, a7, a3
|
|
; RV32IA-NEXT: bge a1, a7, .LBB50_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB50_1 Depth=1
|
|
; RV32IA-NEXT: xor a6, a5, a1
|
|
; RV32IA-NEXT: and a6, a6, a4
|
|
; RV32IA-NEXT: xor a6, a5, a6
|
|
; RV32IA-NEXT: .LBB50_3: # in Loop: Header=BB50_1 Depth=1
|
|
; RV32IA-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-NEXT: bnez a6, .LBB50_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a5, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i8_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 56
|
|
; RV64I-NEXT: srai s2, a0, 56
|
|
; RV64I-NEXT: j .LBB50_2
|
|
; RV64I-NEXT: .LBB50_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB50_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB50_4
|
|
; RV64I-NEXT: .LBB50_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 56
|
|
; RV64I-NEXT: srai a0, a0, 56
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s2, a0, .LBB50_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB50_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB50_1
|
|
; RV64I-NEXT: .LBB50_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_min_i8_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: andi a3, a0, 24
|
|
; RV64IA-NEXT: li a4, 255
|
|
; RV64IA-NEXT: sllw a4, a4, a0
|
|
; RV64IA-NEXT: slli a1, a1, 56
|
|
; RV64IA-NEXT: srai a1, a1, 56
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: xori a3, a3, 56
|
|
; RV64IA-NEXT: .LBB50_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a5, (a2)
|
|
; RV64IA-NEXT: and a7, a5, a4
|
|
; RV64IA-NEXT: mv a6, a5
|
|
; RV64IA-NEXT: sll a7, a7, a3
|
|
; RV64IA-NEXT: sra a7, a7, a3
|
|
; RV64IA-NEXT: bge a1, a7, .LBB50_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB50_1 Depth=1
|
|
; RV64IA-NEXT: xor a6, a5, a1
|
|
; RV64IA-NEXT: and a6, a6, a4
|
|
; RV64IA-NEXT: xor a6, a5, a6
|
|
; RV64IA-NEXT: .LBB50_3: # in Loop: Header=BB50_1 Depth=1
|
|
; RV64IA-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-NEXT: bnez a6, .LBB50_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a5, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i8 %b monotonic
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_min_i8_acquire(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i8_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 24
|
|
; RV32I-NEXT: srai s2, a0, 24
|
|
; RV32I-NEXT: j .LBB51_2
|
|
; RV32I-NEXT: .LBB51_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB51_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB51_4
|
|
; RV32I-NEXT: .LBB51_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 24
|
|
; RV32I-NEXT: srai a0, a0, 24
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bge s2, a0, .LBB51_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB51_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB51_1
|
|
; RV32I-NEXT: .LBB51_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_min_i8_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV32IA-WMO-NEXT: li a4, 255
|
|
; RV32IA-WMO-NEXT: sll a4, a4, a0
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 24
|
|
; RV32IA-WMO-NEXT: srai a1, a1, 24
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: xori a3, a3, 24
|
|
; RV32IA-WMO-NEXT: .LBB51_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a5, (a2)
|
|
; RV32IA-WMO-NEXT: and a7, a5, a4
|
|
; RV32IA-WMO-NEXT: mv a6, a5
|
|
; RV32IA-WMO-NEXT: sll a7, a7, a3
|
|
; RV32IA-WMO-NEXT: sra a7, a7, a3
|
|
; RV32IA-WMO-NEXT: bge a1, a7, .LBB51_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB51_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a6, a5, a1
|
|
; RV32IA-WMO-NEXT: and a6, a6, a4
|
|
; RV32IA-WMO-NEXT: xor a6, a5, a6
|
|
; RV32IA-WMO-NEXT: .LBB51_3: # in Loop: Header=BB51_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a6, .LBB51_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a5, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_min_i8_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV32IA-TSO-NEXT: li a4, 255
|
|
; RV32IA-TSO-NEXT: sll a4, a4, a0
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 24
|
|
; RV32IA-TSO-NEXT: srai a1, a1, 24
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: xori a3, a3, 24
|
|
; RV32IA-TSO-NEXT: .LBB51_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a5, (a2)
|
|
; RV32IA-TSO-NEXT: and a7, a5, a4
|
|
; RV32IA-TSO-NEXT: mv a6, a5
|
|
; RV32IA-TSO-NEXT: sll a7, a7, a3
|
|
; RV32IA-TSO-NEXT: sra a7, a7, a3
|
|
; RV32IA-TSO-NEXT: bge a1, a7, .LBB51_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB51_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a6, a5, a1
|
|
; RV32IA-TSO-NEXT: and a6, a6, a4
|
|
; RV32IA-TSO-NEXT: xor a6, a5, a6
|
|
; RV32IA-TSO-NEXT: .LBB51_3: # in Loop: Header=BB51_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a6, .LBB51_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a5, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i8_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 56
|
|
; RV64I-NEXT: srai s2, a0, 56
|
|
; RV64I-NEXT: j .LBB51_2
|
|
; RV64I-NEXT: .LBB51_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB51_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB51_4
|
|
; RV64I-NEXT: .LBB51_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 56
|
|
; RV64I-NEXT: srai a0, a0, 56
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s2, a0, .LBB51_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB51_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB51_1
|
|
; RV64I-NEXT: .LBB51_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_min_i8_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV64IA-WMO-NEXT: li a4, 255
|
|
; RV64IA-WMO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 56
|
|
; RV64IA-WMO-NEXT: srai a1, a1, 56
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: xori a3, a3, 56
|
|
; RV64IA-WMO-NEXT: .LBB51_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a5, (a2)
|
|
; RV64IA-WMO-NEXT: and a7, a5, a4
|
|
; RV64IA-WMO-NEXT: mv a6, a5
|
|
; RV64IA-WMO-NEXT: sll a7, a7, a3
|
|
; RV64IA-WMO-NEXT: sra a7, a7, a3
|
|
; RV64IA-WMO-NEXT: bge a1, a7, .LBB51_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB51_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a6, a5, a1
|
|
; RV64IA-WMO-NEXT: and a6, a6, a4
|
|
; RV64IA-WMO-NEXT: xor a6, a5, a6
|
|
; RV64IA-WMO-NEXT: .LBB51_3: # in Loop: Header=BB51_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a6, .LBB51_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a5, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_min_i8_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV64IA-TSO-NEXT: li a4, 255
|
|
; RV64IA-TSO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 56
|
|
; RV64IA-TSO-NEXT: srai a1, a1, 56
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: xori a3, a3, 56
|
|
; RV64IA-TSO-NEXT: .LBB51_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a5, (a2)
|
|
; RV64IA-TSO-NEXT: and a7, a5, a4
|
|
; RV64IA-TSO-NEXT: mv a6, a5
|
|
; RV64IA-TSO-NEXT: sll a7, a7, a3
|
|
; RV64IA-TSO-NEXT: sra a7, a7, a3
|
|
; RV64IA-TSO-NEXT: bge a1, a7, .LBB51_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB51_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a6, a5, a1
|
|
; RV64IA-TSO-NEXT: and a6, a6, a4
|
|
; RV64IA-TSO-NEXT: xor a6, a5, a6
|
|
; RV64IA-TSO-NEXT: .LBB51_3: # in Loop: Header=BB51_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a6, .LBB51_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a5, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i8 %b acquire
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_min_i8_release(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i8_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 24
|
|
; RV32I-NEXT: srai s2, a0, 24
|
|
; RV32I-NEXT: j .LBB52_2
|
|
; RV32I-NEXT: .LBB52_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB52_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB52_4
|
|
; RV32I-NEXT: .LBB52_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 24
|
|
; RV32I-NEXT: srai a0, a0, 24
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bge s2, a0, .LBB52_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB52_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB52_1
|
|
; RV32I-NEXT: .LBB52_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_min_i8_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV32IA-WMO-NEXT: li a4, 255
|
|
; RV32IA-WMO-NEXT: sll a4, a4, a0
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 24
|
|
; RV32IA-WMO-NEXT: srai a1, a1, 24
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: xori a3, a3, 24
|
|
; RV32IA-WMO-NEXT: .LBB52_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a5, (a2)
|
|
; RV32IA-WMO-NEXT: and a7, a5, a4
|
|
; RV32IA-WMO-NEXT: mv a6, a5
|
|
; RV32IA-WMO-NEXT: sll a7, a7, a3
|
|
; RV32IA-WMO-NEXT: sra a7, a7, a3
|
|
; RV32IA-WMO-NEXT: bge a1, a7, .LBB52_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB52_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a6, a5, a1
|
|
; RV32IA-WMO-NEXT: and a6, a6, a4
|
|
; RV32IA-WMO-NEXT: xor a6, a5, a6
|
|
; RV32IA-WMO-NEXT: .LBB52_3: # in Loop: Header=BB52_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a6, .LBB52_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a5, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_min_i8_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV32IA-TSO-NEXT: li a4, 255
|
|
; RV32IA-TSO-NEXT: sll a4, a4, a0
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 24
|
|
; RV32IA-TSO-NEXT: srai a1, a1, 24
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: xori a3, a3, 24
|
|
; RV32IA-TSO-NEXT: .LBB52_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a5, (a2)
|
|
; RV32IA-TSO-NEXT: and a7, a5, a4
|
|
; RV32IA-TSO-NEXT: mv a6, a5
|
|
; RV32IA-TSO-NEXT: sll a7, a7, a3
|
|
; RV32IA-TSO-NEXT: sra a7, a7, a3
|
|
; RV32IA-TSO-NEXT: bge a1, a7, .LBB52_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB52_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a6, a5, a1
|
|
; RV32IA-TSO-NEXT: and a6, a6, a4
|
|
; RV32IA-TSO-NEXT: xor a6, a5, a6
|
|
; RV32IA-TSO-NEXT: .LBB52_3: # in Loop: Header=BB52_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a6, .LBB52_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a5, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i8_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 56
|
|
; RV64I-NEXT: srai s2, a0, 56
|
|
; RV64I-NEXT: j .LBB52_2
|
|
; RV64I-NEXT: .LBB52_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB52_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB52_4
|
|
; RV64I-NEXT: .LBB52_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 56
|
|
; RV64I-NEXT: srai a0, a0, 56
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s2, a0, .LBB52_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB52_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB52_1
|
|
; RV64I-NEXT: .LBB52_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_min_i8_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV64IA-WMO-NEXT: li a4, 255
|
|
; RV64IA-WMO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 56
|
|
; RV64IA-WMO-NEXT: srai a1, a1, 56
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: xori a3, a3, 56
|
|
; RV64IA-WMO-NEXT: .LBB52_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a5, (a2)
|
|
; RV64IA-WMO-NEXT: and a7, a5, a4
|
|
; RV64IA-WMO-NEXT: mv a6, a5
|
|
; RV64IA-WMO-NEXT: sll a7, a7, a3
|
|
; RV64IA-WMO-NEXT: sra a7, a7, a3
|
|
; RV64IA-WMO-NEXT: bge a1, a7, .LBB52_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB52_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a6, a5, a1
|
|
; RV64IA-WMO-NEXT: and a6, a6, a4
|
|
; RV64IA-WMO-NEXT: xor a6, a5, a6
|
|
; RV64IA-WMO-NEXT: .LBB52_3: # in Loop: Header=BB52_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a6, .LBB52_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a5, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_min_i8_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV64IA-TSO-NEXT: li a4, 255
|
|
; RV64IA-TSO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 56
|
|
; RV64IA-TSO-NEXT: srai a1, a1, 56
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: xori a3, a3, 56
|
|
; RV64IA-TSO-NEXT: .LBB52_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a5, (a2)
|
|
; RV64IA-TSO-NEXT: and a7, a5, a4
|
|
; RV64IA-TSO-NEXT: mv a6, a5
|
|
; RV64IA-TSO-NEXT: sll a7, a7, a3
|
|
; RV64IA-TSO-NEXT: sra a7, a7, a3
|
|
; RV64IA-TSO-NEXT: bge a1, a7, .LBB52_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB52_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a6, a5, a1
|
|
; RV64IA-TSO-NEXT: and a6, a6, a4
|
|
; RV64IA-TSO-NEXT: xor a6, a5, a6
|
|
; RV64IA-TSO-NEXT: .LBB52_3: # in Loop: Header=BB52_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a6, .LBB52_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a5, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i8 %b release
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_min_i8_acq_rel(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i8_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 24
|
|
; RV32I-NEXT: srai s2, a0, 24
|
|
; RV32I-NEXT: j .LBB53_2
|
|
; RV32I-NEXT: .LBB53_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB53_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB53_4
|
|
; RV32I-NEXT: .LBB53_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 24
|
|
; RV32I-NEXT: srai a0, a0, 24
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bge s2, a0, .LBB53_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB53_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB53_1
|
|
; RV32I-NEXT: .LBB53_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_min_i8_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV32IA-WMO-NEXT: li a4, 255
|
|
; RV32IA-WMO-NEXT: sll a4, a4, a0
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 24
|
|
; RV32IA-WMO-NEXT: srai a1, a1, 24
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: xori a3, a3, 24
|
|
; RV32IA-WMO-NEXT: .LBB53_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a5, (a2)
|
|
; RV32IA-WMO-NEXT: and a7, a5, a4
|
|
; RV32IA-WMO-NEXT: mv a6, a5
|
|
; RV32IA-WMO-NEXT: sll a7, a7, a3
|
|
; RV32IA-WMO-NEXT: sra a7, a7, a3
|
|
; RV32IA-WMO-NEXT: bge a1, a7, .LBB53_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB53_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a6, a5, a1
|
|
; RV32IA-WMO-NEXT: and a6, a6, a4
|
|
; RV32IA-WMO-NEXT: xor a6, a5, a6
|
|
; RV32IA-WMO-NEXT: .LBB53_3: # in Loop: Header=BB53_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a6, .LBB53_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a5, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_min_i8_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV32IA-TSO-NEXT: li a4, 255
|
|
; RV32IA-TSO-NEXT: sll a4, a4, a0
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 24
|
|
; RV32IA-TSO-NEXT: srai a1, a1, 24
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: xori a3, a3, 24
|
|
; RV32IA-TSO-NEXT: .LBB53_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a5, (a2)
|
|
; RV32IA-TSO-NEXT: and a7, a5, a4
|
|
; RV32IA-TSO-NEXT: mv a6, a5
|
|
; RV32IA-TSO-NEXT: sll a7, a7, a3
|
|
; RV32IA-TSO-NEXT: sra a7, a7, a3
|
|
; RV32IA-TSO-NEXT: bge a1, a7, .LBB53_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB53_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a6, a5, a1
|
|
; RV32IA-TSO-NEXT: and a6, a6, a4
|
|
; RV32IA-TSO-NEXT: xor a6, a5, a6
|
|
; RV32IA-TSO-NEXT: .LBB53_3: # in Loop: Header=BB53_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a6, .LBB53_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a5, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i8_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 56
|
|
; RV64I-NEXT: srai s2, a0, 56
|
|
; RV64I-NEXT: j .LBB53_2
|
|
; RV64I-NEXT: .LBB53_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB53_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB53_4
|
|
; RV64I-NEXT: .LBB53_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 56
|
|
; RV64I-NEXT: srai a0, a0, 56
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s2, a0, .LBB53_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB53_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB53_1
|
|
; RV64I-NEXT: .LBB53_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_min_i8_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV64IA-WMO-NEXT: li a4, 255
|
|
; RV64IA-WMO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 56
|
|
; RV64IA-WMO-NEXT: srai a1, a1, 56
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: xori a3, a3, 56
|
|
; RV64IA-WMO-NEXT: .LBB53_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a5, (a2)
|
|
; RV64IA-WMO-NEXT: and a7, a5, a4
|
|
; RV64IA-WMO-NEXT: mv a6, a5
|
|
; RV64IA-WMO-NEXT: sll a7, a7, a3
|
|
; RV64IA-WMO-NEXT: sra a7, a7, a3
|
|
; RV64IA-WMO-NEXT: bge a1, a7, .LBB53_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB53_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a6, a5, a1
|
|
; RV64IA-WMO-NEXT: and a6, a6, a4
|
|
; RV64IA-WMO-NEXT: xor a6, a5, a6
|
|
; RV64IA-WMO-NEXT: .LBB53_3: # in Loop: Header=BB53_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a6, .LBB53_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a5, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_min_i8_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV64IA-TSO-NEXT: li a4, 255
|
|
; RV64IA-TSO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 56
|
|
; RV64IA-TSO-NEXT: srai a1, a1, 56
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: xori a3, a3, 56
|
|
; RV64IA-TSO-NEXT: .LBB53_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a5, (a2)
|
|
; RV64IA-TSO-NEXT: and a7, a5, a4
|
|
; RV64IA-TSO-NEXT: mv a6, a5
|
|
; RV64IA-TSO-NEXT: sll a7, a7, a3
|
|
; RV64IA-TSO-NEXT: sra a7, a7, a3
|
|
; RV64IA-TSO-NEXT: bge a1, a7, .LBB53_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB53_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a6, a5, a1
|
|
; RV64IA-TSO-NEXT: and a6, a6, a4
|
|
; RV64IA-TSO-NEXT: xor a6, a5, a6
|
|
; RV64IA-TSO-NEXT: .LBB53_3: # in Loop: Header=BB53_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a6, .LBB53_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a5, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i8 %b acq_rel
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_min_i8_seq_cst(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i8_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 24
|
|
; RV32I-NEXT: srai s2, a0, 24
|
|
; RV32I-NEXT: j .LBB54_2
|
|
; RV32I-NEXT: .LBB54_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB54_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB54_4
|
|
; RV32I-NEXT: .LBB54_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 24
|
|
; RV32I-NEXT: srai a0, a0, 24
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bge s2, a0, .LBB54_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB54_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB54_1
|
|
; RV32I-NEXT: .LBB54_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_min_i8_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: andi a3, a0, 24
|
|
; RV32IA-NEXT: li a4, 255
|
|
; RV32IA-NEXT: sll a4, a4, a0
|
|
; RV32IA-NEXT: slli a1, a1, 24
|
|
; RV32IA-NEXT: srai a1, a1, 24
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: xori a3, a3, 24
|
|
; RV32IA-NEXT: .LBB54_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a5, (a2)
|
|
; RV32IA-NEXT: and a7, a5, a4
|
|
; RV32IA-NEXT: mv a6, a5
|
|
; RV32IA-NEXT: sll a7, a7, a3
|
|
; RV32IA-NEXT: sra a7, a7, a3
|
|
; RV32IA-NEXT: bge a1, a7, .LBB54_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB54_1 Depth=1
|
|
; RV32IA-NEXT: xor a6, a5, a1
|
|
; RV32IA-NEXT: and a6, a6, a4
|
|
; RV32IA-NEXT: xor a6, a5, a6
|
|
; RV32IA-NEXT: .LBB54_3: # in Loop: Header=BB54_1 Depth=1
|
|
; RV32IA-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV32IA-NEXT: bnez a6, .LBB54_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a5, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i8_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 56
|
|
; RV64I-NEXT: srai s2, a0, 56
|
|
; RV64I-NEXT: j .LBB54_2
|
|
; RV64I-NEXT: .LBB54_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB54_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB54_4
|
|
; RV64I-NEXT: .LBB54_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 56
|
|
; RV64I-NEXT: srai a0, a0, 56
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s2, a0, .LBB54_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB54_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB54_1
|
|
; RV64I-NEXT: .LBB54_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_min_i8_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: andi a3, a0, 24
|
|
; RV64IA-NEXT: li a4, 255
|
|
; RV64IA-NEXT: sllw a4, a4, a0
|
|
; RV64IA-NEXT: slli a1, a1, 56
|
|
; RV64IA-NEXT: srai a1, a1, 56
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: xori a3, a3, 56
|
|
; RV64IA-NEXT: .LBB54_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a5, (a2)
|
|
; RV64IA-NEXT: and a7, a5, a4
|
|
; RV64IA-NEXT: mv a6, a5
|
|
; RV64IA-NEXT: sll a7, a7, a3
|
|
; RV64IA-NEXT: sra a7, a7, a3
|
|
; RV64IA-NEXT: bge a1, a7, .LBB54_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB54_1 Depth=1
|
|
; RV64IA-NEXT: xor a6, a5, a1
|
|
; RV64IA-NEXT: and a6, a6, a4
|
|
; RV64IA-NEXT: xor a6, a5, a6
|
|
; RV64IA-NEXT: .LBB54_3: # in Loop: Header=BB54_1 Depth=1
|
|
; RV64IA-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV64IA-NEXT: bnez a6, .LBB54_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a5, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i8 %b seq_cst
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_umax_i8_monotonic(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i8_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: andi s2, a1, 255
|
|
; RV32I-NEXT: j .LBB55_2
|
|
; RV32I-NEXT: .LBB55_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB55_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB55_4
|
|
; RV32I-NEXT: .LBB55_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: andi a0, a3, 255
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bltu s2, a0, .LBB55_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB55_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB55_1
|
|
; RV32I-NEXT: .LBB55_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umax_i8_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: li a3, 255
|
|
; RV32IA-NEXT: sll a3, a3, a0
|
|
; RV32IA-NEXT: andi a1, a1, 255
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB55_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a4, (a2)
|
|
; RV32IA-NEXT: and a6, a4, a3
|
|
; RV32IA-NEXT: mv a5, a4
|
|
; RV32IA-NEXT: bgeu a6, a1, .LBB55_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB55_1 Depth=1
|
|
; RV32IA-NEXT: xor a5, a4, a1
|
|
; RV32IA-NEXT: and a5, a5, a3
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: .LBB55_3: # in Loop: Header=BB55_1 Depth=1
|
|
; RV32IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB55_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a4, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i8_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: andi s2, a1, 255
|
|
; RV64I-NEXT: j .LBB55_2
|
|
; RV64I-NEXT: .LBB55_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB55_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB55_4
|
|
; RV64I-NEXT: .LBB55_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: andi a0, a3, 255
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bltu s2, a0, .LBB55_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB55_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB55_1
|
|
; RV64I-NEXT: .LBB55_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_umax_i8_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: li a3, 255
|
|
; RV64IA-NEXT: sllw a3, a3, a0
|
|
; RV64IA-NEXT: andi a1, a1, 255
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB55_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a4, (a2)
|
|
; RV64IA-NEXT: and a6, a4, a3
|
|
; RV64IA-NEXT: mv a5, a4
|
|
; RV64IA-NEXT: bgeu a6, a1, .LBB55_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB55_1 Depth=1
|
|
; RV64IA-NEXT: xor a5, a4, a1
|
|
; RV64IA-NEXT: and a5, a5, a3
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: .LBB55_3: # in Loop: Header=BB55_1 Depth=1
|
|
; RV64IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB55_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a4, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i8 %b monotonic
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_umax_i8_acquire(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i8_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: andi s2, a1, 255
|
|
; RV32I-NEXT: j .LBB56_2
|
|
; RV32I-NEXT: .LBB56_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB56_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB56_4
|
|
; RV32I-NEXT: .LBB56_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: andi a0, a3, 255
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bltu s2, a0, .LBB56_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB56_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB56_1
|
|
; RV32I-NEXT: .LBB56_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umax_i8_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB56_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV32IA-WMO-NEXT: and a6, a4, a3
|
|
; RV32IA-WMO-NEXT: mv a5, a4
|
|
; RV32IA-WMO-NEXT: bgeu a6, a1, .LBB56_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB56_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a1
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: .LBB56_3: # in Loop: Header=BB56_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB56_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umax_i8_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB56_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: and a6, a4, a3
|
|
; RV32IA-TSO-NEXT: mv a5, a4
|
|
; RV32IA-TSO-NEXT: bgeu a6, a1, .LBB56_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB56_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a1
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: .LBB56_3: # in Loop: Header=BB56_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB56_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i8_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: andi s2, a1, 255
|
|
; RV64I-NEXT: j .LBB56_2
|
|
; RV64I-NEXT: .LBB56_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB56_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB56_4
|
|
; RV64I-NEXT: .LBB56_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: andi a0, a3, 255
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bltu s2, a0, .LBB56_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB56_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB56_1
|
|
; RV64I-NEXT: .LBB56_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umax_i8_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB56_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV64IA-WMO-NEXT: and a6, a4, a3
|
|
; RV64IA-WMO-NEXT: mv a5, a4
|
|
; RV64IA-WMO-NEXT: bgeu a6, a1, .LBB56_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB56_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a1
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: .LBB56_3: # in Loop: Header=BB56_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB56_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umax_i8_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB56_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: and a6, a4, a3
|
|
; RV64IA-TSO-NEXT: mv a5, a4
|
|
; RV64IA-TSO-NEXT: bgeu a6, a1, .LBB56_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB56_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a1
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: .LBB56_3: # in Loop: Header=BB56_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB56_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i8 %b acquire
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_umax_i8_release(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i8_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: andi s2, a1, 255
|
|
; RV32I-NEXT: j .LBB57_2
|
|
; RV32I-NEXT: .LBB57_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB57_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB57_4
|
|
; RV32I-NEXT: .LBB57_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: andi a0, a3, 255
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bltu s2, a0, .LBB57_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB57_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB57_1
|
|
; RV32I-NEXT: .LBB57_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umax_i8_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB57_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-WMO-NEXT: and a6, a4, a3
|
|
; RV32IA-WMO-NEXT: mv a5, a4
|
|
; RV32IA-WMO-NEXT: bgeu a6, a1, .LBB57_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB57_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a1
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: .LBB57_3: # in Loop: Header=BB57_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB57_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umax_i8_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB57_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: and a6, a4, a3
|
|
; RV32IA-TSO-NEXT: mv a5, a4
|
|
; RV32IA-TSO-NEXT: bgeu a6, a1, .LBB57_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB57_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a1
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: .LBB57_3: # in Loop: Header=BB57_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB57_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i8_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: andi s2, a1, 255
|
|
; RV64I-NEXT: j .LBB57_2
|
|
; RV64I-NEXT: .LBB57_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB57_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB57_4
|
|
; RV64I-NEXT: .LBB57_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: andi a0, a3, 255
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bltu s2, a0, .LBB57_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB57_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB57_1
|
|
; RV64I-NEXT: .LBB57_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umax_i8_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB57_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-WMO-NEXT: and a6, a4, a3
|
|
; RV64IA-WMO-NEXT: mv a5, a4
|
|
; RV64IA-WMO-NEXT: bgeu a6, a1, .LBB57_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB57_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a1
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: .LBB57_3: # in Loop: Header=BB57_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB57_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umax_i8_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB57_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: and a6, a4, a3
|
|
; RV64IA-TSO-NEXT: mv a5, a4
|
|
; RV64IA-TSO-NEXT: bgeu a6, a1, .LBB57_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB57_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a1
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: .LBB57_3: # in Loop: Header=BB57_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB57_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i8 %b release
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_umax_i8_acq_rel(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i8_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: andi s2, a1, 255
|
|
; RV32I-NEXT: j .LBB58_2
|
|
; RV32I-NEXT: .LBB58_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB58_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB58_4
|
|
; RV32I-NEXT: .LBB58_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: andi a0, a3, 255
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bltu s2, a0, .LBB58_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB58_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB58_1
|
|
; RV32I-NEXT: .LBB58_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umax_i8_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB58_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV32IA-WMO-NEXT: and a6, a4, a3
|
|
; RV32IA-WMO-NEXT: mv a5, a4
|
|
; RV32IA-WMO-NEXT: bgeu a6, a1, .LBB58_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB58_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a1
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: .LBB58_3: # in Loop: Header=BB58_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB58_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umax_i8_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB58_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: and a6, a4, a3
|
|
; RV32IA-TSO-NEXT: mv a5, a4
|
|
; RV32IA-TSO-NEXT: bgeu a6, a1, .LBB58_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB58_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a1
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: .LBB58_3: # in Loop: Header=BB58_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB58_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i8_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: andi s2, a1, 255
|
|
; RV64I-NEXT: j .LBB58_2
|
|
; RV64I-NEXT: .LBB58_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB58_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB58_4
|
|
; RV64I-NEXT: .LBB58_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: andi a0, a3, 255
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bltu s2, a0, .LBB58_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB58_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB58_1
|
|
; RV64I-NEXT: .LBB58_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umax_i8_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB58_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV64IA-WMO-NEXT: and a6, a4, a3
|
|
; RV64IA-WMO-NEXT: mv a5, a4
|
|
; RV64IA-WMO-NEXT: bgeu a6, a1, .LBB58_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB58_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a1
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: .LBB58_3: # in Loop: Header=BB58_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB58_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umax_i8_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB58_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: and a6, a4, a3
|
|
; RV64IA-TSO-NEXT: mv a5, a4
|
|
; RV64IA-TSO-NEXT: bgeu a6, a1, .LBB58_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB58_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a1
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: .LBB58_3: # in Loop: Header=BB58_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB58_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i8 %b acq_rel
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_umax_i8_seq_cst(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i8_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: andi s2, a1, 255
|
|
; RV32I-NEXT: j .LBB59_2
|
|
; RV32I-NEXT: .LBB59_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB59_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB59_4
|
|
; RV32I-NEXT: .LBB59_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: andi a0, a3, 255
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bltu s2, a0, .LBB59_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB59_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB59_1
|
|
; RV32I-NEXT: .LBB59_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umax_i8_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: li a3, 255
|
|
; RV32IA-NEXT: sll a3, a3, a0
|
|
; RV32IA-NEXT: andi a1, a1, 255
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB59_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a4, (a2)
|
|
; RV32IA-NEXT: and a6, a4, a3
|
|
; RV32IA-NEXT: mv a5, a4
|
|
; RV32IA-NEXT: bgeu a6, a1, .LBB59_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB59_1 Depth=1
|
|
; RV32IA-NEXT: xor a5, a4, a1
|
|
; RV32IA-NEXT: and a5, a5, a3
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: .LBB59_3: # in Loop: Header=BB59_1 Depth=1
|
|
; RV32IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB59_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a4, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i8_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: andi s2, a1, 255
|
|
; RV64I-NEXT: j .LBB59_2
|
|
; RV64I-NEXT: .LBB59_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB59_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB59_4
|
|
; RV64I-NEXT: .LBB59_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: andi a0, a3, 255
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bltu s2, a0, .LBB59_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB59_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB59_1
|
|
; RV64I-NEXT: .LBB59_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_umax_i8_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: li a3, 255
|
|
; RV64IA-NEXT: sllw a3, a3, a0
|
|
; RV64IA-NEXT: andi a1, a1, 255
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB59_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a4, (a2)
|
|
; RV64IA-NEXT: and a6, a4, a3
|
|
; RV64IA-NEXT: mv a5, a4
|
|
; RV64IA-NEXT: bgeu a6, a1, .LBB59_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB59_1 Depth=1
|
|
; RV64IA-NEXT: xor a5, a4, a1
|
|
; RV64IA-NEXT: and a5, a5, a3
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: .LBB59_3: # in Loop: Header=BB59_1 Depth=1
|
|
; RV64IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB59_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a4, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i8 %b seq_cst
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_umin_i8_monotonic(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i8_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: andi s2, a1, 255
|
|
; RV32I-NEXT: j .LBB60_2
|
|
; RV32I-NEXT: .LBB60_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB60_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB60_4
|
|
; RV32I-NEXT: .LBB60_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: andi a0, a3, 255
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bgeu s2, a0, .LBB60_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB60_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB60_1
|
|
; RV32I-NEXT: .LBB60_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umin_i8_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: li a3, 255
|
|
; RV32IA-NEXT: sll a3, a3, a0
|
|
; RV32IA-NEXT: andi a1, a1, 255
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB60_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a4, (a2)
|
|
; RV32IA-NEXT: and a6, a4, a3
|
|
; RV32IA-NEXT: mv a5, a4
|
|
; RV32IA-NEXT: bgeu a1, a6, .LBB60_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB60_1 Depth=1
|
|
; RV32IA-NEXT: xor a5, a4, a1
|
|
; RV32IA-NEXT: and a5, a5, a3
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: .LBB60_3: # in Loop: Header=BB60_1 Depth=1
|
|
; RV32IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB60_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a4, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i8_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: andi s2, a1, 255
|
|
; RV64I-NEXT: j .LBB60_2
|
|
; RV64I-NEXT: .LBB60_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB60_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB60_4
|
|
; RV64I-NEXT: .LBB60_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: andi a0, a3, 255
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bgeu s2, a0, .LBB60_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB60_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB60_1
|
|
; RV64I-NEXT: .LBB60_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_umin_i8_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: li a3, 255
|
|
; RV64IA-NEXT: sllw a3, a3, a0
|
|
; RV64IA-NEXT: andi a1, a1, 255
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB60_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a4, (a2)
|
|
; RV64IA-NEXT: and a6, a4, a3
|
|
; RV64IA-NEXT: mv a5, a4
|
|
; RV64IA-NEXT: bgeu a1, a6, .LBB60_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB60_1 Depth=1
|
|
; RV64IA-NEXT: xor a5, a4, a1
|
|
; RV64IA-NEXT: and a5, a5, a3
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: .LBB60_3: # in Loop: Header=BB60_1 Depth=1
|
|
; RV64IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB60_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a4, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i8 %b monotonic
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_umin_i8_acquire(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i8_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: andi s2, a1, 255
|
|
; RV32I-NEXT: j .LBB61_2
|
|
; RV32I-NEXT: .LBB61_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB61_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB61_4
|
|
; RV32I-NEXT: .LBB61_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: andi a0, a3, 255
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bgeu s2, a0, .LBB61_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB61_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB61_1
|
|
; RV32I-NEXT: .LBB61_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umin_i8_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB61_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV32IA-WMO-NEXT: and a6, a4, a3
|
|
; RV32IA-WMO-NEXT: mv a5, a4
|
|
; RV32IA-WMO-NEXT: bgeu a1, a6, .LBB61_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB61_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a1
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: .LBB61_3: # in Loop: Header=BB61_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB61_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umin_i8_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB61_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: and a6, a4, a3
|
|
; RV32IA-TSO-NEXT: mv a5, a4
|
|
; RV32IA-TSO-NEXT: bgeu a1, a6, .LBB61_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB61_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a1
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: .LBB61_3: # in Loop: Header=BB61_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB61_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i8_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: andi s2, a1, 255
|
|
; RV64I-NEXT: j .LBB61_2
|
|
; RV64I-NEXT: .LBB61_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB61_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB61_4
|
|
; RV64I-NEXT: .LBB61_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: andi a0, a3, 255
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bgeu s2, a0, .LBB61_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB61_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB61_1
|
|
; RV64I-NEXT: .LBB61_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umin_i8_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB61_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV64IA-WMO-NEXT: and a6, a4, a3
|
|
; RV64IA-WMO-NEXT: mv a5, a4
|
|
; RV64IA-WMO-NEXT: bgeu a1, a6, .LBB61_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB61_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a1
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: .LBB61_3: # in Loop: Header=BB61_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB61_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umin_i8_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB61_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: and a6, a4, a3
|
|
; RV64IA-TSO-NEXT: mv a5, a4
|
|
; RV64IA-TSO-NEXT: bgeu a1, a6, .LBB61_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB61_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a1
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: .LBB61_3: # in Loop: Header=BB61_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB61_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i8 %b acquire
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_umin_i8_release(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i8_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: andi s2, a1, 255
|
|
; RV32I-NEXT: j .LBB62_2
|
|
; RV32I-NEXT: .LBB62_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB62_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB62_4
|
|
; RV32I-NEXT: .LBB62_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: andi a0, a3, 255
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bgeu s2, a0, .LBB62_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB62_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB62_1
|
|
; RV32I-NEXT: .LBB62_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umin_i8_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB62_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-WMO-NEXT: and a6, a4, a3
|
|
; RV32IA-WMO-NEXT: mv a5, a4
|
|
