Skip to content

AMDGPU: Fold frame indexes into s_or_b32 and s_and_b32 #102345

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

AMDGPU: Fold frame indexes into s_or_b32 and s_and_b32 #102345

wants to merge 1 commit into from
+32 −35

Conversation

@arsenm
Copy link
Contributor

@arsenm arsenm commented Aug 7, 2024

Some pointer adds get turned into ors, and sometimes and is
performed on pointers for masking.

@arsenm arsenm mentioned this pull request Aug 7, 2024
Merged
@arsenm Graphite App
Copy link
Contributor Author

arsenm commented Aug 7, 2024 •
edited
Loading

This stack of pull requests is managed by Graphite. Learn more about stacking.

@arsenm arsenm mentioned this pull request Aug 7, 2024
Merged
@arsenm arsenm marked this pull request as ready for review August 7, 2024 18:06
@llvmbot
Copy link
Member

llvmbot commented Aug 7, 2024 •
edited
Loading

@llvm/pr-subscribers-llvm-globalisel

@llvm/pr-subscribers-backend-amdgpu

Author: Matt Arsenault (arsenm)

Changes

Some pointer adds get turned into ors, and sometimes and is
performed on pointers for masking.

Full diff: https://github.com/llvm/llvm-project/pull/102345.diff

2 Files Affected:

  • (modified) llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp (+3-2)
  • (modified) llvm/test/CodeGen/AMDGPU/eliminate-frame-index-scalar-bit-ops.mir (+42-46)
diff --git a/llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp b/llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp
index e8c2cbd3dd671..76da1f0eb4f7d 100644
--- a/llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp
+++ b/llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp
@@ -2268,8 +2268,9 @@ bool SIRegisterInfo::eliminateFrameIndex(MachineBasicBlock::iterator MI,
       MI->eraseFromParent();
       return true;
     }
-    case AMDGPU::S_ADD_I32: {
-      // TODO: Handle s_or_b32, s_and_b32.
+    case AMDGPU::S_ADD_I32:
+    case AMDGPU::S_OR_B32:
+    case AMDGPU::S_AND_B32: {
       MachineOperand &OtherOp = MI->getOperand(FIOperandNum == 1 ? 2 : 1);
 
