[RISCV][GISel] Support f32/f64 powi. #117937
Merged
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
Need to force libcall legalization to treat the integer argument as signed so that it can be promoted to XLen in call lowering for RV64. Alternatively we could promote the operand before converting to libcall, but going through call lowering is closer to what SelectionDAG does.
Member
|
@llvm/pr-subscribers-backend-risc-v Author: Craig Topper (topperc) ChangesNeed to force libcall legalization to treat the integer argument as signed so that it can be promoted to XLen in call lowering for RV64. Alternatively we could promote the operand before converting to libcall, but going through call lowering is closer to what SelectionDAG does. Full diff: https://github.com/llvm/llvm-project/pull/117937.diff 4 Files Affected:
diff --git a/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp b/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp
index 709eb2525a6577..dd424afd0a909f 100644
--- a/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp
+++ b/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp
@@ -1290,9 +1290,10 @@ LegalizerHelper::libcall(MachineInstr &MI, LostDebugLocObserver &LocObserver) {
return UnableToLegalize;
}
auto Libcall = getRTLibDesc(MI.getOpcode(), Size);
- std::initializer_list<CallLowering::ArgInfo> Args = {
+ SmallVector<CallLowering::ArgInfo, 2> Args = {
{MI.getOperand(1).getReg(), HLTy, 0},
{MI.getOperand(2).getReg(), ITy, 1}};
+ Args[1].Flags[0].setSExt();
LegalizeResult Status =
createLibcall(MIRBuilder, Libcall, {MI.getOperand(0).getReg(), HLTy, 0},
Args, LocObserver, &MI);
diff --git a/llvm/lib/Target/RISCV/GISel/RISCVLegalizerInfo.cpp b/llvm/lib/Target/RISCV/GISel/RISCVLegalizerInfo.cpp
index a4e03be98c696d..005d88a1731232 100644
--- a/llvm/lib/Target/RISCV/GISel/RISCVLegalizerInfo.cpp
+++ b/llvm/lib/Target/RISCV/GISel/RISCVLegalizerInfo.cpp
@@ -579,6 +579,8 @@ RISCVLegalizerInfo::RISCVLegalizerInfo(const RISCVSubtarget &ST)
G_FASIN, G_FATAN, G_FATAN2, G_FCOSH, G_FSINH,
G_FTANH})
.libcallFor({s32, s64});
+ getActionDefinitionsBuilder(G_FPOWI)
+ .libcallFor({{s32, s32}});
getActionDefinitionsBuilder(G_VASTART).customFor({p0});
diff --git a/llvm/test/CodeGen/RISCV/GlobalISel/float-intrinsics.ll b/llvm/test/CodeGen/RISCV/GlobalISel/float-intrinsics.ll
index 16b0a2fc2f396d..7098493b2b56d3 100644
--- a/llvm/test/CodeGen/RISCV/GlobalISel/float-intrinsics.ll
+++ b/llvm/test/CodeGen/RISCV/GlobalISel/float-intrinsics.ll
@@ -48,6 +48,48 @@ define float @sqrt_f32(float %a) nounwind {
ret float %1
}
+define float @powi_f32(float %a, i32 %b) nounwind {
+; RV32IF-LABEL: powi_f32:
+; RV32IF: # %bb.0:
+; RV32IF-NEXT: addi sp, sp, -16
+; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
+; RV32IF-NEXT: call __powisf2
+; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
+; RV32IF-NEXT: addi sp, sp, 16
+; RV32IF-NEXT: ret
+;
+; RV64IF-LABEL: powi_f32:
+; RV64IF: # %bb.0:
+; RV64IF-NEXT: addi sp, sp, -16
+; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
+; RV64IF-NEXT: sext.w a0, a0
+; RV64IF-NEXT: call __powisf2
+; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
+; RV64IF-NEXT: addi sp, sp, 16
+; RV64IF-NEXT: ret
+;
+; RV32I-LABEL: powi_f32:
+; RV32I: # %bb.0:
+; RV32I-NEXT: addi sp, sp, -16
+; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
+; RV32I-NEXT: call __powisf2
+; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
+; RV32I-NEXT: addi sp, sp, 16
+; RV32I-NEXT: ret
+;
+; RV64I-LABEL: powi_f32:
+; RV64I: # %bb.0:
+; RV64I-NEXT: addi sp, sp, -16
+; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
+; RV64I-NEXT: sext.w a1, a1
+; RV64I-NEXT: call __powisf2
+; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