; RV32IA-WMO-NEXT: bgeu a1, a6, .LBB62_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB62_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a1
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: .LBB62_3: # in Loop: Header=BB62_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB62_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umin_i8_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB62_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: and a6, a4, a3
|
|
; RV32IA-TSO-NEXT: mv a5, a4
|
|
; RV32IA-TSO-NEXT: bgeu a1, a6, .LBB62_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB62_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a1
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: .LBB62_3: # in Loop: Header=BB62_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB62_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i8_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: andi s2, a1, 255
|
|
; RV64I-NEXT: j .LBB62_2
|
|
; RV64I-NEXT: .LBB62_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB62_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB62_4
|
|
; RV64I-NEXT: .LBB62_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: andi a0, a3, 255
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bgeu s2, a0, .LBB62_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB62_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB62_1
|
|
; RV64I-NEXT: .LBB62_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umin_i8_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB62_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-WMO-NEXT: and a6, a4, a3
|
|
; RV64IA-WMO-NEXT: mv a5, a4
|
|
; RV64IA-WMO-NEXT: bgeu a1, a6, .LBB62_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB62_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a1
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: .LBB62_3: # in Loop: Header=BB62_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB62_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umin_i8_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB62_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: and a6, a4, a3
|
|
; RV64IA-TSO-NEXT: mv a5, a4
|
|
; RV64IA-TSO-NEXT: bgeu a1, a6, .LBB62_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB62_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a1
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: .LBB62_3: # in Loop: Header=BB62_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB62_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i8 %b release
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_umin_i8_acq_rel(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i8_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: andi s2, a1, 255
|
|
; RV32I-NEXT: j .LBB63_2
|
|
; RV32I-NEXT: .LBB63_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB63_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB63_4
|
|
; RV32I-NEXT: .LBB63_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: andi a0, a3, 255
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bgeu s2, a0, .LBB63_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB63_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB63_1
|
|
; RV32I-NEXT: .LBB63_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umin_i8_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: li a3, 255
|
|
; RV32IA-WMO-NEXT: sll a3, a3, a0
|
|
; RV32IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB63_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV32IA-WMO-NEXT: and a6, a4, a3
|
|
; RV32IA-WMO-NEXT: mv a5, a4
|
|
; RV32IA-WMO-NEXT: bgeu a1, a6, .LBB63_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB63_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a1
|
|
; RV32IA-WMO-NEXT: and a5, a5, a3
|
|
; RV32IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV32IA-WMO-NEXT: .LBB63_3: # in Loop: Header=BB63_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB63_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a4, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umin_i8_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: li a3, 255
|
|
; RV32IA-TSO-NEXT: sll a3, a3, a0
|
|
; RV32IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB63_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV32IA-TSO-NEXT: and a6, a4, a3
|
|
; RV32IA-TSO-NEXT: mv a5, a4
|
|
; RV32IA-TSO-NEXT: bgeu a1, a6, .LBB63_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB63_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a1
|
|
; RV32IA-TSO-NEXT: and a5, a5, a3
|
|
; RV32IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV32IA-TSO-NEXT: .LBB63_3: # in Loop: Header=BB63_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB63_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a4, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i8_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: andi s2, a1, 255
|
|
; RV64I-NEXT: j .LBB63_2
|
|
; RV64I-NEXT: .LBB63_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB63_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB63_4
|
|
; RV64I-NEXT: .LBB63_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: andi a0, a3, 255
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bgeu s2, a0, .LBB63_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB63_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB63_1
|
|
; RV64I-NEXT: .LBB63_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umin_i8_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: li a3, 255
|
|
; RV64IA-WMO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-WMO-NEXT: andi a1, a1, 255
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB63_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a4, (a2)
|
|
; RV64IA-WMO-NEXT: and a6, a4, a3
|
|
; RV64IA-WMO-NEXT: mv a5, a4
|
|
; RV64IA-WMO-NEXT: bgeu a1, a6, .LBB63_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB63_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a1
|
|
; RV64IA-WMO-NEXT: and a5, a5, a3
|
|
; RV64IA-WMO-NEXT: xor a5, a4, a5
|
|
; RV64IA-WMO-NEXT: .LBB63_3: # in Loop: Header=BB63_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB63_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umin_i8_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: li a3, 255
|
|
; RV64IA-TSO-NEXT: sllw a3, a3, a0
|
|
; RV64IA-TSO-NEXT: andi a1, a1, 255
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB63_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a4, (a2)
|
|
; RV64IA-TSO-NEXT: and a6, a4, a3
|
|
; RV64IA-TSO-NEXT: mv a5, a4
|
|
; RV64IA-TSO-NEXT: bgeu a1, a6, .LBB63_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB63_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a1
|
|
; RV64IA-TSO-NEXT: and a5, a5, a3
|
|
; RV64IA-TSO-NEXT: xor a5, a4, a5
|
|
; RV64IA-TSO-NEXT: .LBB63_3: # in Loop: Header=BB63_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB63_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a4, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i8 %b acq_rel
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @atomicrmw_umin_i8_seq_cst(ptr %a, i8 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i8_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lbu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: andi s2, a1, 255
|
|
; RV32I-NEXT: j .LBB64_2
|
|
; RV32I-NEXT: .LBB64_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB64_2 Depth=1
|
|
; RV32I-NEXT: sb a3, 15(sp)
|
|
; RV32I-NEXT: addi a1, sp, 15
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV32I-NEXT: lbu a3, 15(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB64_4
|
|
; RV32I-NEXT: .LBB64_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: andi a0, a3, 255
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bgeu s2, a0, .LBB64_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB64_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB64_1
|
|
; RV32I-NEXT: .LBB64_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umin_i8_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: li a3, 255
|
|
; RV32IA-NEXT: sll a3, a3, a0
|
|
; RV32IA-NEXT: andi a1, a1, 255
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB64_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a4, (a2)
|
|
; RV32IA-NEXT: and a6, a4, a3
|
|
; RV32IA-NEXT: mv a5, a4
|
|
; RV32IA-NEXT: bgeu a1, a6, .LBB64_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB64_1 Depth=1
|
|
; RV32IA-NEXT: xor a5, a4, a1
|
|
; RV32IA-NEXT: and a5, a5, a3
|
|
; RV32IA-NEXT: xor a5, a4, a5
|
|
; RV32IA-NEXT: .LBB64_3: # in Loop: Header=BB64_1 Depth=1
|
|
; RV32IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB64_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a4, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i8_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lbu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: andi s2, a1, 255
|
|
; RV64I-NEXT: j .LBB64_2
|
|
; RV64I-NEXT: .LBB64_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB64_2 Depth=1
|
|
; RV64I-NEXT: sb a3, 15(sp)
|
|
; RV64I-NEXT: addi a1, sp, 15
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_1@plt
|
|
; RV64I-NEXT: lbu a3, 15(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB64_4
|
|
; RV64I-NEXT: .LBB64_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: andi a0, a3, 255
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bgeu s2, a0, .LBB64_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB64_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB64_1
|
|
; RV64I-NEXT: .LBB64_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_umin_i8_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: li a3, 255
|
|
; RV64IA-NEXT: sllw a3, a3, a0
|
|
; RV64IA-NEXT: andi a1, a1, 255
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB64_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a4, (a2)
|
|
; RV64IA-NEXT: and a6, a4, a3
|
|
; RV64IA-NEXT: mv a5, a4
|
|
; RV64IA-NEXT: bgeu a1, a6, .LBB64_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB64_1 Depth=1
|
|
; RV64IA-NEXT: xor a5, a4, a1
|
|
; RV64IA-NEXT: and a5, a5, a3
|
|
; RV64IA-NEXT: xor a5, a4, a5
|
|
; RV64IA-NEXT: .LBB64_3: # in Loop: Header=BB64_1 Depth=1
|
|
; RV64IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB64_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a4, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i8 %b seq_cst
|
|
ret i8 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xchg_i16_monotonic(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i16_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_exchange_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xchg_i16_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: lui a3, 16
|
|
; RV32IA-NEXT: addi a3, a3, -1
|
|
; RV32IA-NEXT: sll a4, a3, a0
|
|
; RV32IA-NEXT: and a1, a1, a3
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB65_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a3, (a2)
|
|
; RV32IA-NEXT: mv a5, a1
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: and a5, a5, a4
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB65_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i16_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_exchange_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_xchg_i16_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: lui a3, 16
|
|
; RV64IA-NEXT: addi a3, a3, -1
|
|
; RV64IA-NEXT: sllw a4, a3, a0
|
|
; RV64IA-NEXT: and a1, a1, a3
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB65_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a3, (a2)
|
|
; RV64IA-NEXT: mv a5, a1
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: and a5, a5, a4
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB65_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i16 %b monotonic
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xchg_i16_acquire(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i16_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_exchange_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_i16_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB66_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: mv a5, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB66_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_i16_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB66_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: mv a5, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB66_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i16_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_exchange_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_i16_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB66_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: mv a5, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB66_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_i16_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB66_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: mv a5, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB66_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i16 %b acquire
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xchg_i16_release(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i16_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_exchange_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_i16_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB67_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-WMO-NEXT: mv a5, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB67_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_i16_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB67_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: mv a5, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB67_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i16_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_exchange_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_i16_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB67_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-WMO-NEXT: mv a5, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB67_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_i16_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB67_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: mv a5, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB67_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i16 %b release
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xchg_i16_acq_rel(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i16_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_exchange_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_i16_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB68_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: mv a5, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB68_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_i16_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB68_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: mv a5, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB68_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i16_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_exchange_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_i16_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB68_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: mv a5, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB68_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_i16_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB68_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: mv a5, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB68_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i16 %b acq_rel
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xchg_i16_seq_cst(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i16_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_exchange_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xchg_i16_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: lui a3, 16
|
|
; RV32IA-NEXT: addi a3, a3, -1
|
|
; RV32IA-NEXT: sll a4, a3, a0
|
|
; RV32IA-NEXT: and a1, a1, a3
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB69_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV32IA-NEXT: mv a5, a1
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: and a5, a5, a4
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB69_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i16_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_exchange_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_xchg_i16_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: lui a3, 16
|
|
; RV64IA-NEXT: addi a3, a3, -1
|
|
; RV64IA-NEXT: sllw a4, a3, a0
|
|
; RV64IA-NEXT: and a1, a1, a3
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB69_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV64IA-NEXT: mv a5, a1
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: and a5, a5, a4
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB69_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i16 %b seq_cst
|
|
ret i16 %1
|
|
}
|
|
|
|
; Ensure the following 'atomicrmw xchg a, {0,-1}` cases are lowered to an
|
|
; amoand or amoor with appropriate mask.
|
|
|
|
define i16 @atomicrmw_xchg_0_i16_monotonic(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_0_i16_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a1, 0
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_exchange_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xchg_0_i16_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a1, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: lui a2, 16
|
|
; RV32IA-NEXT: addi a2, a2, -1
|
|
; RV32IA-NEXT: sll a2, a2, a0
|
|
; RV32IA-NEXT: not a2, a2
|
|
; RV32IA-NEXT: amoand.w a1, a2, (a1)
|
|
; RV32IA-NEXT: srl a0, a1, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_0_i16_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a1, 0
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_exchange_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_xchg_0_i16_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a1, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: lui a2, 16
|
|
; RV64IA-NEXT: addi a2, a2, -1
|
|
; RV64IA-NEXT: sllw a2, a2, a0
|
|
; RV64IA-NEXT: not a2, a2
|
|
; RV64IA-NEXT: amoand.w a1, a2, (a1)
|
|
; RV64IA-NEXT: srlw a0, a1, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i16 0 monotonic
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xchg_0_i16_acquire(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_0_i16_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: li a1, 0
|
|
; RV32I-NEXT: call __atomic_exchange_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_0_i16_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a2, 16
|
|
; RV32IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: not a2, a2
|
|
; RV32IA-WMO-NEXT: amoand.w.aq a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_0_i16_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a2, 16
|
|
; RV32IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: not a2, a2
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_0_i16_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: li a1, 0
|
|
; RV64I-NEXT: call __atomic_exchange_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_0_i16_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a2, 16
|
|
; RV64IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: not a2, a2
|
|
; RV64IA-WMO-NEXT: amoand.w.aq a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_0_i16_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a2, 16
|
|
; RV64IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: not a2, a2
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i16 0 acquire
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xchg_0_i16_release(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_0_i16_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: li a1, 0
|
|
; RV32I-NEXT: call __atomic_exchange_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_0_i16_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a2, 16
|
|
; RV32IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: not a2, a2
|
|
; RV32IA-WMO-NEXT: amoand.w.rl a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_0_i16_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a2, 16
|
|
; RV32IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: not a2, a2
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_0_i16_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: li a1, 0
|
|
; RV64I-NEXT: call __atomic_exchange_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_0_i16_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a2, 16
|
|
; RV64IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: not a2, a2
|
|
; RV64IA-WMO-NEXT: amoand.w.rl a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_0_i16_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a2, 16
|
|
; RV64IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: not a2, a2
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i16 0 release
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xchg_0_i16_acq_rel(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_0_i16_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: li a1, 0
|
|
; RV32I-NEXT: call __atomic_exchange_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_0_i16_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a2, 16
|
|
; RV32IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: not a2, a2
|
|
; RV32IA-WMO-NEXT: amoand.w.aqrl a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_0_i16_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a2, 16
|
|
; RV32IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: not a2, a2
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_0_i16_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: li a1, 0
|
|
; RV64I-NEXT: call __atomic_exchange_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_0_i16_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a2, 16
|
|
; RV64IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: not a2, a2
|
|
; RV64IA-WMO-NEXT: amoand.w.aqrl a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_0_i16_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a2, 16
|
|
; RV64IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: not a2, a2
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i16 0 acq_rel
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xchg_0_i16_seq_cst(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_0_i16_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: li a1, 0
|
|
; RV32I-NEXT: call __atomic_exchange_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_0_i16_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a2, 16
|
|
; RV32IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: not a2, a2
|
|
; RV32IA-WMO-NEXT: amoand.w.aqrl a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_0_i16_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a2, 16
|
|
; RV32IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: not a2, a2
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_0_i16_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: li a1, 0
|
|
; RV64I-NEXT: call __atomic_exchange_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_0_i16_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a2, 16
|
|
; RV64IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: not a2, a2
|
|
; RV64IA-WMO-NEXT: amoand.w.aqrl a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_0_i16_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a2, 16
|
|
; RV64IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: not a2, a2
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i16 0 seq_cst
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xchg_minus_1_i16_monotonic(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_minus_1_i16_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: lui a1, 16
|
|
; RV32I-NEXT: addi a1, a1, -1
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_exchange_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xchg_minus_1_i16_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a1, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: lui a2, 16
|
|
; RV32IA-NEXT: addi a2, a2, -1
|
|
; RV32IA-NEXT: sll a2, a2, a0
|
|
; RV32IA-NEXT: amoor.w a1, a2, (a1)
|
|
; RV32IA-NEXT: srl a0, a1, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_minus_1_i16_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: lui a1, 16
|
|
; RV64I-NEXT: addiw a1, a1, -1
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_exchange_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_xchg_minus_1_i16_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a1, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: lui a2, 16
|
|
; RV64IA-NEXT: addi a2, a2, -1
|
|
; RV64IA-NEXT: sllw a2, a2, a0
|
|
; RV64IA-NEXT: amoor.w a1, a2, (a1)
|
|
; RV64IA-NEXT: srlw a0, a1, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i16 -1 monotonic
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xchg_minus_1_i16_acquire(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_minus_1_i16_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: lui a1, 16
|
|
; RV32I-NEXT: addi a1, a1, -1
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_exchange_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_minus_1_i16_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a2, 16
|
|
; RV32IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.aq a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_minus_1_i16_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a2, 16
|
|
; RV32IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_minus_1_i16_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: lui a1, 16
|
|
; RV64I-NEXT: addiw a1, a1, -1
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_exchange_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_minus_1_i16_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a2, 16
|
|
; RV64IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.aq a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_minus_1_i16_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a2, 16
|
|
; RV64IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i16 -1 acquire
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xchg_minus_1_i16_release(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_minus_1_i16_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: lui a1, 16
|
|
; RV32I-NEXT: addi a1, a1, -1
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_exchange_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_minus_1_i16_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a2, 16
|
|
; RV32IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.rl a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_minus_1_i16_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a2, 16
|
|
; RV32IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_minus_1_i16_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: lui a1, 16
|
|
; RV64I-NEXT: addiw a1, a1, -1
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_exchange_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_minus_1_i16_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a2, 16
|
|
; RV64IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.rl a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_minus_1_i16_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a2, 16
|
|
; RV64IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i16 -1 release
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xchg_minus_1_i16_acq_rel(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_minus_1_i16_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: lui a1, 16
|
|
; RV32I-NEXT: addi a1, a1, -1
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_exchange_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_minus_1_i16_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a2, 16
|
|
; RV32IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.aqrl a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_minus_1_i16_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a2, 16
|
|
; RV32IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_minus_1_i16_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: lui a1, 16
|
|
; RV64I-NEXT: addiw a1, a1, -1
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_exchange_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_minus_1_i16_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a2, 16
|
|
; RV64IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.aqrl a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_minus_1_i16_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a2, 16
|
|
; RV64IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i16 -1 acq_rel
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xchg_minus_1_i16_seq_cst(ptr %a) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_minus_1_i16_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: lui a1, 16
|
|
; RV32I-NEXT: addi a1, a1, -1
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_exchange_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_minus_1_i16_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a2, 16
|
|
; RV32IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV32IA-WMO-NEXT: sll a2, a2, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.aqrl a1, a2, (a1)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_minus_1_i16_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a2, 16
|
|
; RV32IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV32IA-TSO-NEXT: sll a2, a2, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_minus_1_i16_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: lui a1, 16
|
|
; RV64I-NEXT: addiw a1, a1, -1
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_exchange_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_minus_1_i16_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a1, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a2, 16
|
|
; RV64IA-WMO-NEXT: addi a2, a2, -1
|
|
; RV64IA-WMO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.aqrl a1, a2, (a1)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_minus_1_i16_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a1, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a2, 16
|
|
; RV64IA-TSO-NEXT: addi a2, a2, -1
|
|
; RV64IA-TSO-NEXT: sllw a2, a2, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a2, (a1)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i16 -1 seq_cst
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_add_i16_monotonic(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i16_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_add_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_add_i16_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: lui a3, 16
|
|
; RV32IA-NEXT: addi a3, a3, -1
|
|
; RV32IA-NEXT: sll a4, a3, a0
|
|
; RV32IA-NEXT: and a1, a1, a3
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB80_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a3, (a2)
|
|
; RV32IA-NEXT: add a5, a3, a1
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: and a5, a5, a4
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB80_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i16_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_add_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_add_i16_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: lui a3, 16
|
|
; RV64IA-NEXT: addi a3, a3, -1
|
|
; RV64IA-NEXT: sllw a4, a3, a0
|
|
; RV64IA-NEXT: and a1, a1, a3
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB80_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a3, (a2)
|
|
; RV64IA-NEXT: add a5, a3, a1
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: and a5, a5, a4
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB80_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i16 %b monotonic
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_add_i16_acquire(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i16_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_add_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_add_i16_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB81_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: add a5, a3, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB81_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_add_i16_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB81_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: add a5, a3, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB81_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i16_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_add_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_add_i16_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB81_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: add a5, a3, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB81_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_add_i16_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB81_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: add a5, a3, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB81_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i16 %b acquire
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_add_i16_release(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i16_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_add_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_add_i16_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB82_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-WMO-NEXT: add a5, a3, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB82_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_add_i16_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB82_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: add a5, a3, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB82_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i16_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_add_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_add_i16_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB82_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-WMO-NEXT: add a5, a3, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB82_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_add_i16_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB82_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: add a5, a3, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB82_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i16 %b release
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_add_i16_acq_rel(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i16_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_add_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_add_i16_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB83_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: add a5, a3, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB83_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_add_i16_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB83_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: add a5, a3, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB83_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i16_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_add_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_add_i16_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB83_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: add a5, a3, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB83_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_add_i16_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB83_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: add a5, a3, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB83_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i16 %b acq_rel
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_add_i16_seq_cst(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i16_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_add_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_add_i16_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: lui a3, 16
|
|
; RV32IA-NEXT: addi a3, a3, -1
|
|
; RV32IA-NEXT: sll a4, a3, a0
|
|
; RV32IA-NEXT: and a1, a1, a3
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB84_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV32IA-NEXT: add a5, a3, a1
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: and a5, a5, a4
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB84_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i16_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_add_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_add_i16_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: lui a3, 16
|
|
; RV64IA-NEXT: addi a3, a3, -1
|
|
; RV64IA-NEXT: sllw a4, a3, a0
|
|
; RV64IA-NEXT: and a1, a1, a3
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB84_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV64IA-NEXT: add a5, a3, a1
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: and a5, a5, a4
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB84_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i16 %b seq_cst
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_sub_i16_monotonic(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i16_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_sub_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_sub_i16_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: lui a3, 16
|
|
; RV32IA-NEXT: addi a3, a3, -1
|
|
; RV32IA-NEXT: sll a4, a3, a0
|
|
; RV32IA-NEXT: and a1, a1, a3
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB85_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a3, (a2)
|
|
; RV32IA-NEXT: sub a5, a3, a1
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: and a5, a5, a4
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB85_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i16_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_sub_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_sub_i16_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: lui a3, 16
|
|
; RV64IA-NEXT: addi a3, a3, -1
|
|
; RV64IA-NEXT: sllw a4, a3, a0
|
|
; RV64IA-NEXT: and a1, a1, a3
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB85_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a3, (a2)
|
|
; RV64IA-NEXT: sub a5, a3, a1
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: and a5, a5, a4
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB85_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i16 %b monotonic
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_sub_i16_acquire(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i16_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_sub_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_sub_i16_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB86_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: sub a5, a3, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB86_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_sub_i16_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB86_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: sub a5, a3, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB86_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i16_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_sub_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_sub_i16_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB86_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: sub a5, a3, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB86_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_sub_i16_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB86_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: sub a5, a3, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB86_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i16 %b acquire
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_sub_i16_release(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i16_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_sub_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_sub_i16_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB87_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-WMO-NEXT: sub a5, a3, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB87_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_sub_i16_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB87_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: sub a5, a3, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB87_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i16_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_sub_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_sub_i16_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB87_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-WMO-NEXT: sub a5, a3, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB87_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_sub_i16_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB87_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: sub a5, a3, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB87_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i16 %b release
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_sub_i16_acq_rel(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i16_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_sub_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_sub_i16_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB88_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: sub a5, a3, a1
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB88_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_sub_i16_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB88_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: sub a5, a3, a1
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB88_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i16_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_sub_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_sub_i16_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB88_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: sub a5, a3, a1
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB88_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_sub_i16_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB88_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: sub a5, a3, a1
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB88_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i16 %b acq_rel
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_sub_i16_seq_cst(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i16_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_sub_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_sub_i16_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: lui a3, 16
|
|
; RV32IA-NEXT: addi a3, a3, -1
|
|
; RV32IA-NEXT: sll a4, a3, a0
|
|
; RV32IA-NEXT: and a1, a1, a3
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB89_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV32IA-NEXT: sub a5, a3, a1
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: and a5, a5, a4
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB89_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i16_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_sub_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_sub_i16_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: lui a3, 16
|
|
; RV64IA-NEXT: addi a3, a3, -1
|
|
; RV64IA-NEXT: sllw a4, a3, a0
|
|
; RV64IA-NEXT: and a1, a1, a3
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB89_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV64IA-NEXT: sub a5, a3, a1
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: and a5, a5, a4
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB89_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i16 %b seq_cst
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_and_i16_monotonic(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i16_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_and_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_and_i16_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: lui a3, 16
|
|
; RV32IA-NEXT: addi a3, a3, -1
|
|
; RV32IA-NEXT: sll a4, a3, a0
|
|
; RV32IA-NEXT: not a4, a4
|
|
; RV32IA-NEXT: and a1, a1, a3
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: or a1, a4, a1
|
|
; RV32IA-NEXT: amoand.w a1, a1, (a2)
|
|
; RV32IA-NEXT: srl a0, a1, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i16_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_and_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_and_i16_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: lui a3, 16
|
|
; RV64IA-NEXT: addi a3, a3, -1
|
|
; RV64IA-NEXT: sllw a4, a3, a0
|
|
; RV64IA-NEXT: not a4, a4
|
|
; RV64IA-NEXT: and a1, a1, a3
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: or a1, a4, a1
