[Mlir-commits] [clang-tools-extra] [libclc] [lld] [flang] [mlir] [libcxx] [libunwind] [clang] [lldb] [libc] [llvm] [compiler-rt] [AMDGPU] Fix broken sign-extended subword buffer load combine (PR #77470)

Jay Foad llvmlistbot at llvm.org
Tue Jan 9 23:02:25 PST 2024


https://github.com/jayfoad updated https://github.com/llvm/llvm-project/pull/77470

>From ae231d88c5b5e2e0996edefd45389992f8e97d05 Mon Sep 17 00:00:00 2001
From: Jay Foad <jay.foad at amd.com>
Date: Tue, 9 Jan 2024 13:16:24 +0000
Subject: [PATCH 1/3] [AMDGPU] Precommit tests for broken combine

Add tests for sign-extending the result of an unsigned subword buffer
load from the wrong width.
---
 .../llvm.amdgcn.struct.buffer.load.ll         | 82 +++++++++++++++++++
 1 file changed, 82 insertions(+)

diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.load.ll b/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.load.ll
index 81c0f7557e6417..fcd7821a86897e 100644
--- a/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.load.ll
+++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.load.ll
@@ -500,6 +500,47 @@ define amdgpu_ps float @struct_buffer_load_i8_sext__sgpr_rsrc__vgpr_vindex__vgpr
   ret float %cast
 }
 
+define amdgpu_ps float @struct_buffer_load_i8_sext_wrong_width(<4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
+  ; GFX8-LABEL: name: struct_buffer_load_i8_sext_wrong_width
+  ; GFX8: bb.1 (%ir-block.0):
+  ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
+  ; GFX8-NEXT: {{  $}}
+  ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
+  ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
+  ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
+  ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
+  ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
+  ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+  ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+  ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
+  ; GFX8-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
+  ; GFX8-NEXT:   [[BUFFER_LOAD_SBYTE_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_SBYTE_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s8), addrspace 8)
+  ; GFX8-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_SBYTE_BOTHEN]]
+  ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
+  ;
+  ; GFX12-LABEL: name: struct_buffer_load_i8_sext_wrong_width
+  ; GFX12: bb.1 (%ir-block.0):
+  ; GFX12-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
+  ; GFX12-NEXT: {{  $}}
+  ; GFX12-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
+  ; GFX12-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
+  ; GFX12-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
+  ; GFX12-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
+  ; GFX12-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
+  ; GFX12-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+  ; GFX12-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+  ; GFX12-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
+  ; GFX12-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
+  ; GFX12-NEXT:   [[BUFFER_LOAD_SBYTE_VBUFFER_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_SBYTE_VBUFFER_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s8), addrspace 8)
+  ; GFX12-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_SBYTE_VBUFFER_BOTHEN]]
+  ; GFX12-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
+  %val = call i8 @llvm.amdgcn.struct.buffer.load.i8(<4 x i32> %rsrc, i32 %vindex, i32 %voffset, i32 %soffset, i32 0)
+  %trunc = trunc i8 %val to i4
+  %ext = sext i4 %trunc to i32
+  %cast = bitcast i32 %ext to float
+  ret float %cast
+}
+
 define amdgpu_ps float @struct_buffer_load_i16_zext__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(<4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
   ; GFX8-LABEL: name: struct_buffer_load_i16_zext__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
   ; GFX8: bb.1 (%ir-block.0):
@@ -580,6 +621,47 @@ define amdgpu_ps float @struct_buffer_load_i16_sext__sgpr_rsrc__vgpr_vindex__vgp
   ret float %cast
 }
 
