[x86] invpcid intrinsic

[GBuella]

Summary

An intrinsic for an old instruction, as described in the Intel SDM.

[craig.topper]

lib/Headers/cpuid.h

158

this should be below ENH_MOVSB to keep the bits in order

lib/Headers/intrin.h

196

@rnk, what's the right thing to do here?

[rnk]

Why do we need a feature flag for this in the first place? The MSVC model for most "instruction" intrinsics is that the exact instruction is emitted regardless of the feature enabled. The target attribute seems like it would get in the way of that.

lib/Headers/intrin.h

196

What problems does this redeclaration cause?

[GBuella]

In D47142#1106664, @rnk wrote:

Why do we need a feature flag for this in the first place? The MSVC model for most "instruction" intrinsics is that the exact instruction is emitted regardless of the feature enabled. The target attribute seems like it would get in the way of that.

I'm fine with getting rid of the feature flag, here, and then probably also for other instructions in similar situations (pconfig, wbnoinvd, clwb, etc...).
But as long as we keep those other feature flags, I guess we should just be consistent.

lib/Headers/intrin.h

196

Now that I think about it, none.

[GBuella]

lib/Headers/intrin.h

196

Also, I think this could be added as TARGET_HEADER_BUILTIN..."intrin.h", ALL_MS_LANGUAGES into BuiltinsX86.def
And removed from here.
Right?

[rnk]

lib/Headers/intrin.h

196

It could be, but then you'd need to implement it as a builtin instead of having invpcidintrin.h, right?

[GBuella]

lib/Headers/intrin.h

196

Meanwhile, I just realized, we add the !defined(_MSC_VER) condition around the inclusion of all such header files, as I did with invpcidintrin.h.
Does clang always define _MSC_VER on Windows? I mean, always in cases where one would want to include the MSVC likes`intrin.h`?
In that case, these are completely independent things. Also, I could just add both builtins _invpcid with the ALL_MS_LANGUAGES thing, and the __builtin_ia32_invpcid with the usual clang/GCC convention of adding separate feature flags (both would be lowered to the same LLVM intrinsic).
BTW, if we stop adding feature flags for such intrinsics, we probably have to convince GCC to do the same, as we love that compatibility.

[GBuella]

Relocated header inclusion from x86intrin.h -> immintrin.h
I gave up on the MSVC intrinsic idea, it wouldn't work in a way compatible with MSVC as long as LLVM requires the feature to be enabled anyways. If once we decide to remove such feature flags for such system programming instructions, we can get back this.

[craig.topper]

LGTM, if you fix the ordering in cpuid.h.

include/clang/Basic/BuiltinsX86.def

 
 // Direct Move
 TARGET_BUILTIN(__builtin_ia32_directstore_u32, "vUi*Ui", "n", "movdiri")
 TARGET_BUILTIN(__builtin_ia32_movdir64b, "vv*vC*", "n", "movdir64b")
 
 // PTWRITE
 TARGET_BUILTIN(__builtin_ia32_ptwrite32, "vUi", "n", "ptwrite")
 
+// INVPCID
+TARGET_BUILTIN(__builtin_ia32_invpcid, "vUiv*", "n", "invpcid")
+
 // MSVC
 TARGET_HEADER_BUILTIN(_BitScanForward, "UcUNi*UNi", "nh", "intrin.h", ALL_MS_LANGUAGES, "")
 TARGET_HEADER_BUILTIN(_BitScanReverse, "UcUNi*UNi", "nh", "intrin.h", ALL_MS_LANGUAGES, "")
 
 TARGET_HEADER_BUILTIN(_ReadWriteBarrier, "v", "nh", "intrin.h", ALL_MS_LANGUAGES, "")
 TARGET_HEADER_BUILTIN(_ReadBarrier,      "v", "nh", "intrin.h", ALL_MS_LANGUAGES, "")
 TARGET_HEADER_BUILTIN(_WriteBarrier,     "v", "nh", "intrin.h", ALL_MS_LANGUAGES, "")
 

