RPCS3/llvm
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm.git synced 2025-02-01 08:12:12 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Fixed a bug in the ExecutionDependencyFix pass that caused dependencies to not propagate through implicit defs.

Browse Source 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@165102 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Silviu Baranga 2012-10-03 08:29:36 +00:00
parent 322e9ba2cb
commit 541a858f1a
2 changed files with 28 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
lib/CodeGen/ExecutionDepsFix.cpp
View File

@ -626,9 +626,12 @@ void ExeDepsFix::visitSoftInstr(MachineInstr *mi, unsigned mask) {
}
dv->Instrs.push_back(mi);
// Finally set all defs and non-collapsed uses to dv.
for (unsigned i = 0, e = mi->getDesc().getNumOperands(); i != e; ++i) {
MachineOperand &mo = mi->getOperand(i);
// Finally set all defs and non-collapsed uses to dv. We must iterate through
// all the operators, including imp-def ones.
for (MachineInstr::mop_iterator ii = mi->operands_begin(),
ee = mi->operands_end();
ii != ee; ++ii) {
MachineOperand &mo = *ii;
if (!mo.isReg()) continue;
int rx = regIndex(mo.getReg());
if (rx < 0) continue;

22
test/CodeGen/ARM/deps-fix.ll Normal file
View File

@ -0,0 +1,22 @@
; RUN: llc < %s -march=arm -mcpu=cortex-a9 -mattr=+neon,+neonfp -float-abi=hard -mtriple armv7-linux-gnueabi | FileCheck %s
;; This test checks that the ExecutionDepsFix pass performs the domain changes
;; even when some dependencies are propagated through implicit definitions.
; CHECK: fun_a
define <4 x float> @fun_a(<4 x float> %in, <4 x float> %x, float %y) nounwind {
; CHECK: vext
; CHECK: vext
; CHECK: vadd.f32
%1 = insertelement <4 x float> %in, float %y, i32 0
%2 = fadd <4 x float> %1, %x
ret <4 x float> %2
}
; CHECK: fun_b
define <4 x i32> @fun_b(<4 x i32> %in, <4 x i32> %x, i32 %y) nounwind {
; CHECK: vmov.32
; CHECK: vadd.i32
%1 = insertelement <4 x i32> %in, i32 %y, i32 0
%2 = add <4 x i32> %1, %x
ret <4 x i32> %2
}