+; RV64I-NEXT: addi sp, sp, 16
+; RV64I-NEXT: ret
+ %1 = call float @llvm.powi.f32.i32(float %a, i32 %b)
+ ret float %1
+}
+
define float @sin_f32(float %a) nounwind {
; RV32IF-LABEL: sin_f32:
; RV32IF: # %bb.0:
diff --git a/llvm/test/CodeGen/RISCV/GlobalISel/legalizer-info-validation.mir b/llvm/test/CodeGen/RISCV/GlobalISel/legalizer-info-validation.mir
index 719ea38cbb9c52..f14c806607c710 100644
--- a/llvm/test/CodeGen/RISCV/GlobalISel/legalizer-info-validation.mir
+++ b/llvm/test/CodeGen/RISCV/GlobalISel/legalizer-info-validation.mir
@@ -481,8 +481,8 @@
# DEBUG-NEXT: .. the first uncovered type index: 1, OK
# DEBUG-NEXT: .. the first uncovered imm index: 0, OK
# DEBUG-NEXT: G_FPOWI (opcode {{[0-9]+}}): 2 type indices, 0 imm indices
-# DEBUG-NEXT: .. type index coverage check SKIPPED: no rules defined
-# DEBUG-NEXT: .. imm index coverage check SKIPPED: no rules defined
+# DEBUG-NEXT: .. the first uncovered type index: 2, OK
+# DEBUG-NEXT: .. the first uncovered imm index: 0, OK
# DEBUG-NEXT: G_FEXP (opcode {{[0-9]+}}): 1 type index, 0 imm indices
# DEBUG-NEXT: .. opcode {{[0-9]+}} is aliased to {{[0-9]+}}
# DEBUG-NEXT: .. the first uncovered type index: 1, OK
|
Member
|
@llvm/pr-subscribers-llvm-globalisel Author: Craig Topper (topperc) ChangesNeed to force libcall legalization to treat the integer argument as signed so that it can be promoted to XLen in call lowering for RV64. Alternatively we could promote the operand before converting to libcall, but going through call lowering is closer to what SelectionDAG does. Full diff: https://github.com/llvm/llvm-project/pull/117937.diff 4 Files Affected:
diff --git a/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp b/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp
index 709eb2525a6577..dd424afd0a909f 100644
--- a/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp
+++ b/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp
@@ -1290,9 +1290,10 @@ LegalizerHelper::libcall(MachineInstr &MI, LostDebugLocObserver &LocObserver) {
return UnableToLegalize;
}
auto Libcall = getRTLibDesc(MI.getOpcode(), Size);
- std::initializer_list<CallLowering::ArgInfo> Args = {
+ SmallVector<CallLowering::ArgInfo, 2> Args = {
{MI.getOperand(1).getReg(), HLTy, 0},
{MI.getOperand(2).getReg(), ITy, 1}};
+ Args[1].Flags[0].setSExt();
LegalizeResult Status =
createLibcall(MIRBuilder, Libcall, {MI.getOperand(0).getReg(), HLTy, 0},
Args, LocObserver, &MI);
diff --git a/llvm/lib/Target/RISCV/GISel/RISCVLegalizerInfo.cpp b/llvm/lib/Target/RISCV/GISel/RISCVLegalizerInfo.cpp
index a4e03be98c696d..005d88a1731232 100644
--- a/llvm/lib/Target/RISCV/GISel/RISCVLegalizerInfo.cpp
+++ b/llvm/lib/Target/RISCV/GISel/RISCVLegalizerInfo.cpp
@@ -579,6 +579,8 @@ RISCVLegalizerInfo::RISCVLegalizerInfo(const RISCVSubtarget &ST)
G_FASIN, G_FATAN, G_FATAN2, G_FCOSH, G_FSINH,
G_FTANH})
.libcallFor({s32, s64});
+ getActionDefinitionsBuilder(G_FPOWI)
+ .libcallFor({{s32, s32}});
getActionDefinitionsBuilder(G_VASTART).customFor({p0});
diff --git a/llvm/test/CodeGen/RISCV/GlobalISel/float-intrinsics.ll b/llvm/test/CodeGen/RISCV/GlobalISel/float-intrinsics.ll
index 16b0a2fc2f396d..7098493b2b56d3 100644
--- a/llvm/test/CodeGen/RISCV/GlobalISel/float-intrinsics.ll
+++ b/llvm/test/CodeGen/RISCV/GlobalISel/float-intrinsics.ll
@@ -48,6 +48,48 @@ define float @sqrt_f32(float %a) nounwind {
ret float %1
}
+define float @powi_f32(float %a, i32 %b) nounwind {
+; RV32IF-LABEL: powi_f32:
+; RV32IF: # %bb.0:
+; RV32IF-NEXT: addi sp, sp, -16
+; RV32IF-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
+; RV32IF-NEXT: call __powisf2
+; RV32IF-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
+; RV32IF-NEXT: addi sp, sp, 16
+; RV32IF-NEXT: ret
