[llvm] [RISCV][VLOPT] Add support for vfclass.v (PR #148246)

[Mikhail R. Gadelha via llvm-commits]

https://github.com/mikhailramalho created https://github.com/llvm/llvm-project/pull/148246

This PR adds support for the vfclass.v instruction in the RISC-V VLOptimizer.

This is the first PR addressing the list at #147647.

>From 4205a330b481cbe76bb52f81ded8a6d827284304 Mon Sep 17 00:00:00 2001
From: "Mikhail R. Gadelha" <mikhail at igalia.com>
Date: Thu, 10 Jul 2025 16:37:34 -0300
Subject: [PATCH 1/2] RISCV][VLOPT] Add support for vfclass.v

Signed-off-by: Mikhail R. Gadelha <mikhail at igalia.com>
---
 llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp b/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp
index 2d9f38221d424..e3b33eb0d7c26 100644
--- a/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp
+++ b/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp
@@ -1121,6 +1121,8 @@ static bool isSupportedInstr(const MachineInstr &MI) {
   case RISCV::VFSGNJN_VF:
   case RISCV::VFSGNJX_VF:
   case RISCV::VFSGNJX_VV:
+  // Vector Floating-Point Classify Instruction
+  case RISCV::VFCLASS_V:
   // Vector Floating-Point Compare Instructions
   case RISCV::VMFEQ_VF:
   case RISCV::VMFEQ_VV:

>From d273827c9863b382910eeaaf0a4797a5fd4d820d Mon Sep 17 00:00:00 2001
From: "Mikhail R. Gadelha" <mikhail at igalia.com>
Date: Thu, 10 Jul 2025 16:57:07 -0300
Subject: [PATCH 2/2] Updated test

Signed-off-by: Mikhail R. Gadelha <mikhail at igalia.com>
---
 llvm/test/CodeGen/RISCV/rvv/vl-opt-instrs.ll | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/llvm/test/CodeGen/RISCV/rvv/vl-opt-instrs.ll b/llvm/test/CodeGen/RISCV/rvv/vl-opt-instrs.ll
index 317ad0c124e73..d97a0c7d3cb36 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vl-opt-instrs.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/vl-opt-instrs.ll
@@ -5455,9 +5455,8 @@ define <vscale x 4 x i32> @vfclass_v(<vscale x 4 x float> %a, <vscale x 4 x i32>
 ;
 ; VLOPT-LABEL: vfclass_v:
 ; VLOPT:       # %bb.0:
-; VLOPT-NEXT:    vsetvli a1, zero, e32, m2, ta, ma
-; VLOPT-NEXT:    vfclass.v v8, v8
 ; VLOPT-NEXT:    vsetvli zero, a0, e32, m2, ta, ma
+; VLOPT-NEXT:    vfclass.v v8, v8
 ; VLOPT-NEXT:    vadd.vv v8, v8, v10
 ; VLOPT-NEXT:    ret
   %1 = call <vscale x 4 x i32> @llvm.riscv.vfclass.nxv4i32(<vscale x 4 x i32> poison, <vscale x 4 x float> %a, iXLen -1)