RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-02-01 05:01:59 +01:00
Code Issues Projects Releases Wiki Activity

AMDGPU/SI: Fix s_waitcnt insertion for flat instructions

Browse Source 
Summary:
This was broken in r260694 which swapped the address and data operands
for flat store instructions.  The code in SIInsertWaits assumes
that the data operand always comes before the address operand, so
we need to add a special case for flat.

Reviewers: arsenm

Subscribers: arsenm, llvm-commits

Differential Revision: http://reviews.llvm.org/D17366

llvm-svn: 261330
This commit is contained in:
Tom Stellard 2016-02-19 15:33:13 +00:00
parent 1f006b6e42
commit db5af50c55
2 changed files with 20 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
lib/Target/AMDGPU/SIInsertWaits.cpp
View File

@ -231,15 +231,17 @@ bool SIInsertWaits::isOpRelevant(MachineOperand &Op) {
return false;
// Check if this operand is the value being stored.
// Special case for DS instructions, since the address
// Special case for DS/FLAT instructions, since the address
// operand comes before the value operand and it may have
// multiple data operands.
if (TII->isDS(MI)) {
if (TII->isDS(MI) || TII->isFLAT(MI)) {
MachineOperand *Data = TII->getNamedOperand(MI, AMDGPU::OpName::data);
if (Data && Op.isIdenticalTo(*Data))
return true;
}
if (TII->isDS(MI)) {
MachineOperand *Data0 = TII->getNamedOperand(MI, AMDGPU::OpName::data0);
if (Data0 && Op.isIdenticalTo(*Data0))
return true;

16
test/CodeGen/AMDGPU/waitcnt-flat.ll Normal file
View File

@ -0,0 +1,16 @@
; RUN: llc < %s -mtriple=amdgcn--amdhsa -mcpu=kaveri | FileCheck --check-prefix=GCN %s
; RUN: llc < %s -mtriple=amdgcn--amdhsa -mcpu=fiji | FileCheck --check-prefix=GCN %s
; If flat_store_dword and flat_load_dword use different registers for the data
; operand, this test is not broken. It just means it is no longer testing
; for the original bug.
; GCN: {{^}}test:
; GCN: flat_store_dword v[{{[0-9]+:[0-9]+}}], [[DATA:v[0-9]+]]
; GCN: s_waitcnt vmcnt(0) lgkmcnt(0)
; GCN: flat_load_dword [[DATA]], v[{{[0-9]+:[0-9]+}}]
define void @test(i32 addrspace(1)* %out, i32 %in) {
store volatile i32 0, i32 addrspace(1)* %out
%val = load volatile i32, i32 addrspace(1)* %out
ret void
}