|
|
; RV64IA-NEXT: amoand.w a1, a1, (a2)
|
|
; RV64IA-NEXT: srlw a0, a1, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i16 %b monotonic
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_and_i16_acquire(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i16_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_and_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_and_i16_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: not a4, a4
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: or a1, a4, a1
|
|
; RV32IA-WMO-NEXT: amoand.w.aq a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_and_i16_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: not a4, a4
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: or a1, a4, a1
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i16_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_and_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i16_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: not a4, a4
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: or a1, a4, a1
|
|
; RV64IA-WMO-NEXT: amoand.w.aq a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i16_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: not a4, a4
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: or a1, a4, a1
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i16 %b acquire
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_and_i16_release(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i16_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_and_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_and_i16_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: not a4, a4
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: or a1, a4, a1
|
|
; RV32IA-WMO-NEXT: amoand.w.rl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_and_i16_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: not a4, a4
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: or a1, a4, a1
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i16_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_and_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i16_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: not a4, a4
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: or a1, a4, a1
|
|
; RV64IA-WMO-NEXT: amoand.w.rl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i16_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: not a4, a4
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: or a1, a4, a1
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i16 %b release
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_and_i16_acq_rel(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i16_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_and_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_and_i16_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: not a4, a4
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: or a1, a4, a1
|
|
; RV32IA-WMO-NEXT: amoand.w.aqrl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_and_i16_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: not a4, a4
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: or a1, a4, a1
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i16_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_and_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i16_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: not a4, a4
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: or a1, a4, a1
|
|
; RV64IA-WMO-NEXT: amoand.w.aqrl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i16_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: not a4, a4
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: or a1, a4, a1
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i16 %b acq_rel
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_and_i16_seq_cst(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i16_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_and_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_and_i16_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: not a4, a4
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: or a1, a4, a1
|
|
; RV32IA-WMO-NEXT: amoand.w.aqrl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_and_i16_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: not a4, a4
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: or a1, a4, a1
|
|
; RV32IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i16_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_and_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i16_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: not a4, a4
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: or a1, a4, a1
|
|
; RV64IA-WMO-NEXT: amoand.w.aqrl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i16_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: not a4, a4
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: or a1, a4, a1
|
|
; RV64IA-TSO-NEXT: amoand.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i16 %b seq_cst
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_nand_i16_monotonic(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i16_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_nand_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_nand_i16_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: lui a3, 16
|
|
; RV32IA-NEXT: addi a3, a3, -1
|
|
; RV32IA-NEXT: sll a4, a3, a0
|
|
; RV32IA-NEXT: and a1, a1, a3
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB95_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a3, (a2)
|
|
; RV32IA-NEXT: and a5, a3, a1
|
|
; RV32IA-NEXT: not a5, a5
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: and a5, a5, a4
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB95_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i16_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_nand_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_nand_i16_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: lui a3, 16
|
|
; RV64IA-NEXT: addi a3, a3, -1
|
|
; RV64IA-NEXT: sllw a4, a3, a0
|
|
; RV64IA-NEXT: and a1, a1, a3
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB95_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a3, (a2)
|
|
; RV64IA-NEXT: and a5, a3, a1
|
|
; RV64IA-NEXT: not a5, a5
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: and a5, a5, a4
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB95_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i16 %b monotonic
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_nand_i16_acquire(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i16_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_nand_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_nand_i16_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB96_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: and a5, a3, a1
|
|
; RV32IA-WMO-NEXT: not a5, a5
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB96_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_nand_i16_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB96_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: and a5, a3, a1
|
|
; RV32IA-TSO-NEXT: not a5, a5
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB96_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i16_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_nand_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_nand_i16_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB96_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: and a5, a3, a1
|
|
; RV64IA-WMO-NEXT: not a5, a5
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB96_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_nand_i16_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB96_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: and a5, a3, a1
|
|
; RV64IA-TSO-NEXT: not a5, a5
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB96_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i16 %b acquire
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_nand_i16_release(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i16_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_nand_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_nand_i16_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB97_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-WMO-NEXT: and a5, a3, a1
|
|
; RV32IA-WMO-NEXT: not a5, a5
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB97_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_nand_i16_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB97_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: and a5, a3, a1
|
|
; RV32IA-TSO-NEXT: not a5, a5
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB97_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i16_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_nand_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_nand_i16_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB97_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-WMO-NEXT: and a5, a3, a1
|
|
; RV64IA-WMO-NEXT: not a5, a5
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB97_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_nand_i16_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB97_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: and a5, a3, a1
|
|
; RV64IA-TSO-NEXT: not a5, a5
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB97_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i16 %b release
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_nand_i16_acq_rel(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i16_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_nand_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_nand_i16_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB98_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: and a5, a3, a1
|
|
; RV32IA-WMO-NEXT: not a5, a5
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB98_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_nand_i16_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB98_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: and a5, a3, a1
|
|
; RV32IA-TSO-NEXT: not a5, a5
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB98_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i16_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_nand_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_nand_i16_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB98_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: and a5, a3, a1
|
|
; RV64IA-WMO-NEXT: not a5, a5
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB98_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_nand_i16_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB98_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: and a5, a3, a1
|
|
; RV64IA-TSO-NEXT: not a5, a5
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB98_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i16 %b acq_rel
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_nand_i16_seq_cst(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i16_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_nand_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_nand_i16_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: lui a3, 16
|
|
; RV32IA-NEXT: addi a3, a3, -1
|
|
; RV32IA-NEXT: sll a4, a3, a0
|
|
; RV32IA-NEXT: and a1, a1, a3
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB99_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV32IA-NEXT: and a5, a3, a1
|
|
; RV32IA-NEXT: not a5, a5
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: and a5, a5, a4
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB99_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i16_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_nand_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_nand_i16_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: lui a3, 16
|
|
; RV64IA-NEXT: addi a3, a3, -1
|
|
; RV64IA-NEXT: sllw a4, a3, a0
|
|
; RV64IA-NEXT: and a1, a1, a3
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB99_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV64IA-NEXT: and a5, a3, a1
|
|
; RV64IA-NEXT: not a5, a5
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: and a5, a5, a4
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB99_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i16 %b seq_cst
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_or_i16_monotonic(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i16_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_or_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_or_i16_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: slli a1, a1, 16
|
|
; RV32IA-NEXT: srli a1, a1, 16
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: amoor.w a1, a1, (a2)
|
|
; RV32IA-NEXT: srl a0, a1, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i16_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_or_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_or_i16_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: slli a1, a1, 48
|
|
; RV64IA-NEXT: srli a1, a1, 48
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: amoor.w a1, a1, (a2)
|
|
; RV64IA-NEXT: srlw a0, a1, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i16 %b monotonic
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_or_i16_acquire(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i16_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_or_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_or_i16_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: srli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.aq a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_or_i16_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: srli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i16_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_or_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i16_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: srli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.aq a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i16_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: srli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i16 %b acquire
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_or_i16_release(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i16_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_or_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_or_i16_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: srli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.rl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_or_i16_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: srli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i16_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_or_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i16_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: srli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.rl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i16_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: srli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i16 %b release
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_or_i16_acq_rel(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i16_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_or_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_or_i16_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: srli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.aqrl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_or_i16_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: srli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i16_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_or_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i16_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: srli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.aqrl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i16_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: srli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i16 %b acq_rel
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_or_i16_seq_cst(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i16_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_or_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_or_i16_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: srli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoor.w.aqrl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_or_i16_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: srli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i16_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_or_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i16_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: srli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoor.w.aqrl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i16_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: srli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i16 %b seq_cst
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xor_i16_monotonic(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i16_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_xor_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xor_i16_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: slli a1, a1, 16
|
|
; RV32IA-NEXT: srli a1, a1, 16
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV32IA-NEXT: srl a0, a1, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i16_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_xor_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_xor_i16_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: slli a1, a1, 48
|
|
; RV64IA-NEXT: srli a1, a1, 48
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV64IA-NEXT: srlw a0, a1, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i16 %b monotonic
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xor_i16_acquire(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i16_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_xor_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xor_i16_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: srli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoxor.w.aq a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xor_i16_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: srli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i16_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_xor_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i16_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: srli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoxor.w.aq a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i16_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: srli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i16 %b acquire
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xor_i16_release(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i16_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_xor_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xor_i16_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: srli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoxor.w.rl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xor_i16_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: srli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i16_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_xor_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i16_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: srli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoxor.w.rl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i16_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: srli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i16 %b release
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xor_i16_acq_rel(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i16_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_xor_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xor_i16_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: srli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoxor.w.aqrl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xor_i16_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: srli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i16_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_xor_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i16_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: srli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoxor.w.aqrl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i16_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: srli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i16 %b acq_rel
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_xor_i16_seq_cst(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i16_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_xor_2@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xor_i16_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: srli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: amoxor.w.aqrl a1, a1, (a2)
|
|
; RV32IA-WMO-NEXT: srl a0, a1, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xor_i16_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: srli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV32IA-TSO-NEXT: srl a0, a1, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i16_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_xor_2@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i16_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: srli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: amoxor.w.aqrl a1, a1, (a2)
|
|
; RV64IA-WMO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i16_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: srli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: amoxor.w a1, a1, (a2)
|
|
; RV64IA-TSO-NEXT: srlw a0, a1, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i16 %b seq_cst
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_max_i16_monotonic(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i16_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lhu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 16
|
|
; RV32I-NEXT: srai s2, a0, 16
|
|
; RV32I-NEXT: j .LBB110_2
|
|
; RV32I-NEXT: .LBB110_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB110_2 Depth=1
|
|
; RV32I-NEXT: sh a3, 14(sp)
|
|
; RV32I-NEXT: addi a1, sp, 14
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a3, 14(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB110_4
|
|
; RV32I-NEXT: .LBB110_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 16
|
|
; RV32I-NEXT: srai a0, a0, 16
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: blt s2, a0, .LBB110_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB110_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB110_1
|
|
; RV32I-NEXT: .LBB110_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_max_i16_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: andi a3, a0, 24
|
|
; RV32IA-NEXT: lui a4, 16
|
|
; RV32IA-NEXT: addi a4, a4, -1
|
|
; RV32IA-NEXT: sll a4, a4, a0
|
|
; RV32IA-NEXT: slli a1, a1, 16
|
|
; RV32IA-NEXT: srai a1, a1, 16
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: li a5, 16
|
|
; RV32IA-NEXT: sub a5, a5, a3
|
|
; RV32IA-NEXT: .LBB110_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a3, (a2)
|
|
; RV32IA-NEXT: and a7, a3, a4
|
|
; RV32IA-NEXT: mv a6, a3
|
|
; RV32IA-NEXT: sll a7, a7, a5
|
|
; RV32IA-NEXT: sra a7, a7, a5
|
|
; RV32IA-NEXT: bge a7, a1, .LBB110_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB110_1 Depth=1
|
|
; RV32IA-NEXT: xor a6, a3, a1
|
|
; RV32IA-NEXT: and a6, a6, a4
|
|
; RV32IA-NEXT: xor a6, a3, a6
|
|
; RV32IA-NEXT: .LBB110_3: # in Loop: Header=BB110_1 Depth=1
|
|
; RV32IA-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-NEXT: bnez a6, .LBB110_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i16_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lhu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 48
|
|
; RV64I-NEXT: srai s2, a0, 48
|
|
; RV64I-NEXT: j .LBB110_2
|
|
; RV64I-NEXT: .LBB110_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB110_2 Depth=1
|
|
; RV64I-NEXT: sh a3, 14(sp)
|
|
; RV64I-NEXT: addi a1, sp, 14
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a3, 14(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB110_4
|
|
; RV64I-NEXT: .LBB110_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 48
|
|
; RV64I-NEXT: srai a0, a0, 48
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s2, a0, .LBB110_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB110_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB110_1
|
|
; RV64I-NEXT: .LBB110_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_max_i16_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: andi a3, a0, 24
|
|
; RV64IA-NEXT: lui a4, 16
|
|
; RV64IA-NEXT: addi a4, a4, -1
|
|
; RV64IA-NEXT: sllw a4, a4, a0
|
|
; RV64IA-NEXT: slli a1, a1, 48
|
|
; RV64IA-NEXT: srai a1, a1, 48
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: li a5, 48
|
|
; RV64IA-NEXT: sub a5, a5, a3
|
|
; RV64IA-NEXT: .LBB110_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a3, (a2)
|
|
; RV64IA-NEXT: and a7, a3, a4
|
|
; RV64IA-NEXT: mv a6, a3
|
|
; RV64IA-NEXT: sll a7, a7, a5
|
|
; RV64IA-NEXT: sra a7, a7, a5
|
|
; RV64IA-NEXT: bge a7, a1, .LBB110_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB110_1 Depth=1
|
|
; RV64IA-NEXT: xor a6, a3, a1
|
|
; RV64IA-NEXT: and a6, a6, a4
|
|
; RV64IA-NEXT: xor a6, a3, a6
|
|
; RV64IA-NEXT: .LBB110_3: # in Loop: Header=BB110_1 Depth=1
|
|
; RV64IA-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-NEXT: bnez a6, .LBB110_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i16 %b monotonic
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_max_i16_acquire(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i16_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lhu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 16
|
|
; RV32I-NEXT: srai s2, a0, 16
|
|
; RV32I-NEXT: j .LBB111_2
|
|
; RV32I-NEXT: .LBB111_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB111_2 Depth=1
|
|
; RV32I-NEXT: sh a3, 14(sp)
|
|
; RV32I-NEXT: addi a1, sp, 14
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a3, 14(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB111_4
|
|
; RV32I-NEXT: .LBB111_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 16
|
|
; RV32I-NEXT: srai a0, a0, 16
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: blt s2, a0, .LBB111_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB111_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB111_1
|
|
; RV32I-NEXT: .LBB111_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_max_i16_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV32IA-WMO-NEXT: lui a4, 16
|
|
; RV32IA-WMO-NEXT: addi a4, a4, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a4, a0
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: srai a1, a1, 16
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: li a5, 16
|
|
; RV32IA-WMO-NEXT: sub a5, a5, a3
|
|
; RV32IA-WMO-NEXT: .LBB111_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: and a7, a3, a4
|
|
; RV32IA-WMO-NEXT: mv a6, a3
|
|
; RV32IA-WMO-NEXT: sll a7, a7, a5
|
|
; RV32IA-WMO-NEXT: sra a7, a7, a5
|
|
; RV32IA-WMO-NEXT: bge a7, a1, .LBB111_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB111_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a6, a3, a1
|
|
; RV32IA-WMO-NEXT: and a6, a6, a4
|
|
; RV32IA-WMO-NEXT: xor a6, a3, a6
|
|
; RV32IA-WMO-NEXT: .LBB111_3: # in Loop: Header=BB111_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a6, .LBB111_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_max_i16_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV32IA-TSO-NEXT: lui a4, 16
|
|
; RV32IA-TSO-NEXT: addi a4, a4, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a4, a0
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: srai a1, a1, 16
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: li a5, 16
|
|
; RV32IA-TSO-NEXT: sub a5, a5, a3
|
|
; RV32IA-TSO-NEXT: .LBB111_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: and a7, a3, a4
|
|
; RV32IA-TSO-NEXT: mv a6, a3
|
|
; RV32IA-TSO-NEXT: sll a7, a7, a5
|
|
; RV32IA-TSO-NEXT: sra a7, a7, a5
|
|
; RV32IA-TSO-NEXT: bge a7, a1, .LBB111_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB111_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a6, a3, a1
|
|
; RV32IA-TSO-NEXT: and a6, a6, a4
|
|
; RV32IA-TSO-NEXT: xor a6, a3, a6
|
|
; RV32IA-TSO-NEXT: .LBB111_3: # in Loop: Header=BB111_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a6, .LBB111_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i16_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lhu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 48
|
|
; RV64I-NEXT: srai s2, a0, 48
|
|
; RV64I-NEXT: j .LBB111_2
|
|
; RV64I-NEXT: .LBB111_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB111_2 Depth=1
|
|
; RV64I-NEXT: sh a3, 14(sp)
|
|
; RV64I-NEXT: addi a1, sp, 14
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a3, 14(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB111_4
|
|
; RV64I-NEXT: .LBB111_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 48
|
|
; RV64I-NEXT: srai a0, a0, 48
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s2, a0, .LBB111_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB111_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB111_1
|
|
; RV64I-NEXT: .LBB111_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_max_i16_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV64IA-WMO-NEXT: lui a4, 16
|
|
; RV64IA-WMO-NEXT: addi a4, a4, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: srai a1, a1, 48
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: li a5, 48
|
|
; RV64IA-WMO-NEXT: sub a5, a5, a3
|
|
; RV64IA-WMO-NEXT: .LBB111_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: and a7, a3, a4
|
|
; RV64IA-WMO-NEXT: mv a6, a3
|
|
; RV64IA-WMO-NEXT: sll a7, a7, a5
|
|
; RV64IA-WMO-NEXT: sra a7, a7, a5
|
|
; RV64IA-WMO-NEXT: bge a7, a1, .LBB111_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB111_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a6, a3, a1
|
|
; RV64IA-WMO-NEXT: and a6, a6, a4
|
|
; RV64IA-WMO-NEXT: xor a6, a3, a6
|
|
; RV64IA-WMO-NEXT: .LBB111_3: # in Loop: Header=BB111_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a6, .LBB111_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_max_i16_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV64IA-TSO-NEXT: lui a4, 16
|
|
; RV64IA-TSO-NEXT: addi a4, a4, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: srai a1, a1, 48
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: li a5, 48
|
|
; RV64IA-TSO-NEXT: sub a5, a5, a3
|
|
; RV64IA-TSO-NEXT: .LBB111_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: and a7, a3, a4
|
|
; RV64IA-TSO-NEXT: mv a6, a3
|
|
; RV64IA-TSO-NEXT: sll a7, a7, a5
|
|
; RV64IA-TSO-NEXT: sra a7, a7, a5
|
|
; RV64IA-TSO-NEXT: bge a7, a1, .LBB111_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB111_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a6, a3, a1
|
|
; RV64IA-TSO-NEXT: and a6, a6, a4
|
|
; RV64IA-TSO-NEXT: xor a6, a3, a6
|
|
; RV64IA-TSO-NEXT: .LBB111_3: # in Loop: Header=BB111_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a6, .LBB111_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i16 %b acquire
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_max_i16_release(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i16_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lhu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 16
|
|
; RV32I-NEXT: srai s2, a0, 16
|
|
; RV32I-NEXT: j .LBB112_2
|
|
; RV32I-NEXT: .LBB112_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB112_2 Depth=1
|
|
; RV32I-NEXT: sh a3, 14(sp)
|
|
; RV32I-NEXT: addi a1, sp, 14
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a3, 14(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB112_4
|
|
; RV32I-NEXT: .LBB112_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 16
|
|
; RV32I-NEXT: srai a0, a0, 16
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: blt s2, a0, .LBB112_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB112_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB112_1
|
|
; RV32I-NEXT: .LBB112_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_max_i16_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV32IA-WMO-NEXT: lui a4, 16
|
|
; RV32IA-WMO-NEXT: addi a4, a4, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a4, a0
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: srai a1, a1, 16
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: li a5, 16
|
|
; RV32IA-WMO-NEXT: sub a5, a5, a3
|
|
; RV32IA-WMO-NEXT: .LBB112_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-WMO-NEXT: and a7, a3, a4
|
|
; RV32IA-WMO-NEXT: mv a6, a3
|
|
; RV32IA-WMO-NEXT: sll a7, a7, a5
|
|
; RV32IA-WMO-NEXT: sra a7, a7, a5
|
|
; RV32IA-WMO-NEXT: bge a7, a1, .LBB112_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB112_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a6, a3, a1
|
|
; RV32IA-WMO-NEXT: and a6, a6, a4
|
|
; RV32IA-WMO-NEXT: xor a6, a3, a6
|
|
; RV32IA-WMO-NEXT: .LBB112_3: # in Loop: Header=BB112_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a6, .LBB112_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_max_i16_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV32IA-TSO-NEXT: lui a4, 16
|
|
; RV32IA-TSO-NEXT: addi a4, a4, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a4, a0
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: srai a1, a1, 16
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: li a5, 16
|
|
; RV32IA-TSO-NEXT: sub a5, a5, a3
|
|
; RV32IA-TSO-NEXT: .LBB112_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: and a7, a3, a4
|
|
; RV32IA-TSO-NEXT: mv a6, a3
|
|
; RV32IA-TSO-NEXT: sll a7, a7, a5
|
|
; RV32IA-TSO-NEXT: sra a7, a7, a5
|
|
; RV32IA-TSO-NEXT: bge a7, a1, .LBB112_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB112_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a6, a3, a1
|
|
; RV32IA-TSO-NEXT: and a6, a6, a4
|
|
; RV32IA-TSO-NEXT: xor a6, a3, a6
|
|
; RV32IA-TSO-NEXT: .LBB112_3: # in Loop: Header=BB112_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a6, .LBB112_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i16_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lhu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 48
|
|
; RV64I-NEXT: srai s2, a0, 48
|
|
; RV64I-NEXT: j .LBB112_2
|
|
; RV64I-NEXT: .LBB112_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB112_2 Depth=1
|
|
; RV64I-NEXT: sh a3, 14(sp)
|
|
; RV64I-NEXT: addi a1, sp, 14
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a3, 14(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB112_4
|
|
; RV64I-NEXT: .LBB112_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 48
|
|
; RV64I-NEXT: srai a0, a0, 48
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s2, a0, .LBB112_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB112_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB112_1
|
|
; RV64I-NEXT: .LBB112_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_max_i16_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV64IA-WMO-NEXT: lui a4, 16
|
|
; RV64IA-WMO-NEXT: addi a4, a4, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: srai a1, a1, 48
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: li a5, 48
|
|
; RV64IA-WMO-NEXT: sub a5, a5, a3
|
|
; RV64IA-WMO-NEXT: .LBB112_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-WMO-NEXT: and a7, a3, a4
|
|
; RV64IA-WMO-NEXT: mv a6, a3
|
|
; RV64IA-WMO-NEXT: sll a7, a7, a5
|
|
; RV64IA-WMO-NEXT: sra a7, a7, a5
|
|
; RV64IA-WMO-NEXT: bge a7, a1, .LBB112_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB112_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a6, a3, a1
|
|
; RV64IA-WMO-NEXT: and a6, a6, a4
|
|
; RV64IA-WMO-NEXT: xor a6, a3, a6
|
|
; RV64IA-WMO-NEXT: .LBB112_3: # in Loop: Header=BB112_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a6, .LBB112_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_max_i16_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV64IA-TSO-NEXT: lui a4, 16
|
|
; RV64IA-TSO-NEXT: addi a4, a4, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: srai a1, a1, 48
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: li a5, 48
|
|
; RV64IA-TSO-NEXT: sub a5, a5, a3
|
|
; RV64IA-TSO-NEXT: .LBB112_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: and a7, a3, a4
|
|
; RV64IA-TSO-NEXT: mv a6, a3
|
|
; RV64IA-TSO-NEXT: sll a7, a7, a5
|
|
; RV64IA-TSO-NEXT: sra a7, a7, a5
|
|
; RV64IA-TSO-NEXT: bge a7, a1, .LBB112_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB112_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a6, a3, a1
|
|
; RV64IA-TSO-NEXT: and a6, a6, a4
|
|
; RV64IA-TSO-NEXT: xor a6, a3, a6
|
|
; RV64IA-TSO-NEXT: .LBB112_3: # in Loop: Header=BB112_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a6, .LBB112_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i16 %b release
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_max_i16_acq_rel(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i16_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lhu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 16
|
|
; RV32I-NEXT: srai s2, a0, 16
|
|
; RV32I-NEXT: j .LBB113_2
|
|
; RV32I-NEXT: .LBB113_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB113_2 Depth=1
|
|
; RV32I-NEXT: sh a3, 14(sp)
|
|
; RV32I-NEXT: addi a1, sp, 14
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a3, 14(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB113_4
|
|
; RV32I-NEXT: .LBB113_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 16
|
|
; RV32I-NEXT: srai a0, a0, 16
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: blt s2, a0, .LBB113_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB113_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB113_1
|
|
; RV32I-NEXT: .LBB113_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_max_i16_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV32IA-WMO-NEXT: lui a4, 16
|
|
; RV32IA-WMO-NEXT: addi a4, a4, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a4, a0
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: srai a1, a1, 16
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: li a5, 16
|
|
; RV32IA-WMO-NEXT: sub a5, a5, a3
|
|
; RV32IA-WMO-NEXT: .LBB113_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: and a7, a3, a4
|
|
; RV32IA-WMO-NEXT: mv a6, a3
|
|
; RV32IA-WMO-NEXT: sll a7, a7, a5
|
|
; RV32IA-WMO-NEXT: sra a7, a7, a5
|
|
; RV32IA-WMO-NEXT: bge a7, a1, .LBB113_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB113_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a6, a3, a1
|
|
; RV32IA-WMO-NEXT: and a6, a6, a4
|
|
; RV32IA-WMO-NEXT: xor a6, a3, a6
|
|
; RV32IA-WMO-NEXT: .LBB113_3: # in Loop: Header=BB113_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a6, .LBB113_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_max_i16_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV32IA-TSO-NEXT: lui a4, 16
|
|
; RV32IA-TSO-NEXT: addi a4, a4, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a4, a0
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: srai a1, a1, 16
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: li a5, 16
|
|
; RV32IA-TSO-NEXT: sub a5, a5, a3
|
|
; RV32IA-TSO-NEXT: .LBB113_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: and a7, a3, a4
|
|
; RV32IA-TSO-NEXT: mv a6, a3
|
|
; RV32IA-TSO-NEXT: sll a7, a7, a5
|
|
; RV32IA-TSO-NEXT: sra a7, a7, a5
|
|
; RV32IA-TSO-NEXT: bge a7, a1, .LBB113_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB113_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a6, a3, a1
|
|
; RV32IA-TSO-NEXT: and a6, a6, a4
|
|
; RV32IA-TSO-NEXT: xor a6, a3, a6
|
|
; RV32IA-TSO-NEXT: .LBB113_3: # in Loop: Header=BB113_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a6, .LBB113_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i16_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lhu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 48
|
|
; RV64I-NEXT: srai s2, a0, 48
|
|
; RV64I-NEXT: j .LBB113_2
|
|
; RV64I-NEXT: .LBB113_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB113_2 Depth=1
|
|
; RV64I-NEXT: sh a3, 14(sp)
|
|
; RV64I-NEXT: addi a1, sp, 14
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a3, 14(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB113_4
|
|
; RV64I-NEXT: .LBB113_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 48
|
|
; RV64I-NEXT: srai a0, a0, 48
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s2, a0, .LBB113_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB113_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB113_1
|
|
; RV64I-NEXT: .LBB113_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_max_i16_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV64IA-WMO-NEXT: lui a4, 16
|
|
; RV64IA-WMO-NEXT: addi a4, a4, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: srai a1, a1, 48
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: li a5, 48
|
|
; RV64IA-WMO-NEXT: sub a5, a5, a3
|
|
; RV64IA-WMO-NEXT: .LBB113_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: and a7, a3, a4
|
|
; RV64IA-WMO-NEXT: mv a6, a3
|
|
; RV64IA-WMO-NEXT: sll a7, a7, a5
|
|
; RV64IA-WMO-NEXT: sra a7, a7, a5
|
|
; RV64IA-WMO-NEXT: bge a7, a1, .LBB113_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB113_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a6, a3, a1
|
|
; RV64IA-WMO-NEXT: and a6, a6, a4
|
|
; RV64IA-WMO-NEXT: xor a6, a3, a6
|
|
; RV64IA-WMO-NEXT: .LBB113_3: # in Loop: Header=BB113_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a6, .LBB113_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_max_i16_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV64IA-TSO-NEXT: lui a4, 16
|
|
; RV64IA-TSO-NEXT: addi a4, a4, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: srai a1, a1, 48
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: li a5, 48
|
|
; RV64IA-TSO-NEXT: sub a5, a5, a3
|
|
; RV64IA-TSO-NEXT: .LBB113_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: and a7, a3, a4
|
|
; RV64IA-TSO-NEXT: mv a6, a3
|
|
; RV64IA-TSO-NEXT: sll a7, a7, a5
|
|
; RV64IA-TSO-NEXT: sra a7, a7, a5
|
|
; RV64IA-TSO-NEXT: bge a7, a1, .LBB113_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB113_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a6, a3, a1
|
|
; RV64IA-TSO-NEXT: and a6, a6, a4
|
|
; RV64IA-TSO-NEXT: xor a6, a3, a6
|
|
; RV64IA-TSO-NEXT: .LBB113_3: # in Loop: Header=BB113_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a6, .LBB113_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i16 %b acq_rel
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_max_i16_seq_cst(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i16_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lhu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 16
|
|
; RV32I-NEXT: srai s2, a0, 16
|
|
; RV32I-NEXT: j .LBB114_2
|
|
; RV32I-NEXT: .LBB114_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB114_2 Depth=1
|
|
; RV32I-NEXT: sh a3, 14(sp)
|
|
; RV32I-NEXT: addi a1, sp, 14
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a3, 14(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB114_4
|
|
; RV32I-NEXT: .LBB114_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 16
|
|
; RV32I-NEXT: srai a0, a0, 16
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: blt s2, a0, .LBB114_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB114_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB114_1
|
|
; RV32I-NEXT: .LBB114_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_max_i16_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: andi a3, a0, 24
|
|
; RV32IA-NEXT: lui a4, 16
|
|
; RV32IA-NEXT: addi a4, a4, -1
|
|
; RV32IA-NEXT: sll a4, a4, a0
|
|
; RV32IA-NEXT: slli a1, a1, 16
|
|
; RV32IA-NEXT: srai a1, a1, 16
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: li a5, 16
|
|
; RV32IA-NEXT: sub a5, a5, a3
|
|
; RV32IA-NEXT: .LBB114_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV32IA-NEXT: and a7, a3, a4
|
|
; RV32IA-NEXT: mv a6, a3
|
|
; RV32IA-NEXT: sll a7, a7, a5
|
|
; RV32IA-NEXT: sra a7, a7, a5
|
|
; RV32IA-NEXT: bge a7, a1, .LBB114_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB114_1 Depth=1
|
|
; RV32IA-NEXT: xor a6, a3, a1
|
|
; RV32IA-NEXT: and a6, a6, a4
|
|
; RV32IA-NEXT: xor a6, a3, a6
|
|
; RV32IA-NEXT: .LBB114_3: # in Loop: Header=BB114_1 Depth=1
|
|
; RV32IA-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV32IA-NEXT: bnez a6, .LBB114_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i16_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lhu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 48
|
|
; RV64I-NEXT: srai s2, a0, 48
|
|
; RV64I-NEXT: j .LBB114_2
|
|
; RV64I-NEXT: .LBB114_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB114_2 Depth=1
