6599961c17
This patch contains following enhancements to SrcRegMap and DstRegMap:
1 In findOnlyInterestingUse not only check if the Reg is two address usage,
but also check after commutation can it be two address usage.
2 If a physical register is clobbered, remove SrcRegMap entries that are
mapped to it.
3 In processTiedPairs, when create a new COPY instruction, add a SrcRegMap
entry only when the COPY instruction is coalescable. (The COPY src is
killed)
With these enhancements isProfitableToCommute can do better commute decision,
and finally more register copies are removed.
Differential Revision: https://reviews.llvm.org/D108731
334 lines
9.1 KiB
LLVM
334 lines
9.1 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc < %s -mtriple=i686-- | FileCheck %s --check-prefixes=X86,X86-BASE
|
|
; RUN: llc < %s -mtriple=i686-- -mattr=movbe | FileCheck %s --check-prefixes=X86,X86-MOVBE
|
|
; RUN: llc < %s -mtriple=x86_64-- | FileCheck %s --check-prefixes=X64,X64-BASE
|
|
; RUN: llc < %s -mtriple=x86_64-- -mattr=movbe | FileCheck %s --check-prefixes=X64,X64-MOVBE
|
|
|
|
define i16 @foo(i16 %x, i16 %y, i16 %z) nounwind {
|
|
; X86-LABEL: foo:
|
|
; X86: # %bb.0:
|
|
; X86-NEXT: movb {{[0-9]+}}(%esp), %cl
|
|
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
|
|
; X86-NEXT: rolw %cl, %ax
|
|
; X86-NEXT: retl
|
|
;
|
|
; X64-LABEL: foo:
|
|
; X64: # %bb.0:
|
|
; X64-NEXT: movl %edx, %ecx
|
|
; X64-NEXT: movl %edi, %eax
|
|
; X64-NEXT: # kill: def $cl killed $cl killed $ecx
|
|
; X64-NEXT: rolw %cl, %ax
|
|
; X64-NEXT: # kill: def $ax killed $ax killed $eax
|
|
; X64-NEXT: retq
|
|
%t0 = shl i16 %x, %z
|
|
%t1 = sub i16 16, %z
|
|
%t2 = lshr i16 %x, %t1
|
|
%t3 = or i16 %t2, %t0
|
|
ret i16 %t3
|
|
}
|
|
|
|
define i16 @bar(i16 %x, i16 %y, i16 %z) nounwind {
|
|
; X86-LABEL: bar:
|
|
; X86: # %bb.0:
|
|
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %edx
|
|
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
|
|
; X86-NEXT: movb {{[0-9]+}}(%esp), %cl
|
|
; X86-NEXT: andb $15, %cl
|
|
; X86-NEXT: shldw %cl, %dx, %ax
|
|
; X86-NEXT: retl
|
|
;
|
|
; X64-LABEL: bar:
|
|
; X64: # %bb.0:
|
|
; X64-NEXT: movl %edx, %ecx
|
|
; X64-NEXT: movl %esi, %eax
|
|
; X64-NEXT: andb $15, %cl
|
|
; X64-NEXT: # kill: def $cl killed $cl killed $ecx
|
|
; X64-NEXT: shldw %cl, %di, %ax
|
|
; X64-NEXT: # kill: def $ax killed $ax killed $eax
|
|
; X64-NEXT: retq
|
|
%t0 = shl i16 %y, %z
|
|
%t1 = sub i16 16, %z
|
|
%t2 = lshr i16 %x, %t1
|
|
%t3 = or i16 %t2, %t0
|
|
ret i16 %t3
|
|
}
|
|
|
|
define i16 @un(i16 %x, i16 %y, i16 %z) nounwind {
|
|
; X86-LABEL: un:
|
|
; X86: # %bb.0:
|
|
; X86-NEXT: movb {{[0-9]+}}(%esp), %cl
|
|
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
|
|
; X86-NEXT: rorw %cl, %ax
