Revert "[AMDGPU] Add buffer.fat.ptr.load.lds intrinsic wrapping raw rsrc version (#133015)" #134871
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…src version (llvm#133015)" This reverts commit d1a0572. There was further discussion on the PR about whether the intinsics should exist in this form.
|
@llvm/pr-subscribers-backend-amdgpu Author: Krzysztof Drewniak (krzysz00) ChangesThis reverts commit d1a0572. There was further discussion on the PR about whether the intinsics should exist in this form. Full diff: https://github.com/llvm/llvm-project/pull/134871.diff 3 Files Affected:
diff --git a/llvm/include/llvm/IR/IntrinsicsAMDGPU.td b/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
index ae2f6e62c0272..217e43fcce4fd 100644
--- a/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
+++ b/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
@@ -1941,27 +1941,6 @@ def int_amdgcn_s_buffer_prefetch_data : DefaultAttrsIntrinsic <
} // defset AMDGPUBufferIntrinsics
-// A wrapper around raw_ptr_buffer_load_lds that takes the global offset
-// from the addrspace(7) pointer argument.
-def int_amdgcn_buffer_fat_ptr_load_lds : Intrinsic <
- [],
- [LLVMQualPointerType<7>, // buffer fat pointer (SGPRx4 + VGPR)
- LLVMQualPointerType<3>, // LDS base offset
- llvm_i32_ty, // Data byte size: 1/2/4 (/12/16 for gfx950)
- llvm_i32_ty, // imm offset(imm, included in bounds checking and swizzling)
- llvm_i32_ty], // auxiliary/cachepolicy(imm):
- // bit 0 = glc, bit 1 = slc, bit 2 = dlc (gfx10/gfx11),
- // bit 3 = swz, bit 4 = scc (gfx90a)
- // gfx942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1
- // gfx12+: bits [0-2] = th, bits [3-4] = scope,
- // bit 6 = swz
- // all: volatile op (bit 31, stripped at lowering)
- [IntrWillReturn, IntrArgMemOnly,
- ReadOnly<ArgIndex<0>>, NoCapture<ArgIndex<0>>,
- WriteOnly<ArgIndex<1>>, NoCapture<ArgIndex<1>>,
- ImmArg<ArgIndex<2>>, ImmArg<ArgIndex<3>>,
- ImmArg<ArgIndex<4>>, IntrNoCallback, IntrNoFree], "", [SDNPMemOperand]>;
-
// Uses that do not set the done bit should set IntrWriteMem on the
// call site.
def int_amdgcn_exp : DefaultAttrsIntrinsic <[], [
diff --git a/llvm/lib/Target/AMDGPU/AMDGPULowerBufferFatPointers.cpp b/llvm/lib/Target/AMDGPU/AMDGPULowerBufferFatPointers.cpp
index 4448570b6b979..766a4ea250942 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPULowerBufferFatPointers.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPULowerBufferFatPointers.cpp
@@ -2167,7 +2167,6 @@ static bool isRemovablePointerIntrinsic(Intrinsic::ID IID) {
case Intrinsic::memset:
case Intrinsic::memset_inline:
case Intrinsic::experimental_memset_pattern:
- case Intrinsic::amdgcn_buffer_fat_ptr_load_lds:
return true;
}
}
@@ -2256,26 +2255,6 @@ PtrParts SplitPtrStructs::visitIntrinsicInst(IntrinsicInst &I) {
SplitUsers.insert(&I);
return {NewRsrc, Off};
}
- case Intrinsic::amdgcn_buffer_fat_ptr_load_lds: {
- Value *BufferPtr = I.getArgOperand(0);
- assert(isSplitFatPtr(BufferPtr->getType()) &&
- "amdgcn.buffer.fat.pointer.load.lds must have a buffer fat pointer "
- "as argument 0");
- IRB.SetInsertPoint(&I);
- auto [Rsrc, Off] = getPtrParts(BufferPtr);
- Value *LDSPtr = I.getArgOperand(1);
- Value *LoadSize = I.getArgOperand(2);
- Value *ImmOff = I.getArgOperand(3);
- Value *Aux = I.getArgOperand(4);
- Value *SOffset = IRB.getInt32(0);
- Instruction *NewLoad = IRB.CreateIntrinsic(
- Intrinsic::amdgcn_raw_ptr_buffer_load_lds, {},
- {Rsrc, LDSPtr, LoadSize, Off, SOffset, ImmOff, Aux});
- copyMetadata(NewLoad, &I);