+define amdgpu_ps float @struct_buffer_load_i16_sext_wrong_width(<4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
+  ; GFX8-LABEL: name: struct_buffer_load_i16_sext_wrong_width
+  ; GFX8: bb.1 (%ir-block.0):
+  ; GFX8-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
+  ; GFX8-NEXT: {{  $}}
+  ; GFX8-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
+  ; GFX8-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
+  ; GFX8-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
+  ; GFX8-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
+  ; GFX8-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
+  ; GFX8-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+  ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+  ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
+  ; GFX8-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
+  ; GFX8-NEXT:   [[BUFFER_LOAD_SSHORT_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_SSHORT_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
+  ; GFX8-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_SSHORT_BOTHEN]]
+  ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
+  ;
+  ; GFX12-LABEL: name: struct_buffer_load_i16_sext_wrong_width
+  ; GFX12: bb.1 (%ir-block.0):
+  ; GFX12-NEXT:   liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
+  ; GFX12-NEXT: {{  $}}
+  ; GFX12-NEXT:   [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
+  ; GFX12-NEXT:   [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
+  ; GFX12-NEXT:   [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
+  ; GFX12-NEXT:   [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
+  ; GFX12-NEXT:   [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
+  ; GFX12-NEXT:   [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+  ; GFX12-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+  ; GFX12-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
+  ; GFX12-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
+  ; GFX12-NEXT:   [[BUFFER_LOAD_SSHORT_VBUFFER_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_SSHORT_VBUFFER_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
+  ; GFX12-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_SSHORT_VBUFFER_BOTHEN]]
+  ; GFX12-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
+  %val = call i16 @llvm.amdgcn.struct.buffer.load.i16(<4 x i32> %rsrc, i32 %vindex, i32 %voffset, i32 %soffset, i32 0)
+  %trunc = trunc i16 %val to i8
+  %ext = sext i8 %trunc to i32
+  %cast = bitcast i32 %ext to float
+  ret float %cast
+}
+
 ; Natural mapping
 define amdgpu_ps half @struct_buffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(<4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
   ; GFX8-LABEL: name: struct_buffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset

>From 32a84fbfc8428cd4f8fb1992f12027d220eb64e9 Mon Sep 17 00:00:00 2001
From: Jay Foad <jay.foad at amd.com>
Date: Tue, 9 Jan 2024 12:58:32 +0000
Subject: [PATCH 2/3] [AMDGPU] Rewrite G_SEXT_INREG combine. NFC.

Simplify the apply function by storing the new opcode in the matchdata.
---
 llvm/lib/Target/AMDGPU/AMDGPUCombine.td       |  2 +-
 .../AMDGPU/AMDGPUPostLegalizerCombiner.cpp    | 45 ++++++++++---------
 2 files changed, 24 insertions(+), 23 deletions(-)

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUCombine.td b/llvm/lib/Target/AMDGPU/AMDGPUCombine.td
index 0c77fe72595880..b9411e2052120d 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUCombine.td
+++ b/llvm/lib/Target/AMDGPU/AMDGPUCombine.td
@@ -111,7 +111,7 @@ def smulu64 : GICombineRule<
          [{ return matchCombine_s_mul_u64(*${smul}, ${matchinfo}); }]),
   (apply [{ applyCombine_s_mul_u64(*${smul}, ${matchinfo}); }])>;
 
-def sign_exension_in_reg_matchdata : GIDefMatchData<"MachineInstr *">;
+def sign_exension_in_reg_matchdata : GIDefMatchData<"std::pair<MachineInstr *, unsigned>">;
 
 def sign_extension_in_reg : GICombineRule<
   (defs root:$sign_inreg, sign_exension_in_reg_matchdata:$matchinfo),
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUPostLegalizerCombiner.cpp b/llvm/lib/Target/AMDGPU/AMDGPUPostLegalizerCombiner.cpp
index 21bfab52c6c4b5..441375ea6e458c 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUPostLegalizerCombiner.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUPostLegalizerCombiner.cpp
@@ -99,10 +99,10 @@ class AMDGPUPostLegalizerCombinerImpl : public Combiner {
 
   // Combine unsigned buffer load and signed extension instructions to generate
   // signed buffer laod instructions.
-  bool matchCombineSignExtendInReg(MachineInstr &MI,
-                                   MachineInstr *&MatchInfo) const;
-  void applyCombineSignExtendInReg(MachineInstr &MI,
-                                   MachineInstr *&MatchInfo) const;
+  bool matchCombineSignExtendInReg(
+      MachineInstr &MI, std::pair<MachineInstr *, unsigned> &MatchInfo) const;
+  void applyCombineSignExtendInReg(
+      MachineInstr &MI, std::pair<MachineInstr *, unsigned> &MatchInfo) const;
 
   // Find the s_mul_u64 instructions where the higher bits are either
   // zero-extended or sign-extended.
@@ -395,34 +395,35 @@ bool AMDGPUPostLegalizerCombinerImpl::matchRemoveFcanonicalize(
 
 // Identify buffer_load_{u8, u16}.
 bool AMDGPUPostLegalizerCombinerImpl::matchCombineSignExtendInReg(
-    MachineInstr &MI, MachineInstr *&SubwordBufferLoad) const {
-  Register Op0Reg = MI.getOperand(1).getReg();
-  SubwordBufferLoad = MRI.getVRegDef(Op0Reg);
-
-  if (!MRI.hasOneNonDBGUse(Op0Reg))
+    MachineInstr &MI, std::pair<MachineInstr *, unsigned> &MatchData) const {
+  Register LoadReg = MI.getOperand(1).getReg();
+  if (!MRI.hasOneNonDBGUse(LoadReg))
     return false;
 