       assert(FrameReg || MFI->isBottomOfStack());
diff --git a/llvm/test/CodeGen/AMDGPU/eliminate-frame-index-scalar-bit-ops.mir b/llvm/test/CodeGen/AMDGPU/eliminate-frame-index-scalar-bit-ops.mir
index 1456bbc369b6a..f7f43b59085de 100644
--- a/llvm/test/CodeGen/AMDGPU/eliminate-frame-index-scalar-bit-ops.mir
+++ b/llvm/test/CodeGen/AMDGPU/eliminate-frame-index-scalar-bit-ops.mir
@@ -21,21 +21,21 @@ machineFunctionInfo:
 body:             |
   bb.0:
     ; MUBUFW64-LABEL: name: s_or_b32__inline_imm__fi_offset0
-    ; MUBUFW64: $sgpr4 = S_LSHR_B32 $sgpr32, 6, implicit-def dead $scc
-    ; MUBUFW64-NEXT: renamable $sgpr7 = S_OR_B32 12, killed $sgpr4, implicit-def $scc
+    ; MUBUFW64: renamable $sgpr4 = S_LSHR_B32 $sgpr32, 6, implicit-def dead $scc
+    ; MUBUFW64-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr4, 12, implicit-def $scc
     ; MUBUFW64-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     ;
     ; MUBUFW32-LABEL: name: s_or_b32__inline_imm__fi_offset0
-    ; MUBUFW32: $sgpr4 = S_LSHR_B32 $sgpr32, 5, implicit-def dead $scc
-    ; MUBUFW32-NEXT: renamable $sgpr7 = S_OR_B32 12, killed $sgpr4, implicit-def $scc
+    ; MUBUFW32: renamable $sgpr4 = S_LSHR_B32 $sgpr32, 5, implicit-def dead $scc
+    ; MUBUFW32-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr4, 12, implicit-def $scc
     ; MUBUFW32-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     ;
     ; FLATSCRW64-LABEL: name: s_or_b32__inline_imm__fi_offset0
-    ; FLATSCRW64: renamable $sgpr7 = S_OR_B32 12, $sgpr32, implicit-def $scc
+    ; FLATSCRW64: renamable $sgpr7 = S_OR_B32 $sgpr32, 12, implicit-def $scc
     ; FLATSCRW64-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     ;
     ; FLATSCRW32-LABEL: name: s_or_b32__inline_imm__fi_offset0
-    ; FLATSCRW32: renamable $sgpr7 = S_OR_B32 12, $sgpr32, implicit-def $scc
+    ; FLATSCRW32: renamable $sgpr7 = S_OR_B32 $sgpr32, 12, implicit-def $scc
     ; FLATSCRW32-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     renamable $sgpr7 = S_OR_B32 12, %stack.0, implicit-def $scc
     SI_RETURN implicit $sgpr7, implicit $scc
@@ -55,25 +55,21 @@ machineFunctionInfo:
 body:             |
   bb.0:
     ; MUBUFW64-LABEL: name: s_or_b32__literal__fi_offset96
-    ; MUBUFW64: $sgpr4 = S_LSHR_B32 $sgpr32, 6, implicit-def $scc
-    ; MUBUFW64-NEXT: $sgpr4 = S_ADD_I32 killed $sgpr4, 96, implicit-def $scc
-    ; MUBUFW64-NEXT: renamable $sgpr7 = S_OR_B32 68, killed $sgpr4, implicit-def $scc
+    ; MUBUFW64: renamable $sgpr4 = S_LSHR_B32 $sgpr32, 6, implicit-def dead $scc
+    ; MUBUFW64-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr4, 164, implicit-def $scc
     ; MUBUFW64-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     ;
     ; MUBUFW32-LABEL: name: s_or_b32__literal__fi_offset96
-    ; MUBUFW32: $sgpr4 = S_LSHR_B32 $sgpr32, 5, implicit-def $scc
-    ; MUBUFW32-NEXT: $sgpr4 = S_ADD_I32 killed $sgpr4, 96, implicit-def $scc
-    ; MUBUFW32-NEXT: renamable $sgpr7 = S_OR_B32 68, killed $sgpr4, implicit-def $scc
+    ; MUBUFW32: renamable $sgpr4 = S_LSHR_B32 $sgpr32, 5, implicit-def dead $scc
