RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-23 11:13:28 +01:00
Code Issues Projects Releases Wiki Activity

[GlobalISel] Don't emit multiply by magic constant for zero memset values.

Browse Source
This commit is contained in:
Amara Emerson 2020-06-15 14:37:55 -07:00
parent 534cba2336
commit b89b9a9eed
2 changed files with 44 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
lib/CodeGen/GlobalISel/CombinerHelper.cpp
View File

@ -936,6 +936,11 @@ static Register getMemsetValue(Register Val, LLT Ty, MachineIRBuilder &MIB) {
// Extend the byte value to the larger type, and then multiply by a magic
// value 0x010101... in order to replicate it across every byte.
// Unless it's zero, in which case just emit a larger G_CONSTANT 0.
if (ValVRegAndVal && ValVRegAndVal->Value == 0) {
return MIB.buildConstant(Ty, 0).getReg(0);
}
LLT ExtType = Ty.getScalarType();
auto ZExt = MIB.buildZExtOrTrunc(ExtType, Val);
if (NumBits > 8) {

39
test/CodeGen/AArch64/GlobalISel/inline-memset.mir
View File

@ -21,6 +21,12 @@
ret void
}
define void @test_zero_const(i8* nocapture %dst) local_unnamed_addr #0 {
entry:
tail call void @llvm.memset.p0i8.i64(i8* align 1 %dst, i8 0, i64 64, i1 false)
ret void
}
define void @test_ms3_const_both(i8* nocapture %dst) local_unnamed_addr #0 {
entry:
tail call void @llvm.memset.p0i8.i64(i8* align 1 %dst, i8 64, i64 16, i1 false)
@ -101,6 +107,39 @@ body: |
RET_ReallyLR
...
---
name: test_zero_const
alignment: 4
tracksRegLiveness: true
body: |
bb.1.entry:
liveins: $w1, $x0
; CHECK-LABEL: name: test_zero_const
; CHECK: liveins: $w1, $x0
; CHECK: [[COPY:%[0-9]+]]:_(p0) = COPY $x0
; CHECK: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
; CHECK: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[C]](s64), [[C]](s64)
; CHECK: G_STORE [[BUILD_VECTOR]](<2 x s64>), [[COPY]](p0) :: (store 16 into %ir.dst, align 1)
; CHECK: [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 16
; CHECK: [[PTR_ADD:%[0-9]+]]:_(p0) = G_PTR_ADD [[COPY]], [[C1]](s64)
; CHECK: G_STORE [[BUILD_VECTOR]](<2 x s64>), [[PTR_ADD]](p0) :: (store 16 into %ir.dst + 16, align 1)
; CHECK: [[C2:%[0-9]+]]:_(s64) = G_CONSTANT i64 32
; CHECK: [[PTR_ADD1:%[0-9]+]]:_(p0) = G_PTR_ADD [[COPY]], [[C2]](s64)
; CHECK: G_STORE [[BUILD_VECTOR]](<2 x s64>), [[PTR_ADD1]](p0) :: (store 16 into %ir.dst + 32, align 1)
; CHECK: [[C3:%[0-9]+]]:_(s64) = G_CONSTANT i64 48
; CHECK: [[PTR_ADD2:%[0-9]+]]:_(p0) = G_PTR_ADD [[COPY]], [[C3]](s64)
; CHECK: G_STORE [[BUILD_VECTOR]](<2 x s64>), [[PTR_ADD2]](p0) :: (store 16 into %ir.dst + 48, align 1)
; CHECK: RET_ReallyLR
%0:_(p0) = COPY $x0
%1:_(s32) = G_CONSTANT i32 0
%3:_(s64) = G_CONSTANT i64 64
%2:_(s8) = G_TRUNC %1(s32)
G_INTRINSIC_W_SIDE_EFFECTS intrinsic(@llvm.memset), %0(p0), %2(s8), %3(s64), 1 :: (store 1 into %ir.dst)
RET_ReallyLR
...
---
name: test_ms3_const_both
alignment: 4