ARM32 vector mul

BUG= https://bugs.chromium.org/p/nativeclient/issues/detail?id=4076
R=kschimpf@google.com

Review URL: https://codereview.chromium.org/1646033002 .

diff --git a/src/IceInstARM32.cpp b/src/IceInstARM32.cpp
index 8eca8cf..78e4e9a 100644 (file)
--- a/src/IceInstARM32.cpp
+++ b/src/IceInstARM32.cpp
@@ -708,13 +708,15 @@ template <> void InstARM32Vmul::emitIAS(const Cfg *Func) const {
   switch (Dest->getType()) {
   default:
     // TODO(kschimpf) Figure if more cases are needed.
-    Asm->setNeedsTextFixup();
+    emitUsingTextFixup(Func);
     break;
   case IceType_f32:
     Asm->vmuls(getDest(), getSrc(0), getSrc(1), CondARM32::AL);
+    assert(!Asm->needsTextFixup());
     break;
   case IceType_f64:
     Asm->vmuld(getDest(), getSrc(0), getSrc(1), CondARM32::AL);
+    assert(!Asm->needsTextFixup());
     break;
   }
   assert(!Asm->needsTextFixup());

diff --git a/src/IceTargetLoweringARM32.cpp b/src/IceTargetLoweringARM32.cpp
index 10e1dbe..f369ff2 100644 (file)
--- a/src/IceTargetLoweringARM32.cpp
+++ b/src/IceTargetLoweringARM32.cpp
@@ -2810,6 +2810,8 @@ void TargetARM32::lowerArithmetic(const InstArithmetic *Instr) {
     case InstArithmetic::And:
     case InstArithmetic::Or:
     case InstArithmetic::Xor:
+    case InstArithmetic::Fmul:
+    case InstArithmetic::Mul:
       break;
     }
   }
@@ -3116,7 +3118,11 @@ void TargetARM32::lowerArithmetic(const InstArithmetic *Instr) {
     }
     Variable *Src0R = Srcs.unswappedSrc0R(this);
     Variable *Src1R = Srcs.unswappedSrc1R(this);
-    _mul(T, Src0R, Src1R);
+    if (isVectorType(DestTy)) {
+      _vmul(T, Src0R, Src1R);
+    } else {
+      _mul(T, Src0R, Src1R);
+    }
     _mov(Dest, T);
     return;
   }

diff --git a/tests_lit/assembler/arm32/mul-vec.ll b/tests_lit/assembler/arm32/mul-vec.ll
new file mode 100644 (file)
index 0000000..534a3ce
--- /dev/null
+++ b/tests_lit/assembler/arm32/mul-vec.ll
@@ -0,0 +1,85 @@
+; Show that we know how to translate vmul vector instructions.
+
+; REQUIRES: allow_dump
+
+; Compile using standalone assembler.
+; RUN: %p2i --filetype=asm -i %s --target=arm32 --args -O2 \
+; RUN:   -reg-use q10,q11 \
+; RUN:   | FileCheck %s --check-prefix=ASM
+
+; Show bytes in assembled standalone code.
+; RUN: %p2i --filetype=asm -i %s --target=arm32 --assemble --disassemble \
+; RUN:   --args -O2 \
+; RUN:   -reg-use q10,q11 \
+; RUN:   | FileCheck %s --check-prefix=DIS
+
+; Compile using integrated assembler.
+; RUN: %p2i --filetype=iasm -i %s --target=arm32 --args -O2 \
+; RUN:   -reg-use q10,q11 \
+; RUN:   | FileCheck %s --check-prefix=IASM
+
+; Show bytes in assembled integrated code.
+; RUN: %p2i --filetype=iasm -i %s --target=arm32 --assemble --disassemble \
+; RUN:   --args -O2 \
+; RUN:   -reg-use q10,q11 \
+; RUN:   | FileCheck %s --check-prefix=DIS
+
+define internal <4 x float> @testVmulFloat4(<4 x float> %v1, <4 x float> %v2) {
+; ASM-LABEL: testVmulFloat4:
+; DIS-LABEL: 00000000 <testVmulFloat4>:
+; IASM-LABEL: testVmulFloat4:
+
+entry:
+  %res = fmul <4 x float> %v1, %v2
+
+; ASM:     vmul.f32        q10, q10, q11
+; DIS:   8:       f3444df6
+; IASM:     vmul.f32
+
+  ret <4 x float> %res
+}
+
+define internal <4 x i32> @testVmul4i32(<4 x i32> %v1, <4 x i32> %v2) {
+; ASM-LABEL: testVmul4i32:
+; DIS-LABEL: 00000020 <testVmul4i32>:
+; IASM-LABEL: testVmul4i32:
+
+entry:
+  %res = mul <4 x i32> %v1, %v2
+
+; ASM:     vmul.i32        q10, q10, q11
+; DIS:   28:       f26449f6
+; IASM:     vmul.i32
+
+  ret <4 x i32> %res
+}
+
+define internal <8 x i16> @testVmul8i16(<8 x i16> %v1, <8 x i16> %v2) {
+; ASM-LABEL: testVmul8i16:
+; DIS-LABEL: 00000040 <testVmul8i16>:
+; IASM-LABEL: testVmul8i16:
+
+entry:
+  %res = mul <8 x i16> %v1, %v2
+
+; ASM:     vmul.i16        q10, q10, q11
+; DIS:   48:       f25449f6
+; IASM:     vmul.i16
+
+  ret <8 x i16> %res
+}
+
+define internal <16 x i8> @testVmul16i8(<16 x i8> %v1, <16 x i8> %v2) {
+; ASM-LABEL: testVmul16i8:
+; DIS-LABEL: 00000060 <testVmul16i8>:
+; IASM-LABEL: testVmul16i8:
+
+entry:
+  %res = mul <16 x i8> %v1, %v2
+
+; ASM:     vmul.i8        q10, q10, q11
+; DIS:   68:       f24449f6
+; IASM:     vmul.i8
+
+  ret <16 x i8> %res
+}