+    ; MUBUFW32-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr4, 164, implicit-def $scc
     ; MUBUFW32-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     ;
     ; FLATSCRW64-LABEL: name: s_or_b32__literal__fi_offset96
-    ; FLATSCRW64: $sgpr4 = S_ADD_I32 $sgpr32, 96, implicit-def $scc
-    ; FLATSCRW64-NEXT: renamable $sgpr7 = S_OR_B32 68, killed $sgpr4, implicit-def $scc
+    ; FLATSCRW64: renamable $sgpr7 = S_OR_B32 $sgpr32, 164, implicit-def $scc
     ; FLATSCRW64-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     ;
     ; FLATSCRW32-LABEL: name: s_or_b32__literal__fi_offset96
-    ; FLATSCRW32: $sgpr4 = S_ADD_I32 $sgpr32, 96, implicit-def $scc
-    ; FLATSCRW32-NEXT: renamable $sgpr7 = S_OR_B32 68, killed $sgpr4, implicit-def $scc
+    ; FLATSCRW32: renamable $sgpr7 = S_OR_B32 $sgpr32, 164, implicit-def $scc
     ; FLATSCRW32-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     renamable $sgpr7 = S_OR_B32 68, %stack.1, implicit-def $scc
     SI_RETURN implicit $sgpr7, implicit $scc
@@ -96,31 +92,31 @@ body:             |
     ; MUBUFW64-LABEL: name: s_or_b32__sgpr__fi_literal_offset
     ; MUBUFW64: liveins: $sgpr8
     ; MUBUFW64-NEXT: {{  $}}
-    ; MUBUFW64-NEXT: $sgpr4 = S_LSHR_B32 $sgpr32, 6, implicit-def $scc
-    ; MUBUFW64-NEXT: $sgpr4 = S_ADD_I32 killed $sgpr4, 80, implicit-def $scc
-    ; MUBUFW64-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr8, killed $sgpr4, implicit-def $scc
+    ; MUBUFW64-NEXT: renamable $sgpr4 = S_LSHR_B32 $sgpr32, 6, implicit-def dead $scc
+    ; MUBUFW64-NEXT: renamable $sgpr7 = S_ADD_I32 $sgpr4, $sgpr8, implicit-def $scc
+    ; MUBUFW64-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr7, 80, implicit-def $scc
     ; MUBUFW64-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     ;
     ; MUBUFW32-LABEL: name: s_or_b32__sgpr__fi_literal_offset
     ; MUBUFW32: liveins: $sgpr8
     ; MUBUFW32-NEXT: {{  $}}
-    ; MUBUFW32-NEXT: $sgpr4 = S_LSHR_B32 $sgpr32, 5, implicit-def $scc
-    ; MUBUFW32-NEXT: $sgpr4 = S_ADD_I32 killed $sgpr4, 80, implicit-def $scc
-    ; MUBUFW32-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr8, killed $sgpr4, implicit-def $scc
+    ; MUBUFW32-NEXT: renamable $sgpr4 = S_LSHR_B32 $sgpr32, 5, implicit-def dead $scc
+    ; MUBUFW32-NEXT: renamable $sgpr7 = S_ADD_I32 $sgpr4, $sgpr8, implicit-def $scc
+    ; MUBUFW32-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr7, 80, implicit-def $scc
     ; MUBUFW32-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     ;
     ; FLATSCRW64-LABEL: name: s_or_b32__sgpr__fi_literal_offset
     ; FLATSCRW64: liveins: $sgpr8
     ; FLATSCRW64-NEXT: {{  $}}
-    ; FLATSCRW64-NEXT: $sgpr4 = S_ADD_I32 $sgpr32, 80, implicit-def $scc
-    ; FLATSCRW64-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr8, killed $sgpr4, implicit-def $scc
+    ; FLATSCRW64-NEXT: renamable $sgpr7 = S_ADD_I32 $sgpr32, $sgpr8, implicit-def $scc
+    ; FLATSCRW64-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr7, 80, implicit-def $scc