|
|
; RV64I-NEXT: sh a3, 14(sp)
|
|
; RV64I-NEXT: addi a1, sp, 14
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a3, 14(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB114_4
|
|
; RV64I-NEXT: .LBB114_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 48
|
|
; RV64I-NEXT: srai a0, a0, 48
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s2, a0, .LBB114_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB114_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB114_1
|
|
; RV64I-NEXT: .LBB114_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_max_i16_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: andi a3, a0, 24
|
|
; RV64IA-NEXT: lui a4, 16
|
|
; RV64IA-NEXT: addi a4, a4, -1
|
|
; RV64IA-NEXT: sllw a4, a4, a0
|
|
; RV64IA-NEXT: slli a1, a1, 48
|
|
; RV64IA-NEXT: srai a1, a1, 48
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: li a5, 48
|
|
; RV64IA-NEXT: sub a5, a5, a3
|
|
; RV64IA-NEXT: .LBB114_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV64IA-NEXT: and a7, a3, a4
|
|
; RV64IA-NEXT: mv a6, a3
|
|
; RV64IA-NEXT: sll a7, a7, a5
|
|
; RV64IA-NEXT: sra a7, a7, a5
|
|
; RV64IA-NEXT: bge a7, a1, .LBB114_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB114_1 Depth=1
|
|
; RV64IA-NEXT: xor a6, a3, a1
|
|
; RV64IA-NEXT: and a6, a6, a4
|
|
; RV64IA-NEXT: xor a6, a3, a6
|
|
; RV64IA-NEXT: .LBB114_3: # in Loop: Header=BB114_1 Depth=1
|
|
; RV64IA-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV64IA-NEXT: bnez a6, .LBB114_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i16 %b seq_cst
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_min_i16_monotonic(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i16_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lhu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 16
|
|
; RV32I-NEXT: srai s2, a0, 16
|
|
; RV32I-NEXT: j .LBB115_2
|
|
; RV32I-NEXT: .LBB115_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB115_2 Depth=1
|
|
; RV32I-NEXT: sh a3, 14(sp)
|
|
; RV32I-NEXT: addi a1, sp, 14
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a3, 14(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB115_4
|
|
; RV32I-NEXT: .LBB115_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 16
|
|
; RV32I-NEXT: srai a0, a0, 16
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bge s2, a0, .LBB115_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB115_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB115_1
|
|
; RV32I-NEXT: .LBB115_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_min_i16_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: andi a3, a0, 24
|
|
; RV32IA-NEXT: lui a4, 16
|
|
; RV32IA-NEXT: addi a4, a4, -1
|
|
; RV32IA-NEXT: sll a4, a4, a0
|
|
; RV32IA-NEXT: slli a1, a1, 16
|
|
; RV32IA-NEXT: srai a1, a1, 16
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: li a5, 16
|
|
; RV32IA-NEXT: sub a5, a5, a3
|
|
; RV32IA-NEXT: .LBB115_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a3, (a2)
|
|
; RV32IA-NEXT: and a7, a3, a4
|
|
; RV32IA-NEXT: mv a6, a3
|
|
; RV32IA-NEXT: sll a7, a7, a5
|
|
; RV32IA-NEXT: sra a7, a7, a5
|
|
; RV32IA-NEXT: bge a1, a7, .LBB115_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB115_1 Depth=1
|
|
; RV32IA-NEXT: xor a6, a3, a1
|
|
; RV32IA-NEXT: and a6, a6, a4
|
|
; RV32IA-NEXT: xor a6, a3, a6
|
|
; RV32IA-NEXT: .LBB115_3: # in Loop: Header=BB115_1 Depth=1
|
|
; RV32IA-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-NEXT: bnez a6, .LBB115_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i16_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lhu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 48
|
|
; RV64I-NEXT: srai s2, a0, 48
|
|
; RV64I-NEXT: j .LBB115_2
|
|
; RV64I-NEXT: .LBB115_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB115_2 Depth=1
|
|
; RV64I-NEXT: sh a3, 14(sp)
|
|
; RV64I-NEXT: addi a1, sp, 14
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a3, 14(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB115_4
|
|
; RV64I-NEXT: .LBB115_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 48
|
|
; RV64I-NEXT: srai a0, a0, 48
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s2, a0, .LBB115_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB115_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB115_1
|
|
; RV64I-NEXT: .LBB115_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_min_i16_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: andi a3, a0, 24
|
|
; RV64IA-NEXT: lui a4, 16
|
|
; RV64IA-NEXT: addi a4, a4, -1
|
|
; RV64IA-NEXT: sllw a4, a4, a0
|
|
; RV64IA-NEXT: slli a1, a1, 48
|
|
; RV64IA-NEXT: srai a1, a1, 48
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: li a5, 48
|
|
; RV64IA-NEXT: sub a5, a5, a3
|
|
; RV64IA-NEXT: .LBB115_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a3, (a2)
|
|
; RV64IA-NEXT: and a7, a3, a4
|
|
; RV64IA-NEXT: mv a6, a3
|
|
; RV64IA-NEXT: sll a7, a7, a5
|
|
; RV64IA-NEXT: sra a7, a7, a5
|
|
; RV64IA-NEXT: bge a1, a7, .LBB115_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB115_1 Depth=1
|
|
; RV64IA-NEXT: xor a6, a3, a1
|
|
; RV64IA-NEXT: and a6, a6, a4
|
|
; RV64IA-NEXT: xor a6, a3, a6
|
|
; RV64IA-NEXT: .LBB115_3: # in Loop: Header=BB115_1 Depth=1
|
|
; RV64IA-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-NEXT: bnez a6, .LBB115_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i16 %b monotonic
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_min_i16_acquire(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i16_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lhu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 16
|
|
; RV32I-NEXT: srai s2, a0, 16
|
|
; RV32I-NEXT: j .LBB116_2
|
|
; RV32I-NEXT: .LBB116_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB116_2 Depth=1
|
|
; RV32I-NEXT: sh a3, 14(sp)
|
|
; RV32I-NEXT: addi a1, sp, 14
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a3, 14(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB116_4
|
|
; RV32I-NEXT: .LBB116_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 16
|
|
; RV32I-NEXT: srai a0, a0, 16
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bge s2, a0, .LBB116_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB116_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB116_1
|
|
; RV32I-NEXT: .LBB116_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_min_i16_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV32IA-WMO-NEXT: lui a4, 16
|
|
; RV32IA-WMO-NEXT: addi a4, a4, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a4, a0
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: srai a1, a1, 16
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: li a5, 16
|
|
; RV32IA-WMO-NEXT: sub a5, a5, a3
|
|
; RV32IA-WMO-NEXT: .LBB116_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: and a7, a3, a4
|
|
; RV32IA-WMO-NEXT: mv a6, a3
|
|
; RV32IA-WMO-NEXT: sll a7, a7, a5
|
|
; RV32IA-WMO-NEXT: sra a7, a7, a5
|
|
; RV32IA-WMO-NEXT: bge a1, a7, .LBB116_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB116_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a6, a3, a1
|
|
; RV32IA-WMO-NEXT: and a6, a6, a4
|
|
; RV32IA-WMO-NEXT: xor a6, a3, a6
|
|
; RV32IA-WMO-NEXT: .LBB116_3: # in Loop: Header=BB116_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a6, .LBB116_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_min_i16_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV32IA-TSO-NEXT: lui a4, 16
|
|
; RV32IA-TSO-NEXT: addi a4, a4, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a4, a0
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: srai a1, a1, 16
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: li a5, 16
|
|
; RV32IA-TSO-NEXT: sub a5, a5, a3
|
|
; RV32IA-TSO-NEXT: .LBB116_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: and a7, a3, a4
|
|
; RV32IA-TSO-NEXT: mv a6, a3
|
|
; RV32IA-TSO-NEXT: sll a7, a7, a5
|
|
; RV32IA-TSO-NEXT: sra a7, a7, a5
|
|
; RV32IA-TSO-NEXT: bge a1, a7, .LBB116_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB116_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a6, a3, a1
|
|
; RV32IA-TSO-NEXT: and a6, a6, a4
|
|
; RV32IA-TSO-NEXT: xor a6, a3, a6
|
|
; RV32IA-TSO-NEXT: .LBB116_3: # in Loop: Header=BB116_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a6, .LBB116_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i16_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lhu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 48
|
|
; RV64I-NEXT: srai s2, a0, 48
|
|
; RV64I-NEXT: j .LBB116_2
|
|
; RV64I-NEXT: .LBB116_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB116_2 Depth=1
|
|
; RV64I-NEXT: sh a3, 14(sp)
|
|
; RV64I-NEXT: addi a1, sp, 14
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a3, 14(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB116_4
|
|
; RV64I-NEXT: .LBB116_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 48
|
|
; RV64I-NEXT: srai a0, a0, 48
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s2, a0, .LBB116_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB116_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB116_1
|
|
; RV64I-NEXT: .LBB116_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_min_i16_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV64IA-WMO-NEXT: lui a4, 16
|
|
; RV64IA-WMO-NEXT: addi a4, a4, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: srai a1, a1, 48
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: li a5, 48
|
|
; RV64IA-WMO-NEXT: sub a5, a5, a3
|
|
; RV64IA-WMO-NEXT: .LBB116_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: and a7, a3, a4
|
|
; RV64IA-WMO-NEXT: mv a6, a3
|
|
; RV64IA-WMO-NEXT: sll a7, a7, a5
|
|
; RV64IA-WMO-NEXT: sra a7, a7, a5
|
|
; RV64IA-WMO-NEXT: bge a1, a7, .LBB116_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB116_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a6, a3, a1
|
|
; RV64IA-WMO-NEXT: and a6, a6, a4
|
|
; RV64IA-WMO-NEXT: xor a6, a3, a6
|
|
; RV64IA-WMO-NEXT: .LBB116_3: # in Loop: Header=BB116_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a6, .LBB116_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_min_i16_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV64IA-TSO-NEXT: lui a4, 16
|
|
; RV64IA-TSO-NEXT: addi a4, a4, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: srai a1, a1, 48
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: li a5, 48
|
|
; RV64IA-TSO-NEXT: sub a5, a5, a3
|
|
; RV64IA-TSO-NEXT: .LBB116_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: and a7, a3, a4
|
|
; RV64IA-TSO-NEXT: mv a6, a3
|
|
; RV64IA-TSO-NEXT: sll a7, a7, a5
|
|
; RV64IA-TSO-NEXT: sra a7, a7, a5
|
|
; RV64IA-TSO-NEXT: bge a1, a7, .LBB116_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB116_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a6, a3, a1
|
|
; RV64IA-TSO-NEXT: and a6, a6, a4
|
|
; RV64IA-TSO-NEXT: xor a6, a3, a6
|
|
; RV64IA-TSO-NEXT: .LBB116_3: # in Loop: Header=BB116_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a6, .LBB116_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i16 %b acquire
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_min_i16_release(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i16_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lhu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 16
|
|
; RV32I-NEXT: srai s2, a0, 16
|
|
; RV32I-NEXT: j .LBB117_2
|
|
; RV32I-NEXT: .LBB117_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB117_2 Depth=1
|
|
; RV32I-NEXT: sh a3, 14(sp)
|
|
; RV32I-NEXT: addi a1, sp, 14
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a3, 14(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB117_4
|
|
; RV32I-NEXT: .LBB117_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 16
|
|
; RV32I-NEXT: srai a0, a0, 16
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bge s2, a0, .LBB117_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB117_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB117_1
|
|
; RV32I-NEXT: .LBB117_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_min_i16_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV32IA-WMO-NEXT: lui a4, 16
|
|
; RV32IA-WMO-NEXT: addi a4, a4, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a4, a0
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: srai a1, a1, 16
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: li a5, 16
|
|
; RV32IA-WMO-NEXT: sub a5, a5, a3
|
|
; RV32IA-WMO-NEXT: .LBB117_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-WMO-NEXT: and a7, a3, a4
|
|
; RV32IA-WMO-NEXT: mv a6, a3
|
|
; RV32IA-WMO-NEXT: sll a7, a7, a5
|
|
; RV32IA-WMO-NEXT: sra a7, a7, a5
|
|
; RV32IA-WMO-NEXT: bge a1, a7, .LBB117_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB117_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a6, a3, a1
|
|
; RV32IA-WMO-NEXT: and a6, a6, a4
|
|
; RV32IA-WMO-NEXT: xor a6, a3, a6
|
|
; RV32IA-WMO-NEXT: .LBB117_3: # in Loop: Header=BB117_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a6, .LBB117_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_min_i16_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV32IA-TSO-NEXT: lui a4, 16
|
|
; RV32IA-TSO-NEXT: addi a4, a4, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a4, a0
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: srai a1, a1, 16
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: li a5, 16
|
|
; RV32IA-TSO-NEXT: sub a5, a5, a3
|
|
; RV32IA-TSO-NEXT: .LBB117_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: and a7, a3, a4
|
|
; RV32IA-TSO-NEXT: mv a6, a3
|
|
; RV32IA-TSO-NEXT: sll a7, a7, a5
|
|
; RV32IA-TSO-NEXT: sra a7, a7, a5
|
|
; RV32IA-TSO-NEXT: bge a1, a7, .LBB117_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB117_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a6, a3, a1
|
|
; RV32IA-TSO-NEXT: and a6, a6, a4
|
|
; RV32IA-TSO-NEXT: xor a6, a3, a6
|
|
; RV32IA-TSO-NEXT: .LBB117_3: # in Loop: Header=BB117_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a6, .LBB117_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i16_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lhu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 48
|
|
; RV64I-NEXT: srai s2, a0, 48
|
|
; RV64I-NEXT: j .LBB117_2
|
|
; RV64I-NEXT: .LBB117_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB117_2 Depth=1
|
|
; RV64I-NEXT: sh a3, 14(sp)
|
|
; RV64I-NEXT: addi a1, sp, 14
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a3, 14(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB117_4
|
|
; RV64I-NEXT: .LBB117_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 48
|
|
; RV64I-NEXT: srai a0, a0, 48
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s2, a0, .LBB117_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB117_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB117_1
|
|
; RV64I-NEXT: .LBB117_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_min_i16_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV64IA-WMO-NEXT: lui a4, 16
|
|
; RV64IA-WMO-NEXT: addi a4, a4, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: srai a1, a1, 48
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: li a5, 48
|
|
; RV64IA-WMO-NEXT: sub a5, a5, a3
|
|
; RV64IA-WMO-NEXT: .LBB117_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-WMO-NEXT: and a7, a3, a4
|
|
; RV64IA-WMO-NEXT: mv a6, a3
|
|
; RV64IA-WMO-NEXT: sll a7, a7, a5
|
|
; RV64IA-WMO-NEXT: sra a7, a7, a5
|
|
; RV64IA-WMO-NEXT: bge a1, a7, .LBB117_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB117_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a6, a3, a1
|
|
; RV64IA-WMO-NEXT: and a6, a6, a4
|
|
; RV64IA-WMO-NEXT: xor a6, a3, a6
|
|
; RV64IA-WMO-NEXT: .LBB117_3: # in Loop: Header=BB117_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a6, .LBB117_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_min_i16_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV64IA-TSO-NEXT: lui a4, 16
|
|
; RV64IA-TSO-NEXT: addi a4, a4, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: srai a1, a1, 48
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: li a5, 48
|
|
; RV64IA-TSO-NEXT: sub a5, a5, a3
|
|
; RV64IA-TSO-NEXT: .LBB117_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: and a7, a3, a4
|
|
; RV64IA-TSO-NEXT: mv a6, a3
|
|
; RV64IA-TSO-NEXT: sll a7, a7, a5
|
|
; RV64IA-TSO-NEXT: sra a7, a7, a5
|
|
; RV64IA-TSO-NEXT: bge a1, a7, .LBB117_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB117_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a6, a3, a1
|
|
; RV64IA-TSO-NEXT: and a6, a6, a4
|
|
; RV64IA-TSO-NEXT: xor a6, a3, a6
|
|
; RV64IA-TSO-NEXT: .LBB117_3: # in Loop: Header=BB117_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a6, .LBB117_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i16 %b release
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_min_i16_acq_rel(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i16_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lhu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 16
|
|
; RV32I-NEXT: srai s2, a0, 16
|
|
; RV32I-NEXT: j .LBB118_2
|
|
; RV32I-NEXT: .LBB118_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB118_2 Depth=1
|
|
; RV32I-NEXT: sh a3, 14(sp)
|
|
; RV32I-NEXT: addi a1, sp, 14
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a3, 14(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB118_4
|
|
; RV32I-NEXT: .LBB118_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 16
|
|
; RV32I-NEXT: srai a0, a0, 16
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bge s2, a0, .LBB118_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB118_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB118_1
|
|
; RV32I-NEXT: .LBB118_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_min_i16_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV32IA-WMO-NEXT: lui a4, 16
|
|
; RV32IA-WMO-NEXT: addi a4, a4, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a4, a0
|
|
; RV32IA-WMO-NEXT: slli a1, a1, 16
|
|
; RV32IA-WMO-NEXT: srai a1, a1, 16
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: li a5, 16
|
|
; RV32IA-WMO-NEXT: sub a5, a5, a3
|
|
; RV32IA-WMO-NEXT: .LBB118_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: and a7, a3, a4
|
|
; RV32IA-WMO-NEXT: mv a6, a3
|
|
; RV32IA-WMO-NEXT: sll a7, a7, a5
|
|
; RV32IA-WMO-NEXT: sra a7, a7, a5
|
|
; RV32IA-WMO-NEXT: bge a1, a7, .LBB118_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB118_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a6, a3, a1
|
|
; RV32IA-WMO-NEXT: and a6, a6, a4
|
|
; RV32IA-WMO-NEXT: xor a6, a3, a6
|
|
; RV32IA-WMO-NEXT: .LBB118_3: # in Loop: Header=BB118_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a6, .LBB118_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_min_i16_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV32IA-TSO-NEXT: lui a4, 16
|
|
; RV32IA-TSO-NEXT: addi a4, a4, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a4, a0
|
|
; RV32IA-TSO-NEXT: slli a1, a1, 16
|
|
; RV32IA-TSO-NEXT: srai a1, a1, 16
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: li a5, 16
|
|
; RV32IA-TSO-NEXT: sub a5, a5, a3
|
|
; RV32IA-TSO-NEXT: .LBB118_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: and a7, a3, a4
|
|
; RV32IA-TSO-NEXT: mv a6, a3
|
|
; RV32IA-TSO-NEXT: sll a7, a7, a5
|
|
; RV32IA-TSO-NEXT: sra a7, a7, a5
|
|
; RV32IA-TSO-NEXT: bge a1, a7, .LBB118_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB118_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a6, a3, a1
|
|
; RV32IA-TSO-NEXT: and a6, a6, a4
|
|
; RV32IA-TSO-NEXT: xor a6, a3, a6
|
|
; RV32IA-TSO-NEXT: .LBB118_3: # in Loop: Header=BB118_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a6, .LBB118_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i16_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lhu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 48
|
|
; RV64I-NEXT: srai s2, a0, 48
|
|
; RV64I-NEXT: j .LBB118_2
|
|
; RV64I-NEXT: .LBB118_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB118_2 Depth=1
|
|
; RV64I-NEXT: sh a3, 14(sp)
|
|
; RV64I-NEXT: addi a1, sp, 14
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a3, 14(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB118_4
|
|
; RV64I-NEXT: .LBB118_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 48
|
|
; RV64I-NEXT: srai a0, a0, 48
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s2, a0, .LBB118_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB118_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB118_1
|
|
; RV64I-NEXT: .LBB118_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_min_i16_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: andi a3, a0, 24
|
|
; RV64IA-WMO-NEXT: lui a4, 16
|
|
; RV64IA-WMO-NEXT: addi a4, a4, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-WMO-NEXT: slli a1, a1, 48
|
|
; RV64IA-WMO-NEXT: srai a1, a1, 48
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: li a5, 48
|
|
; RV64IA-WMO-NEXT: sub a5, a5, a3
|
|
; RV64IA-WMO-NEXT: .LBB118_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: and a7, a3, a4
|
|
; RV64IA-WMO-NEXT: mv a6, a3
|
|
; RV64IA-WMO-NEXT: sll a7, a7, a5
|
|
; RV64IA-WMO-NEXT: sra a7, a7, a5
|
|
; RV64IA-WMO-NEXT: bge a1, a7, .LBB118_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB118_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a6, a3, a1
|
|
; RV64IA-WMO-NEXT: and a6, a6, a4
|
|
; RV64IA-WMO-NEXT: xor a6, a3, a6
|
|
; RV64IA-WMO-NEXT: .LBB118_3: # in Loop: Header=BB118_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a6, .LBB118_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_min_i16_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: andi a3, a0, 24
|
|
; RV64IA-TSO-NEXT: lui a4, 16
|
|
; RV64IA-TSO-NEXT: addi a4, a4, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a4, a0
|
|
; RV64IA-TSO-NEXT: slli a1, a1, 48
|
|
; RV64IA-TSO-NEXT: srai a1, a1, 48
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: li a5, 48
|
|
; RV64IA-TSO-NEXT: sub a5, a5, a3
|
|
; RV64IA-TSO-NEXT: .LBB118_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: and a7, a3, a4
|
|
; RV64IA-TSO-NEXT: mv a6, a3
|
|
; RV64IA-TSO-NEXT: sll a7, a7, a5
|
|
; RV64IA-TSO-NEXT: sra a7, a7, a5
|
|
; RV64IA-TSO-NEXT: bge a1, a7, .LBB118_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB118_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a6, a3, a1
|
|
; RV64IA-TSO-NEXT: and a6, a6, a4
|
|
; RV64IA-TSO-NEXT: xor a6, a3, a6
|
|
; RV64IA-TSO-NEXT: .LBB118_3: # in Loop: Header=BB118_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a6, a6, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a6, .LBB118_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i16 %b acq_rel
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_min_i16_seq_cst(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i16_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lhu a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: slli a0, a1, 16
|
|
; RV32I-NEXT: srai s2, a0, 16
|
|
; RV32I-NEXT: j .LBB119_2
|
|
; RV32I-NEXT: .LBB119_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB119_2 Depth=1
|
|
; RV32I-NEXT: sh a3, 14(sp)
|
|
; RV32I-NEXT: addi a1, sp, 14
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a3, 14(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB119_4
|
|
; RV32I-NEXT: .LBB119_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: slli a0, a3, 16
|
|
; RV32I-NEXT: srai a0, a0, 16
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bge s2, a0, .LBB119_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB119_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB119_1
|
|
; RV32I-NEXT: .LBB119_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_min_i16_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: andi a3, a0, 24
|
|
; RV32IA-NEXT: lui a4, 16
|
|
; RV32IA-NEXT: addi a4, a4, -1
|
|
; RV32IA-NEXT: sll a4, a4, a0
|
|
; RV32IA-NEXT: slli a1, a1, 16
|
|
; RV32IA-NEXT: srai a1, a1, 16
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: li a5, 16
|
|
; RV32IA-NEXT: sub a5, a5, a3
|
|
; RV32IA-NEXT: .LBB119_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV32IA-NEXT: and a7, a3, a4
|
|
; RV32IA-NEXT: mv a6, a3
|
|
; RV32IA-NEXT: sll a7, a7, a5
|
|
; RV32IA-NEXT: sra a7, a7, a5
|
|
; RV32IA-NEXT: bge a1, a7, .LBB119_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB119_1 Depth=1
|
|
; RV32IA-NEXT: xor a6, a3, a1
|
|
; RV32IA-NEXT: and a6, a6, a4
|
|
; RV32IA-NEXT: xor a6, a3, a6
|
|
; RV32IA-NEXT: .LBB119_3: # in Loop: Header=BB119_1 Depth=1
|
|
; RV32IA-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV32IA-NEXT: bnez a6, .LBB119_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i16_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lhu a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: slli a0, a1, 48
|
|
; RV64I-NEXT: srai s2, a0, 48
|
|
; RV64I-NEXT: j .LBB119_2
|
|
; RV64I-NEXT: .LBB119_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB119_2 Depth=1
|
|
; RV64I-NEXT: sh a3, 14(sp)
|
|
; RV64I-NEXT: addi a1, sp, 14
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a3, 14(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB119_4
|
|
; RV64I-NEXT: .LBB119_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: slli a0, a3, 48
|
|
; RV64I-NEXT: srai a0, a0, 48
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s2, a0, .LBB119_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB119_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB119_1
|
|
; RV64I-NEXT: .LBB119_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_min_i16_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: andi a3, a0, 24
|
|
; RV64IA-NEXT: lui a4, 16
|
|
; RV64IA-NEXT: addi a4, a4, -1
|
|
; RV64IA-NEXT: sllw a4, a4, a0
|
|
; RV64IA-NEXT: slli a1, a1, 48
|
|
; RV64IA-NEXT: srai a1, a1, 48
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: li a5, 48
|
|
; RV64IA-NEXT: sub a5, a5, a3
|
|
; RV64IA-NEXT: .LBB119_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV64IA-NEXT: and a7, a3, a4
|
|
; RV64IA-NEXT: mv a6, a3
|
|
; RV64IA-NEXT: sll a7, a7, a5
|
|
; RV64IA-NEXT: sra a7, a7, a5
|
|
; RV64IA-NEXT: bge a1, a7, .LBB119_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB119_1 Depth=1
|
|
; RV64IA-NEXT: xor a6, a3, a1
|
|
; RV64IA-NEXT: and a6, a6, a4
|
|
; RV64IA-NEXT: xor a6, a3, a6
|
|
; RV64IA-NEXT: .LBB119_3: # in Loop: Header=BB119_1 Depth=1
|
|
; RV64IA-NEXT: sc.w.rl a6, a6, (a2)
|
|
; RV64IA-NEXT: bnez a6, .LBB119_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i16 %b seq_cst
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_umax_i16_monotonic(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i16_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a1
|
|
; RV32I-NEXT: mv s1, a0
|
|
; RV32I-NEXT: lhu a1, 0(a0)
|
|
; RV32I-NEXT: lui s2, 16
|
|
; RV32I-NEXT: addi s2, s2, -1
|
|
; RV32I-NEXT: and s3, s0, s2
|
|
; RV32I-NEXT: j .LBB120_2
|
|
; RV32I-NEXT: .LBB120_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB120_2 Depth=1
|
|
; RV32I-NEXT: sh a1, 10(sp)
|
|
; RV32I-NEXT: addi a1, sp, 10
|
|
; RV32I-NEXT: mv a0, s1
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a1, 10(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB120_4
|
|
; RV32I-NEXT: .LBB120_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: and a0, a1, s2
|
|
; RV32I-NEXT: mv a2, a1
|
|
; RV32I-NEXT: bltu s3, a0, .LBB120_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB120_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s0
|
|
; RV32I-NEXT: j .LBB120_1
|
|
; RV32I-NEXT: .LBB120_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a1
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umax_i16_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: lui a3, 16
|
|
; RV32IA-NEXT: addi a3, a3, -1
|
|
; RV32IA-NEXT: sll a4, a3, a0
|
|
; RV32IA-NEXT: and a1, a1, a3
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB120_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a3, (a2)
|
|
; RV32IA-NEXT: and a6, a3, a4
|
|
; RV32IA-NEXT: mv a5, a3
|
|
; RV32IA-NEXT: bgeu a6, a1, .LBB120_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB120_1 Depth=1
|
|
; RV32IA-NEXT: xor a5, a3, a1
|
|
; RV32IA-NEXT: and a5, a5, a4
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: .LBB120_3: # in Loop: Header=BB120_1 Depth=1
|
|
; RV32IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB120_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i16_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s3, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a1
|
|
; RV64I-NEXT: mv s1, a0
|
|
; RV64I-NEXT: lhu a1, 0(a0)
|
|
; RV64I-NEXT: lui s2, 16
|
|
; RV64I-NEXT: addiw s2, s2, -1
|
|
; RV64I-NEXT: and s3, s0, s2
|
|
; RV64I-NEXT: j .LBB120_2
|
|
; RV64I-NEXT: .LBB120_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB120_2 Depth=1
|
|
; RV64I-NEXT: sh a1, 6(sp)
|
|
; RV64I-NEXT: addi a1, sp, 6
|
|
; RV64I-NEXT: mv a0, s1
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a1, 6(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB120_4
|
|
; RV64I-NEXT: .LBB120_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: and a0, a1, s2
|
|
; RV64I-NEXT: mv a2, a1
|
|
; RV64I-NEXT: bltu s3, a0, .LBB120_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB120_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s0
|
|
; RV64I-NEXT: j .LBB120_1
|
|
; RV64I-NEXT: .LBB120_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a1
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s3, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_umax_i16_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: lui a3, 16
|
|
; RV64IA-NEXT: addi a3, a3, -1
|
|
; RV64IA-NEXT: sllw a4, a3, a0
|
|
; RV64IA-NEXT: and a1, a1, a3
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB120_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a3, (a2)
|
|
; RV64IA-NEXT: and a6, a3, a4
|
|
; RV64IA-NEXT: mv a5, a3
|
|
; RV64IA-NEXT: bgeu a6, a1, .LBB120_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB120_1 Depth=1
|
|
; RV64IA-NEXT: xor a5, a3, a1
|
|
; RV64IA-NEXT: and a5, a5, a4
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: .LBB120_3: # in Loop: Header=BB120_1 Depth=1
|
|
; RV64IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB120_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i16 %b monotonic
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_umax_i16_acquire(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i16_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a1
|
|
; RV32I-NEXT: mv s1, a0
|
|
; RV32I-NEXT: lhu a1, 0(a0)
|
|
; RV32I-NEXT: lui s2, 16
|
|
; RV32I-NEXT: addi s2, s2, -1
|
|
; RV32I-NEXT: and s3, s0, s2
|
|
; RV32I-NEXT: j .LBB121_2
|
|
; RV32I-NEXT: .LBB121_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB121_2 Depth=1
|
|
; RV32I-NEXT: sh a1, 10(sp)
|
|
; RV32I-NEXT: addi a1, sp, 10
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s1
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a1, 10(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB121_4
|
|
; RV32I-NEXT: .LBB121_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: and a0, a1, s2
|
|
; RV32I-NEXT: mv a2, a1
|
|
; RV32I-NEXT: bltu s3, a0, .LBB121_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB121_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s0
|
|
; RV32I-NEXT: j .LBB121_1
|
|
; RV32I-NEXT: .LBB121_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a1
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umax_i16_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB121_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: and a6, a3, a4
|
|
; RV32IA-WMO-NEXT: mv a5, a3
|
|
; RV32IA-WMO-NEXT: bgeu a6, a1, .LBB121_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB121_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a1
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: .LBB121_3: # in Loop: Header=BB121_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB121_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umax_i16_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB121_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: and a6, a3, a4
|
|
; RV32IA-TSO-NEXT: mv a5, a3
|
|
; RV32IA-TSO-NEXT: bgeu a6, a1, .LBB121_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB121_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a1
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: .LBB121_3: # in Loop: Header=BB121_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB121_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i16_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s3, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a1
|
|
; RV64I-NEXT: mv s1, a0
|
|
; RV64I-NEXT: lhu a1, 0(a0)
|
|
; RV64I-NEXT: lui s2, 16
|
|
; RV64I-NEXT: addiw s2, s2, -1
|
|
; RV64I-NEXT: and s3, s0, s2
|
|
; RV64I-NEXT: j .LBB121_2
|
|
; RV64I-NEXT: .LBB121_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB121_2 Depth=1
|
|
; RV64I-NEXT: sh a1, 6(sp)
|
|
; RV64I-NEXT: addi a1, sp, 6
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s1
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a1, 6(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB121_4
|
|
; RV64I-NEXT: .LBB121_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: and a0, a1, s2
|
|
; RV64I-NEXT: mv a2, a1
|
|
; RV64I-NEXT: bltu s3, a0, .LBB121_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB121_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s0
|
|
; RV64I-NEXT: j .LBB121_1
|
|
; RV64I-NEXT: .LBB121_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a1
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s3, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umax_i16_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB121_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: and a6, a3, a4
|
|
; RV64IA-WMO-NEXT: mv a5, a3
|
|
; RV64IA-WMO-NEXT: bgeu a6, a1, .LBB121_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB121_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a1
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: .LBB121_3: # in Loop: Header=BB121_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB121_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umax_i16_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB121_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: and a6, a3, a4
|
|
; RV64IA-TSO-NEXT: mv a5, a3
|
|
; RV64IA-TSO-NEXT: bgeu a6, a1, .LBB121_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB121_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a1
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: .LBB121_3: # in Loop: Header=BB121_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB121_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i16 %b acquire
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_umax_i16_release(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i16_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a1
|
|
; RV32I-NEXT: mv s1, a0
|
|
; RV32I-NEXT: lhu a1, 0(a0)
|
|
; RV32I-NEXT: lui s2, 16
|
|
; RV32I-NEXT: addi s2, s2, -1
|
|
; RV32I-NEXT: and s3, s0, s2
|
|
; RV32I-NEXT: j .LBB122_2
|
|
; RV32I-NEXT: .LBB122_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB122_2 Depth=1
|
|
; RV32I-NEXT: sh a1, 10(sp)
|
|
; RV32I-NEXT: addi a1, sp, 10
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: mv a0, s1
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a1, 10(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB122_4
|
|
; RV32I-NEXT: .LBB122_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: and a0, a1, s2
|
|
; RV32I-NEXT: mv a2, a1
|
|
; RV32I-NEXT: bltu s3, a0, .LBB122_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB122_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s0
|
|
; RV32I-NEXT: j .LBB122_1
|
|
; RV32I-NEXT: .LBB122_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a1
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umax_i16_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB122_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-WMO-NEXT: and a6, a3, a4
|
|
; RV32IA-WMO-NEXT: mv a5, a3
|
|
; RV32IA-WMO-NEXT: bgeu a6, a1, .LBB122_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB122_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a1
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: .LBB122_3: # in Loop: Header=BB122_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB122_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umax_i16_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB122_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: and a6, a3, a4
|
|
; RV32IA-TSO-NEXT: mv a5, a3
|
|
; RV32IA-TSO-NEXT: bgeu a6, a1, .LBB122_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB122_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a1
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: .LBB122_3: # in Loop: Header=BB122_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB122_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i16_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s3, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a1
|
|
; RV64I-NEXT: mv s1, a0
|
|
; RV64I-NEXT: lhu a1, 0(a0)
|
|
; RV64I-NEXT: lui s2, 16
|
|
; RV64I-NEXT: addiw s2, s2, -1
|
|
; RV64I-NEXT: and s3, s0, s2
|
|
; RV64I-NEXT: j .LBB122_2
|
|
; RV64I-NEXT: .LBB122_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB122_2 Depth=1
|
|
; RV64I-NEXT: sh a1, 6(sp)
|
|
; RV64I-NEXT: addi a1, sp, 6
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s1
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a1, 6(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB122_4
|
|
; RV64I-NEXT: .LBB122_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: and a0, a1, s2
|
|
; RV64I-NEXT: mv a2, a1
|
|
; RV64I-NEXT: bltu s3, a0, .LBB122_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB122_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s0
|
|
; RV64I-NEXT: j .LBB122_1
|
|
; RV64I-NEXT: .LBB122_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a1
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s3, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umax_i16_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB122_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-WMO-NEXT: and a6, a3, a4
|
|
; RV64IA-WMO-NEXT: mv a5, a3
|
|
; RV64IA-WMO-NEXT: bgeu a6, a1, .LBB122_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB122_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a1
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: .LBB122_3: # in Loop: Header=BB122_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB122_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umax_i16_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB122_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: and a6, a3, a4
|
|
; RV64IA-TSO-NEXT: mv a5, a3
|
|
; RV64IA-TSO-NEXT: bgeu a6, a1, .LBB122_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB122_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a1
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: .LBB122_3: # in Loop: Header=BB122_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB122_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i16 %b release
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_umax_i16_acq_rel(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i16_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a1
|
|
; RV32I-NEXT: mv s1, a0
|
|
; RV32I-NEXT: lhu a1, 0(a0)
|
|
; RV32I-NEXT: lui s2, 16
|
|
; RV32I-NEXT: addi s2, s2, -1
|
|
; RV32I-NEXT: and s3, s0, s2
|
|
; RV32I-NEXT: j .LBB123_2
|
|
; RV32I-NEXT: .LBB123_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB123_2 Depth=1
|
|
; RV32I-NEXT: sh a1, 10(sp)
|
|
; RV32I-NEXT: addi a1, sp, 10
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s1
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a1, 10(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB123_4
|
|
; RV32I-NEXT: .LBB123_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: and a0, a1, s2
|
|
; RV32I-NEXT: mv a2, a1
|
|
; RV32I-NEXT: bltu s3, a0, .LBB123_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB123_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s0
|
|
; RV32I-NEXT: j .LBB123_1
|
|
; RV32I-NEXT: .LBB123_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a1
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umax_i16_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB123_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: and a6, a3, a4
|
|
; RV32IA-WMO-NEXT: mv a5, a3
|
|
; RV32IA-WMO-NEXT: bgeu a6, a1, .LBB123_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB123_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a1
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: .LBB123_3: # in Loop: Header=BB123_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB123_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umax_i16_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB123_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: and a6, a3, a4
|
|
; RV32IA-TSO-NEXT: mv a5, a3
|
|
; RV32IA-TSO-NEXT: bgeu a6, a1, .LBB123_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB123_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a1
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: .LBB123_3: # in Loop: Header=BB123_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB123_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i16_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s3, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a1
|
|
; RV64I-NEXT: mv s1, a0
|
|
; RV64I-NEXT: lhu a1, 0(a0)
|
|
; RV64I-NEXT: lui s2, 16
|
|
; RV64I-NEXT: addiw s2, s2, -1
|
|
; RV64I-NEXT: and s3, s0, s2
|
|
; RV64I-NEXT: j .LBB123_2
|
|
; RV64I-NEXT: .LBB123_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB123_2 Depth=1
|
|
; RV64I-NEXT: sh a1, 6(sp)
|
|
; RV64I-NEXT: addi a1, sp, 6
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s1
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a1, 6(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB123_4
|
|
; RV64I-NEXT: .LBB123_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: and a0, a1, s2
|
|
; RV64I-NEXT: mv a2, a1
|
|
; RV64I-NEXT: bltu s3, a0, .LBB123_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB123_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s0
|
|
; RV64I-NEXT: j .LBB123_1
|
|
; RV64I-NEXT: .LBB123_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a1
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s3, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umax_i16_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB123_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: and a6, a3, a4
|
|
; RV64IA-WMO-NEXT: mv a5, a3
|
|
; RV64IA-WMO-NEXT: bgeu a6, a1, .LBB123_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB123_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a1
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: .LBB123_3: # in Loop: Header=BB123_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB123_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umax_i16_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB123_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: and a6, a3, a4
|
|
; RV64IA-TSO-NEXT: mv a5, a3
|
|
; RV64IA-TSO-NEXT: bgeu a6, a1, .LBB123_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB123_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a1
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: .LBB123_3: # in Loop: Header=BB123_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB123_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i16 %b acq_rel
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_umax_i16_seq_cst(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i16_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a1
|
|
; RV32I-NEXT: mv s1, a0
|
|
; RV32I-NEXT: lhu a1, 0(a0)
|
|
; RV32I-NEXT: lui s2, 16
|
|
; RV32I-NEXT: addi s2, s2, -1
|
|
; RV32I-NEXT: and s3, s0, s2
|
|
; RV32I-NEXT: j .LBB124_2
|
|
; RV32I-NEXT: .LBB124_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB124_2 Depth=1
|
|
; RV32I-NEXT: sh a1, 10(sp)
|
|
; RV32I-NEXT: addi a1, sp, 10
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: mv a0, s1
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a1, 10(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB124_4
|
|
; RV32I-NEXT: .LBB124_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: and a0, a1, s2
|
|
; RV32I-NEXT: mv a2, a1
|
|
; RV32I-NEXT: bltu s3, a0, .LBB124_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB124_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s0
|
|
; RV32I-NEXT: j .LBB124_1
|
|
; RV32I-NEXT: .LBB124_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a1
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umax_i16_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: lui a3, 16
|
|
; RV32IA-NEXT: addi a3, a3, -1
|
|
; RV32IA-NEXT: sll a4, a3, a0
|
|
; RV32IA-NEXT: and a1, a1, a3
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB124_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV32IA-NEXT: and a6, a3, a4
|
|
; RV32IA-NEXT: mv a5, a3
|
|
; RV32IA-NEXT: bgeu a6, a1, .LBB124_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB124_1 Depth=1
|
|
; RV32IA-NEXT: xor a5, a3, a1
|
|
; RV32IA-NEXT: and a5, a5, a4
