S390 [liftoff]: Implement simd I8x16Popcnt

Change-Id: Ide34fe0b2e5c12640a084b001ae17ff316de3f3e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3423578Reviewed-by: Junliang Yan <junyan@redhat.com>
Commit-Queue: Milad Farazmand <mfarazma@redhat.com>
Cr-Commit-Position: refs/heads/main@{#78877}

src/codegen/s390/macro-assembler-s390.cc

@@ -5262,7 +5262,8 @@ void TurboAssembler::S128Not(Simd128Register dst, Simd128Register src) {
   V(I16x8UConvertI8x16Low, vupll, 0, 0, 0)  \
   V(I16x8UConvertI8x16High, vuplh, 0, 0, 0) \
   V(I8x16Abs, vlp, 0, 0, 0)                 \
-  V(I8x16Neg, vlc, 0, 0, 0)
+  V(I8x16Neg, vlc, 0, 0, 0)                 \
+  V(I8x16Popcnt, vpopct, 0, 0, 0)
 
 #define EMIT_SIMD_UNOP_VRR_A(name, op, c1, c2, c3)                      \
   void TurboAssembler::name(Simd128Register dst, Simd128Register src) { \

src/codegen/s390/macro-assembler-s390.h

@@ -1152,6 +1152,7 @@ class V8_EXPORT_PRIVATE TurboAssembler : public TurboAssemblerBase {
   V(I16x8UConvertI8x16High) \
   V(I8x16Abs)               \
   V(I8x16Neg)               \
+  V(I8x16Popcnt)            \
   V(S128Not)
 
 #define PROTOTYPE_SIMD_UNOP(name) \

src/compiler/backend/s390/code-generator-s390.cc

@@ -3195,8 +3195,7 @@ CodeGenerator::CodeGenResult CodeGenerator::AssembleArchInstruction(
     }
 #undef Q15_MUL_ROAUND
     case kS390_I8x16Popcnt: {
-      __ vpopct(i.OutputSimd128Register(), i.InputSimd128Register(0),
-                Condition(0), Condition(0), Condition(0));
+      __ I8x16Popcnt(i.OutputSimd128Register(), i.InputSimd128Register(0));
       break;
     }
     case kS390_F64x2ConvertLowI32x4S: {

src/wasm/baseline/s390/liftoff-assembler-s390.h

@@ -2631,7 +2631,7 @@ void LiftoffAssembler::emit_i8x16_shuffle(LiftoffRegister dst,
 
 void LiftoffAssembler::emit_i8x16_popcnt(LiftoffRegister dst,
                                          LiftoffRegister src) {
-  bailout(kSimd, "i8x16.popcnt");
+  I8x16Popcnt(dst.fp(), src.fp());
 }
 
 void LiftoffAssembler::emit_v128_anytrue(LiftoffRegister dst,