mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2025-02-01 13:11:39 +01:00
f1558cfb9d
When SIFixSGPRCopies attempts to fix an illegal copy from vector to scalar register it calls moveToVALU(). A copy from an agpr to sgpr becomes a copy from agpr to agpr, which may result in the illegal register class at a use of this copy. Solution is to copy it always into a vgpr. This may result in a subsequent copy into an agpr if that is what really needed, however should not happen too often and likely will be folded later. The opposite situation may not happen because an sgpr is always illegal where agpr is legal, so such user instructions may not exist. Differential Revision: https://reviews.llvm.org/D68358 llvm-svn: 373544
16 lines
511 B
YAML
16 lines
511 B
YAML
# RUN: llc -march=amdgcn -mcpu=gfx908 -o - -run-pass si-fix-sgpr-copies -verify-machineinstrs %s | FileCheck -check-prefix=GCN %s
|
|
|
|
# GCN-LABEL: fold_acc_copy_into_valu
|
|
# GCN: [[COPY:%[0-9]+]]:vgpr_32 = COPY %0.sub0
|
|
# GCN: %2:vgpr_32 = V_AND_B32_e32 [[COPY]], undef %3:vgpr_32, implicit $exec
|
|
---
|
|
name: fold_acc_copy_into_valu
|
|
body: |
|
|
bb.0.entry:
|
|
|
|
%0:areg_1024 = IMPLICIT_DEF
|
|
%1:sreg_32_xm0 = COPY %0.sub0
|
|
%3:vgpr_32 = V_AND_B32_e32 %1, undef %2:vgpr_32, implicit $exec
|
|
|
|
...
|