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: .LBB124_3: # in Loop: Header=BB124_1 Depth=1
|
|
; RV32IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB124_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i16_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s3, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a1
|
|
; RV64I-NEXT: mv s1, a0
|
|
; RV64I-NEXT: lhu a1, 0(a0)
|
|
; RV64I-NEXT: lui s2, 16
|
|
; RV64I-NEXT: addiw s2, s2, -1
|
|
; RV64I-NEXT: and s3, s0, s2
|
|
; RV64I-NEXT: j .LBB124_2
|
|
; RV64I-NEXT: .LBB124_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB124_2 Depth=1
|
|
; RV64I-NEXT: sh a1, 6(sp)
|
|
; RV64I-NEXT: addi a1, sp, 6
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s1
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a1, 6(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB124_4
|
|
; RV64I-NEXT: .LBB124_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: and a0, a1, s2
|
|
; RV64I-NEXT: mv a2, a1
|
|
; RV64I-NEXT: bltu s3, a0, .LBB124_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB124_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s0
|
|
; RV64I-NEXT: j .LBB124_1
|
|
; RV64I-NEXT: .LBB124_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a1
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s3, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_umax_i16_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: lui a3, 16
|
|
; RV64IA-NEXT: addi a3, a3, -1
|
|
; RV64IA-NEXT: sllw a4, a3, a0
|
|
; RV64IA-NEXT: and a1, a1, a3
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB124_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV64IA-NEXT: and a6, a3, a4
|
|
; RV64IA-NEXT: mv a5, a3
|
|
; RV64IA-NEXT: bgeu a6, a1, .LBB124_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB124_1 Depth=1
|
|
; RV64IA-NEXT: xor a5, a3, a1
|
|
; RV64IA-NEXT: and a5, a5, a4
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: .LBB124_3: # in Loop: Header=BB124_1 Depth=1
|
|
; RV64IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB124_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i16 %b seq_cst
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_umin_i16_monotonic(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i16_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a1
|
|
; RV32I-NEXT: mv s1, a0
|
|
; RV32I-NEXT: lhu a1, 0(a0)
|
|
; RV32I-NEXT: lui s2, 16
|
|
; RV32I-NEXT: addi s2, s2, -1
|
|
; RV32I-NEXT: and s3, s0, s2
|
|
; RV32I-NEXT: j .LBB125_2
|
|
; RV32I-NEXT: .LBB125_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB125_2 Depth=1
|
|
; RV32I-NEXT: sh a1, 10(sp)
|
|
; RV32I-NEXT: addi a1, sp, 10
|
|
; RV32I-NEXT: mv a0, s1
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a1, 10(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB125_4
|
|
; RV32I-NEXT: .LBB125_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: and a0, a1, s2
|
|
; RV32I-NEXT: mv a2, a1
|
|
; RV32I-NEXT: bgeu s3, a0, .LBB125_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB125_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s0
|
|
; RV32I-NEXT: j .LBB125_1
|
|
; RV32I-NEXT: .LBB125_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a1
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umin_i16_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: lui a3, 16
|
|
; RV32IA-NEXT: addi a3, a3, -1
|
|
; RV32IA-NEXT: sll a4, a3, a0
|
|
; RV32IA-NEXT: and a1, a1, a3
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB125_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a3, (a2)
|
|
; RV32IA-NEXT: and a6, a3, a4
|
|
; RV32IA-NEXT: mv a5, a3
|
|
; RV32IA-NEXT: bgeu a1, a6, .LBB125_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB125_1 Depth=1
|
|
; RV32IA-NEXT: xor a5, a3, a1
|
|
; RV32IA-NEXT: and a5, a5, a4
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: .LBB125_3: # in Loop: Header=BB125_1 Depth=1
|
|
; RV32IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB125_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i16_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s3, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a1
|
|
; RV64I-NEXT: mv s1, a0
|
|
; RV64I-NEXT: lhu a1, 0(a0)
|
|
; RV64I-NEXT: lui s2, 16
|
|
; RV64I-NEXT: addiw s2, s2, -1
|
|
; RV64I-NEXT: and s3, s0, s2
|
|
; RV64I-NEXT: j .LBB125_2
|
|
; RV64I-NEXT: .LBB125_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB125_2 Depth=1
|
|
; RV64I-NEXT: sh a1, 6(sp)
|
|
; RV64I-NEXT: addi a1, sp, 6
|
|
; RV64I-NEXT: mv a0, s1
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a1, 6(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB125_4
|
|
; RV64I-NEXT: .LBB125_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: and a0, a1, s2
|
|
; RV64I-NEXT: mv a2, a1
|
|
; RV64I-NEXT: bgeu s3, a0, .LBB125_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB125_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s0
|
|
; RV64I-NEXT: j .LBB125_1
|
|
; RV64I-NEXT: .LBB125_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a1
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s3, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_umin_i16_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: lui a3, 16
|
|
; RV64IA-NEXT: addi a3, a3, -1
|
|
; RV64IA-NEXT: sllw a4, a3, a0
|
|
; RV64IA-NEXT: and a1, a1, a3
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB125_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a3, (a2)
|
|
; RV64IA-NEXT: and a6, a3, a4
|
|
; RV64IA-NEXT: mv a5, a3
|
|
; RV64IA-NEXT: bgeu a1, a6, .LBB125_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB125_1 Depth=1
|
|
; RV64IA-NEXT: xor a5, a3, a1
|
|
; RV64IA-NEXT: and a5, a5, a4
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: .LBB125_3: # in Loop: Header=BB125_1 Depth=1
|
|
; RV64IA-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB125_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i16 %b monotonic
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_umin_i16_acquire(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i16_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a1
|
|
; RV32I-NEXT: mv s1, a0
|
|
; RV32I-NEXT: lhu a1, 0(a0)
|
|
; RV32I-NEXT: lui s2, 16
|
|
; RV32I-NEXT: addi s2, s2, -1
|
|
; RV32I-NEXT: and s3, s0, s2
|
|
; RV32I-NEXT: j .LBB126_2
|
|
; RV32I-NEXT: .LBB126_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB126_2 Depth=1
|
|
; RV32I-NEXT: sh a1, 10(sp)
|
|
; RV32I-NEXT: addi a1, sp, 10
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s1
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a1, 10(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB126_4
|
|
; RV32I-NEXT: .LBB126_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: and a0, a1, s2
|
|
; RV32I-NEXT: mv a2, a1
|
|
; RV32I-NEXT: bgeu s3, a0, .LBB126_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB126_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s0
|
|
; RV32I-NEXT: j .LBB126_1
|
|
; RV32I-NEXT: .LBB126_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a1
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umin_i16_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB126_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: and a6, a3, a4
|
|
; RV32IA-WMO-NEXT: mv a5, a3
|
|
; RV32IA-WMO-NEXT: bgeu a1, a6, .LBB126_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB126_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a1
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: .LBB126_3: # in Loop: Header=BB126_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB126_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umin_i16_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB126_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: and a6, a3, a4
|
|
; RV32IA-TSO-NEXT: mv a5, a3
|
|
; RV32IA-TSO-NEXT: bgeu a1, a6, .LBB126_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB126_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a1
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: .LBB126_3: # in Loop: Header=BB126_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB126_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i16_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s3, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a1
|
|
; RV64I-NEXT: mv s1, a0
|
|
; RV64I-NEXT: lhu a1, 0(a0)
|
|
; RV64I-NEXT: lui s2, 16
|
|
; RV64I-NEXT: addiw s2, s2, -1
|
|
; RV64I-NEXT: and s3, s0, s2
|
|
; RV64I-NEXT: j .LBB126_2
|
|
; RV64I-NEXT: .LBB126_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB126_2 Depth=1
|
|
; RV64I-NEXT: sh a1, 6(sp)
|
|
; RV64I-NEXT: addi a1, sp, 6
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s1
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a1, 6(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB126_4
|
|
; RV64I-NEXT: .LBB126_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: and a0, a1, s2
|
|
; RV64I-NEXT: mv a2, a1
|
|
; RV64I-NEXT: bgeu s3, a0, .LBB126_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB126_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s0
|
|
; RV64I-NEXT: j .LBB126_1
|
|
; RV64I-NEXT: .LBB126_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a1
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s3, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umin_i16_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB126_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: and a6, a3, a4
|
|
; RV64IA-WMO-NEXT: mv a5, a3
|
|
; RV64IA-WMO-NEXT: bgeu a1, a6, .LBB126_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB126_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a1
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: .LBB126_3: # in Loop: Header=BB126_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB126_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umin_i16_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB126_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: and a6, a3, a4
|
|
; RV64IA-TSO-NEXT: mv a5, a3
|
|
; RV64IA-TSO-NEXT: bgeu a1, a6, .LBB126_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB126_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a1
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: .LBB126_3: # in Loop: Header=BB126_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB126_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i16 %b acquire
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_umin_i16_release(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i16_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a1
|
|
; RV32I-NEXT: mv s1, a0
|
|
; RV32I-NEXT: lhu a1, 0(a0)
|
|
; RV32I-NEXT: lui s2, 16
|
|
; RV32I-NEXT: addi s2, s2, -1
|
|
; RV32I-NEXT: and s3, s0, s2
|
|
; RV32I-NEXT: j .LBB127_2
|
|
; RV32I-NEXT: .LBB127_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB127_2 Depth=1
|
|
; RV32I-NEXT: sh a1, 10(sp)
|
|
; RV32I-NEXT: addi a1, sp, 10
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: mv a0, s1
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a1, 10(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB127_4
|
|
; RV32I-NEXT: .LBB127_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: and a0, a1, s2
|
|
; RV32I-NEXT: mv a2, a1
|
|
; RV32I-NEXT: bgeu s3, a0, .LBB127_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB127_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s0
|
|
; RV32I-NEXT: j .LBB127_1
|
|
; RV32I-NEXT: .LBB127_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a1
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umin_i16_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB127_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-WMO-NEXT: and a6, a3, a4
|
|
; RV32IA-WMO-NEXT: mv a5, a3
|
|
; RV32IA-WMO-NEXT: bgeu a1, a6, .LBB127_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB127_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a1
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: .LBB127_3: # in Loop: Header=BB127_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB127_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umin_i16_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB127_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: and a6, a3, a4
|
|
; RV32IA-TSO-NEXT: mv a5, a3
|
|
; RV32IA-TSO-NEXT: bgeu a1, a6, .LBB127_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB127_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a1
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: .LBB127_3: # in Loop: Header=BB127_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB127_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i16_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s3, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a1
|
|
; RV64I-NEXT: mv s1, a0
|
|
; RV64I-NEXT: lhu a1, 0(a0)
|
|
; RV64I-NEXT: lui s2, 16
|
|
; RV64I-NEXT: addiw s2, s2, -1
|
|
; RV64I-NEXT: and s3, s0, s2
|
|
; RV64I-NEXT: j .LBB127_2
|
|
; RV64I-NEXT: .LBB127_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB127_2 Depth=1
|
|
; RV64I-NEXT: sh a1, 6(sp)
|
|
; RV64I-NEXT: addi a1, sp, 6
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s1
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a1, 6(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB127_4
|
|
; RV64I-NEXT: .LBB127_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: and a0, a1, s2
|
|
; RV64I-NEXT: mv a2, a1
|
|
; RV64I-NEXT: bgeu s3, a0, .LBB127_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB127_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s0
|
|
; RV64I-NEXT: j .LBB127_1
|
|
; RV64I-NEXT: .LBB127_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a1
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s3, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umin_i16_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB127_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-WMO-NEXT: and a6, a3, a4
|
|
; RV64IA-WMO-NEXT: mv a5, a3
|
|
; RV64IA-WMO-NEXT: bgeu a1, a6, .LBB127_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB127_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a1
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: .LBB127_3: # in Loop: Header=BB127_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB127_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umin_i16_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB127_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: and a6, a3, a4
|
|
; RV64IA-TSO-NEXT: mv a5, a3
|
|
; RV64IA-TSO-NEXT: bgeu a1, a6, .LBB127_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB127_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a1
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: .LBB127_3: # in Loop: Header=BB127_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB127_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i16 %b release
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_umin_i16_acq_rel(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i16_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a1
|
|
; RV32I-NEXT: mv s1, a0
|
|
; RV32I-NEXT: lhu a1, 0(a0)
|
|
; RV32I-NEXT: lui s2, 16
|
|
; RV32I-NEXT: addi s2, s2, -1
|
|
; RV32I-NEXT: and s3, s0, s2
|
|
; RV32I-NEXT: j .LBB128_2
|
|
; RV32I-NEXT: .LBB128_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB128_2 Depth=1
|
|
; RV32I-NEXT: sh a1, 10(sp)
|
|
; RV32I-NEXT: addi a1, sp, 10
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s1
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a1, 10(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB128_4
|
|
; RV32I-NEXT: .LBB128_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: and a0, a1, s2
|
|
; RV32I-NEXT: mv a2, a1
|
|
; RV32I-NEXT: bgeu s3, a0, .LBB128_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB128_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s0
|
|
; RV32I-NEXT: j .LBB128_1
|
|
; RV32I-NEXT: .LBB128_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a1
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umin_i16_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV32IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV32IA-WMO-NEXT: lui a3, 16
|
|
; RV32IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV32IA-WMO-NEXT: sll a4, a3, a0
|
|
; RV32IA-WMO-NEXT: and a1, a1, a3
|
|
; RV32IA-WMO-NEXT: sll a1, a1, a0
|
|
; RV32IA-WMO-NEXT: .LBB128_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV32IA-WMO-NEXT: and a6, a3, a4
|
|
; RV32IA-WMO-NEXT: mv a5, a3
|
|
; RV32IA-WMO-NEXT: bgeu a1, a6, .LBB128_3
|
|
; RV32IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB128_1 Depth=1
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a1
|
|
; RV32IA-WMO-NEXT: and a5, a5, a4
|
|
; RV32IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV32IA-WMO-NEXT: .LBB128_3: # in Loop: Header=BB128_1 Depth=1
|
|
; RV32IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-WMO-NEXT: bnez a5, .LBB128_1
|
|
; RV32IA-WMO-NEXT: # %bb.4:
|
|
; RV32IA-WMO-NEXT: srl a0, a3, a0
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umin_i16_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV32IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV32IA-TSO-NEXT: lui a3, 16
|
|
; RV32IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV32IA-TSO-NEXT: sll a4, a3, a0
|
|
; RV32IA-TSO-NEXT: and a1, a1, a3
|
|
; RV32IA-TSO-NEXT: sll a1, a1, a0
|
|
; RV32IA-TSO-NEXT: .LBB128_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV32IA-TSO-NEXT: and a6, a3, a4
|
|
; RV32IA-TSO-NEXT: mv a5, a3
|
|
; RV32IA-TSO-NEXT: bgeu a1, a6, .LBB128_3
|
|
; RV32IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB128_1 Depth=1
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a1
|
|
; RV32IA-TSO-NEXT: and a5, a5, a4
|
|
; RV32IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV32IA-TSO-NEXT: .LBB128_3: # in Loop: Header=BB128_1 Depth=1
|
|
; RV32IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV32IA-TSO-NEXT: bnez a5, .LBB128_1
|
|
; RV32IA-TSO-NEXT: # %bb.4:
|
|
; RV32IA-TSO-NEXT: srl a0, a3, a0
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i16_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s3, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a1
|
|
; RV64I-NEXT: mv s1, a0
|
|
; RV64I-NEXT: lhu a1, 0(a0)
|
|
; RV64I-NEXT: lui s2, 16
|
|
; RV64I-NEXT: addiw s2, s2, -1
|
|
; RV64I-NEXT: and s3, s0, s2
|
|
; RV64I-NEXT: j .LBB128_2
|
|
; RV64I-NEXT: .LBB128_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB128_2 Depth=1
|
|
; RV64I-NEXT: sh a1, 6(sp)
|
|
; RV64I-NEXT: addi a1, sp, 6
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s1
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a1, 6(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB128_4
|
|
; RV64I-NEXT: .LBB128_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: and a0, a1, s2
|
|
; RV64I-NEXT: mv a2, a1
|
|
; RV64I-NEXT: bgeu s3, a0, .LBB128_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB128_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s0
|
|
; RV64I-NEXT: j .LBB128_1
|
|
; RV64I-NEXT: .LBB128_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a1
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s3, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umin_i16_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: andi a2, a0, -4
|
|
; RV64IA-WMO-NEXT: slli a0, a0, 3
|
|
; RV64IA-WMO-NEXT: lui a3, 16
|
|
; RV64IA-WMO-NEXT: addi a3, a3, -1
|
|
; RV64IA-WMO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-WMO-NEXT: and a1, a1, a3
|
|
; RV64IA-WMO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-WMO-NEXT: .LBB128_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a3, (a2)
|
|
; RV64IA-WMO-NEXT: and a6, a3, a4
|
|
; RV64IA-WMO-NEXT: mv a5, a3
|
|
; RV64IA-WMO-NEXT: bgeu a1, a6, .LBB128_3
|
|
; RV64IA-WMO-NEXT: # %bb.2: # in Loop: Header=BB128_1 Depth=1
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a1
|
|
; RV64IA-WMO-NEXT: and a5, a5, a4
|
|
; RV64IA-WMO-NEXT: xor a5, a3, a5
|
|
; RV64IA-WMO-NEXT: .LBB128_3: # in Loop: Header=BB128_1 Depth=1
|
|
; RV64IA-WMO-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-WMO-NEXT: bnez a5, .LBB128_1
|
|
; RV64IA-WMO-NEXT: # %bb.4:
|
|
; RV64IA-WMO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umin_i16_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: andi a2, a0, -4
|
|
; RV64IA-TSO-NEXT: slli a0, a0, 3
|
|
; RV64IA-TSO-NEXT: lui a3, 16
|
|
; RV64IA-TSO-NEXT: addi a3, a3, -1
|
|
; RV64IA-TSO-NEXT: sllw a4, a3, a0
|
|
; RV64IA-TSO-NEXT: and a1, a1, a3
|
|
; RV64IA-TSO-NEXT: sllw a1, a1, a0
|
|
; RV64IA-TSO-NEXT: .LBB128_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a3, (a2)
|
|
; RV64IA-TSO-NEXT: and a6, a3, a4
|
|
; RV64IA-TSO-NEXT: mv a5, a3
|
|
; RV64IA-TSO-NEXT: bgeu a1, a6, .LBB128_3
|
|
; RV64IA-TSO-NEXT: # %bb.2: # in Loop: Header=BB128_1 Depth=1
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a1
|
|
; RV64IA-TSO-NEXT: and a5, a5, a4
|
|
; RV64IA-TSO-NEXT: xor a5, a3, a5
|
|
; RV64IA-TSO-NEXT: .LBB128_3: # in Loop: Header=BB128_1 Depth=1
|
|
; RV64IA-TSO-NEXT: sc.w a5, a5, (a2)
|
|
; RV64IA-TSO-NEXT: bnez a5, .LBB128_1
|
|
; RV64IA-TSO-NEXT: # %bb.4:
|
|
; RV64IA-TSO-NEXT: srlw a0, a3, a0
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i16 %b acq_rel
|
|
ret i16 %1
|
|
}
|
|
|
|
define i16 @atomicrmw_umin_i16_seq_cst(ptr %a, i16 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i16_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s3, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a1
|
|
; RV32I-NEXT: mv s1, a0
|
|
; RV32I-NEXT: lhu a1, 0(a0)
|
|
; RV32I-NEXT: lui s2, 16
|
|
; RV32I-NEXT: addi s2, s2, -1
|
|
; RV32I-NEXT: and s3, s0, s2
|
|
; RV32I-NEXT: j .LBB129_2
|
|
; RV32I-NEXT: .LBB129_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB129_2 Depth=1
|
|
; RV32I-NEXT: sh a1, 10(sp)
|
|
; RV32I-NEXT: addi a1, sp, 10
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: mv a0, s1
|
|
; RV32I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV32I-NEXT: lh a1, 10(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB129_4
|
|
; RV32I-NEXT: .LBB129_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: and a0, a1, s2
|
|
; RV32I-NEXT: mv a2, a1
|
|
; RV32I-NEXT: bgeu s3, a0, .LBB129_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB129_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s0
|
|
; RV32I-NEXT: j .LBB129_1
|
|
; RV32I-NEXT: .LBB129_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a1
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s3, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umin_i16_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: andi a2, a0, -4
|
|
; RV32IA-NEXT: slli a0, a0, 3
|
|
; RV32IA-NEXT: lui a3, 16
|
|
; RV32IA-NEXT: addi a3, a3, -1
|
|
; RV32IA-NEXT: sll a4, a3, a0
|
|
; RV32IA-NEXT: and a1, a1, a3
|
|
; RV32IA-NEXT: sll a1, a1, a0
|
|
; RV32IA-NEXT: .LBB129_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV32IA-NEXT: and a6, a3, a4
|
|
; RV32IA-NEXT: mv a5, a3
|
|
; RV32IA-NEXT: bgeu a1, a6, .LBB129_3
|
|
; RV32IA-NEXT: # %bb.2: # in Loop: Header=BB129_1 Depth=1
|
|
; RV32IA-NEXT: xor a5, a3, a1
|
|
; RV32IA-NEXT: and a5, a5, a4
|
|
; RV32IA-NEXT: xor a5, a3, a5
|
|
; RV32IA-NEXT: .LBB129_3: # in Loop: Header=BB129_1 Depth=1
|
|
; RV32IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV32IA-NEXT: bnez a5, .LBB129_1
|
|
; RV32IA-NEXT: # %bb.4:
|
|
; RV32IA-NEXT: srl a0, a3, a0
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i16_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s3, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a1
|
|
; RV64I-NEXT: mv s1, a0
|
|
; RV64I-NEXT: lhu a1, 0(a0)
|
|
; RV64I-NEXT: lui s2, 16
|
|
; RV64I-NEXT: addiw s2, s2, -1
|
|
; RV64I-NEXT: and s3, s0, s2
|
|
; RV64I-NEXT: j .LBB129_2
|
|
; RV64I-NEXT: .LBB129_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB129_2 Depth=1
|
|
; RV64I-NEXT: sh a1, 6(sp)
|
|
; RV64I-NEXT: addi a1, sp, 6
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s1
|
|
; RV64I-NEXT: call __atomic_compare_exchange_2@plt
|
|
; RV64I-NEXT: lh a1, 6(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB129_4
|
|
; RV64I-NEXT: .LBB129_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: and a0, a1, s2
|
|
; RV64I-NEXT: mv a2, a1
|
|
; RV64I-NEXT: bgeu s3, a0, .LBB129_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB129_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s0
|
|
; RV64I-NEXT: j .LBB129_1
|
|
; RV64I-NEXT: .LBB129_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a1
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s3, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_umin_i16_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: andi a2, a0, -4
|
|
; RV64IA-NEXT: slli a0, a0, 3
|
|
; RV64IA-NEXT: lui a3, 16
|
|
; RV64IA-NEXT: addi a3, a3, -1
|
|
; RV64IA-NEXT: sllw a4, a3, a0
|
|
; RV64IA-NEXT: and a1, a1, a3
|
|
; RV64IA-NEXT: sllw a1, a1, a0
|
|
; RV64IA-NEXT: .LBB129_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a3, (a2)
|
|
; RV64IA-NEXT: and a6, a3, a4
|
|
; RV64IA-NEXT: mv a5, a3
|
|
; RV64IA-NEXT: bgeu a1, a6, .LBB129_3
|
|
; RV64IA-NEXT: # %bb.2: # in Loop: Header=BB129_1 Depth=1
|
|
; RV64IA-NEXT: xor a5, a3, a1
|
|
; RV64IA-NEXT: and a5, a5, a4
|
|
; RV64IA-NEXT: xor a5, a3, a5
|
|
; RV64IA-NEXT: .LBB129_3: # in Loop: Header=BB129_1 Depth=1
|
|
; RV64IA-NEXT: sc.w.rl a5, a5, (a2)
|
|
; RV64IA-NEXT: bnez a5, .LBB129_1
|
|
; RV64IA-NEXT: # %bb.4:
|
|
; RV64IA-NEXT: srlw a0, a3, a0
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i16 %b seq_cst
|
|
ret i16 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_xchg_i32_monotonic(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i32_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_exchange_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xchg_i32_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: amoswap.w a0, a1, (a0)
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i32_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_exchange_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_xchg_i32_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amoswap.w a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i32 %b monotonic
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_xchg_i32_acquire(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i32_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_exchange_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_i32_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoswap.w.aq a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_i32_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoswap.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i32_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_exchange_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_i32_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoswap.w.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_i32_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoswap.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i32 %b acquire
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_xchg_i32_release(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i32_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_exchange_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_i32_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoswap.w.rl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_i32_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoswap.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i32_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_exchange_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_i32_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoswap.w.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_i32_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoswap.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i32 %b release
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_xchg_i32_acq_rel(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i32_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_exchange_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_i32_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoswap.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_i32_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoswap.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i32_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_exchange_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_i32_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoswap.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_i32_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoswap.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i32 %b acq_rel
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_xchg_i32_seq_cst(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i32_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_exchange_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xchg_i32_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoswap.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xchg_i32_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoswap.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i32_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_exchange_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_i32_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoswap.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_i32_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoswap.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i32 %b seq_cst
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_add_i32_monotonic(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i32_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_add_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_add_i32_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i32_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_add_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_add_i32_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i32 %b monotonic
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_add_i32_acquire(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i32_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_add_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_add_i32_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoadd.w.aq a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_add_i32_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i32_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_add_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_add_i32_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoadd.w.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_add_i32_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i32 %b acquire
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_add_i32_release(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i32_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_add_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_add_i32_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoadd.w.rl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_add_i32_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i32_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_add_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_add_i32_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoadd.w.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_add_i32_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i32 %b release
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_add_i32_acq_rel(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i32_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_add_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_add_i32_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoadd.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_add_i32_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i32_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_add_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_add_i32_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoadd.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_add_i32_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i32 %b acq_rel
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_add_i32_seq_cst(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i32_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_add_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_add_i32_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoadd.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_add_i32_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i32_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_add_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_add_i32_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoadd.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_add_i32_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i32 %b seq_cst
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_sub_i32_monotonic(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i32_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_sub_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_sub_i32_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: neg a1, a1
|
|
; RV32IA-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i32_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_sub_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_sub_i32_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: neg a1, a1
|
|
; RV64IA-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i32 %b monotonic
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_sub_i32_acquire(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i32_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_sub_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_sub_i32_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: neg a1, a1
|
|
; RV32IA-WMO-NEXT: amoadd.w.aq a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_sub_i32_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: neg a1, a1
|
|
; RV32IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i32_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_sub_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_sub_i32_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: neg a1, a1
|
|
; RV64IA-WMO-NEXT: amoadd.w.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_sub_i32_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: neg a1, a1
|
|
; RV64IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i32 %b acquire
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_sub_i32_release(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i32_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_sub_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_sub_i32_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: neg a1, a1
|
|
; RV32IA-WMO-NEXT: amoadd.w.rl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_sub_i32_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: neg a1, a1
|
|
; RV32IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i32_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_sub_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_sub_i32_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: neg a1, a1
|
|
; RV64IA-WMO-NEXT: amoadd.w.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_sub_i32_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: neg a1, a1
|
|
; RV64IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i32 %b release
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_sub_i32_acq_rel(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i32_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_sub_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_sub_i32_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: neg a1, a1
|
|
; RV32IA-WMO-NEXT: amoadd.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_sub_i32_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: neg a1, a1
|
|
; RV32IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i32_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_sub_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_sub_i32_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: neg a1, a1
|
|
; RV64IA-WMO-NEXT: amoadd.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_sub_i32_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: neg a1, a1
|
|
; RV64IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i32 %b acq_rel
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_sub_i32_seq_cst(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i32_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_sub_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_sub_i32_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: neg a1, a1
|
|
; RV32IA-WMO-NEXT: amoadd.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_sub_i32_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: neg a1, a1
|
|
; RV32IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i32_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_sub_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_sub_i32_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: neg a1, a1
|
|
; RV64IA-WMO-NEXT: amoadd.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_sub_i32_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: neg a1, a1
|
|
; RV64IA-TSO-NEXT: amoadd.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i32 %b seq_cst
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_and_i32_monotonic(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i32_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_and_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_and_i32_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: amoand.w a0, a1, (a0)
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i32_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_and_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_and_i32_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amoand.w a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i32 %b monotonic
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_and_i32_acquire(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i32_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_and_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_and_i32_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoand.w.aq a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_and_i32_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoand.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i32_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_and_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i32_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoand.w.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i32_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoand.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i32 %b acquire
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_and_i32_release(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i32_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_and_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_and_i32_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoand.w.rl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_and_i32_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoand.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i32_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_and_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i32_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoand.w.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i32_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoand.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i32 %b release
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_and_i32_acq_rel(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i32_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_and_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_and_i32_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoand.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_and_i32_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoand.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i32_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_and_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i32_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoand.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i32_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoand.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i32 %b acq_rel
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_and_i32_seq_cst(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i32_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_and_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_and_i32_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoand.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_and_i32_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoand.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i32_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_and_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i32_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoand.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i32_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoand.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i32 %b seq_cst
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_nand_i32_monotonic(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i32_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_nand_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_nand_i32_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: .LBB150_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w a2, (a0)
|
|
; RV32IA-NEXT: and a3, a2, a1
|
|
; RV32IA-NEXT: not a3, a3
|
|
; RV32IA-NEXT: sc.w a3, a3, (a0)
|
|
; RV32IA-NEXT: bnez a3, .LBB150_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: mv a0, a2
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i32_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_nand_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_nand_i32_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: .LBB150_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w a2, (a0)
|
|
; RV64IA-NEXT: and a3, a2, a1
|
|
; RV64IA-NEXT: not a3, a3
|
|
; RV64IA-NEXT: sc.w a3, a3, (a0)
|
|
; RV64IA-NEXT: bnez a3, .LBB150_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: mv a0, a2
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i32 %b monotonic
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_nand_i32_acquire(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i32_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_nand_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_nand_i32_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: .LBB151_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a2, (a0)
|
|
; RV32IA-WMO-NEXT: and a3, a2, a1
|
|
; RV32IA-WMO-NEXT: not a3, a3
|
|
; RV32IA-WMO-NEXT: sc.w a3, a3, (a0)
|
|
; RV32IA-WMO-NEXT: bnez a3, .LBB151_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: mv a0, a2
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_nand_i32_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: .LBB151_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a2, (a0)
|
|
; RV32IA-TSO-NEXT: and a3, a2, a1
|
|
; RV32IA-TSO-NEXT: not a3, a3
|
|
; RV32IA-TSO-NEXT: sc.w a3, a3, (a0)
|
|
; RV32IA-TSO-NEXT: bnez a3, .LBB151_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: mv a0, a2
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i32_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_nand_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_nand_i32_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: .LBB151_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a2, (a0)
|
|
; RV64IA-WMO-NEXT: and a3, a2, a1
|
|
; RV64IA-WMO-NEXT: not a3, a3
|
|
; RV64IA-WMO-NEXT: sc.w a3, a3, (a0)
|
|
; RV64IA-WMO-NEXT: bnez a3, .LBB151_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: mv a0, a2
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_nand_i32_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: .LBB151_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a2, (a0)
|
|
; RV64IA-TSO-NEXT: and a3, a2, a1
|
|
; RV64IA-TSO-NEXT: not a3, a3
|
|
; RV64IA-TSO-NEXT: sc.w a3, a3, (a0)
|
|
; RV64IA-TSO-NEXT: bnez a3, .LBB151_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: mv a0, a2
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i32 %b acquire
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_nand_i32_release(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i32_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_nand_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_nand_i32_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: .LBB152_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w a2, (a0)
|
|
; RV32IA-WMO-NEXT: and a3, a2, a1
|
|
; RV32IA-WMO-NEXT: not a3, a3
|
|
; RV32IA-WMO-NEXT: sc.w.rl a3, a3, (a0)
|
|
; RV32IA-WMO-NEXT: bnez a3, .LBB152_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: mv a0, a2