include/clang/Driver/Options.td

 def mfma : Flag<["-"], "mfma">, Group<m_x86_Features_Group>;
 def mno_fma : Flag<["-"], "mno-fma">, Group<m_x86_Features_Group>;
 def mfma4 : Flag<["-"], "mfma4">, Group<m_x86_Features_Group>;
 def mno_fma4 : Flag<["-"], "mno-fma4">, Group<m_x86_Features_Group>;
 def mfsgsbase : Flag<["-"], "mfsgsbase">, Group<m_x86_Features_Group>;
 def mno_fsgsbase : Flag<["-"], "mno-fsgsbase">, Group<m_x86_Features_Group>;
 def mfxsr : Flag<["-"], "mfxsr">, Group<m_x86_Features_Group>;
 def mno_fxsr : Flag<["-"], "mno-fxsr">, Group<m_x86_Features_Group>;
+def minvpcid : Flag<["-"], "minvpcid">, Group<m_x86_Features_Group>;
+def mno_invpcid : Flag<["-"], "mno-invpcid">, Group<m_x86_Features_Group>;
 def mgfni : Flag<["-"], "mgfni">, Group<m_x86_Features_Group>;
 def mno_gfni : Flag<["-"], "mno-gfni">, Group<m_x86_Features_Group>;
 def mlwp : Flag<["-"], "mlwp">, Group<m_x86_Features_Group>;
 def mno_lwp : Flag<["-"], "mno-lwp">, Group<m_x86_Features_Group>;
 def mlzcnt : Flag<["-"], "mlzcnt">, Group<m_x86_Features_Group>;
 def mno_lzcnt : Flag<["-"], "mno-lzcnt">, Group<m_x86_Features_Group>;
 def mmovbe : Flag<["-"], "mmovbe">, Group<m_x86_Features_Group>;
 def mno_movbe : Flag<["-"], "mno-movbe">, Group<m_x86_Features_Group>;

lib/Basic/Targets/X86.h

class LLVM_LIBRARY_VISIBILITY X86TargetInfo : public TargetInfo {
   bool HasRetpoline = false;
   bool HasRetpolineExternalThunk = false;
   bool HasLAHFSAHF = false;
   bool HasWBNOINVD = false;
   bool HasWAITPKG = false;
   bool HasMOVDIRI = false;
   bool HasMOVDIR64B = false;
   bool HasPTWRITE = false;
+  bool HasINVPCID = false;
 
 protected:
   /// Enumeration of all of the X86 CPUs supported by Clang.
   ///
   /// Each enumeration represents a particular CPU supported by Clang. These
   /// loosely correspond to the options passed to '-march' or '-mtune' flags.
   enum CPUKind {
     CK_Generic,

lib/Basic/Targets/X86.cpp

case CK_Broadwell:
     setFeatureEnabledImpl(Features, "prfchw", true);
     LLVM_FALLTHROUGH;
   case CK_Haswell:
     setFeatureEnabledImpl(Features, "avx2", true);
     setFeatureEnabledImpl(Features, "lzcnt", true);
     setFeatureEnabledImpl(Features, "bmi", true);
     setFeatureEnabledImpl(Features, "bmi2", true);
     setFeatureEnabledImpl(Features, "fma", true);
+    setFeatureEnabledImpl(Features, "invpcid", true);
     setFeatureEnabledImpl(Features, "movbe", true);
     LLVM_FALLTHROUGH;
   case CK_IvyBridge:
     setFeatureEnabledImpl(Features, "rdrnd", true);
     setFeatureEnabledImpl(Features, "f16c", true);
     setFeatureEnabledImpl(Features, "fsgsbase", true);
     LLVM_FALLTHROUGH;
   case CK_SandyBridge:
for (const auto &Feature : Features) {
     } else if (Feature == "+movdiri") {
       HasMOVDIRI = true;
     } else if (Feature == "+movdir64b") {
       HasMOVDIR64B = true;
     } else if (Feature == "+pconfig") {
       HasPCONFIG = true;
     } else if (Feature == "+ptwrite") {
       HasPTWRITE = true;
+    } else if (Feature == "+invpcid") {
+      HasINVPCID = true;
     }
 
     X86SSEEnum Level = llvm::StringSwitch<X86SSEEnum>(Feature)
                            .Case("+avx512f", AVX512F)
                            .Case("+avx2", AVX2)
                            .Case("+avx", AVX)
                            .Case("+sse4.2", SSE42)
                            .Case("+sse4.1", SSE41)
void X86TargetInfo::getTargetDefines(const LangOptions &Opts,
   if (HasMOVDIRI)
     Builder.defineMacro("__MOVDIRI__");
   if (HasMOVDIR64B)
     Builder.defineMacro("__MOVDIR64B__");
   if (HasPCONFIG)
     Builder.defineMacro("__PCONFIG__");
   if (HasPTWRITE)
     Builder.defineMacro("__PTWRITE__");
+  if (HasINVPCID)
+    Builder.defineMacro("__INVPCID__");
 
