428f36401b
This reverts commit37b8f09a4b, and returns commit1bd0b82e50. The miscompile was in InstCombine, and it has been addressed. This tries to approach the problem noted by @arsenm: terrible codegen for `__builtin_fpclassify()`: https://godbolt.org/z/388zqdE37 Just because the PHI in the common successor happens to have different incoming values for these two blocks, doesn't mean we have to give up. It's quite easy to deal with this, we just need to produce a select: https://alive2.llvm.org/ce/z/000srb Now, the cost model for this transform is rather overly strict, so this will basically never fire. We tally all (over all preds) the selects needed to the NumBonusInsts Differential Revision: https://reviews.llvm.org/D139275
69 lines
2.5 KiB
LLVM
69 lines
2.5 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc -relocation-model=pic -verify-machineinstrs < %s \
|
|
; RUN: -mtriple=powerpc64le-unknown-linux-gnu -mcpu=pwr8 | FileCheck %s
|
|
%struct.e.0.12.28.44.104.108.112.188 = type { i32 }
|
|
%struct.t.1.13.29.45.105.109.113.189 = type { i64, i64 }
|
|
|
|
@g = external local_unnamed_addr global %struct.e.0.12.28.44.104.108.112.188, align 4
|
|
@aj = external thread_local local_unnamed_addr global %struct.t.1.13.29.45.105.109.113.189, align 8
|
|
|
|
define void @_ZNK1q1rEv() local_unnamed_addr #0 align 2 {
|
|
; CHECK-LABEL: _ZNK1q1rEv:
|
|
; CHECK: # %bb.0: # %entry
|
|
; CHECK-NEXT: mflr 0
|
|
; CHECK-NEXT: std 30, -16(1) # 8-byte Folded Spill
|
|
; CHECK-NEXT: stdu 1, -48(1)
|
|
; CHECK-NEXT: std 0, 64(1)
|
|
; CHECK-NEXT: addis 4, 2, .LC0@toc@ha
|
|
; CHECK-NEXT: lwz 3, 0(3)
|
|
; CHECK-NEXT: ld 4, .LC0@toc@l(4)
|
|
; CHECK-NEXT: addi 3, 3, -1
|
|
; CHECK-NEXT: .p2align 5
|
|
; CHECK-NEXT: .LBB0_1: # %monotonic.i
|
|
; CHECK-NEXT: #
|
|
; CHECK-NEXT: lwz 5, 0(4)
|
|
; CHECK-NEXT: cmpwi 1, 3, 0
|
|
; CHECK-NEXT: addi 3, 3, -1
|
|
; CHECK-NEXT: andi. 5, 5, 255
|
|
; CHECK-NEXT: crorc 20, 6, 2
|
|
; CHECK-NEXT: bc 4, 20, .LBB0_1
|
|
; CHECK-NEXT: # %bb.2: # %if.end
|
|
; CHECK-NEXT: crnot 20, 2
|
|
; CHECK-NEXT: li 3, 0
|
|
; CHECK-NEXT: li 4, 8
|
|
; CHECK-NEXT: isel 30, 4, 3, 20
|
|
; CHECK-NEXT: addis 3, 2, aj@got@tlsgd@ha
|
|
; CHECK-NEXT: addi 3, 3, aj@got@tlsgd@l
|
|
; CHECK-NEXT: bl __tls_get_addr(aj@tlsgd)
|
|
; CHECK-NEXT: nop
|
|
; CHECK-NEXT: li 4, 1
|
|
; CHECK-NEXT: stdx 4, 3, 30
|
|
; CHECK-NEXT: addi 1, 1, 48
|
|
; CHECK-NEXT: ld 0, 16(1)
|
|
; CHECK-NEXT: ld 30, -16(1) # 8-byte Folded Reload
|
|
; CHECK-NEXT: mtlr 0
|
|
; CHECK-NEXT: blr
|
|
entry:
|
|
%0 = load i32, ptr undef, align 4
|
|
br label %monotonic.i
|
|
|
|
for.cond.i: ; preds = %monotonic.i
|
|
%exitcond.not = icmp eq i32 %inc.i, %0
|
|
br i1 %exitcond.not, label %if.end, label %monotonic.i
|
|
|
|
monotonic.i: ; preds = %for.cond.i, %entry
|
|
%i.018.i = phi i32 [ %inc.i, %for.cond.i ], [ 0, %entry ]
|
|
%1 = load atomic i32, ptr @g monotonic, align 4
|
|
%conv.i = trunc i32 %1 to i8
|
|
%tobool.not.i = icmp eq i8 %conv.i, 0
|
|
%inc.i = add nuw nsw i32 %i.018.i, 1
|
|
br i1 %tobool.not.i, label %for.cond.i, label %if.end
|
|
|
|
if.end: ; preds = %monotonic.i, %for.cond.i
|
|
%.sink = phi ptr [ getelementptr inbounds (%struct.t.1.13.29.45.105.109.113.189, ptr @aj, i64 0, i32 1), %monotonic.i ], [ @aj, %for.cond.i ]
|
|
store i64 1, ptr %.sink, align 8
|
|
ret void
|
|
}
|
|
|
|
attributes #0 = { nounwind }
|