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_nand_i32_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: .LBB152_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a2, (a0)
|
|
; RV32IA-TSO-NEXT: and a3, a2, a1
|
|
; RV32IA-TSO-NEXT: not a3, a3
|
|
; RV32IA-TSO-NEXT: sc.w a3, a3, (a0)
|
|
; RV32IA-TSO-NEXT: bnez a3, .LBB152_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: mv a0, a2
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i32_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_nand_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_nand_i32_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: .LBB152_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w a2, (a0)
|
|
; RV64IA-WMO-NEXT: and a3, a2, a1
|
|
; RV64IA-WMO-NEXT: not a3, a3
|
|
; RV64IA-WMO-NEXT: sc.w.rl a3, a3, (a0)
|
|
; RV64IA-WMO-NEXT: bnez a3, .LBB152_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: mv a0, a2
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_nand_i32_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: .LBB152_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a2, (a0)
|
|
; RV64IA-TSO-NEXT: and a3, a2, a1
|
|
; RV64IA-TSO-NEXT: not a3, a3
|
|
; RV64IA-TSO-NEXT: sc.w a3, a3, (a0)
|
|
; RV64IA-TSO-NEXT: bnez a3, .LBB152_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: mv a0, a2
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i32 %b release
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_nand_i32_acq_rel(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i32_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_nand_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_nand_i32_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: .LBB153_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-WMO-NEXT: lr.w.aq a2, (a0)
|
|
; RV32IA-WMO-NEXT: and a3, a2, a1
|
|
; RV32IA-WMO-NEXT: not a3, a3
|
|
; RV32IA-WMO-NEXT: sc.w.rl a3, a3, (a0)
|
|
; RV32IA-WMO-NEXT: bnez a3, .LBB153_1
|
|
; RV32IA-WMO-NEXT: # %bb.2:
|
|
; RV32IA-WMO-NEXT: mv a0, a2
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_nand_i32_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: .LBB153_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-TSO-NEXT: lr.w a2, (a0)
|
|
; RV32IA-TSO-NEXT: and a3, a2, a1
|
|
; RV32IA-TSO-NEXT: not a3, a3
|
|
; RV32IA-TSO-NEXT: sc.w a3, a3, (a0)
|
|
; RV32IA-TSO-NEXT: bnez a3, .LBB153_1
|
|
; RV32IA-TSO-NEXT: # %bb.2:
|
|
; RV32IA-TSO-NEXT: mv a0, a2
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i32_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_nand_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_nand_i32_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: .LBB153_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.w.aq a2, (a0)
|
|
; RV64IA-WMO-NEXT: and a3, a2, a1
|
|
; RV64IA-WMO-NEXT: not a3, a3
|
|
; RV64IA-WMO-NEXT: sc.w.rl a3, a3, (a0)
|
|
; RV64IA-WMO-NEXT: bnez a3, .LBB153_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: mv a0, a2
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_nand_i32_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: .LBB153_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.w a2, (a0)
|
|
; RV64IA-TSO-NEXT: and a3, a2, a1
|
|
; RV64IA-TSO-NEXT: not a3, a3
|
|
; RV64IA-TSO-NEXT: sc.w a3, a3, (a0)
|
|
; RV64IA-TSO-NEXT: bnez a3, .LBB153_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: mv a0, a2
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i32 %b acq_rel
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_nand_i32_seq_cst(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i32_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_nand_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_nand_i32_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: .LBB154_1: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: lr.w.aqrl a2, (a0)
|
|
; RV32IA-NEXT: and a3, a2, a1
|
|
; RV32IA-NEXT: not a3, a3
|
|
; RV32IA-NEXT: sc.w.rl a3, a3, (a0)
|
|
; RV32IA-NEXT: bnez a3, .LBB154_1
|
|
; RV32IA-NEXT: # %bb.2:
|
|
; RV32IA-NEXT: mv a0, a2
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i32_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_nand_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_nand_i32_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: .LBB154_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.w.aqrl a2, (a0)
|
|
; RV64IA-NEXT: and a3, a2, a1
|
|
; RV64IA-NEXT: not a3, a3
|
|
; RV64IA-NEXT: sc.w.rl a3, a3, (a0)
|
|
; RV64IA-NEXT: bnez a3, .LBB154_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: mv a0, a2
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i32 %b seq_cst
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_or_i32_monotonic(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i32_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_or_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_or_i32_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: amoor.w a0, a1, (a0)
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i32_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_or_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_or_i32_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amoor.w a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i32 %b monotonic
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_or_i32_acquire(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i32_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_or_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_or_i32_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoor.w.aq a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_or_i32_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoor.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i32_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_or_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i32_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoor.w.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i32_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoor.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i32 %b acquire
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_or_i32_release(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i32_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_or_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_or_i32_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoor.w.rl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_or_i32_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoor.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i32_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_or_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i32_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoor.w.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i32_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoor.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i32 %b release
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_or_i32_acq_rel(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i32_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_or_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_or_i32_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoor.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_or_i32_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoor.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i32_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_or_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i32_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoor.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i32_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoor.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i32 %b acq_rel
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_or_i32_seq_cst(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i32_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_or_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_or_i32_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoor.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_or_i32_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoor.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i32_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_or_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i32_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoor.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i32_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoor.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i32 %b seq_cst
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_xor_i32_monotonic(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i32_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 0
|
|
; RV32I-NEXT: call __atomic_fetch_xor_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xor_i32_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: amoxor.w a0, a1, (a0)
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i32_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_xor_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_xor_i32_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amoxor.w a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i32 %b monotonic
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_xor_i32_acquire(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i32_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 2
|
|
; RV32I-NEXT: call __atomic_fetch_xor_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xor_i32_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoxor.w.aq a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xor_i32_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoxor.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i32_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_xor_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i32_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoxor.w.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i32_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoxor.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i32 %b acquire
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_xor_i32_release(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i32_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 3
|
|
; RV32I-NEXT: call __atomic_fetch_xor_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xor_i32_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoxor.w.rl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xor_i32_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoxor.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i32_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_xor_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i32_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoxor.w.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i32_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoxor.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i32 %b release
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_xor_i32_acq_rel(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i32_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 4
|
|
; RV32I-NEXT: call __atomic_fetch_xor_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xor_i32_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoxor.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xor_i32_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoxor.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i32_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_xor_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i32_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoxor.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i32_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoxor.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i32 %b acq_rel
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_xor_i32_seq_cst(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i32_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a2, 5
|
|
; RV32I-NEXT: call __atomic_fetch_xor_4@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_xor_i32_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amoxor.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_xor_i32_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amoxor.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i32_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_xor_4@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i32_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoxor.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i32_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoxor.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i32 %b seq_cst
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_max_i32_monotonic(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i32_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB165_2
|
|
; RV32I-NEXT: .LBB165_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB165_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB165_4
|
|
; RV32I-NEXT: .LBB165_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: blt s1, a3, .LBB165_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB165_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB165_1
|
|
; RV32I-NEXT: .LBB165_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_max_i32_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: amomax.w a0, a1, (a0)
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i32_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB165_2
|
|
; RV64I-NEXT: .LBB165_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB165_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB165_4
|
|
; RV64I-NEXT: .LBB165_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s2, a3, .LBB165_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB165_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB165_1
|
|
; RV64I-NEXT: .LBB165_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_max_i32_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amomax.w a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i32 %b monotonic
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_max_i32_acquire(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i32_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB166_2
|
|
; RV32I-NEXT: .LBB166_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB166_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB166_4
|
|
; RV32I-NEXT: .LBB166_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: blt s1, a3, .LBB166_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB166_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB166_1
|
|
; RV32I-NEXT: .LBB166_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_max_i32_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amomax.w.aq a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_max_i32_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amomax.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i32_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB166_2
|
|
; RV64I-NEXT: .LBB166_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB166_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB166_4
|
|
; RV64I-NEXT: .LBB166_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s2, a3, .LBB166_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB166_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB166_1
|
|
; RV64I-NEXT: .LBB166_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_max_i32_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomax.w.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_max_i32_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomax.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i32 %b acquire
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_max_i32_release(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i32_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB167_2
|
|
; RV32I-NEXT: .LBB167_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB167_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB167_4
|
|
; RV32I-NEXT: .LBB167_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: blt s1, a3, .LBB167_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB167_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB167_1
|
|
; RV32I-NEXT: .LBB167_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_max_i32_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amomax.w.rl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_max_i32_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amomax.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i32_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB167_2
|
|
; RV64I-NEXT: .LBB167_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB167_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB167_4
|
|
; RV64I-NEXT: .LBB167_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s2, a3, .LBB167_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB167_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB167_1
|
|
; RV64I-NEXT: .LBB167_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_max_i32_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomax.w.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_max_i32_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomax.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i32 %b release
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_max_i32_acq_rel(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i32_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB168_2
|
|
; RV32I-NEXT: .LBB168_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB168_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB168_4
|
|
; RV32I-NEXT: .LBB168_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: blt s1, a3, .LBB168_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB168_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB168_1
|
|
; RV32I-NEXT: .LBB168_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_max_i32_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amomax.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_max_i32_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amomax.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i32_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB168_2
|
|
; RV64I-NEXT: .LBB168_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB168_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB168_4
|
|
; RV64I-NEXT: .LBB168_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s2, a3, .LBB168_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB168_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB168_1
|
|
; RV64I-NEXT: .LBB168_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_max_i32_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomax.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_max_i32_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomax.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i32 %b acq_rel
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_max_i32_seq_cst(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i32_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB169_2
|
|
; RV32I-NEXT: .LBB169_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB169_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB169_4
|
|
; RV32I-NEXT: .LBB169_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: blt s1, a3, .LBB169_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB169_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB169_1
|
|
; RV32I-NEXT: .LBB169_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_max_i32_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amomax.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_max_i32_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amomax.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i32_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB169_2
|
|
; RV64I-NEXT: .LBB169_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB169_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB169_4
|
|
; RV64I-NEXT: .LBB169_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s2, a3, .LBB169_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB169_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB169_1
|
|
; RV64I-NEXT: .LBB169_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_max_i32_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomax.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_max_i32_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomax.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i32 %b seq_cst
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_min_i32_monotonic(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i32_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB170_2
|
|
; RV32I-NEXT: .LBB170_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB170_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB170_4
|
|
; RV32I-NEXT: .LBB170_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bge s1, a3, .LBB170_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB170_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB170_1
|
|
; RV32I-NEXT: .LBB170_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_min_i32_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: amomin.w a0, a1, (a0)
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i32_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB170_2
|
|
; RV64I-NEXT: .LBB170_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB170_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB170_4
|
|
; RV64I-NEXT: .LBB170_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s2, a3, .LBB170_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB170_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB170_1
|
|
; RV64I-NEXT: .LBB170_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_min_i32_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amomin.w a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i32 %b monotonic
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_min_i32_acquire(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i32_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB171_2
|
|
; RV32I-NEXT: .LBB171_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB171_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB171_4
|
|
; RV32I-NEXT: .LBB171_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bge s1, a3, .LBB171_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB171_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB171_1
|
|
; RV32I-NEXT: .LBB171_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_min_i32_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amomin.w.aq a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_min_i32_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amomin.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i32_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB171_2
|
|
; RV64I-NEXT: .LBB171_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB171_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB171_4
|
|
; RV64I-NEXT: .LBB171_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s2, a3, .LBB171_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB171_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB171_1
|
|
; RV64I-NEXT: .LBB171_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_min_i32_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomin.w.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_min_i32_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomin.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i32 %b acquire
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_min_i32_release(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i32_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB172_2
|
|
; RV32I-NEXT: .LBB172_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB172_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB172_4
|
|
; RV32I-NEXT: .LBB172_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bge s1, a3, .LBB172_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB172_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB172_1
|
|
; RV32I-NEXT: .LBB172_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_min_i32_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amomin.w.rl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_min_i32_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amomin.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i32_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB172_2
|
|
; RV64I-NEXT: .LBB172_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB172_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB172_4
|
|
; RV64I-NEXT: .LBB172_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s2, a3, .LBB172_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB172_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB172_1
|
|
; RV64I-NEXT: .LBB172_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_min_i32_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomin.w.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_min_i32_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomin.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i32 %b release
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_min_i32_acq_rel(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i32_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB173_2
|
|
; RV32I-NEXT: .LBB173_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB173_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB173_4
|
|
; RV32I-NEXT: .LBB173_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bge s1, a3, .LBB173_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB173_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB173_1
|
|
; RV32I-NEXT: .LBB173_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_min_i32_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amomin.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_min_i32_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amomin.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i32_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB173_2
|
|
; RV64I-NEXT: .LBB173_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB173_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB173_4
|
|
; RV64I-NEXT: .LBB173_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s2, a3, .LBB173_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB173_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB173_1
|
|
; RV64I-NEXT: .LBB173_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_min_i32_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomin.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_min_i32_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomin.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i32 %b acq_rel
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_min_i32_seq_cst(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i32_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB174_2
|
|
; RV32I-NEXT: .LBB174_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB174_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB174_4
|
|
; RV32I-NEXT: .LBB174_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bge s1, a3, .LBB174_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB174_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB174_1
|
|
; RV32I-NEXT: .LBB174_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_min_i32_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amomin.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_min_i32_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amomin.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i32_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB174_2
|
|
; RV64I-NEXT: .LBB174_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB174_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB174_4
|
|
; RV64I-NEXT: .LBB174_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s2, a3, .LBB174_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB174_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB174_1
|
|
; RV64I-NEXT: .LBB174_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_min_i32_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomin.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_min_i32_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomin.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i32 %b seq_cst
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_umax_i32_monotonic(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i32_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB175_2
|
|
; RV32I-NEXT: .LBB175_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB175_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB175_4
|
|
; RV32I-NEXT: .LBB175_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bltu s1, a3, .LBB175_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB175_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB175_1
|
|
; RV32I-NEXT: .LBB175_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umax_i32_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: amomaxu.w a0, a1, (a0)
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i32_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB175_2
|
|
; RV64I-NEXT: .LBB175_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB175_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB175_4
|
|
; RV64I-NEXT: .LBB175_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bltu s2, a3, .LBB175_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB175_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB175_1
|
|
; RV64I-NEXT: .LBB175_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_umax_i32_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amomaxu.w a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i32 %b monotonic
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_umax_i32_acquire(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i32_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB176_2
|
|
; RV32I-NEXT: .LBB176_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB176_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB176_4
|
|
; RV32I-NEXT: .LBB176_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bltu s1, a3, .LBB176_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB176_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB176_1
|
|
; RV32I-NEXT: .LBB176_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umax_i32_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amomaxu.w.aq a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umax_i32_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amomaxu.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i32_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB176_2
|
|
; RV64I-NEXT: .LBB176_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB176_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB176_4
|
|
; RV64I-NEXT: .LBB176_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bltu s2, a3, .LBB176_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB176_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB176_1
|
|
; RV64I-NEXT: .LBB176_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umax_i32_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomaxu.w.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umax_i32_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomaxu.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i32 %b acquire
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_umax_i32_release(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i32_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB177_2
|
|
; RV32I-NEXT: .LBB177_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB177_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB177_4
|
|
; RV32I-NEXT: .LBB177_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bltu s1, a3, .LBB177_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB177_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB177_1
|
|
; RV32I-NEXT: .LBB177_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umax_i32_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amomaxu.w.rl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umax_i32_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amomaxu.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i32_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB177_2
|
|
; RV64I-NEXT: .LBB177_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB177_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB177_4
|
|
; RV64I-NEXT: .LBB177_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bltu s2, a3, .LBB177_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB177_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB177_1
|
|
; RV64I-NEXT: .LBB177_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umax_i32_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomaxu.w.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umax_i32_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomaxu.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i32 %b release
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_umax_i32_acq_rel(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i32_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB178_2
|
|
; RV32I-NEXT: .LBB178_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB178_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB178_4
|
|
; RV32I-NEXT: .LBB178_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bltu s1, a3, .LBB178_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB178_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB178_1
|
|
; RV32I-NEXT: .LBB178_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umax_i32_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amomaxu.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umax_i32_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amomaxu.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i32_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB178_2
|
|
; RV64I-NEXT: .LBB178_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB178_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB178_4
|
|
; RV64I-NEXT: .LBB178_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bltu s2, a3, .LBB178_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB178_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB178_1
|
|
; RV64I-NEXT: .LBB178_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umax_i32_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomaxu.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umax_i32_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomaxu.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i32 %b acq_rel
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_umax_i32_seq_cst(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i32_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB179_2
|
|
; RV32I-NEXT: .LBB179_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB179_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB179_4
|
|
; RV32I-NEXT: .LBB179_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bltu s1, a3, .LBB179_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB179_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB179_1
|
|
; RV32I-NEXT: .LBB179_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umax_i32_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amomaxu.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umax_i32_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amomaxu.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i32_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB179_2
|
|
; RV64I-NEXT: .LBB179_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB179_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB179_4
|
|
; RV64I-NEXT: .LBB179_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bltu s2, a3, .LBB179_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB179_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB179_1
|
|
; RV64I-NEXT: .LBB179_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umax_i32_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomaxu.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umax_i32_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomaxu.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i32 %b seq_cst
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_umin_i32_monotonic(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i32_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB180_2
|
|
; RV32I-NEXT: .LBB180_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB180_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB180_4
|
|
; RV32I-NEXT: .LBB180_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bgeu s1, a3, .LBB180_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB180_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB180_1
|
|
; RV32I-NEXT: .LBB180_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umin_i32_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: amominu.w a0, a1, (a0)
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i32_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB180_2
|
|
; RV64I-NEXT: .LBB180_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB180_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB180_4
|
|
; RV64I-NEXT: .LBB180_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bgeu s2, a3, .LBB180_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB180_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB180_1
|
|
; RV64I-NEXT: .LBB180_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_umin_i32_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amominu.w a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i32 %b monotonic
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_umin_i32_acquire(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i32_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB181_2
|
|
; RV32I-NEXT: .LBB181_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB181_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB181_4
|
|
; RV32I-NEXT: .LBB181_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bgeu s1, a3, .LBB181_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB181_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB181_1
|
|
; RV32I-NEXT: .LBB181_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umin_i32_acquire:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amominu.w.aq a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umin_i32_acquire:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amominu.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i32_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB181_2
|
|
; RV64I-NEXT: .LBB181_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB181_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB181_4
|
|
; RV64I-NEXT: .LBB181_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bgeu s2, a3, .LBB181_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB181_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB181_1
|
|
; RV64I-NEXT: .LBB181_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umin_i32_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amominu.w.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umin_i32_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amominu.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i32 %b acquire
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_umin_i32_release(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i32_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB182_2
|
|
; RV32I-NEXT: .LBB182_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB182_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB182_4
|
|
; RV32I-NEXT: .LBB182_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bgeu s1, a3, .LBB182_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB182_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB182_1
|
|
; RV32I-NEXT: .LBB182_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umin_i32_release:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amominu.w.rl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umin_i32_release:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amominu.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i32_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB182_2
|
|
; RV64I-NEXT: .LBB182_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB182_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB182_4
|
|
; RV64I-NEXT: .LBB182_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bgeu s2, a3, .LBB182_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB182_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB182_1
|
|
; RV64I-NEXT: .LBB182_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umin_i32_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amominu.w.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umin_i32_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amominu.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i32 %b release
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_umin_i32_acq_rel(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i32_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB183_2
|
|
; RV32I-NEXT: .LBB183_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB183_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB183_4
|
|
; RV32I-NEXT: .LBB183_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bgeu s1, a3, .LBB183_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB183_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB183_1
|
|
; RV32I-NEXT: .LBB183_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umin_i32_acq_rel:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amominu.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umin_i32_acq_rel:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amominu.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i32_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB183_2
|
|
; RV64I-NEXT: .LBB183_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB183_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB183_4
|
|
; RV64I-NEXT: .LBB183_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bgeu s2, a3, .LBB183_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB183_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB183_1
|
|
; RV64I-NEXT: .LBB183_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umin_i32_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amominu.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umin_i32_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amominu.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i32 %b acq_rel
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @atomicrmw_umin_i32_seq_cst(ptr %a, i32 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i32_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a3, 0(a0)
|
|
; RV32I-NEXT: mv s1, a1
|
|
; RV32I-NEXT: j .LBB184_2
|
|
; RV32I-NEXT: .LBB184_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB184_2 Depth=1
|
|
; RV32I-NEXT: sw a3, 0(sp)
|
|
; RV32I-NEXT: mv a1, sp
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV32I-NEXT: lw a3, 0(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB184_4
|
|
; RV32I-NEXT: .LBB184_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: mv a2, a3
|
|
; RV32I-NEXT: bgeu s1, a3, .LBB184_1
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB184_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s1
|
|
; RV32I-NEXT: j .LBB184_1
|
|
; RV32I-NEXT: .LBB184_4: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a3
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-WMO-LABEL: atomicrmw_umin_i32_seq_cst:
|
|
; RV32IA-WMO: # %bb.0:
|
|
; RV32IA-WMO-NEXT: amominu.w.aqrl a0, a1, (a0)
|
|
; RV32IA-WMO-NEXT: ret
|
|
;
|
|
; RV32IA-TSO-LABEL: atomicrmw_umin_i32_seq_cst:
|
|
; RV32IA-TSO: # %bb.0:
|
|
; RV32IA-TSO-NEXT: amominu.w a0, a1, (a0)
|
|
; RV32IA-TSO-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i32_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -48
|
|
; RV64I-NEXT: sd ra, 40(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 32(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s2, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: lw a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: sext.w s2, a1
|
|
; RV64I-NEXT: j .LBB184_2
|
|
; RV64I-NEXT: .LBB184_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB184_2 Depth=1
|
|
; RV64I-NEXT: sw a3, 12(sp)
|
|
; RV64I-NEXT: addi a1, sp, 12
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_4@plt
|
|
; RV64I-NEXT: lw a3, 12(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB184_4
|
|
; RV64I-NEXT: .LBB184_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bgeu s2, a3, .LBB184_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB184_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB184_1