   // Each case falls through to the previous one here.
   switch (SSELevel) {
   case AVX512F:
     Builder.defineMacro("__AVX512F__");
     LLVM_FALLTHROUGH;
   case AVX2:
     Builder.defineMacro("__AVX2__");
return llvm::StringSwitch<bool>(Name)
       .Case("clzero", true)
       .Case("cx16", true)
       .Case("f16c", true)
       .Case("fma", true)
       .Case("fma4", true)
       .Case("fsgsbase", true)
       .Case("fxsr", true)
       .Case("gfni", true)
+      .Case("invpcid", true)
       .Case("lwp", true)
       .Case("lzcnt", true)
       .Case("mmx", true)
       .Case("movbe", true)
       .Case("movdiri", true)
       .Case("movdir64b", true)
       .Case("mpx", true)
       .Case("mwaitx", true)
return llvm::StringSwitch<bool>(Feature)
       .Case("clzero", HasCLZERO)
       .Case("cx16", HasCX16)
       .Case("f16c", HasF16C)
       .Case("fma", HasFMA)
       .Case("fma4", XOPLevel >= FMA4)
       .Case("fsgsbase", HasFSGSBASE)
       .Case("fxsr", HasFXSR)
       .Case("gfni", HasGFNI)
+      .Case("invpcid", HasINVPCID)
       .Case("lwp", HasLWP)
       .Case("lzcnt", HasLZCNT)
       .Case("mm3dnow", MMX3DNowLevel >= AMD3DNow)
       .Case("mm3dnowa", MMX3DNowLevel >= AMD3DNowAthlon)
       .Case("mmx", MMX3DNowLevel >= MMX)
       .Case("movbe", HasMOVBE)
       .Case("movdiri", HasMOVDIRI)
       .Case("movdir64b", HasMOVDIR64B)

lib/Headers/CMakeLists.txt

set(files
   fxsrintrin.h
   gfniintrin.h
   htmintrin.h
   htmxlintrin.h
   ia32intrin.h
   immintrin.h
   intrin.h
   inttypes.h
+  invpcidintrin.h
   iso646.h
   limits.h
   lwpintrin.h
   lzcntintrin.h
   mm3dnow.h
   mmintrin.h
   mm_malloc.h
   module.modulemap

lib/Headers/cpuid.h

 /* Features in %ebx for leaf 7 sub-leaf 0 */
 #define bit_FSGSBASE    0x00000001
 #define bit_SGX         0x00000004
 #define bit_BMI         0x00000008
 #define bit_HLE         0x00000010
 #define bit_AVX2        0x00000020
 #define bit_SMEP        0x00000080
 #define bit_BMI2        0x00000100
+#define bit_INVCPID     0x00000400
 #define bit_ENH_MOVSB   0x00000200
 #define bit_RTM         0x00000800
 #define bit_MPX         0x00004000
 #define bit_AVX512F     0x00010000
 #define bit_AVX512DQ    0x00020000
 #define bit_RDSEED      0x00040000
 #define bit_ADX         0x00080000
 #define bit_AVX512IFMA  0x00200000

lib/Headers/intrin.h

 void *_InterlockedCompareExchangePointer_HLEAcquire(void *volatile *, void *,
                                                     void *);
 void *_InterlockedCompareExchangePointer_HLERelease(void *volatile *, void *,
                                                     void *);
 long _InterlockedExchangeAdd_HLEAcquire(long volatile *, long);
 long _InterlockedExchangeAdd_HLERelease(long volatile *, long);
 __int64 _InterlockedExchangeAdd64_HLEAcquire(__int64 volatile *, __int64);
 __int64 _InterlockedExchangeAdd64_HLERelease(__int64 volatile *, __int64);
+#ifndef __INVPCIDINTRIN_H
+/*
+ * This might be already defined, when including x86intrin.h at the
+ * beginning of this header.
+ */
 void __cdecl _invpcid(unsigned int, void *);
+#endif
 static __inline__ void
 __attribute__((__deprecated__("use other intrinsics or C++11 atomics instead")))
 _ReadBarrier(void);
 static __inline__ void
 __attribute__((__deprecated__("use other intrinsics or C++11 atomics instead")))
 _ReadWriteBarrier(void);
 unsigned int _rorx_u32(unsigned int, const unsigned int);
 int _sarx_i32(int, unsigned int);

lib/Headers/invpcidintrin.h

+/*===------------- invpcidintrin.h - INVPCID intrinsic ---------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __X86INTRIN_H
+#error "Never use <invpcidintrin.h> directly; include <x86intrin.h> instead."
+#endif
+
+#ifndef __INVPCIDINTRIN_H
+#define __INVPCIDINTRIN_H
+
+static __inline__ void
+  __attribute__((__always_inline__, __nodebug__,  __target__("invpcid")))
+_invpcid(unsigned int __type, void *__descriptor) {
+  __builtin_ia32_invpcid(__type, __descriptor);
+}
+
+#endif /* __INVPCIDINTRIN_H */

lib/Headers/module.modulemap

explicit module intel {
     textual header "clzerointrin.h"
     textual header "wbnoinvdintrin.h"
     textual header "cldemoteintrin.h"
     textual header "waitpkgintrin.h"
     textual header "movdirintrin.h"
     textual header "pconfigintrin.h"
     textual header "sgxintrin.h"
     textual header "ptwriteintrin.h"
+    textual header "invpcidintrin.h"
 
     explicit module mm_malloc {
       requires !freestanding
       header "mm_malloc.h"
       export * // note: for <stdlib.h> dependency
     }
 
     explicit module cpuid {

lib/Headers/x86intrin.h

 #if !defined(_MSC_VER) || __has_feature(modules) || defined(__SGX__)
 #include <sgxintrin.h>
 #endif
 