- SplitUsers.insert(&I);
- I.replaceAllUsesWith(NewLoad);
- return {nullptr, nullptr};
- }
}
return {nullptr, nullptr};
}
diff --git a/llvm/test/CodeGen/AMDGPU/lower-buffer-fat-pointers-mem-transfer.ll b/llvm/test/CodeGen/AMDGPU/lower-buffer-fat-pointers-mem-transfer.ll
index 56d0cdd29ffb2..ee51b0b84554e 100644
--- a/llvm/test/CodeGen/AMDGPU/lower-buffer-fat-pointers-mem-transfer.ll
+++ b/llvm/test/CodeGen/AMDGPU/lower-buffer-fat-pointers-mem-transfer.ll
@@ -1724,21 +1724,3 @@ define void @memset_pattern_unknown(ptr addrspace(7) inreg %ptr, i32 inreg %leng
call void @llvm.experimental.memset.pattern.p7.i32.i32(ptr addrspace(7) %ptr, i32 1, i32 %length, i1 false)
ret void
}
-
-;;; Buffer load to LDS
-
-declare void @llvm.amdgcn.buffer.fat.ptr.load.lds(ptr addrspace(7), ptr addrspace(3), i32 immarg, i32 immarg, i32 immarg)
-
-define void @llvm_amdgcn_buffer_fat_ptr_load_lds(ptr addrspace(7) inreg %p, ptr addrspace(3) inreg %l, i32 %idx) {
-; CHECK-LABEL: define void @llvm_amdgcn_buffer_fat_ptr_load_lds(
-; CHECK-SAME: { ptr addrspace(8), i32 } inreg [[P:%.*]], ptr addrspace(3) inreg [[L:%.*]], i32 [[IDX:%.*]]) #[[ATTR0]] {
-; CHECK-NEXT: [[P_RSRC:%.*]] = extractvalue { ptr addrspace(8), i32 } [[P]], 0
-; CHECK-NEXT: [[P_OFF:%.*]] = extractvalue { ptr addrspace(8), i32 } [[P]], 1
-; CHECK-NEXT: [[Q:%.*]] = add i32 [[P_OFF]], [[IDX]]
-; CHECK-NEXT: call void @llvm.amdgcn.raw.ptr.buffer.load.lds(ptr addrspace(8) [[P_RSRC]], ptr addrspace(3) [[L]], i32 4, i32 [[Q]], i32 0, i32 16, i32 0)
-; CHECK-NEXT: ret void
-;
- %q = getelementptr i8, ptr addrspace(7) %p, i32 %idx
- call void @llvm.amdgcn.buffer.fat.ptr.load.lds(ptr addrspace(7) %q, ptr addrspace(3) %l, i32 4, i32 16, i32 0)
- ret void
-}
|
|
@llvm/pr-subscribers-llvm-ir Author: Krzysztof Drewniak (krzysz00) ChangesThis reverts commit d1a0572. There was further discussion on the PR about whether the intinsics should exist in this form. Full diff: https://github.com/llvm/llvm-project/pull/134871.diff 3 Files Affected:
diff --git a/llvm/include/llvm/IR/IntrinsicsAMDGPU.td b/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
index ae2f6e62c0272..217e43fcce4fd 100644
--- a/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
+++ b/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
@@ -1941,27 +1941,6 @@ def int_amdgcn_s_buffer_prefetch_data : DefaultAttrsIntrinsic <
} // defset AMDGPUBufferIntrinsics
-// A wrapper around raw_ptr_buffer_load_lds that takes the global offset
-// from the addrspace(7) pointer argument.
-def int_amdgcn_buffer_fat_ptr_load_lds : Intrinsic <
- [],
- [LLVMQualPointerType<7>, // buffer fat pointer (SGPRx4 + VGPR)
- LLVMQualPointerType<3>, // LDS base offset
- llvm_i32_ty, // Data byte size: 1/2/4 (/12/16 for gfx950)
- llvm_i32_ty, // imm offset(imm, included in bounds checking and swizzling)
- llvm_i32_ty], // auxiliary/cachepolicy(imm):
- // bit 0 = glc, bit 1 = slc, bit 2 = dlc (gfx10/gfx11),
- // bit 3 = swz, bit 4 = scc (gfx90a)
- // gfx942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1
- // gfx12+: bits [0-2] = th, bits [3-4] = scope,
- // bit 6 = swz
- // all: volatile op (bit 31, stripped at lowering)
- [IntrWillReturn, IntrArgMemOnly,
- ReadOnly<ArgIndex<0>>, NoCapture<ArgIndex<0>>,
- WriteOnly<ArgIndex<1>>, NoCapture<ArgIndex<1>>,
- ImmArg<ArgIndex<2>>, ImmArg<ArgIndex<3>>,
- ImmArg<ArgIndex<4>>, IntrNoCallback, IntrNoFree], "", [SDNPMemOperand]>;
-
// Uses that do not set the done bit should set IntrWriteMem on the
// call site.