|
|
; RV64I-NEXT: .LBB184_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 40(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 32(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s2, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umin_i32_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amominu.w.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umin_i32_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amominu.w a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i32 %b seq_cst
|
|
ret i32 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_xchg_i64_monotonic(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i64_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: call __atomic_exchange_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xchg_i64_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 0
|
|
; RV32IA-NEXT: call __atomic_exchange_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i64_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_exchange_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_xchg_i64_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amoswap.d a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i64 %b monotonic
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_xchg_i64_acquire(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i64_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: call __atomic_exchange_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xchg_i64_acquire:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 2
|
|
; RV32IA-NEXT: call __atomic_exchange_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i64_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_exchange_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_i64_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoswap.d.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_i64_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoswap.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i64 %b acquire
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_xchg_i64_release(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i64_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: call __atomic_exchange_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xchg_i64_release:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 3
|
|
; RV32IA-NEXT: call __atomic_exchange_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i64_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_exchange_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_i64_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoswap.d.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_i64_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoswap.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i64 %b release
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_xchg_i64_acq_rel(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i64_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: call __atomic_exchange_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xchg_i64_acq_rel:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 4
|
|
; RV32IA-NEXT: call __atomic_exchange_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i64_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_exchange_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_i64_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoswap.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_i64_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoswap.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i64 %b acq_rel
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_xchg_i64_seq_cst(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xchg_i64_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: call __atomic_exchange_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xchg_i64_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 5
|
|
; RV32IA-NEXT: call __atomic_exchange_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xchg_i64_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_exchange_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xchg_i64_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoswap.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xchg_i64_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoswap.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xchg ptr %a, i64 %b seq_cst
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_add_i64_monotonic(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i64_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: call __atomic_fetch_add_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_add_i64_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 0
|
|
; RV32IA-NEXT: call __atomic_fetch_add_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i64_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_add_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_add_i64_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amoadd.d a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i64 %b monotonic
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_add_i64_acquire(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i64_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: call __atomic_fetch_add_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_add_i64_acquire:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 2
|
|
; RV32IA-NEXT: call __atomic_fetch_add_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i64_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_add_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_add_i64_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoadd.d.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_add_i64_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoadd.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i64 %b acquire
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_add_i64_release(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i64_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: call __atomic_fetch_add_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_add_i64_release:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 3
|
|
; RV32IA-NEXT: call __atomic_fetch_add_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i64_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_add_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_add_i64_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoadd.d.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_add_i64_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoadd.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i64 %b release
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_add_i64_acq_rel(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i64_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: call __atomic_fetch_add_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_add_i64_acq_rel:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 4
|
|
; RV32IA-NEXT: call __atomic_fetch_add_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i64_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_add_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_add_i64_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoadd.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_add_i64_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoadd.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i64 %b acq_rel
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_add_i64_seq_cst(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_add_i64_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: call __atomic_fetch_add_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_add_i64_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 5
|
|
; RV32IA-NEXT: call __atomic_fetch_add_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_add_i64_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_add_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_add_i64_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoadd.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_add_i64_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoadd.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw add ptr %a, i64 %b seq_cst
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_sub_i64_monotonic(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i64_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: call __atomic_fetch_sub_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_sub_i64_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 0
|
|
; RV32IA-NEXT: call __atomic_fetch_sub_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i64_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_sub_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_sub_i64_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: neg a1, a1
|
|
; RV64IA-NEXT: amoadd.d a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i64 %b monotonic
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_sub_i64_acquire(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i64_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: call __atomic_fetch_sub_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_sub_i64_acquire:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 2
|
|
; RV32IA-NEXT: call __atomic_fetch_sub_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i64_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_sub_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_sub_i64_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: neg a1, a1
|
|
; RV64IA-WMO-NEXT: amoadd.d.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_sub_i64_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: neg a1, a1
|
|
; RV64IA-TSO-NEXT: amoadd.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i64 %b acquire
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_sub_i64_release(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i64_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: call __atomic_fetch_sub_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_sub_i64_release:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 3
|
|
; RV32IA-NEXT: call __atomic_fetch_sub_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i64_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_sub_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_sub_i64_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: neg a1, a1
|
|
; RV64IA-WMO-NEXT: amoadd.d.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_sub_i64_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: neg a1, a1
|
|
; RV64IA-TSO-NEXT: amoadd.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i64 %b release
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_sub_i64_acq_rel(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i64_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: call __atomic_fetch_sub_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_sub_i64_acq_rel:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 4
|
|
; RV32IA-NEXT: call __atomic_fetch_sub_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i64_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_sub_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_sub_i64_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: neg a1, a1
|
|
; RV64IA-WMO-NEXT: amoadd.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_sub_i64_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: neg a1, a1
|
|
; RV64IA-TSO-NEXT: amoadd.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i64 %b acq_rel
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_sub_i64_seq_cst(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_sub_i64_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: call __atomic_fetch_sub_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_sub_i64_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 5
|
|
; RV32IA-NEXT: call __atomic_fetch_sub_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_sub_i64_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_sub_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_sub_i64_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: neg a1, a1
|
|
; RV64IA-WMO-NEXT: amoadd.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_sub_i64_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: neg a1, a1
|
|
; RV64IA-TSO-NEXT: amoadd.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw sub ptr %a, i64 %b seq_cst
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_and_i64_monotonic(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i64_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: call __atomic_fetch_and_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_and_i64_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 0
|
|
; RV32IA-NEXT: call __atomic_fetch_and_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i64_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_and_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_and_i64_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amoand.d a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i64 %b monotonic
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_and_i64_acquire(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i64_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: call __atomic_fetch_and_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_and_i64_acquire:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 2
|
|
; RV32IA-NEXT: call __atomic_fetch_and_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i64_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_and_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i64_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoand.d.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i64_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoand.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i64 %b acquire
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_and_i64_release(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i64_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: call __atomic_fetch_and_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_and_i64_release:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 3
|
|
; RV32IA-NEXT: call __atomic_fetch_and_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i64_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_and_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i64_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoand.d.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i64_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoand.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i64 %b release
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_and_i64_acq_rel(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i64_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: call __atomic_fetch_and_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_and_i64_acq_rel:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 4
|
|
; RV32IA-NEXT: call __atomic_fetch_and_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i64_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_and_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i64_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoand.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i64_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoand.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i64 %b acq_rel
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_and_i64_seq_cst(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_and_i64_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: call __atomic_fetch_and_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_and_i64_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 5
|
|
; RV32IA-NEXT: call __atomic_fetch_and_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_and_i64_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_and_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_and_i64_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoand.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_and_i64_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoand.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw and ptr %a, i64 %b seq_cst
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_nand_i64_monotonic(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i64_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: call __atomic_fetch_nand_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_nand_i64_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 0
|
|
; RV32IA-NEXT: call __atomic_fetch_nand_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i64_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_nand_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_nand_i64_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: .LBB205_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.d a2, (a0)
|
|
; RV64IA-NEXT: and a3, a2, a1
|
|
; RV64IA-NEXT: not a3, a3
|
|
; RV64IA-NEXT: sc.d a3, a3, (a0)
|
|
; RV64IA-NEXT: bnez a3, .LBB205_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: mv a0, a2
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i64 %b monotonic
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_nand_i64_acquire(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i64_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: call __atomic_fetch_nand_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_nand_i64_acquire:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 2
|
|
; RV32IA-NEXT: call __atomic_fetch_nand_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i64_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_nand_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_nand_i64_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: .LBB206_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.d.aq a2, (a0)
|
|
; RV64IA-WMO-NEXT: and a3, a2, a1
|
|
; RV64IA-WMO-NEXT: not a3, a3
|
|
; RV64IA-WMO-NEXT: sc.d a3, a3, (a0)
|
|
; RV64IA-WMO-NEXT: bnez a3, .LBB206_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: mv a0, a2
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_nand_i64_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: .LBB206_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.d a2, (a0)
|
|
; RV64IA-TSO-NEXT: and a3, a2, a1
|
|
; RV64IA-TSO-NEXT: not a3, a3
|
|
; RV64IA-TSO-NEXT: sc.d a3, a3, (a0)
|
|
; RV64IA-TSO-NEXT: bnez a3, .LBB206_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: mv a0, a2
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i64 %b acquire
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_nand_i64_release(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i64_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: call __atomic_fetch_nand_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_nand_i64_release:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 3
|
|
; RV32IA-NEXT: call __atomic_fetch_nand_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i64_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_nand_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_nand_i64_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: .LBB207_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.d a2, (a0)
|
|
; RV64IA-WMO-NEXT: and a3, a2, a1
|
|
; RV64IA-WMO-NEXT: not a3, a3
|
|
; RV64IA-WMO-NEXT: sc.d.rl a3, a3, (a0)
|
|
; RV64IA-WMO-NEXT: bnez a3, .LBB207_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: mv a0, a2
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_nand_i64_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: .LBB207_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.d a2, (a0)
|
|
; RV64IA-TSO-NEXT: and a3, a2, a1
|
|
; RV64IA-TSO-NEXT: not a3, a3
|
|
; RV64IA-TSO-NEXT: sc.d a3, a3, (a0)
|
|
; RV64IA-TSO-NEXT: bnez a3, .LBB207_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: mv a0, a2
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i64 %b release
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_nand_i64_acq_rel(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i64_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: call __atomic_fetch_nand_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_nand_i64_acq_rel:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 4
|
|
; RV32IA-NEXT: call __atomic_fetch_nand_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i64_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_nand_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_nand_i64_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: .LBB208_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-WMO-NEXT: lr.d.aq a2, (a0)
|
|
; RV64IA-WMO-NEXT: and a3, a2, a1
|
|
; RV64IA-WMO-NEXT: not a3, a3
|
|
; RV64IA-WMO-NEXT: sc.d.rl a3, a3, (a0)
|
|
; RV64IA-WMO-NEXT: bnez a3, .LBB208_1
|
|
; RV64IA-WMO-NEXT: # %bb.2:
|
|
; RV64IA-WMO-NEXT: mv a0, a2
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_nand_i64_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: .LBB208_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-TSO-NEXT: lr.d a2, (a0)
|
|
; RV64IA-TSO-NEXT: and a3, a2, a1
|
|
; RV64IA-TSO-NEXT: not a3, a3
|
|
; RV64IA-TSO-NEXT: sc.d a3, a3, (a0)
|
|
; RV64IA-TSO-NEXT: bnez a3, .LBB208_1
|
|
; RV64IA-TSO-NEXT: # %bb.2:
|
|
; RV64IA-TSO-NEXT: mv a0, a2
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i64 %b acq_rel
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_nand_i64_seq_cst(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_nand_i64_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: call __atomic_fetch_nand_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_nand_i64_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 5
|
|
; RV32IA-NEXT: call __atomic_fetch_nand_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_nand_i64_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_nand_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_nand_i64_seq_cst:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: .LBB209_1: # =>This Inner Loop Header: Depth=1
|
|
; RV64IA-NEXT: lr.d.aqrl a2, (a0)
|
|
; RV64IA-NEXT: and a3, a2, a1
|
|
; RV64IA-NEXT: not a3, a3
|
|
; RV64IA-NEXT: sc.d.rl a3, a3, (a0)
|
|
; RV64IA-NEXT: bnez a3, .LBB209_1
|
|
; RV64IA-NEXT: # %bb.2:
|
|
; RV64IA-NEXT: mv a0, a2
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw nand ptr %a, i64 %b seq_cst
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_or_i64_monotonic(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i64_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: call __atomic_fetch_or_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_or_i64_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 0
|
|
; RV32IA-NEXT: call __atomic_fetch_or_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i64_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_or_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_or_i64_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amoor.d a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i64 %b monotonic
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_or_i64_acquire(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i64_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: call __atomic_fetch_or_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_or_i64_acquire:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 2
|
|
; RV32IA-NEXT: call __atomic_fetch_or_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i64_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_or_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i64_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoor.d.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i64_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoor.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i64 %b acquire
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_or_i64_release(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i64_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: call __atomic_fetch_or_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_or_i64_release:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 3
|
|
; RV32IA-NEXT: call __atomic_fetch_or_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i64_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_or_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i64_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoor.d.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i64_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoor.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i64 %b release
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_or_i64_acq_rel(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i64_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: call __atomic_fetch_or_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_or_i64_acq_rel:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 4
|
|
; RV32IA-NEXT: call __atomic_fetch_or_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i64_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_or_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i64_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoor.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i64_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoor.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i64 %b acq_rel
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_or_i64_seq_cst(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_or_i64_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: call __atomic_fetch_or_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_or_i64_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 5
|
|
; RV32IA-NEXT: call __atomic_fetch_or_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_or_i64_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_or_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_or_i64_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoor.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_or_i64_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoor.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw or ptr %a, i64 %b seq_cst
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_xor_i64_monotonic(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i64_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 0
|
|
; RV32I-NEXT: call __atomic_fetch_xor_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xor_i64_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 0
|
|
; RV32IA-NEXT: call __atomic_fetch_xor_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i64_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 0
|
|
; RV64I-NEXT: call __atomic_fetch_xor_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_xor_i64_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amoxor.d a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i64 %b monotonic
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_xor_i64_acquire(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i64_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 2
|
|
; RV32I-NEXT: call __atomic_fetch_xor_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xor_i64_acquire:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 2
|
|
; RV32IA-NEXT: call __atomic_fetch_xor_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i64_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 2
|
|
; RV64I-NEXT: call __atomic_fetch_xor_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i64_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoxor.d.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i64_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoxor.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i64 %b acquire
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_xor_i64_release(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i64_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 3
|
|
; RV32I-NEXT: call __atomic_fetch_xor_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xor_i64_release:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 3
|
|
; RV32IA-NEXT: call __atomic_fetch_xor_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i64_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 3
|
|
; RV64I-NEXT: call __atomic_fetch_xor_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i64_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoxor.d.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i64_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoxor.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i64 %b release
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_xor_i64_acq_rel(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i64_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 4
|
|
; RV32I-NEXT: call __atomic_fetch_xor_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xor_i64_acq_rel:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 4
|
|
; RV32IA-NEXT: call __atomic_fetch_xor_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i64_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 4
|
|
; RV64I-NEXT: call __atomic_fetch_xor_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i64_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoxor.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i64_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoxor.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i64 %b acq_rel
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_xor_i64_seq_cst(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_xor_i64_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -16
|
|
; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: li a3, 5
|
|
; RV32I-NEXT: call __atomic_fetch_xor_8@plt
|
|
; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_xor_i64_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -16
|
|
; RV32IA-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: li a3, 5
|
|
; RV32IA-NEXT: call __atomic_fetch_xor_8@plt
|
|
; RV32IA-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 16
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_xor_i64_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -16
|
|
; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: li a2, 5
|
|
; RV64I-NEXT: call __atomic_fetch_xor_8@plt
|
|
; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 16
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_xor_i64_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amoxor.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_xor_i64_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amoxor.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw xor ptr %a, i64 %b seq_cst
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_max_i64_monotonic(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i64_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB220_2
|
|
; RV32I-NEXT: .LBB220_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB220_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: li a5, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB220_7
|
|
; RV32I-NEXT: .LBB220_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB220_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB220_2 Depth=1
|
|
; RV32I-NEXT: slt a0, s1, a5
|
|
; RV32I-NEXT: j .LBB220_5
|
|
; RV32I-NEXT: .LBB220_4: # in Loop: Header=BB220_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB220_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB220_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: bnez a0, .LBB220_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB220_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB220_1
|
|
; RV32I-NEXT: .LBB220_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_max_i64_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB220_2
|
|
; RV32IA-NEXT: .LBB220_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB220_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: li a4, 0
|
|
; RV32IA-NEXT: li a5, 0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB220_7
|
|
; RV32IA-NEXT: .LBB220_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB220_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB220_2 Depth=1
|
|
; RV32IA-NEXT: slt a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB220_5
|
|
; RV32IA-NEXT: .LBB220_4: # in Loop: Header=BB220_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB220_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB220_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: bnez a0, .LBB220_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB220_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB220_1
|
|
; RV32IA-NEXT: .LBB220_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i64_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB220_2
|
|
; RV64I-NEXT: .LBB220_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB220_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB220_4
|
|
; RV64I-NEXT: .LBB220_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s1, a3, .LBB220_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB220_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB220_1
|
|
; RV64I-NEXT: .LBB220_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_max_i64_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amomax.d a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i64 %b monotonic
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_max_i64_acquire(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i64_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB221_2
|
|
; RV32I-NEXT: .LBB221_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB221_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: li a5, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB221_7
|
|
; RV32I-NEXT: .LBB221_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB221_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB221_2 Depth=1
|
|
; RV32I-NEXT: slt a0, s1, a5
|
|
; RV32I-NEXT: j .LBB221_5
|
|
; RV32I-NEXT: .LBB221_4: # in Loop: Header=BB221_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB221_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB221_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: bnez a0, .LBB221_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB221_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB221_1
|
|
; RV32I-NEXT: .LBB221_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_max_i64_acquire:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB221_2
|
|
; RV32IA-NEXT: .LBB221_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB221_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 2
|
|
; RV32IA-NEXT: li a5, 2
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB221_7
|
|
; RV32IA-NEXT: .LBB221_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB221_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB221_2 Depth=1
|
|
; RV32IA-NEXT: slt a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB221_5
|
|
; RV32IA-NEXT: .LBB221_4: # in Loop: Header=BB221_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB221_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB221_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: bnez a0, .LBB221_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB221_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB221_1
|
|
; RV32IA-NEXT: .LBB221_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i64_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB221_2
|
|
; RV64I-NEXT: .LBB221_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB221_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB221_4
|
|
; RV64I-NEXT: .LBB221_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s1, a3, .LBB221_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB221_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB221_1
|
|
; RV64I-NEXT: .LBB221_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_max_i64_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomax.d.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_max_i64_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomax.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i64 %b acquire
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_max_i64_release(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i64_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB222_2
|
|
; RV32I-NEXT: .LBB222_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB222_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 3
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a5, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB222_7
|
|
; RV32I-NEXT: .LBB222_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB222_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB222_2 Depth=1
|
|
; RV32I-NEXT: slt a0, s1, a5
|
|
; RV32I-NEXT: j .LBB222_5
|
|
; RV32I-NEXT: .LBB222_4: # in Loop: Header=BB222_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB222_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB222_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: bnez a0, .LBB222_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB222_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB222_1
|
|
; RV32I-NEXT: .LBB222_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_max_i64_release:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB222_2
|
|
; RV32IA-NEXT: .LBB222_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB222_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 3
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: li a5, 0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB222_7
|
|
; RV32IA-NEXT: .LBB222_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB222_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB222_2 Depth=1
|
|
; RV32IA-NEXT: slt a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB222_5
|
|
; RV32IA-NEXT: .LBB222_4: # in Loop: Header=BB222_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB222_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB222_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: bnez a0, .LBB222_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB222_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB222_1
|
|
; RV32IA-NEXT: .LBB222_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i64_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB222_2
|
|
; RV64I-NEXT: .LBB222_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB222_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB222_4
|
|
; RV64I-NEXT: .LBB222_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s1, a3, .LBB222_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB222_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB222_1
|
|
; RV64I-NEXT: .LBB222_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_max_i64_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomax.d.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_max_i64_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomax.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i64 %b release
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_max_i64_acq_rel(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i64_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB223_2
|
|
; RV32I-NEXT: .LBB223_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB223_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 4
|
|
; RV32I-NEXT: li a5, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB223_7
|
|
; RV32I-NEXT: .LBB223_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB223_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB223_2 Depth=1
|
|
; RV32I-NEXT: slt a0, s1, a5
|
|
; RV32I-NEXT: j .LBB223_5
|
|
; RV32I-NEXT: .LBB223_4: # in Loop: Header=BB223_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB223_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB223_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: bnez a0, .LBB223_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB223_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB223_1
|
|
; RV32I-NEXT: .LBB223_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_max_i64_acq_rel:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB223_2
|
|
; RV32IA-NEXT: .LBB223_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB223_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 4
|
|
; RV32IA-NEXT: li a5, 2
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB223_7
|
|
; RV32IA-NEXT: .LBB223_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB223_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB223_2 Depth=1
|
|
; RV32IA-NEXT: slt a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB223_5
|
|
; RV32IA-NEXT: .LBB223_4: # in Loop: Header=BB223_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB223_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB223_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: bnez a0, .LBB223_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB223_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB223_1
|
|
; RV32IA-NEXT: .LBB223_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i64_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB223_2
|
|
; RV64I-NEXT: .LBB223_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB223_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB223_4
|
|
; RV64I-NEXT: .LBB223_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s1, a3, .LBB223_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB223_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB223_1
|
|
; RV64I-NEXT: .LBB223_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_max_i64_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomax.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_max_i64_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomax.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i64 %b acq_rel
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_max_i64_seq_cst(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_max_i64_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB224_2
|
|
; RV32I-NEXT: .LBB224_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB224_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: li a5, 5
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB224_7
|
|
; RV32I-NEXT: .LBB224_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB224_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB224_2 Depth=1
|
|
; RV32I-NEXT: slt a0, s1, a5
|
|
; RV32I-NEXT: j .LBB224_5
|
|
; RV32I-NEXT: .LBB224_4: # in Loop: Header=BB224_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB224_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB224_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: bnez a0, .LBB224_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB224_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB224_1
|
|
; RV32I-NEXT: .LBB224_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_max_i64_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB224_2
|
|
; RV32IA-NEXT: .LBB224_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB224_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 5
|
|
; RV32IA-NEXT: li a5, 5
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB224_7
|
|
; RV32IA-NEXT: .LBB224_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB224_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB224_2 Depth=1
|
|
; RV32IA-NEXT: slt a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB224_5
|
|
; RV32IA-NEXT: .LBB224_4: # in Loop: Header=BB224_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB224_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB224_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: bnez a0, .LBB224_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB224_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB224_1
|
|
; RV32IA-NEXT: .LBB224_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_max_i64_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB224_2
|
|
; RV64I-NEXT: .LBB224_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB224_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB224_4
|
|
; RV64I-NEXT: .LBB224_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: blt s1, a3, .LBB224_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB224_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB224_1
|
|
; RV64I-NEXT: .LBB224_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_max_i64_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomax.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_max_i64_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomax.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw max ptr %a, i64 %b seq_cst
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_min_i64_monotonic(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i64_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB225_2
|
|
