[RISCV][llvm] Correct shamt in P extension EXTRACT_VECTOR_ELT lowering #169823
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
During operation legalization, element type should have been turn into XLenVT which makes the SHL a no-op. We need to use exact vector element type instead.
Member
|
@llvm/pr-subscribers-backend-risc-v Author: Brandon Wu (4vtomat) ChangesDuring operation legalization, element type should have been turn into Full diff: https://github.com/llvm/llvm-project/pull/169823.diff 3 Files Affected:
diff --git a/llvm/lib/Target/RISCV/RISCVISelLowering.cpp b/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
index be53f51afe79f..0d9b8bddbcc8f 100644
--- a/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
+++ b/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
@@ -10699,7 +10699,7 @@ SDValue RISCVTargetLowering::lowerEXTRACT_VECTOR_ELT(SDValue Op,
VecVT != MVT::v4i8 && VecVT != MVT::v2i32)
return SDValue();
SDValue Extracted = DAG.getBitcast(XLenVT, Vec);
- unsigned ElemWidth = EltVT.getSizeInBits();
+ unsigned ElemWidth = VecVT.getVectorElementType().getSizeInBits();
SDValue Shamt = DAG.getNode(ISD::MUL, DL, XLenVT, Idx,
DAG.getConstant(ElemWidth, DL, XLenVT));
return DAG.getNode(ISD::SRL, DL, XLenVT, Extracted, Shamt);
diff --git a/llvm/test/CodeGen/RISCV/rvp-ext-rv32.ll b/llvm/test/CodeGen/RISCV/rvp-ext-rv32.ll
index d4ea9e6c3def0..f803f6aa09652 100644
--- a/llvm/test/CodeGen/RISCV/rvp-ext-rv32.ll
+++ b/llvm/test/CodeGen/RISCV/rvp-ext-rv32.ll
@@ -484,6 +484,25 @@ define void @test_extract_vector_16(ptr %ret_ptr, ptr %a_ptr) {
ret void
}
+define void @test_extract_vector_16_elem1(ptr %ret_ptr, ptr %a_ptr) {
+; CHECK-RV32-LABEL: test_extract_vector_16_elem1:
+; CHECK-RV32: # %bb.0:
+; CHECK-RV32-NEXT: lhu a1, 2(a1)
+; CHECK-RV32-NEXT: sh a1, 0(a0)
+; CHECK-RV32-NEXT: ret
+;
+; CHECK-RV64-LABEL: test_extract_vector_16_elem1:
+; CHECK-RV64: # %bb.0:
+; CHECK-RV64-NEXT: lw a1, 0(a1)
+; CHECK-RV64-NEXT: srli a1, a1, 16
+; CHECK-RV64-NEXT: sh a1, 0(a0)
+; CHECK-RV64-NEXT: ret
+ %a = load <2 x i16>, ptr %a_ptr
+ %extracted = extractelement <2 x i16> %a, i32 1
+ store i16 %extracted, ptr %ret_ptr
+ ret void
+}
+
define void @test_extract_vector_8(ptr %ret_ptr, ptr %a_ptr) {
; CHECK-LABEL: test_extract_vector_8:
; CHECK: # %bb.0:
@@ -496,6 +515,19 @@ define void @test_extract_vector_8(ptr %ret_ptr, ptr %a_ptr) {
ret void
}
+define void @test_extract_vector_8_elem1(ptr %ret_ptr, ptr %a_ptr) {
+; CHECK-LABEL: test_extract_vector_8_elem1:
+; CHECK: # %bb.0:
+; CHECK-NEXT: lw a1, 0(a1)
+; CHECK-NEXT: srli a1, a1, 8
+; CHECK-NEXT: sb a1, 0(a0)
+; CHECK-NEXT: ret
+ %a = load <4 x i8>, ptr %a_ptr
+ %extracted = extractelement <4 x i8> %a, i32 1
+ store i8 %extracted, ptr %ret_ptr
+ ret void
+}
+
; Test for splat
define void @test_non_const_splat_i8(ptr %ret_ptr, ptr %a_ptr, i8 %elt) {
; CHECK-LABEL: test_non_const_splat_i8:
diff --git a/llvm/test/CodeGen/RISCV/rvp-ext-rv64.ll b/llvm/test/CodeGen/RISCV/rvp-ext-rv64.ll
index b39b807d43154..9b021df8dd452 100644
--- a/llvm/test/CodeGen/RISCV/rvp-ext-rv64.ll
+++ b/llvm/test/CodeGen/RISCV/rvp-ext-rv64.ll
@@ -495,6 +495,18 @@ define void @test_extract_vector_32(ptr %ret_ptr, ptr %a_ptr) {
ret void
}
+define void @test_extract_vector_32_elem1(ptr %ret_ptr, ptr %a_ptr) {
+; CHECK-LABEL: test_extract_vector_32_elem1:
+; CHECK: # %bb.0:
+; CHECK-NEXT: lw a1, 4(a1)
+; CHECK-NEXT: sw a1, 0(a0)
+; CHECK-NEXT: ret
+ %a = load <2 x i32>, ptr %a_ptr
+ %extracted = extractelement <2 x i32> %a, i32 1
+ store i32 %extracted, ptr %ret_ptr
+ ret void
+}
+
; Test basic add/sub operations for v2i32 (RV64 only)
define void @test_padd_w(ptr %ret_ptr, ptr %a_ptr, ptr %b_ptr) {
; CHECK-LABEL: test_padd_w:
|
Collaborator
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/169/builds/17590 Here is the relevant piece of the build log for the reference |
aahrun
pushed a commit
to aahrun/llvm-project
that referenced
this pull request
Dec 1, 2025
llvm#169823) During operation legalization, element type should have been turn into XLenVT which makes the SHL a no-op. We need to use exact vector element type instead.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 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.
During operation legalization, element type should have been turn into
XLenVT which makes the SHL a no-op. We need to use exact vector element
type instead.