RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-23 03:02:36 +01:00
Code Issues Projects Releases Wiki Activity

[SelectionDAG] memcpy expansion of const volatile struct ignores const zero

Browse Source 
In getMemcpyLoadsAndStores(), a memcpy where the source is a zero constant is expanded to a MemOp::Set instead of a MemOp::Copy, even when the memcpy is volatile.
This is incorrect.

The fix is to add a check for volatile, and expand to MemOp::Copy in the volatile case.

Reviewed By: chill

Differential Revision: https://reviews.llvm.org/D87134
This commit is contained in:
Simon Wallis 2020-09-07 13:21:27 +01:00
parent 2beb2ca87b
commit 63f9b9c858
2 changed files with 20 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
lib/CodeGen/SelectionDAG/SelectionDAG.cpp
View File

@ -6037,7 +6037,8 @@ static SDValue getMemcpyLoadsAndStores(SelectionDAG &DAG, const SDLoc &dl,
SrcAlign = Alignment;
assert(SrcAlign && "SrcAlign must be set");
ConstantDataArraySlice Slice;
bool CopyFromConstant = isMemSrcFromConstant(Src, Slice);
// If marked as volatile, perform a copy even when marked as constant.
bool CopyFromConstant = !isVol && isMemSrcFromConstant(Src, Slice);
bool isZeroConstant = CopyFromConstant && Slice.Array == nullptr;
unsigned Limit = AlwaysInline ? ~0U : TLI.getMaxStoresPerMemcpy(OptSize);
const MemOp Op = isZeroConstant

18
test/CodeGen/ARM/memcpy-const-vol-struct.ll Normal file
View File

@ -0,0 +1,18 @@
; RUN: llc -mtriple=armv7-arm-none-eabi -o - %s | FileCheck %s
%struct.sMyType = type { i32 }
@val = hidden constant %struct.sMyType zeroinitializer, align 4
@v = internal global %struct.sMyType zeroinitializer, align 4
define hidden void @InitVal() local_unnamed_addr {
entry:
call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 4 bitcast (%struct.sMyType* @v to i8*), i8* align 4 bitcast (%struct.sMyType* @val to i8*), i32 4, i1 true)
; The last argument is the isvolatile argument. This is a volatile memcpy.
; Test that the memcpy expansion does not optimize away the load.
; CHECK: ldr
; CHECK: str
ret void
}
declare void @llvm.memcpy.p0i8.p0i8.i32(i8* noalias nocapture writeonly, i8* noalias nocapture readonly, i32, i1 immarg)