From 92823947fd0989d52e0ab9b6fca52bc67bf7d59f Mon Sep 17 00:00:00 2001
From: Shengchen Kan <shengchen.kan@intel.com>
Date: Wed, 24 Jan 2024 17:10:28 +0800
Subject: [PATCH] [X86][CodeGen] Fix crash when commute operands of Instruction
 for code size (#79245)

Reported in 134fcc62786d31ab73439201dce2d73808d1785a

Incorrect opcode is used  b/c there is a `[[fallthrough]]` at line 2386.

(cherry picked from commit 33ecef9812e2c9bfadef035b8e34a949acae2abc)
---
 llvm/lib/Target/X86/X86InstrInfo.cpp        | 35 +++++++++------------
 llvm/test/CodeGen/X86/commute-blend-avx2.ll |  9 ++++++
 2 files changed, 23 insertions(+), 21 deletions(-)

diff --git a/llvm/lib/Target/X86/X86InstrInfo.cpp b/llvm/lib/Target/X86/X86InstrInfo.cpp
index d6f9aa6d6acec0..9ac1f783b7f0a1 100644
--- a/llvm/lib/Target/X86/X86InstrInfo.cpp
+++ b/llvm/lib/Target/X86/X86InstrInfo.cpp
@@ -2354,33 +2354,26 @@ MachineInstr *X86InstrInfo::commuteInstructionImpl(MachineInstr &MI, bool NewMI,
   case X86::VBLENDPSrri:
     // If we're optimizing for size, try to use MOVSD/MOVSS.
     if (MI.getParent()->getParent()->getFunction().hasOptSize()) {
-      unsigned Mask;
-      switch (Opc) {
-      default:
-        llvm_unreachable("Unreachable!");
-      case X86::BLENDPDrri:
-        Opc = X86::MOVSDrr;
-        Mask = 0x03;
-        break;
-      case X86::BLENDPSrri:
-        Opc = X86::MOVSSrr;
-        Mask = 0x0F;
-        break;
-      case X86::VBLENDPDrri:
-        Opc = X86::VMOVSDrr;
-        Mask = 0x03;
-        break;
-      case X86::VBLENDPSrri:
-        Opc = X86::VMOVSSrr;
-        Mask = 0x0F;
-        break;
-      }
+      unsigned Mask = (Opc == X86::BLENDPDrri || Opc == X86::VBLENDPDrri) ? 0x03: 0x0F;
       if ((MI.getOperand(3).getImm() ^ Mask) == 1) {
+#define FROM_TO(FROM, TO)                                                      \
+  case X86::FROM:                                                              \
+    Opc = X86::TO;                                                             \
+    break;
+        switch (Opc) {
+        default:
+          llvm_unreachable("Unreachable!");
+        FROM_TO(BLENDPDrri, MOVSDrr)
+        FROM_TO(BLENDPSrri, MOVSSrr)
+        FROM_TO(VBLENDPDrri, VMOVSDrr)
+        FROM_TO(VBLENDPSrri, VMOVSSrr)
+        }
         WorkingMI = CloneIfNew(MI);
         WorkingMI->setDesc(get(Opc));
         WorkingMI->removeOperand(3);
         break;
       }
+#undef FROM_TO
     }
     [[fallthrough]];
   case X86::PBLENDWrri:
diff --git a/llvm/test/CodeGen/X86/commute-blend-avx2.ll b/llvm/test/CodeGen/X86/commute-blend-avx2.ll
index b5ffe78d29a610..75511104580e90 100644
--- a/llvm/test/CodeGen/X86/commute-blend-avx2.ll
+++ b/llvm/test/CodeGen/X86/commute-blend-avx2.ll
@@ -88,3 +88,12 @@ define <4 x double> @commute_fold_vblendpd_256(<4 x double> %a, ptr %b) #0 {
   ret <4 x double> %2
 }
 declare <4 x double> @llvm.x86.avx.blend.pd.256(<4 x double>, <4 x double>, i8) nounwind readnone
+
+define <4 x float> @commute_vblendpd_128_for_code_size(<4 x float> %a, <4 x float> %b) optsize {
+; CHECK-LABEL: commute_vblendpd_128_for_code_size:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vblendps {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2,3]
+; CHECK-NEXT:    retq
+  %r = shufflevector <4 x float> %b, <4 x float> %a, <4 x i32> <i32 0, i32 5, i32 2, i32 3>
+  ret <4 x float> %r
+}