   // Check if the first operand of the sign extension is a subword buffer load
   // instruction.
-  return SubwordBufferLoad->getOpcode() == AMDGPU::G_AMDGPU_BUFFER_LOAD_UBYTE ||
-         SubwordBufferLoad->getOpcode() == AMDGPU::G_AMDGPU_BUFFER_LOAD_USHORT;
+  MachineInstr *LoadMI = MRI.getVRegDef(LoadReg);
+  switch (LoadMI->getOpcode()) {
+  case AMDGPU::G_AMDGPU_BUFFER_LOAD_UBYTE:
+    MatchData = {LoadMI, AMDGPU::G_AMDGPU_BUFFER_LOAD_SBYTE};
+    return true;
+  case AMDGPU::G_AMDGPU_BUFFER_LOAD_USHORT:
+    MatchData = {LoadMI, AMDGPU::G_AMDGPU_BUFFER_LOAD_SSHORT};
+    return true;
+  }
+  return false;
 }
 
 // Combine buffer_load_{u8, u16} and the sign extension instruction to generate
 // buffer_load_{i8, i16}.
 void AMDGPUPostLegalizerCombinerImpl::applyCombineSignExtendInReg(
-    MachineInstr &MI, MachineInstr *&SubwordBufferLoad) const {
-  // Modify the opcode and the destination of buffer_load_{u8, u16}:
-  // Replace the opcode.
-  unsigned Opc =
-      SubwordBufferLoad->getOpcode() == AMDGPU::G_AMDGPU_BUFFER_LOAD_UBYTE
-          ? AMDGPU::G_AMDGPU_BUFFER_LOAD_SBYTE
-          : AMDGPU::G_AMDGPU_BUFFER_LOAD_SSHORT;
-  SubwordBufferLoad->setDesc(TII.get(Opc));
-  // Update the destination register of SubwordBufferLoad with the destination
-  // register of the sign extension.
+    MachineInstr &MI, std::pair<MachineInstr *, unsigned> &MatchData) const {
+  auto [LoadMI, NewOpcode] = MatchData;
+  LoadMI->setDesc(TII.get(NewOpcode));
+  // Update the destination register of the load with the destination register
+  // of the sign extension.
   Register SignExtendInsnDst = MI.getOperand(0).getReg();
-  SubwordBufferLoad->getOperand(0).setReg(SignExtendInsnDst);
+  LoadMI->getOperand(0).setReg(SignExtendInsnDst);
   // Remove the sign extension.
   MI.eraseFromParent();
 }

>From e85ff44794f0c1cd9b341ad720d4df2c89c46b57 Mon Sep 17 00:00:00 2001
From: Jay Foad <jay.foad at amd.com>
Date: Tue, 9 Jan 2024 13:20:30 +0000
Subject: [PATCH 3/3] [AMDGPU] Fix broken sign-extended subword buffer load
 combine

---
 .../AMDGPU/AMDGPUPostLegalizerCombiner.cpp    |  5 +++--
 .../llvm.amdgcn.struct.buffer.load.ll         | 20 +++++++++++--------
 2 files changed, 15 insertions(+), 10 deletions(-)