; RV32I-NEXT: .LBB225_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB225_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: li a5, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB225_7
|
|
; RV32I-NEXT: .LBB225_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB225_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB225_2 Depth=1
|
|
; RV32I-NEXT: slt a0, s1, a5
|
|
; RV32I-NEXT: j .LBB225_5
|
|
; RV32I-NEXT: .LBB225_4: # in Loop: Header=BB225_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB225_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB225_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: beqz a0, .LBB225_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB225_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB225_1
|
|
; RV32I-NEXT: .LBB225_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_min_i64_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB225_2
|
|
; RV32IA-NEXT: .LBB225_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB225_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: li a4, 0
|
|
; RV32IA-NEXT: li a5, 0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB225_7
|
|
; RV32IA-NEXT: .LBB225_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB225_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB225_2 Depth=1
|
|
; RV32IA-NEXT: slt a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB225_5
|
|
; RV32IA-NEXT: .LBB225_4: # in Loop: Header=BB225_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB225_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB225_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: beqz a0, .LBB225_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB225_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB225_1
|
|
; RV32IA-NEXT: .LBB225_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i64_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB225_2
|
|
; RV64I-NEXT: .LBB225_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB225_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB225_4
|
|
; RV64I-NEXT: .LBB225_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s1, a3, .LBB225_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB225_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB225_1
|
|
; RV64I-NEXT: .LBB225_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_min_i64_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amomin.d a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i64 %b monotonic
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_min_i64_acquire(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i64_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB226_2
|
|
; RV32I-NEXT: .LBB226_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB226_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: li a5, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB226_7
|
|
; RV32I-NEXT: .LBB226_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB226_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB226_2 Depth=1
|
|
; RV32I-NEXT: slt a0, s1, a5
|
|
; RV32I-NEXT: j .LBB226_5
|
|
; RV32I-NEXT: .LBB226_4: # in Loop: Header=BB226_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB226_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB226_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: beqz a0, .LBB226_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB226_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB226_1
|
|
; RV32I-NEXT: .LBB226_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_min_i64_acquire:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB226_2
|
|
; RV32IA-NEXT: .LBB226_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB226_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 2
|
|
; RV32IA-NEXT: li a5, 2
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB226_7
|
|
; RV32IA-NEXT: .LBB226_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB226_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB226_2 Depth=1
|
|
; RV32IA-NEXT: slt a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB226_5
|
|
; RV32IA-NEXT: .LBB226_4: # in Loop: Header=BB226_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB226_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB226_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: beqz a0, .LBB226_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB226_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB226_1
|
|
; RV32IA-NEXT: .LBB226_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i64_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB226_2
|
|
; RV64I-NEXT: .LBB226_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB226_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB226_4
|
|
; RV64I-NEXT: .LBB226_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s1, a3, .LBB226_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB226_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB226_1
|
|
; RV64I-NEXT: .LBB226_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_min_i64_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomin.d.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_min_i64_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomin.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i64 %b acquire
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_min_i64_release(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i64_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB227_2
|
|
; RV32I-NEXT: .LBB227_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB227_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 3
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a5, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB227_7
|
|
; RV32I-NEXT: .LBB227_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB227_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB227_2 Depth=1
|
|
; RV32I-NEXT: slt a0, s1, a5
|
|
; RV32I-NEXT: j .LBB227_5
|
|
; RV32I-NEXT: .LBB227_4: # in Loop: Header=BB227_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB227_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB227_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: beqz a0, .LBB227_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB227_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB227_1
|
|
; RV32I-NEXT: .LBB227_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_min_i64_release:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB227_2
|
|
; RV32IA-NEXT: .LBB227_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB227_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 3
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: li a5, 0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB227_7
|
|
; RV32IA-NEXT: .LBB227_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB227_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB227_2 Depth=1
|
|
; RV32IA-NEXT: slt a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB227_5
|
|
; RV32IA-NEXT: .LBB227_4: # in Loop: Header=BB227_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB227_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB227_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: beqz a0, .LBB227_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB227_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB227_1
|
|
; RV32IA-NEXT: .LBB227_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i64_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB227_2
|
|
; RV64I-NEXT: .LBB227_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB227_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB227_4
|
|
; RV64I-NEXT: .LBB227_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s1, a3, .LBB227_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB227_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB227_1
|
|
; RV64I-NEXT: .LBB227_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_min_i64_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomin.d.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_min_i64_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomin.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i64 %b release
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_min_i64_acq_rel(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i64_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB228_2
|
|
; RV32I-NEXT: .LBB228_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB228_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 4
|
|
; RV32I-NEXT: li a5, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB228_7
|
|
; RV32I-NEXT: .LBB228_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB228_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB228_2 Depth=1
|
|
; RV32I-NEXT: slt a0, s1, a5
|
|
; RV32I-NEXT: j .LBB228_5
|
|
; RV32I-NEXT: .LBB228_4: # in Loop: Header=BB228_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB228_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB228_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: beqz a0, .LBB228_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB228_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB228_1
|
|
; RV32I-NEXT: .LBB228_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_min_i64_acq_rel:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB228_2
|
|
; RV32IA-NEXT: .LBB228_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB228_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 4
|
|
; RV32IA-NEXT: li a5, 2
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB228_7
|
|
; RV32IA-NEXT: .LBB228_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB228_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB228_2 Depth=1
|
|
; RV32IA-NEXT: slt a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB228_5
|
|
; RV32IA-NEXT: .LBB228_4: # in Loop: Header=BB228_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB228_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB228_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: beqz a0, .LBB228_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB228_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB228_1
|
|
; RV32IA-NEXT: .LBB228_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i64_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB228_2
|
|
; RV64I-NEXT: .LBB228_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB228_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB228_4
|
|
; RV64I-NEXT: .LBB228_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s1, a3, .LBB228_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB228_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB228_1
|
|
; RV64I-NEXT: .LBB228_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_min_i64_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomin.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_min_i64_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomin.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i64 %b acq_rel
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_min_i64_seq_cst(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_min_i64_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB229_2
|
|
; RV32I-NEXT: .LBB229_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB229_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: li a5, 5
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB229_7
|
|
; RV32I-NEXT: .LBB229_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB229_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB229_2 Depth=1
|
|
; RV32I-NEXT: slt a0, s1, a5
|
|
; RV32I-NEXT: j .LBB229_5
|
|
; RV32I-NEXT: .LBB229_4: # in Loop: Header=BB229_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB229_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB229_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: beqz a0, .LBB229_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB229_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB229_1
|
|
; RV32I-NEXT: .LBB229_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_min_i64_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB229_2
|
|
; RV32IA-NEXT: .LBB229_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB229_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 5
|
|
; RV32IA-NEXT: li a5, 5
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB229_7
|
|
; RV32IA-NEXT: .LBB229_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB229_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB229_2 Depth=1
|
|
; RV32IA-NEXT: slt a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB229_5
|
|
; RV32IA-NEXT: .LBB229_4: # in Loop: Header=BB229_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB229_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB229_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: beqz a0, .LBB229_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB229_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB229_1
|
|
; RV32IA-NEXT: .LBB229_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_min_i64_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB229_2
|
|
; RV64I-NEXT: .LBB229_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB229_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB229_4
|
|
; RV64I-NEXT: .LBB229_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bge s1, a3, .LBB229_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB229_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB229_1
|
|
; RV64I-NEXT: .LBB229_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_min_i64_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomin.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_min_i64_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomin.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw min ptr %a, i64 %b seq_cst
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_umax_i64_monotonic(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i64_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB230_2
|
|
; RV32I-NEXT: .LBB230_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB230_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: li a5, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB230_7
|
|
; RV32I-NEXT: .LBB230_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB230_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB230_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s1, a5
|
|
; RV32I-NEXT: j .LBB230_5
|
|
; RV32I-NEXT: .LBB230_4: # in Loop: Header=BB230_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB230_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB230_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: bnez a0, .LBB230_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB230_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB230_1
|
|
; RV32I-NEXT: .LBB230_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umax_i64_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB230_2
|
|
; RV32IA-NEXT: .LBB230_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB230_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: li a4, 0
|
|
; RV32IA-NEXT: li a5, 0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB230_7
|
|
; RV32IA-NEXT: .LBB230_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB230_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB230_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB230_5
|
|
; RV32IA-NEXT: .LBB230_4: # in Loop: Header=BB230_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB230_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB230_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: bnez a0, .LBB230_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB230_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB230_1
|
|
; RV32IA-NEXT: .LBB230_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i64_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB230_2
|
|
; RV64I-NEXT: .LBB230_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB230_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB230_4
|
|
; RV64I-NEXT: .LBB230_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bltu s1, a3, .LBB230_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB230_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB230_1
|
|
; RV64I-NEXT: .LBB230_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_umax_i64_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amomaxu.d a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i64 %b monotonic
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_umax_i64_acquire(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i64_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB231_2
|
|
; RV32I-NEXT: .LBB231_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB231_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: li a5, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB231_7
|
|
; RV32I-NEXT: .LBB231_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB231_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB231_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s1, a5
|
|
; RV32I-NEXT: j .LBB231_5
|
|
; RV32I-NEXT: .LBB231_4: # in Loop: Header=BB231_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB231_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB231_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: bnez a0, .LBB231_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB231_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB231_1
|
|
; RV32I-NEXT: .LBB231_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umax_i64_acquire:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB231_2
|
|
; RV32IA-NEXT: .LBB231_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB231_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 2
|
|
; RV32IA-NEXT: li a5, 2
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB231_7
|
|
; RV32IA-NEXT: .LBB231_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB231_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB231_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB231_5
|
|
; RV32IA-NEXT: .LBB231_4: # in Loop: Header=BB231_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB231_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB231_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: bnez a0, .LBB231_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB231_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB231_1
|
|
; RV32IA-NEXT: .LBB231_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i64_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB231_2
|
|
; RV64I-NEXT: .LBB231_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB231_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB231_4
|
|
; RV64I-NEXT: .LBB231_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bltu s1, a3, .LBB231_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB231_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB231_1
|
|
; RV64I-NEXT: .LBB231_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umax_i64_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomaxu.d.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umax_i64_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomaxu.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i64 %b acquire
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_umax_i64_release(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i64_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB232_2
|
|
; RV32I-NEXT: .LBB232_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB232_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 3
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a5, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB232_7
|
|
; RV32I-NEXT: .LBB232_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB232_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB232_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s1, a5
|
|
; RV32I-NEXT: j .LBB232_5
|
|
; RV32I-NEXT: .LBB232_4: # in Loop: Header=BB232_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB232_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB232_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: bnez a0, .LBB232_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB232_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB232_1
|
|
; RV32I-NEXT: .LBB232_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umax_i64_release:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB232_2
|
|
; RV32IA-NEXT: .LBB232_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB232_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 3
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: li a5, 0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB232_7
|
|
; RV32IA-NEXT: .LBB232_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB232_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB232_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB232_5
|
|
; RV32IA-NEXT: .LBB232_4: # in Loop: Header=BB232_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB232_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB232_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: bnez a0, .LBB232_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB232_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB232_1
|
|
; RV32IA-NEXT: .LBB232_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i64_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB232_2
|
|
; RV64I-NEXT: .LBB232_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB232_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB232_4
|
|
; RV64I-NEXT: .LBB232_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bltu s1, a3, .LBB232_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB232_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB232_1
|
|
; RV64I-NEXT: .LBB232_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umax_i64_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomaxu.d.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umax_i64_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomaxu.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i64 %b release
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_umax_i64_acq_rel(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i64_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB233_2
|
|
; RV32I-NEXT: .LBB233_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB233_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 4
|
|
; RV32I-NEXT: li a5, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB233_7
|
|
; RV32I-NEXT: .LBB233_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB233_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB233_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s1, a5
|
|
; RV32I-NEXT: j .LBB233_5
|
|
; RV32I-NEXT: .LBB233_4: # in Loop: Header=BB233_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB233_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB233_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: bnez a0, .LBB233_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB233_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB233_1
|
|
; RV32I-NEXT: .LBB233_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umax_i64_acq_rel:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB233_2
|
|
; RV32IA-NEXT: .LBB233_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB233_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 4
|
|
; RV32IA-NEXT: li a5, 2
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB233_7
|
|
; RV32IA-NEXT: .LBB233_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB233_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB233_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB233_5
|
|
; RV32IA-NEXT: .LBB233_4: # in Loop: Header=BB233_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB233_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB233_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: bnez a0, .LBB233_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB233_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB233_1
|
|
; RV32IA-NEXT: .LBB233_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i64_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB233_2
|
|
; RV64I-NEXT: .LBB233_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB233_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB233_4
|
|
; RV64I-NEXT: .LBB233_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bltu s1, a3, .LBB233_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB233_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB233_1
|
|
; RV64I-NEXT: .LBB233_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umax_i64_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomaxu.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umax_i64_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomaxu.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i64 %b acq_rel
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_umax_i64_seq_cst(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umax_i64_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB234_2
|
|
; RV32I-NEXT: .LBB234_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB234_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: li a5, 5
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB234_7
|
|
; RV32I-NEXT: .LBB234_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB234_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB234_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s1, a5
|
|
; RV32I-NEXT: j .LBB234_5
|
|
; RV32I-NEXT: .LBB234_4: # in Loop: Header=BB234_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB234_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB234_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: bnez a0, .LBB234_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB234_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB234_1
|
|
; RV32I-NEXT: .LBB234_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umax_i64_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB234_2
|
|
; RV32IA-NEXT: .LBB234_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB234_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 5
|
|
; RV32IA-NEXT: li a5, 5
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB234_7
|
|
; RV32IA-NEXT: .LBB234_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB234_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB234_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB234_5
|
|
; RV32IA-NEXT: .LBB234_4: # in Loop: Header=BB234_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB234_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB234_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: bnez a0, .LBB234_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB234_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB234_1
|
|
; RV32IA-NEXT: .LBB234_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umax_i64_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB234_2
|
|
; RV64I-NEXT: .LBB234_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB234_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB234_4
|
|
; RV64I-NEXT: .LBB234_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bltu s1, a3, .LBB234_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB234_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB234_1
|
|
; RV64I-NEXT: .LBB234_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umax_i64_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amomaxu.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umax_i64_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amomaxu.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umax ptr %a, i64 %b seq_cst
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_umin_i64_monotonic(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i64_monotonic:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB235_2
|
|
; RV32I-NEXT: .LBB235_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB235_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a4, 0
|
|
; RV32I-NEXT: li a5, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB235_7
|
|
; RV32I-NEXT: .LBB235_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB235_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB235_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s1, a5
|
|
; RV32I-NEXT: j .LBB235_5
|
|
; RV32I-NEXT: .LBB235_4: # in Loop: Header=BB235_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB235_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB235_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: beqz a0, .LBB235_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB235_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB235_1
|
|
; RV32I-NEXT: .LBB235_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umin_i64_monotonic:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB235_2
|
|
; RV32IA-NEXT: .LBB235_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB235_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: li a4, 0
|
|
; RV32IA-NEXT: li a5, 0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB235_7
|
|
; RV32IA-NEXT: .LBB235_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB235_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB235_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB235_5
|
|
; RV32IA-NEXT: .LBB235_4: # in Loop: Header=BB235_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB235_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB235_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: beqz a0, .LBB235_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB235_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB235_1
|
|
; RV32IA-NEXT: .LBB235_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i64_monotonic:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB235_2
|
|
; RV64I-NEXT: .LBB235_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB235_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a3, 0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB235_4
|
|
; RV64I-NEXT: .LBB235_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bgeu s1, a3, .LBB235_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB235_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB235_1
|
|
; RV64I-NEXT: .LBB235_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-LABEL: atomicrmw_umin_i64_monotonic:
|
|
; RV64IA: # %bb.0:
|
|
; RV64IA-NEXT: amominu.d a0, a1, (a0)
|
|
; RV64IA-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i64 %b monotonic
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_umin_i64_acquire(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i64_acquire:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB236_2
|
|
; RV32I-NEXT: .LBB236_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB236_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 2
|
|
; RV32I-NEXT: li a5, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB236_7
|
|
; RV32I-NEXT: .LBB236_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB236_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB236_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s1, a5
|
|
; RV32I-NEXT: j .LBB236_5
|
|
; RV32I-NEXT: .LBB236_4: # in Loop: Header=BB236_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB236_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB236_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: beqz a0, .LBB236_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB236_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB236_1
|
|
; RV32I-NEXT: .LBB236_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umin_i64_acquire:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB236_2
|
|
; RV32IA-NEXT: .LBB236_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB236_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 2
|
|
; RV32IA-NEXT: li a5, 2
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB236_7
|
|
; RV32IA-NEXT: .LBB236_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB236_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB236_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB236_5
|
|
; RV32IA-NEXT: .LBB236_4: # in Loop: Header=BB236_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB236_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB236_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: beqz a0, .LBB236_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB236_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB236_1
|
|
; RV32IA-NEXT: .LBB236_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i64_acquire:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB236_2
|
|
; RV64I-NEXT: .LBB236_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB236_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 2
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB236_4
|
|
; RV64I-NEXT: .LBB236_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bgeu s1, a3, .LBB236_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB236_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB236_1
|
|
; RV64I-NEXT: .LBB236_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umin_i64_acquire:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amominu.d.aq a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umin_i64_acquire:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amominu.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i64 %b acquire
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_umin_i64_release(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i64_release:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB237_2
|
|
; RV32I-NEXT: .LBB237_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB237_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 3
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: li a5, 0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB237_7
|
|
; RV32I-NEXT: .LBB237_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB237_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB237_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s1, a5
|
|
; RV32I-NEXT: j .LBB237_5
|
|
; RV32I-NEXT: .LBB237_4: # in Loop: Header=BB237_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB237_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB237_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: beqz a0, .LBB237_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB237_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB237_1
|
|
; RV32I-NEXT: .LBB237_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umin_i64_release:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB237_2
|
|
; RV32IA-NEXT: .LBB237_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB237_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 3
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: li a5, 0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB237_7
|
|
; RV32IA-NEXT: .LBB237_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB237_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB237_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB237_5
|
|
; RV32IA-NEXT: .LBB237_4: # in Loop: Header=BB237_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB237_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB237_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: beqz a0, .LBB237_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB237_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB237_1
|
|
; RV32IA-NEXT: .LBB237_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i64_release:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB237_2
|
|
; RV64I-NEXT: .LBB237_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB237_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 3
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: li a4, 0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB237_4
|
|
; RV64I-NEXT: .LBB237_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bgeu s1, a3, .LBB237_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB237_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB237_1
|
|
; RV64I-NEXT: .LBB237_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umin_i64_release:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amominu.d.rl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umin_i64_release:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amominu.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i64 %b release
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_umin_i64_acq_rel(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i64_acq_rel:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB238_2
|
|
; RV32I-NEXT: .LBB238_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB238_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 4
|
|
; RV32I-NEXT: li a5, 2
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB238_7
|
|
; RV32I-NEXT: .LBB238_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB238_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB238_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s1, a5
|
|
; RV32I-NEXT: j .LBB238_5
|
|
; RV32I-NEXT: .LBB238_4: # in Loop: Header=BB238_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB238_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB238_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: beqz a0, .LBB238_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB238_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB238_1
|
|
; RV32I-NEXT: .LBB238_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umin_i64_acq_rel:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB238_2
|
|
; RV32IA-NEXT: .LBB238_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB238_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 4
|
|
; RV32IA-NEXT: li a5, 2
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB238_7
|
|
; RV32IA-NEXT: .LBB238_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB238_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB238_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB238_5
|
|
; RV32IA-NEXT: .LBB238_4: # in Loop: Header=BB238_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB238_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB238_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: beqz a0, .LBB238_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB238_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB238_1
|
|
; RV32IA-NEXT: .LBB238_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i64_acq_rel:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB238_2
|
|
; RV64I-NEXT: .LBB238_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB238_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 4
|
|
; RV64I-NEXT: li a4, 2
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB238_4
|
|
; RV64I-NEXT: .LBB238_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bgeu s1, a3, .LBB238_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB238_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB238_1
|
|
; RV64I-NEXT: .LBB238_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umin_i64_acq_rel:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amominu.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umin_i64_acq_rel:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amominu.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i64 %b acq_rel
|
|
ret i64 %1
|
|
}
|
|
|
|
define i64 @atomicrmw_umin_i64_seq_cst(ptr %a, i64 %b) nounwind {
|
|
; RV32I-LABEL: atomicrmw_umin_i64_seq_cst:
|
|
; RV32I: # %bb.0:
|
|
; RV32I-NEXT: addi sp, sp, -32
|
|
; RV32I-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32I-NEXT: mv s0, a0
|
|
; RV32I-NEXT: lw a5, 4(a0)
|
|
; RV32I-NEXT: lw a4, 0(a0)
|
|
; RV32I-NEXT: mv s1, a2
|
|
; RV32I-NEXT: mv s2, a1
|
|
; RV32I-NEXT: j .LBB239_2
|
|
; RV32I-NEXT: .LBB239_1: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB239_2 Depth=1
|
|
; RV32I-NEXT: sw a4, 8(sp)
|
|
; RV32I-NEXT: sw a5, 12(sp)
|
|
; RV32I-NEXT: addi a1, sp, 8
|
|
; RV32I-NEXT: li a4, 5
|
|
; RV32I-NEXT: li a5, 5
|
|
; RV32I-NEXT: mv a0, s0
|
|
; RV32I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32I-NEXT: lw a5, 12(sp)
|
|
; RV32I-NEXT: lw a4, 8(sp)
|
|
; RV32I-NEXT: bnez a0, .LBB239_7
|
|
; RV32I-NEXT: .LBB239_2: # %atomicrmw.start
|
|
; RV32I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32I-NEXT: beq a5, s1, .LBB239_4
|
|
; RV32I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB239_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s1, a5
|
|
; RV32I-NEXT: j .LBB239_5
|
|
; RV32I-NEXT: .LBB239_4: # in Loop: Header=BB239_2 Depth=1
|
|
; RV32I-NEXT: sltu a0, s2, a4
|
|
; RV32I-NEXT: .LBB239_5: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB239_2 Depth=1
|
|
; RV32I-NEXT: mv a2, a4
|
|
; RV32I-NEXT: mv a3, a5
|
|
; RV32I-NEXT: beqz a0, .LBB239_1
|
|
; RV32I-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32I-NEXT: # in Loop: Header=BB239_2 Depth=1
|
|
; RV32I-NEXT: mv a2, s2
|
|
; RV32I-NEXT: mv a3, s1
|
|
; RV32I-NEXT: j .LBB239_1
|
|
; RV32I-NEXT: .LBB239_7: # %atomicrmw.end
|
|
; RV32I-NEXT: mv a0, a4
|
|
; RV32I-NEXT: mv a1, a5
|
|
; RV32I-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32I-NEXT: addi sp, sp, 32
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV32IA-LABEL: atomicrmw_umin_i64_seq_cst:
|
|
; RV32IA: # %bb.0:
|
|
; RV32IA-NEXT: addi sp, sp, -32
|
|
; RV32IA-NEXT: sw ra, 28(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s0, 24(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s1, 20(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: sw s2, 16(sp) # 4-byte Folded Spill
|
|
; RV32IA-NEXT: mv s0, a0
|
|
; RV32IA-NEXT: lw a5, 4(a0)
|
|
; RV32IA-NEXT: lw a4, 0(a0)
|
|
; RV32IA-NEXT: mv s1, a2
|
|
; RV32IA-NEXT: mv s2, a1
|
|
; RV32IA-NEXT: j .LBB239_2
|
|
; RV32IA-NEXT: .LBB239_1: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB239_2 Depth=1
|
|
; RV32IA-NEXT: sw a4, 8(sp)
|
|
; RV32IA-NEXT: sw a5, 12(sp)
|
|
; RV32IA-NEXT: addi a1, sp, 8
|
|
; RV32IA-NEXT: li a4, 5
|
|
; RV32IA-NEXT: li a5, 5
|
|
; RV32IA-NEXT: mv a0, s0
|
|
; RV32IA-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV32IA-NEXT: lw a5, 12(sp)
|
|
; RV32IA-NEXT: lw a4, 8(sp)
|
|
; RV32IA-NEXT: bnez a0, .LBB239_7
|
|
; RV32IA-NEXT: .LBB239_2: # %atomicrmw.start
|
|
; RV32IA-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV32IA-NEXT: beq a5, s1, .LBB239_4
|
|
; RV32IA-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB239_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s1, a5
|
|
; RV32IA-NEXT: j .LBB239_5
|
|
; RV32IA-NEXT: .LBB239_4: # in Loop: Header=BB239_2 Depth=1
|
|
; RV32IA-NEXT: sltu a0, s2, a4
|
|
; RV32IA-NEXT: .LBB239_5: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB239_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, a4
|
|
; RV32IA-NEXT: mv a3, a5
|
|
; RV32IA-NEXT: beqz a0, .LBB239_1
|
|
; RV32IA-NEXT: # %bb.6: # %atomicrmw.start
|
|
; RV32IA-NEXT: # in Loop: Header=BB239_2 Depth=1
|
|
; RV32IA-NEXT: mv a2, s2
|
|
; RV32IA-NEXT: mv a3, s1
|
|
; RV32IA-NEXT: j .LBB239_1
|
|
; RV32IA-NEXT: .LBB239_7: # %atomicrmw.end
|
|
; RV32IA-NEXT: mv a0, a4
|
|
; RV32IA-NEXT: mv a1, a5
|
|
; RV32IA-NEXT: lw ra, 28(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s0, 24(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s1, 20(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: lw s2, 16(sp) # 4-byte Folded Reload
|
|
; RV32IA-NEXT: addi sp, sp, 32
|
|
; RV32IA-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: atomicrmw_umin_i64_seq_cst:
|
|
; RV64I: # %bb.0:
|
|
; RV64I-NEXT: addi sp, sp, -32
|
|
; RV64I-NEXT: sd ra, 24(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s0, 16(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: sd s1, 8(sp) # 8-byte Folded Spill
|
|
; RV64I-NEXT: mv s0, a0
|
|
; RV64I-NEXT: ld a3, 0(a0)
|
|
; RV64I-NEXT: mv s1, a1
|
|
; RV64I-NEXT: j .LBB239_2
|
|
; RV64I-NEXT: .LBB239_1: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB239_2 Depth=1
|
|
; RV64I-NEXT: sd a3, 0(sp)
|
|
; RV64I-NEXT: mv a1, sp
|
|
; RV64I-NEXT: li a3, 5
|
|
; RV64I-NEXT: li a4, 5
|
|
; RV64I-NEXT: mv a0, s0
|
|
; RV64I-NEXT: call __atomic_compare_exchange_8@plt
|
|
; RV64I-NEXT: ld a3, 0(sp)
|
|
; RV64I-NEXT: bnez a0, .LBB239_4
|
|
; RV64I-NEXT: .LBB239_2: # %atomicrmw.start
|
|
; RV64I-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; RV64I-NEXT: mv a2, a3
|
|
; RV64I-NEXT: bgeu s1, a3, .LBB239_1
|
|
; RV64I-NEXT: # %bb.3: # %atomicrmw.start
|
|
; RV64I-NEXT: # in Loop: Header=BB239_2 Depth=1
|
|
; RV64I-NEXT: mv a2, s1
|
|
; RV64I-NEXT: j .LBB239_1
|
|
; RV64I-NEXT: .LBB239_4: # %atomicrmw.end
|
|
; RV64I-NEXT: mv a0, a3
|
|
; RV64I-NEXT: ld ra, 24(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s0, 16(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: ld s1, 8(sp) # 8-byte Folded Reload
|
|
; RV64I-NEXT: addi sp, sp, 32
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV64IA-WMO-LABEL: atomicrmw_umin_i64_seq_cst:
|
|
; RV64IA-WMO: # %bb.0:
|
|
; RV64IA-WMO-NEXT: amominu.d.aqrl a0, a1, (a0)
|
|
; RV64IA-WMO-NEXT: ret
|
|
;
|
|
; RV64IA-TSO-LABEL: atomicrmw_umin_i64_seq_cst:
|
|
; RV64IA-TSO: # %bb.0:
|
|
; RV64IA-TSO-NEXT: amominu.d a0, a1, (a0)
|
|
; RV64IA-TSO-NEXT: ret
|
|
%1 = atomicrmw umin ptr %a, i64 %b seq_cst
|
|
ret i64 %1
|
|
}
|