|
|
; X86-NEXT: retl
|
|
;
|
|
; X64-LABEL: un:
|
|
; X64: # %bb.0:
|
|
; X64-NEXT: movl %edx, %ecx
|
|
; X64-NEXT: movl %edi, %eax
|
|
; X64-NEXT: # kill: def $cl killed $cl killed $ecx
|
|
; X64-NEXT: rorw %cl, %ax
|
|
; X64-NEXT: # kill: def $ax killed $ax killed $eax
|
|
; X64-NEXT: retq
|
|
%t0 = lshr i16 %x, %z
|
|
%t1 = sub i16 16, %z
|
|
%t2 = shl i16 %x, %t1
|
|
%t3 = or i16 %t2, %t0
|
|
ret i16 %t3
|
|
}
|
|
|
|
define i16 @bu(i16 %x, i16 %y, i16 %z) nounwind {
|
|
; X86-LABEL: bu:
|
|
; X86: # %bb.0:
|
|
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %edx
|
|
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
|
|
; X86-NEXT: movb {{[0-9]+}}(%esp), %cl
|
|
; X86-NEXT: andb $15, %cl
|
|
; X86-NEXT: shrdw %cl, %dx, %ax
|
|
; X86-NEXT: retl
|
|
;
|
|
; X64-LABEL: bu:
|
|
; X64: # %bb.0:
|
|
; X64-NEXT: movl %edx, %ecx
|
|
; X64-NEXT: movl %esi, %eax
|
|
; X64-NEXT: andb $15, %cl
|
|
; X64-NEXT: # kill: def $cl killed $cl killed $ecx
|
|
; X64-NEXT: shrdw %cl, %di, %ax
|
|
; X64-NEXT: # kill: def $ax killed $ax killed $eax
|
|
; X64-NEXT: retq
|
|
%t0 = lshr i16 %y, %z
|
|
%t1 = sub i16 16, %z
|
|
%t2 = shl i16 %x, %t1
|
|
%t3 = or i16 %t2, %t0
|
|
ret i16 %t3
|
|
}
|
|
|
|
define i16 @xfoo(i16 %x, i16 %y, i16 %z) nounwind {
|
|
; X86-LABEL: xfoo:
|
|
; X86: # %bb.0:
|
|
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
|
|
; X86-NEXT: rolw $5, %ax
|
|
; X86-NEXT: retl
|
|
;
|
|
; X64-LABEL: xfoo:
|
|
; X64: # %bb.0:
|
|
; X64-NEXT: movl %edi, %eax
|
|
; X64-NEXT: rolw $5, %ax
|
|
; X64-NEXT: # kill: def $ax killed $ax killed $eax
|
|
; X64-NEXT: retq
|
|
%t0 = lshr i16 %x, 11
|
|
%t1 = shl i16 %x, 5
|
|
%t2 = or i16 %t0, %t1
|
|
ret i16 %t2
|
|
}
|
|
|
|
define i16 @xbar(i16 %x, i16 %y, i16 %z) nounwind {
|
|
; X86-LABEL: xbar:
|
|
; X86: # %bb.0:
|
|
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %ecx
|
|
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
|
|
; X86-NEXT: shldw $5, %cx, %ax
|
|
; X86-NEXT: retl
|
|
;
|
|
; X64-LABEL: xbar:
|
|
; X64: # %bb.0:
|
|
; X64-NEXT: movl %esi, %eax
|
|
; X64-NEXT: shldw $5, %di, %ax
|
|
; X64-NEXT: # kill: def $ax killed $ax killed $eax
|
|
; X64-NEXT: retq
|
|
%t0 = shl i16 %y, 5
|
|
%t1 = lshr i16 %x, 11
|
|
%t2 = or i16 %t0, %t1
|
|
ret i16 %t2
|
|
}
|
|
|
|
define i16 @xun(i16 %x, i16 %y, i16 %z) nounwind {
|
|
; X86-LABEL: xun:
|
|
; X86: # %bb.0:
|
|
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
|
|
; X86-NEXT: rolw $11, %ax
|
|
; X86-NEXT: retl
|
|
;
|
|
; X64-LABEL: xun:
|
|
; X64: # %bb.0:
|
|
; X64-NEXT: movl %edi, %eax
|
|
; X64-NEXT: rolw $11, %ax
|
|
; X64-NEXT: # kill: def $ax killed $ax killed $eax
|
|
; X64-NEXT: retq
|
|
%t0 = lshr i16 %x, 5
|
|
%t1 = shl i16 %x, 11
|
|
%t2 = or i16 %t0, %t1
|