def int_amdgcn_exp : DefaultAttrsIntrinsic <[], [
diff --git a/llvm/lib/Target/AMDGPU/AMDGPULowerBufferFatPointers.cpp b/llvm/lib/Target/AMDGPU/AMDGPULowerBufferFatPointers.cpp
index 4448570b6b979..766a4ea250942 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPULowerBufferFatPointers.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPULowerBufferFatPointers.cpp
@@ -2167,7 +2167,6 @@ static bool isRemovablePointerIntrinsic(Intrinsic::ID IID) {
case Intrinsic::memset:
case Intrinsic::memset_inline:
case Intrinsic::experimental_memset_pattern:
- case Intrinsic::amdgcn_buffer_fat_ptr_load_lds:
return true;
}
}
@@ -2256,26 +2255,6 @@ PtrParts SplitPtrStructs::visitIntrinsicInst(IntrinsicInst &I) {
SplitUsers.insert(&I);
return {NewRsrc, Off};
}
- case Intrinsic::amdgcn_buffer_fat_ptr_load_lds: {
- Value *BufferPtr = I.getArgOperand(0);
- assert(isSplitFatPtr(BufferPtr->getType()) &&
- "amdgcn.buffer.fat.pointer.load.lds must have a buffer fat pointer "
- "as argument 0");
- IRB.SetInsertPoint(&I);
- auto [Rsrc, Off] = getPtrParts(BufferPtr);
- Value *LDSPtr = I.getArgOperand(1);
- Value *LoadSize = I.getArgOperand(2);
- Value *ImmOff = I.getArgOperand(3);
- Value *Aux = I.getArgOperand(4);
- Value *SOffset = IRB.getInt32(0);
- Instruction *NewLoad = IRB.CreateIntrinsic(
- Intrinsic::amdgcn_raw_ptr_buffer_load_lds, {},
- {Rsrc, LDSPtr, LoadSize, Off, SOffset, ImmOff, Aux});
- copyMetadata(NewLoad, &I);
- SplitUsers.insert(&I);
- I.replaceAllUsesWith(NewLoad);
- return {nullptr, nullptr};
- }
}
return {nullptr, nullptr};
}
diff --git a/llvm/test/CodeGen/AMDGPU/lower-buffer-fat-pointers-mem-transfer.ll b/llvm/test/CodeGen/AMDGPU/lower-buffer-fat-pointers-mem-transfer.ll
index 56d0cdd29ffb2..ee51b0b84554e 100644
--- a/llvm/test/CodeGen/AMDGPU/lower-buffer-fat-pointers-mem-transfer.ll
+++ b/llvm/test/CodeGen/AMDGPU/lower-buffer-fat-pointers-mem-transfer.ll
@@ -1724,21 +1724,3 @@ define void @memset_pattern_unknown(ptr addrspace(7) inreg %ptr, i32 inreg %leng
call void @llvm.experimental.memset.pattern.p7.i32.i32(ptr addrspace(7) %ptr, i32 1, i32 %length, i1 false)
ret void
}
-
-;;; Buffer load to LDS
-
-declare void @llvm.amdgcn.buffer.fat.ptr.load.lds(ptr addrspace(7), ptr addrspace(3), i32 immarg, i32 immarg, i32 immarg)
-
-define void @llvm_amdgcn_buffer_fat_ptr_load_lds(ptr addrspace(7) inreg %p, ptr addrspace(3) inreg %l, i32 %idx) {
-; CHECK-LABEL: define void @llvm_amdgcn_buffer_fat_ptr_load_lds(
-; CHECK-SAME: { ptr addrspace(8), i32 } inreg [[P:%.*]], ptr addrspace(3) inreg [[L:%.*]], i32 [[IDX:%.*]]) #[[ATTR0]] {
-; CHECK-NEXT: [[P_RSRC:%.*]] = extractvalue { ptr addrspace(8), i32 } [[P]], 0
-; CHECK-NEXT: [[P_OFF:%.*]] = extractvalue { ptr addrspace(8), i32 } [[P]], 1
-; CHECK-NEXT: [[Q:%.*]] = add i32 [[P_OFF]], [[IDX]]
-; CHECK-NEXT: call void @llvm.amdgcn.raw.ptr.buffer.load.lds(ptr addrspace(8) [[P_RSRC]], ptr addrspace(3) [[L]], i32 4, i32 [[Q]], i32 0, i32 16, i32 0)
-; CHECK-NEXT: ret void
-;
- %q = getelementptr i8, ptr addrspace(7) %p, i32 %idx
- call void @llvm.amdgcn.buffer.fat.ptr.load.lds(ptr addrspace(7) %q, ptr addrspace(3) %l, i32 4, i32 16, i32 0)
- ret void
-}
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This reverts commit d1a0572.
There was further discussion on the PR about whether the intinsics should exist in this form.