RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-01-31 20:51:52 +01:00
Code Issues Projects Releases Wiki Activity

[SystemZ] Bugfix in combineSTORE().

Browse Source 
Remember to check if store is truncating before calling
combineTruncateExtract().

Review: Ulrich Weigand
llvm-svn: 333262
This commit is contained in:
Jonas Paulsson 2018-05-25 09:01:23 +00:00
parent 6c63bb753f
commit adc691d8dc
2 changed files with 20 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
lib/Target/SystemZ/SystemZISelLowering.cpp
View File

@ -5357,7 +5357,7 @@ SDValue SystemZTargetLowering::combineSTORE(
// for the extraction to be done on a vMiN value, so that we can use VSTE. // for the extraction to be done on a vMiN value, so that we can use VSTE.
// If X has wider elements then convert it to: // If X has wider elements then convert it to:
// (truncstoreiN (extract_vector_elt (bitcast X), Y2), Z). // (truncstoreiN (extract_vector_elt (bitcast X), Y2), Z).
if (MemVT.isInteger()) { if (MemVT.isInteger() && SN->isTruncatingStore()) {
if (SDValue Value = if (SDValue Value =
combineTruncateExtract(SDLoc(N), MemVT, SN->getValue(), DCI)) { combineTruncateExtract(SDLoc(N), MemVT, SN->getValue(), DCI)) {
DCI.AddToWorklist(Value.getNode()); DCI.AddToWorklist(Value.getNode());

19
test/CodeGen/SystemZ/dag-combine-04.ll Normal file
View File

@ -0,0 +1,19 @@
; Test that SystemZTargetLowering::combineSTORE() does not crash due to not
; checking if store is actually a truncating store before calling
; combineTruncateExtract().
;
; RUN: llc -mtriple=s390x-linux-gnu -mcpu=z13 < %s
@g_348 = external dso_local unnamed_addr global [6 x [10 x i16]], align 2
define void @main() local_unnamed_addr {
bb:
%tmp = load i16, i16* getelementptr inbounds ([6 x [10 x i16]], [6 x [10 x i16]]* @g_348, i64 0, i64 1, i64 6), align 2
%tmp1 = xor i16 %tmp, 0
%tmp2 = insertelement <2 x i16> <i16 undef, i16 0>, i16 %tmp1, i32 0
%tmp3 = shufflevector <2 x i16> %tmp2, <2 x i16> undef, <2 x i32> <i32 1, i32 undef>
%tmp4 = xor <2 x i16> %tmp2, %tmp3
%tmp5 = extractelement <2 x i16> %tmp4, i32 0
store i16 %tmp5, i16* getelementptr inbounds ([6 x [10 x i16]], [6 x [10 x i16]]* @g_348, i64 0, i64 1, i64 6), align 2
unreachable
}