|
ret i16 %t2
|
|
}
|
|
|
|
define i16 @xbu(i16 %x, i16 %y, i16 %z) nounwind {
|
|
; X86-LABEL: xbu:
|
|
; X86: # %bb.0:
|
|
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %ecx
|
|
; X86-NEXT: movzwl {{[0-9]+}}(%esp), %eax
|
|
; X86-NEXT: shldw $11, %cx, %ax
|
|
; X86-NEXT: retl
|
|
;
|
|
; X64-LABEL: xbu:
|
|
; X64: # %bb.0:
|
|
; X64-NEXT: movl %edi, %eax
|
|
; X64-NEXT: shldw $11, %si, %ax
|
|
; X64-NEXT: # kill: def $ax killed $ax killed $eax
|
|
; X64-NEXT: retq
|
|
%t0 = lshr i16 %y, 5
|
|
%t1 = shl i16 %x, 11
|
|
%t2 = or i16 %t0, %t1
|
|
ret i16 %t2
|
|
}
|
|
|
|
define i32 @rot16_demandedbits(i32 %x, i32 %y) nounwind {
|
|
; X86-LABEL: rot16_demandedbits:
|
|
; X86: # %bb.0:
|
|
; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
|
|
; X86-NEXT: movl %eax, %ecx
|
|
; X86-NEXT: shrl $11, %ecx
|
|
; X86-NEXT: shll $5, %eax
|
|
; X86-NEXT: orl %ecx, %eax
|
|
; X86-NEXT: movzwl %ax, %eax
|
|
; X86-NEXT: retl
|
|
;
|
|
; X64-LABEL: rot16_demandedbits:
|
|
; X64: # %bb.0:
|
|
; X64-NEXT: movl %edi, %eax
|
|
; X64-NEXT: shrl $11, %eax
|
|
; X64-NEXT: shll $5, %edi
|
|
; X64-NEXT: orl %eax, %edi
|
|
; X64-NEXT: movzwl %di, %eax
|
|
; X64-NEXT: retq
|
|
%t0 = lshr i32 %x, 11
|
|
%t1 = shl i32 %x, 5
|
|
%t2 = or i32 %t0, %t1
|
|
%t3 = and i32 %t2, 65535
|
|
ret i32 %t3
|
|
}
|
|
|
|
define i16 @rot16_trunc(i32 %x, i32 %y) nounwind {
|
|
; X86-LABEL: rot16_trunc:
|
|
; X86: # %bb.0:
|
|
; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
|
|
; X86-NEXT: movl %eax, %ecx
|
|
; X86-NEXT: shrl $11, %ecx
|
|
; X86-NEXT: shll $5, %eax
|
|
; X86-NEXT: orl %ecx, %eax
|
|
; X86-NEXT: # kill: def $ax killed $ax killed $eax
|
|
; X86-NEXT: retl
|
|
;
|
|
; X64-LABEL: rot16_trunc:
|
|
; X64: # %bb.0:
|
|
; X64-NEXT: movl %edi, %eax
|
|
; X64-NEXT: shrl $11, %eax
|
|
; X64-NEXT: shll $5, %edi
|
|
; X64-NEXT: orl %edi, %eax
|
|
; X64-NEXT: # kill: def $ax killed $ax killed $eax
|
|
; X64-NEXT: retq
|
|
%t0 = lshr i32 %x, 11
|
|
%t1 = shl i32 %x, 5
|
|
%t2 = or i32 %t0, %t1
|
|
%t3 = trunc i32 %t2 to i16
|
|
ret i16 %t3
|
|
}
|
|
|
|
define i16 @rotate16(i16 %x) {
|
|
; X86-BASE-LABEL: rotate16:
|
|
; X86-BASE: # %bb.0:
|
|
; X86-BASE-NEXT: movzwl {{[0-9]+}}(%esp), %eax
|
|
; X86-BASE-NEXT: rolw $8, %ax
|
|
; X86-BASE-NEXT: retl
|
|
;
|
|
; X86-MOVBE-LABEL: rotate16:
|
|
; X86-MOVBE: # %bb.0:
|
|
; X86-MOVBE-NEXT: movbew {{[0-9]+}}(%esp), %ax
|
|
; X86-MOVBE-NEXT: retl
|
|
;
|
|
; X64-LABEL: rotate16:
|
|
; X64: # %bb.0:
|
|
; X64-NEXT: movl %edi, %eax
|
|
; X64-NEXT: rolw $8, %ax
|
|
; X64-NEXT: # kill: def $ax killed $ax killed $eax
|
|
; X64-NEXT: retq
|
|
%r = call i16 @llvm.fshl.i16(i16 %x, i16 %x, i16 8)
|
|
ret i16 %r
|
|
}
|
|
|
|
; TODO: Should this always be rolw with memory operand?
