[llvm] eab9a4f - [AMDGPU] Don't assert on partial exec copy

Austin Kerbow via llvm-commits llvm-commits at lists.llvm.org
Sun Apr 12 21:15:02 PDT 2020


Author: Austin Kerbow
Date: 2020-04-12T21:14:36-07:00
New Revision: eab9a4f1192a42ded26e6d7e6a0eb51ee8256ebb

URL: https://github.com/llvm/llvm-project/commit/eab9a4f1192a42ded26e6d7e6a0eb51ee8256ebb
DIFF: https://github.com/llvm/llvm-project/commit/eab9a4f1192a42ded26e6d7e6a0eb51ee8256ebb.diff

LOG: [AMDGPU] Don't assert on partial exec copy

After Machine CSE and coalescing we can end up with copies of exec to
subregister SGPRs.

Differential Revision: https://reviews.llvm.org/D77992

Added: 
    

Modified: 
    llvm/lib/Target/AMDGPU/SIOptimizeExecMaskingPreRA.cpp
    llvm/test/CodeGen/AMDGPU/optimize-exec-masking-pre-ra.mir

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/AMDGPU/SIOptimizeExecMaskingPreRA.cpp b/llvm/lib/Target/AMDGPU/SIOptimizeExecMaskingPreRA.cpp
index 03e722105feb..7eb1ec941dbd 100644
--- a/llvm/lib/Target/AMDGPU/SIOptimizeExecMaskingPreRA.cpp
+++ b/llvm/lib/Target/AMDGPU/SIOptimizeExecMaskingPreRA.cpp
@@ -71,10 +71,8 @@ FunctionPass *llvm::createSIOptimizeExecMaskingPreRAPass() {
 static bool isFullExecCopy(const MachineInstr& MI, const GCNSubtarget& ST) {
   unsigned Exec = ST.isWave32() ? AMDGPU::EXEC_LO : AMDGPU::EXEC;
 
-  if (MI.isCopy() && MI.getOperand(1).getReg() == Exec) {
-    assert(MI.isFullCopy());
+  if (MI.isFullCopy() && MI.getOperand(1).getReg() == Exec)
     return true;
-  }
 
   return false;
 }

diff  --git a/llvm/test/CodeGen/AMDGPU/optimize-exec-masking-pre-ra.mir b/llvm/test/CodeGen/AMDGPU/optimize-exec-masking-pre-ra.mir
index fee6b52d1a11..ec774291067a 100644
--- a/llvm/test/CodeGen/AMDGPU/optimize-exec-masking-pre-ra.mir
+++ b/llvm/test/CodeGen/AMDGPU/optimize-exec-masking-pre-ra.mir
@@ -117,3 +117,25 @@ body:             |
   bb.1:
 
 ...
+
+# Don't crash on exec copy to SGPR subregister.
+---
+name: exec_copy_to_subreg
+tracksRegLiveness: true
+body:             |
+  ; GCN-LABEL: name: exec_copy_to_subreg
+  ; GCN: bb.0:
+  ; GCN:   successors: %bb.1(0x80000000)
+  ; GCN:   dead undef %0.sub0:sreg_256 = COPY $exec
+  ; GCN:   dead %1:vgpr_32 = V_CNDMASK_B32_e64 0, 0, 0, 1, undef %2:sreg_64_xexec, implicit $exec
+  ; GCN:   S_BRANCH %bb.1
+  ; GCN: bb.1:
+  bb.0:
+
+    undef %0.sub0:sreg_256 = COPY $exec
+    %2:vgpr_32 = V_CNDMASK_B32_e64 0, 0, 0, 1, undef %1:sreg_64_xexec, implicit $exec
+    S_BRANCH %bb.1
+
+  bb.1:
+
+...


        


More information about the llvm-commits mailing list