[clang] [ARM] Explicitly enable NEON for Windows/Darwin targets (llvm#122095)

Upstream LLVM implicitly enables NEON for any ARMv7 target.

Many platform ABIs with an ARMv7 baseline also include NEON in that,
this is the case on e.g. Windows and iOS. On Linux, however, things are
not quite as clearly defined. Some distributions target an ARMv7
baseline without NEON available (this is the case for e.g. Debian/Ubuntu
for the "armhf" architecture).

To achieve this, Debian/Ubuntu patch LLVM downstream to make ARMv7 only
implicitly enable VPFv3-D16, not NEON - see [1].

That patch has the (unintended) effect that NEON no longer is available
by default for Windows/ARMv7 and iOS/ARMv7.

In practice, when compiling C for Windows/ARMv7 with Debian patched
Clang, NEON actually still is available, but not when compiling assembly
files. This is due to ARM::getARMCPUForArch (in
llvm/lib/TargetParser/ARMTargetParser.cpp) returning "cortex-a9" for
Windows. This difference, between C and assembly, is due to how
getARMTargetCPU is called in getARMTargetFeatures (in
clang/lib/Driver/ToolChains/Arch/ARM.cpp) to get defaults, only when
ForAS is not set - see [2].

There is an existing case in getARMTargetFeatures, for Android, which
explicitly enables NEON when targeting ARM >= v7. As Windows and iOS
have NEON as part of their ABI baseline just like Android does these
days (see [3] for where this default was added for Android), add the
implicit default in a similar way.

However, first do the general lookup of getARMTargetCPU (unless ForAS);
this makes sure that we get the same default CPU as before ("cortex-a9"
for Windows and "swift" for the "armv7s" architecture on Darwin).

[1] https://salsa.debian.org/pkg-llvm-team/llvm-toolchain/-/blob/19/debian/patches/clang-arm-default-vfp3-on-armv7a.patch?ref_type=heads
[2] llvm@b8baa2a
[3] llvm@d0fbef9

Author: mstorsjo

clang/lib/Driver/ToolChains/Arch/ARM.cpp

@@ -659,13 +659,21 @@ llvm::ARM::FPUKind arm::getARMTargetFeatures(const Driver &D,
         CPUArgFPUKind != llvm::ARM::FK_INVALID ? CPUArgFPUKind : ArchArgFPUKind;
     (void)llvm::ARM::getFPUFeatures(FPUKind, Features);
   } else {
+    bool Generic = true;
     if (!ForAS) {
       std::string CPU = arm::getARMTargetCPU(CPUName, ArchName, Triple);
+      if (CPU != "generic")
+        Generic = false;
       llvm::ARM::ArchKind ArchKind =
           arm::getLLVMArchKindForARM(CPU, ArchName, Triple);
       FPUKind = llvm::ARM::getDefaultFPU(CPU, ArchKind);
       (void)llvm::ARM::getFPUFeatures(FPUKind, Features);
     }
+    if (Generic && (Triple.isOSWindows() || Triple.isOSDarwin()) &&
+        getARMSubArchVersionNumber(Triple) >= 7) {
+      FPUKind = llvm::ARM::parseFPU("neon");
+      (void)llvm::ARM::getFPUFeatures(FPUKind, Features);
+    }
   }
 
   // Now we've finished accumulating features from arch, cpu and fpu,

clang/test/Driver/arm-mfpu.c

@@ -356,8 +356,10 @@
 // CHECK-HF-DAG: "-target-cpu" "arm1176jzf-s"
 
 // RUN: %clang -target armv7-apple-darwin -x assembler %s -### -c 2>&1 \
-// RUN:   | FileCheck --check-prefix=ASM %s
-// ASM-NOT: -target-feature
+// RUN:   | FileCheck --check-prefix=ASM-NEON %s
+// RUN: %clang -target armv7-windows -x assembler %s -### -c 2>&1 \
+// RUN:   | FileCheck --check-prefix=ASM-NEON %s
+// ASM-NEON: "-target-feature" "+neon"
 