     ; FLATSCRW64-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     ;
     ; FLATSCRW32-LABEL: name: s_or_b32__sgpr__fi_literal_offset
     ; FLATSCRW32: liveins: $sgpr8
     ; FLATSCRW32-NEXT: {{  $}}
-    ; FLATSCRW32-NEXT: $sgpr4 = S_ADD_I32 $sgpr32, 80, implicit-def $scc
-    ; FLATSCRW32-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr8, killed $sgpr4, implicit-def $scc
+    ; FLATSCRW32-NEXT: renamable $sgpr7 = S_ADD_I32 $sgpr32, $sgpr8, implicit-def $scc
+    ; FLATSCRW32-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr7, 80, implicit-def $scc
     ; FLATSCRW32-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     renamable $sgpr7 = S_OR_B32 $sgpr8, %stack.1, implicit-def $scc
     SI_RETURN implicit $sgpr7, implicit $scc
@@ -143,31 +139,31 @@ body:             |
     ; MUBUFW64-LABEL: name: s_or_b32__sgpr__fi_inlineimm_offset
     ; MUBUFW64: liveins: $sgpr8
     ; MUBUFW64-NEXT: {{  $}}
-    ; MUBUFW64-NEXT: $sgpr4 = S_LSHR_B32 $sgpr32, 6, implicit-def $scc
-    ; MUBUFW64-NEXT: $sgpr4 = S_ADD_I32 killed $sgpr4, 32, implicit-def $scc
-    ; MUBUFW64-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr8, killed $sgpr4, implicit-def $scc
+    ; MUBUFW64-NEXT: renamable $sgpr4 = S_LSHR_B32 $sgpr32, 6, implicit-def dead $scc
+    ; MUBUFW64-NEXT: renamable $sgpr7 = S_ADD_I32 $sgpr4, $sgpr8, implicit-def $scc
+    ; MUBUFW64-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr7, 32, implicit-def $scc
     ; MUBUFW64-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     ;
     ; MUBUFW32-LABEL: name: s_or_b32__sgpr__fi_inlineimm_offset
     ; MUBUFW32: liveins: $sgpr8
     ; MUBUFW32-NEXT: {{  $}}
-    ; MUBUFW32-NEXT: $sgpr4 = S_LSHR_B32 $sgpr32, 5, implicit-def $scc
-    ; MUBUFW32-NEXT: $sgpr4 = S_ADD_I32 killed $sgpr4, 32, implicit-def $scc
-    ; MUBUFW32-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr8, killed $sgpr4, implicit-def $scc
+    ; MUBUFW32-NEXT: renamable $sgpr4 = S_LSHR_B32 $sgpr32, 5, implicit-def dead $scc
+    ; MUBUFW32-NEXT: renamable $sgpr7 = S_ADD_I32 $sgpr4, $sgpr8, implicit-def $scc
+    ; MUBUFW32-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr7, 32, implicit-def $scc
     ; MUBUFW32-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     ;
     ; FLATSCRW64-LABEL: name: s_or_b32__sgpr__fi_inlineimm_offset
     ; FLATSCRW64: liveins: $sgpr8
     ; FLATSCRW64-NEXT: {{  $}}
-    ; FLATSCRW64-NEXT: $sgpr4 = S_ADD_I32 $sgpr32, 32, implicit-def $scc
-    ; FLATSCRW64-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr8, killed $sgpr4, implicit-def $scc
+    ; FLATSCRW64-NEXT: renamable $sgpr7 = S_ADD_I32 $sgpr32, $sgpr8, implicit-def $scc
+    ; FLATSCRW64-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr7, 32, implicit-def $scc
     ; FLATSCRW64-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     ;
     ; FLATSCRW32-LABEL: name: s_or_b32__sgpr__fi_inlineimm_offset
     ; FLATSCRW32: liveins: $sgpr8
     ; FLATSCRW32-NEXT: {{  $}}
-    ; FLATSCRW32-NEXT: $sgpr4 = S_ADD_I32 $sgpr32, 32, implicit-def $scc