|
|
|
|
define void @rotate16_in_place_memory(i8* %p) {
|
|
; X86-BASE-LABEL: rotate16_in_place_memory:
|
|
; X86-BASE: # %bb.0:
|
|
; X86-BASE-NEXT: movl {{[0-9]+}}(%esp), %eax
|
|
; X86-BASE-NEXT: rolw $8, (%eax)
|
|
; X86-BASE-NEXT: retl
|
|
;
|
|
; X86-MOVBE-LABEL: rotate16_in_place_memory:
|
|
; X86-MOVBE: # %bb.0:
|
|
; X86-MOVBE-NEXT: movl {{[0-9]+}}(%esp), %eax
|
|
; X86-MOVBE-NEXT: movzwl (%eax), %ecx
|
|
; X86-MOVBE-NEXT: movbew %cx, (%eax)
|
|
; X86-MOVBE-NEXT: retl
|
|
;
|
|
; X64-BASE-LABEL: rotate16_in_place_memory:
|
|
; X64-BASE: # %bb.0:
|
|
; X64-BASE-NEXT: rolw $8, (%rdi)
|
|
; X64-BASE-NEXT: retq
|
|
;
|
|
; X64-MOVBE-LABEL: rotate16_in_place_memory:
|
|
; X64-MOVBE: # %bb.0:
|
|
; X64-MOVBE-NEXT: movzwl (%rdi), %eax
|
|
; X64-MOVBE-NEXT: movbew %ax, (%rdi)
|
|
; X64-MOVBE-NEXT: retq
|
|
%p0 = getelementptr i8, i8* %p, i64 0
|
|
%p1 = getelementptr i8, i8* %p, i64 1
|
|
%i0 = load i8, i8* %p0, align 1
|
|
%i1 = load i8, i8* %p1, align 1
|
|
store i8 %i1, i8* %p0, align 1
|
|
store i8 %i0, i8* %p1, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @rotate16_memory(i8* %p, i8* %q) {
|
|
; X86-BASE-LABEL: rotate16_memory:
|
|
; X86-BASE: # %bb.0:
|
|
; X86-BASE-NEXT: movl {{[0-9]+}}(%esp), %eax
|
|
; X86-BASE-NEXT: movl {{[0-9]+}}(%esp), %ecx
|
|
; X86-BASE-NEXT: movzwl (%ecx), %ecx
|
|
; X86-BASE-NEXT: rolw $8, %cx
|
|
; X86-BASE-NEXT: movw %cx, (%eax)
|
|
; X86-BASE-NEXT: retl
|
|
;
|
|
; X86-MOVBE-LABEL: rotate16_memory:
|
|
; X86-MOVBE: # %bb.0:
|
|
; X86-MOVBE-NEXT: movl {{[0-9]+}}(%esp), %eax
|
|
; X86-MOVBE-NEXT: movl {{[0-9]+}}(%esp), %ecx
|
|
; X86-MOVBE-NEXT: movzwl (%ecx), %ecx
|
|
; X86-MOVBE-NEXT: movbew %cx, (%eax)
|
|
; X86-MOVBE-NEXT: retl
|
|
;
|
|
; X64-BASE-LABEL: rotate16_memory:
|
|
; X64-BASE: # %bb.0:
|
|
; X64-BASE-NEXT: movzwl (%rdi), %eax
|
|
; X64-BASE-NEXT: rolw $8, %ax
|
|
; X64-BASE-NEXT: movw %ax, (%rsi)
|
|
; X64-BASE-NEXT: retq
|
|
;
|
|
; X64-MOVBE-LABEL: rotate16_memory:
|
|
; X64-MOVBE: # %bb.0:
|
|
; X64-MOVBE-NEXT: movzwl (%rdi), %eax
|
|
; X64-MOVBE-NEXT: movbew %ax, (%rsi)
|
|
; X64-MOVBE-NEXT: retq
|
|
%p0 = getelementptr i8, i8* %p, i64 0
|
|
%p1 = getelementptr i8, i8* %p, i64 1
|
|
%q0 = getelementptr i8, i8* %q, i64 0
|
|
%q1 = getelementptr i8, i8* %q, i64 1
|
|
%i0 = load i8, i8* %p0, align 1
|
|
%i1 = load i8, i8* %p1, align 1
|
|
store i8 %i1, i8* %q0, align 1
|
|
store i8 %i0, i8* %q1, align 1
|
|
ret void
|
|
}
|
|
|
|
declare i16 @llvm.fshl.i16(i16, i16, i16)
|