 // RUN: %clang -target armv8-linux-gnueabi -mfloat-abi=soft -mfpu=none %s -### -c 2>&1 \
 // RUN:   | FileCheck --check-prefix=CHECK-SOFT-ABI-FP %s

clang/test/Preprocessor/arm-target-features.c

@@ -132,6 +132,30 @@
 // CHECK-V7VE-DEFAULT-ABI-SOFT: #define __ARM_ARCH_EXT_IDIV__ 1
 // CHECK-V7VE-DEFAULT-ABI-SOFT: #define __ARM_FP 0xc
 
+// RUN: %clang -target x86_64-apple-macosx10.10 -arch armv7 -x c -E -dM %s -o - | FileCheck -match-full-lines --check-prefix=CHECK-DARWIN-V7 %s
+// CHECK-DARWIN-V7: #define __ARMEL__ 1
+// CHECK-DARWIN-V7: #define __ARM_ARCH 7
+// CHECK-DARWIN-V7: #define __ARM_ARCH_7A__ 1
+// CHECK-DARWIN-V7-NOT: __ARM_FEATURE_CRC32
+// CHECK-DARWIN-V7-NOT: __ARM_FEATURE_NUMERIC_MAXMIN
+// CHECK-DARWIN-V7-NOT: __ARM_FEATURE_DIRECTED_ROUNDING
+// CHECK-DARWIN-V7: #define __ARM_FP 0xc
+// CHECK-DARWIN-V7: #define __ARM_NEON 1
+// CHECK-DARWIN-V7: #define __ARM_NEON_FP 0x4
+// CHECK-DARWIN-V7: #define __ARM_NEON__ 1
+
+// RUN: %clang -target armv7-windows -x c -E -dM %s -o - | FileCheck -match-full-lines --check-prefix=CHECK-WINDOWS-V7 %s
+// CHECK-WINDOWS-V7: #define __ARMEL__ 1
+// CHECK-WINDOWS-V7: #define __ARM_ARCH 7
+// CHECK-WINDOWS-V7: #define __ARM_ARCH_7A__ 1
+// CHECK-WINDOWS-V7-NOT: __ARM_FEATURE_CRC32
+// CHECK-WINDOWS-V7-NOT: __ARM_FEATURE_NUMERIC_MAXMIN
+// CHECK-WINDOWS-V7-NOT: __ARM_FEATURE_DIRECTED_ROUNDING
+// CHECK-WINDOWS-V7: #define __ARM_FP 0xe
+// CHECK-WINDOWS-V7: #define __ARM_NEON 1
+// CHECK-WINDOWS-V7: #define __ARM_NEON_FP 0x6
+// CHECK-WINDOWS-V7: #define __ARM_NEON__ 1
+
 // RUN: %clang -target x86_64-apple-macosx10.10 -arch armv7s -x c -E -dM %s -o - | FileCheck -match-full-lines --check-prefix=CHECK-V7S %s
 // CHECK-V7S: #define __ARMEL__ 1
 // CHECK-V7S: #define __ARM_ARCH 7
@@ -140,6 +164,9 @@
 // CHECK-V7S-NOT: __ARM_FEATURE_NUMERIC_MAXMIN
 // CHECK-V7S-NOT: __ARM_FEATURE_DIRECTED_ROUNDING
 // CHECK-V7S: #define __ARM_FP 0xe
+// CHECK-V7S: #define __ARM_NEON 1
+// CHECK-V7S: #define __ARM_NEON_FP 0x6
+// CHECK-V7S: #define __ARM_NEON__ 1
 
 // RUN: %clang -target arm-arm-none-eabi -march=armv7-m -mfloat-abi=soft -x c -E -dM %s | FileCheck -match-full-lines --check-prefix=CHECK-VFP-FP %s
 // RUN: %clang -target arm-arm-none-eabi -march=armv7-m -mfloat-abi=softfp -x c -E -dM %s | FileCheck -match-full-lines --check-prefix=CHECK-VFP-FP %s