-    ; FLATSCRW32-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr8, killed $sgpr4, implicit-def $scc
+    ; FLATSCRW32-NEXT: renamable $sgpr7 = S_ADD_I32 $sgpr32, $sgpr8, implicit-def $scc
+    ; FLATSCRW32-NEXT: renamable $sgpr7 = S_OR_B32 $sgpr7, 32, implicit-def $scc
     ; FLATSCRW32-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     renamable $sgpr7 = S_OR_B32 $sgpr8, %stack.1, implicit-def $scc
     SI_RETURN implicit $sgpr7, implicit $scc
@@ -190,31 +186,31 @@ body:             |
     ; MUBUFW64-LABEL: name: s_and_b32__sgpr__fi_literal_offset
     ; MUBUFW64: liveins: $sgpr8
     ; MUBUFW64-NEXT: {{  $}}
-    ; MUBUFW64-NEXT: $sgpr4 = S_LSHR_B32 $sgpr32, 6, implicit-def $scc
-    ; MUBUFW64-NEXT: $sgpr4 = S_ADD_I32 killed $sgpr4, 80, implicit-def $scc
-    ; MUBUFW64-NEXT: renamable $sgpr7 = S_AND_B32 $sgpr8, killed $sgpr4, implicit-def $scc
+    ; MUBUFW64-NEXT: renamable $sgpr4 = S_LSHR_B32 $sgpr32, 6, implicit-def dead $scc
+    ; MUBUFW64-NEXT: renamable $sgpr7 = S_ADD_I32 $sgpr4, $sgpr8, implicit-def $scc
+    ; MUBUFW64-NEXT: renamable $sgpr7 = S_AND_B32 $sgpr7, 80, implicit-def $scc
     ; MUBUFW64-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     ;
     ; MUBUFW32-LABEL: name: s_and_b32__sgpr__fi_literal_offset
     ; MUBUFW32: liveins: $sgpr8
     ; MUBUFW32-NEXT: {{  $}}
-    ; MUBUFW32-NEXT: $sgpr4 = S_LSHR_B32 $sgpr32, 5, implicit-def $scc
-    ; MUBUFW32-NEXT: $sgpr4 = S_ADD_I32 killed $sgpr4, 80, implicit-def $scc
-    ; MUBUFW32-NEXT: renamable $sgpr7 = S_AND_B32 $sgpr8, killed $sgpr4, implicit-def $scc
+    ; MUBUFW32-NEXT: renamable $sgpr4 = S_LSHR_B32 $sgpr32, 5, implicit-def dead $scc
+    ; MUBUFW32-NEXT: renamable $sgpr7 = S_ADD_I32 $sgpr4, $sgpr8, implicit-def $scc
+    ; MUBUFW32-NEXT: renamable $sgpr7 = S_AND_B32 $sgpr7, 80, implicit-def $scc
     ; MUBUFW32-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     ;
     ; FLATSCRW64-LABEL: name: s_and_b32__sgpr__fi_literal_offset
     ; FLATSCRW64: liveins: $sgpr8
     ; FLATSCRW64-NEXT: {{  $}}
-    ; FLATSCRW64-NEXT: $sgpr4 = S_ADD_I32 $sgpr32, 80, implicit-def $scc
-    ; FLATSCRW64-NEXT: renamable $sgpr7 = S_AND_B32 $sgpr8, killed $sgpr4, implicit-def $scc
+    ; FLATSCRW64-NEXT: renamable $sgpr7 = S_ADD_I32 $sgpr32, $sgpr8, implicit-def $scc
+    ; FLATSCRW64-NEXT: renamable $sgpr7 = S_AND_B32 $sgpr7, 80, implicit-def $scc
     ; FLATSCRW64-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     ;
     ; FLATSCRW32-LABEL: name: s_and_b32__sgpr__fi_literal_offset
     ; FLATSCRW32: liveins: $sgpr8
     ; FLATSCRW32-NEXT: {{  $}}
-    ; FLATSCRW32-NEXT: $sgpr4 = S_ADD_I32 $sgpr32, 80, implicit-def $scc
-    ; FLATSCRW32-NEXT: renamable $sgpr7 = S_AND_B32 $sgpr8, killed $sgpr4, implicit-def $scc
+    ; FLATSCRW32-NEXT: renamable $sgpr7 = S_ADD_I32 $sgpr32, $sgpr8, implicit-def $scc
+    ; FLATSCRW32-NEXT: renamable $sgpr7 = S_AND_B32 $sgpr7, 80, implicit-def $scc
     ; FLATSCRW32-NEXT: SI_RETURN implicit $sgpr7, implicit $scc
     renamable $sgpr7 = S_AND_B32 $sgpr8, %stack.1, implicit-def $scc
     SI_RETURN implicit $sgpr7, implicit $scc