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUPostLegalizerCombiner.cpp b/llvm/lib/Target/AMDGPU/AMDGPUPostLegalizerCombiner.cpp
index 441375ea6e458c..bb1d6cb72e8071 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUPostLegalizerCombiner.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUPostLegalizerCombiner.cpp
@@ -403,13 +403,14 @@ bool AMDGPUPostLegalizerCombinerImpl::matchCombineSignExtendInReg(
   // Check if the first operand of the sign extension is a subword buffer load
   // instruction.
   MachineInstr *LoadMI = MRI.getVRegDef(LoadReg);
+  int64_t Width = MI.getOperand(2).getImm();
   switch (LoadMI->getOpcode()) {
   case AMDGPU::G_AMDGPU_BUFFER_LOAD_UBYTE:
     MatchData = {LoadMI, AMDGPU::G_AMDGPU_BUFFER_LOAD_SBYTE};
-    return true;
+    return Width == 8;
   case AMDGPU::G_AMDGPU_BUFFER_LOAD_USHORT:
     MatchData = {LoadMI, AMDGPU::G_AMDGPU_BUFFER_LOAD_SSHORT};
-    return true;
+    return Width == 16;
   }
   return false;
 }
diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.load.ll b/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.load.ll
index fcd7821a86897e..94ce8aac8a4c68 100644
--- a/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.load.ll
+++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.load.ll
@@ -514,8 +514,9 @@ define amdgpu_ps float @struct_buffer_load_i8_sext_wrong_width(<4 x i32> inreg %
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX8-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
-  ; GFX8-NEXT:   [[BUFFER_LOAD_SBYTE_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_SBYTE_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s8), addrspace 8)
-  ; GFX8-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_SBYTE_BOTHEN]]
+  ; GFX8-NEXT:   [[BUFFER_LOAD_UBYTE_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_UBYTE_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s8), addrspace 8)
+  ; GFX8-NEXT:   [[V_BFE_I32_e64_:%[0-9]+]]:vgpr_32 = V_BFE_I32_e64 [[BUFFER_LOAD_UBYTE_BOTHEN]], 0, 4, implicit $exec
+  ; GFX8-NEXT:   $vgpr0 = COPY [[V_BFE_I32_e64_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ;
   ; GFX12-LABEL: name: struct_buffer_load_i8_sext_wrong_width
@@ -531,8 +532,9 @@ define amdgpu_ps float @struct_buffer_load_i8_sext_wrong_width(<4 x i32> inreg %
   ; GFX12-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
   ; GFX12-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX12-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
-  ; GFX12-NEXT:   [[BUFFER_LOAD_SBYTE_VBUFFER_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_SBYTE_VBUFFER_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s8), addrspace 8)
-  ; GFX12-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_SBYTE_VBUFFER_BOTHEN]]
+  ; GFX12-NEXT:   [[BUFFER_LOAD_UBYTE_VBUFFER_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_UBYTE_VBUFFER_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s8), addrspace 8)
+  ; GFX12-NEXT:   [[V_BFE_I32_e64_:%[0-9]+]]:vgpr_32 = V_BFE_I32_e64 [[BUFFER_LOAD_UBYTE_VBUFFER_BOTHEN]], 0, 4, implicit $exec
+  ; GFX12-NEXT:   $vgpr0 = COPY [[V_BFE_I32_e64_]]
   ; GFX12-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %val = call i8 @llvm.amdgcn.struct.buffer.load.i8(<4 x i32> %rsrc, i32 %vindex, i32 %voffset, i32 %soffset, i32 0)
   %trunc = trunc i8 %val to i4
@@ -635,8 +637,9 @@ define amdgpu_ps float @struct_buffer_load_i16_sext_wrong_width(<4 x i32> inreg
   ; GFX8-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
   ; GFX8-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX8-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
-  ; GFX8-NEXT:   [[BUFFER_LOAD_SSHORT_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_SSHORT_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
-  ; GFX8-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_SSHORT_BOTHEN]]
+  ; GFX8-NEXT:   [[BUFFER_LOAD_USHORT_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_USHORT_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
+  ; GFX8-NEXT:   [[V_BFE_I32_e64_:%[0-9]+]]:vgpr_32 = V_BFE_I32_e64 [[BUFFER_LOAD_USHORT_BOTHEN]], 0, 8, implicit $exec
+  ; GFX8-NEXT:   $vgpr0 = COPY [[V_BFE_I32_e64_]]
   ; GFX8-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   ;
   ; GFX12-LABEL: name: struct_buffer_load_i16_sext_wrong_width
@@ -652,8 +655,9 @@ define amdgpu_ps float @struct_buffer_load_i16_sext_wrong_width(<4 x i32> inreg
   ; GFX12-NEXT:   [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
   ; GFX12-NEXT:   [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
   ; GFX12-NEXT:   [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
-  ; GFX12-NEXT:   [[BUFFER_LOAD_SSHORT_VBUFFER_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_SSHORT_VBUFFER_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
-  ; GFX12-NEXT:   $vgpr0 = COPY [[BUFFER_LOAD_SSHORT_VBUFFER_BOTHEN]]
+  ; GFX12-NEXT:   [[BUFFER_LOAD_USHORT_VBUFFER_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_USHORT_VBUFFER_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
+  ; GFX12-NEXT:   [[V_BFE_I32_e64_:%[0-9]+]]:vgpr_32 = V_BFE_I32_e64 [[BUFFER_LOAD_USHORT_VBUFFER_BOTHEN]], 0, 8, implicit $exec
+  ; GFX12-NEXT:   $vgpr0 = COPY [[V_BFE_I32_e64_]]
   ; GFX12-NEXT:   SI_RETURN_TO_EPILOG implicit $vgpr0
   %val = call i16 @llvm.amdgcn.struct.buffer.load.i16(<4 x i32> %rsrc, i32 %vindex, i32 %voffset, i32 %soffset, i32 0)
   %trunc = trunc i16 %val to i8



More information about the Mlir-commits mailing list