+;
+; RV64IF-LABEL: powi_f32:
+; RV64IF: # %bb.0:
+; RV64IF-NEXT: addi sp, sp, -16
+; RV64IF-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
+; RV64IF-NEXT: sext.w a0, a0
+; RV64IF-NEXT: call __powisf2
+; RV64IF-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
+; RV64IF-NEXT: addi sp, sp, 16
+; RV64IF-NEXT: ret
+;
+; RV32I-LABEL: powi_f32:
+; RV32I: # %bb.0:
+; RV32I-NEXT: addi sp, sp, -16
+; RV32I-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
+; RV32I-NEXT: call __powisf2
+; RV32I-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
+; RV32I-NEXT: addi sp, sp, 16
+; RV32I-NEXT: ret
+;
+; RV64I-LABEL: powi_f32:
+; RV64I: # %bb.0:
+; RV64I-NEXT: addi sp, sp, -16
+; RV64I-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
+; RV64I-NEXT: sext.w a1, a1
+; RV64I-NEXT: call __powisf2
+; RV64I-NEXT: ld ra, 8(sp) # 8-byte Folded Reload
+; RV64I-NEXT: addi sp, sp, 16
+; RV64I-NEXT: ret
+ %1 = call float @llvm.powi.f32.i32(float %a, i32 %b)
+ ret float %1
+}
+
define float @sin_f32(float %a) nounwind {
; RV32IF-LABEL: sin_f32:
; RV32IF: # %bb.0:
diff --git a/llvm/test/CodeGen/RISCV/GlobalISel/legalizer-info-validation.mir b/llvm/test/CodeGen/RISCV/GlobalISel/legalizer-info-validation.mir
index 719ea38cbb9c52..f14c806607c710 100644
--- a/llvm/test/CodeGen/RISCV/GlobalISel/legalizer-info-validation.mir
+++ b/llvm/test/CodeGen/RISCV/GlobalISel/legalizer-info-validation.mir
@@ -481,8 +481,8 @@
# DEBUG-NEXT: .. the first uncovered type index: 1, OK
# DEBUG-NEXT: .. the first uncovered imm index: 0, OK
# DEBUG-NEXT: G_FPOWI (opcode {{[0-9]+}}): 2 type indices, 0 imm indices
-# DEBUG-NEXT: .. type index coverage check SKIPPED: no rules defined
-# DEBUG-NEXT: .. imm index coverage check SKIPPED: no rules defined
+# DEBUG-NEXT: .. the first uncovered type index: 2, OK
+# DEBUG-NEXT: .. the first uncovered imm index: 0, OK
# DEBUG-NEXT: G_FEXP (opcode {{[0-9]+}}): 1 type index, 0 imm indices
# DEBUG-NEXT: .. opcode {{[0-9]+}} is aliased to {{[0-9]+}}
# DEBUG-NEXT: .. the first uncovered type index: 1, OK
|
|
✅ With the latest revision this PR passed the C/C++ code formatter. |
lenary
reviewed
Nov 28, 2024
| SmallVector<CallLowering::ArgInfo, 2> Args = { | ||
| {MI.getOperand(1).getReg(), HLTy, 0}, | ||
| {MI.getOperand(2).getReg(), ITy, 1}}; | ||
| Args[1].Flags[0].setSExt(); |
Member
There was a problem hiding this comment.
Actually, sorry, I'm not 100% sure about this - should we be using things like TargetLowering::shouldSignExtendTypeInLibCall/TargetLowering::shouldExtendTypeInLibCall to choose whether to (sign) extend? It's not clear to me we have a GISel equivalent to those callbacks, but I think we might need something like that?
Collaborator
Author
There was a problem hiding this comment.
Eventually yes we probably need something like those hooks. A signed integer argument is maybe the least interesting case for those hooks.
shouldExtendTypeInLibCall defaults to true and only RISC-V and LoongArch override it to return false for FP types.
shouldSignExtendTypeInLibCall defaults to returning true for signed types and only RISC-V, LoongArch, and Mips override it to return true for some unsigned integers.
lenary
approved these changes
Nov 28, 2024
| SmallVector<CallLowering::ArgInfo, 2> Args = { | ||
| {MI.getOperand(1).getReg(), HLTy, 0}, | ||
| {MI.getOperand(2).getReg(), ITy, 1}}; | ||
| Args[1].Flags[0].setSExt(); |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
Need to force libcall legalization to treat the integer argument as signed so that it can be promoted to XLen in call lowering for RV64. Alternatively we could promote the operand before converting to libcall, but going through call lowering is closer to what SelectionDAG does.