rampitec
rampitec reviewed Aug 7, 2024
View reviewed changes
llvm/test/CodeGen/AMDGPU/eliminate-frame-index-scalar-bit-ops.mir Outdated
; FLATSCRW32-NEXT: {{ $}}
; FLATSCRW32-NEXT: $sgpr4 = S_ADD_I32 $sgpr32, 80, implicit-def $scc
; FLATSCRW32-NEXT: renamable $sgpr7 = S_AND_B32 $sgpr8, killed $sgpr4, implicit-def $scc
; FLATSCRW32-NEXT: renamable $sgpr7 = S_ADD_I32 $sgpr32, $sgpr8, implicit-def $scc
Copy link
Collaborator

@rampitec rampitec Aug 7, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I do not understand this. The transformation is (s8 & (sp + 80)) ->((s8 + sp) & 80) does not look immediately obvious.

@arsenm arsenm force-pushed the users/arsenm/amdgpu-eliminateframeindex-s-add-i32 branch 3 times, most recently from 8726c2b to b9a85ce Compare August 20, 2024 11:54
Base automatically changed from users/arsenm/amdgpu-eliminateframeindex-s-add-i32 to main August 22, 2024 05:16
@arsenm
Copy link
Contributor Author

arsenm commented Sep 30, 2024

For the or case, I think we need to check the disjoint flag on the instruction (and also transform the final or into an add, if we can fold the offsets)

arsenm added a commit that referenced this pull request Mar 4, 2025
@arsenm
AMDGPU: Make frame index folding logic consistent with eliminateFrame…
259d50b 
…Index

This adds handling of s_add_u32, which is handled and removes handling of
s_or_b32 and s_and_b32, which are not. I was working on handling them
in #102345, but need to finish that patch. This fixes a regression
exposed by a316539 where the
final instruction would use two literals.
@arsenm arsenm mentioned this pull request Mar 4, 2025
Merged
arsenm added a commit that referenced this pull request Mar 5, 2025
@arsenm
AMDGPU: Make frame index folding logic consistent with eliminateFrame…
d48682d 
…Index

This adds handling of s_add_u32, which is handled and removes handling of
s_or_b32 and s_and_b32, which are not. I was working on handling them
in #102345, but need to finish that patch. This fixes a regression
exposed by a316539 where the
final instruction would use two literals.
arsenm added a commit that referenced this pull request Mar 5, 2025
@arsenm
AMDGPU: Make frame index folding logic consistent with eliminateFrame…
3e53aea 
…Index (#129633)

This adds handling of s_add_u32, which is handled and removes handling of
s_or_b32 and s_and_b32, which are not. I was working on handling them
in #102345, but need to finish that patch. This fixes a regression
exposed by a316539 where the
final instruction would use two literals.
@arsenm
AMDGPU: Fold frame indexes into s_or_b32
f811338 
Some pointer adds get turned into ors, and sometimes and is
performed on pointers for masking.
@arsenm arsenm force-pushed the users/arsenm/amdgpu-fold-frameindex-s-or-and branch from ac17eed to f811338 Compare June 20, 2025 11:49
changpeng
changpeng reviewed Jun 21, 2025
View reviewed changes
llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp
case AMDGPU::S_ADD_U32: {
// TODO: Handle s_or_b32, s_and_b32.
case AMDGPU::S_OR_B32: {
// TODO: Handle s_and_b32 for ptrmask
Copy link
Contributor

@changpeng changpeng Jun 21, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I do not fully understand how can the Offset be folded into the other operand the same way as add:
OtherOp.setImm(OtherOp.getImm() + Offset);

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@changpeng changpeng changpeng left review comments

@rampitec rampitec rampitec left review comments

@cdevadas cdevadas Awaiting requested review from cdevadas

@jayfoad jayfoad Awaiting requested review from jayfoad

@jwanggit86 jwanggit86 Awaiting requested review from jwanggit86

@mariusz-sikora-at-amd mariusz-sikora-at-amd Awaiting requested review from mariusz-sikora-at-amd

@pgodeq pgodeq Awaiting requested review from pgodeq

@Pierre-vh Pierre-vh Awaiting requested review from Pierre-vh

@rovka rovka Awaiting requested review from rovka

@Sisyph Sisyph Awaiting requested review from Sisyph

Assignees

No one assigned

Labels

backend:AMDGPU llvm:globalisel

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

5 participants

@arsenm @llvmbot @changpeng @rampitec