 #if !defined(_MSC_VER) || __has_feature(modules) || defined(__PTWRITE__)
 #include <ptwriteintrin.h>
 #endif
 
+#if !defined(_MSC_VER) || __has_feature(modules) || defined(__INVPCID__)
+#include <invpcidintrin.h>
+#endif
+
 #endif /* __X86INTRIN_H */

test/CodeGen/invpcid.c

+// RUN: %clang_cc1 %s -ffreestanding -triple=x86_64-unknown-unknown -target-feature +invpcid -emit-llvm -o - -Wall -Werror -pedantic | FileCheck %s
+// RUN: %clang_cc1 %s -ffreestanding -triple=i386-unknown-unknown -target-feature +invpcid -emit-llvm -o - -Wall -Werror -pedantic | FileCheck %s
+
+#include <x86intrin.h>
+
+#include <stdint.h>
+
+void test_invpcid(uint32_t type, void *descriptor) {
+  //CHECK-LABEL: @test_invpcid
+  //CHECK: call void @llvm.x86.invpcid(i32 %{{.*}}, i8* %{{.*}})
+  _invpcid(type, descriptor);
+}

test/Driver/x86-target-features.c

 // RUN: %clang -target i386-unknown-linux-gnu -march=i386 -mno-pconfig %s -### -o %t.o 2>&1 | FileCheck -check-prefix=NO-PCONFIG %s
 // PCONFIG: "-target-feature" "+pconfig"
 // NO-PCONFIG: "-target-feature" "-pconfig"
 
 // RUN: %clang -target i386-unknown-linux-gnu -march=i386 -mptwrite %s -### -o %t.o 2>&1 | FileCheck -check-prefix=PTWRITE %s
 // RUN: %clang -target i386-unknown-linux-gnu -march=i386 -mno-ptwrite %s -### -o %t.o 2>&1 | FileCheck -check-prefix=NO-PTWRITE %s
 // PTWRITE: "-target-feature" "+ptwrite"
 // NO-PTWRITE: "-target-feature" "-ptwrite"
+
+// RUN: %clang -target i386-unknown-linux-gnu -march=i386 -minvpcid %s -### -o %t.o 2>&1 | FileCheck -check-prefix=INVPCID %s
+// RUN: %clang -target i386-unknown-linux-gnu -march=i386 -mno-invpcid %s -### -o %t.o 2>&1 | FileCheck -check-prefix=NO-INVPCID %s
+// INVPCID: "-target-feature" "+invpcid"
+// NO-INVPCID: "-target-feature" "-invpcid"

test/Preprocessor/predefined-arch-macros.c

 // RUN:   | FileCheck -match-full-lines %s -check-prefix=CHECK_CORE_AVX2_M32
 // CHECK_CORE_AVX2_M32: #define __AES__ 1
 // CHECK_CORE_AVX2_M32: #define __AVX2__ 1
 // CHECK_CORE_AVX2_M32: #define __AVX__ 1
 // CHECK_CORE_AVX2_M32: #define __BMI2__ 1
 // CHECK_CORE_AVX2_M32: #define __BMI__ 1
 // CHECK_CORE_AVX2_M32: #define __F16C__ 1
 // CHECK_CORE_AVX2_M32: #define __FMA__ 1
+// CHECK_CORE_AVX2_M32: #define __INVPCID__ 1
 // CHECK_CORE_AVX2_M32: #define __LZCNT__ 1
 // CHECK_CORE_AVX2_M32: #define __MMX__ 1
 // CHECK_CORE_AVX2_M32: #define __PCLMUL__ 1
 // CHECK_CORE_AVX2_M32: #define __POPCNT__ 1
 // CHECK_CORE_AVX2_M32: #define __RDRND__ 1
 // CHECK_CORE_AVX2_M32: #define __SSE2__ 1
 // CHECK_CORE_AVX2_M32: #define __SSE3__ 1
 // CHECK_CORE_AVX2_M32: #define __SSE4_1__ 1

 // RUN:   | FileCheck -match-full-lines %s -check-prefix=CHECK_CORE_AVX2_M64
 // CHECK_CORE_AVX2_M64: #define __AES__ 1
 // CHECK_CORE_AVX2_M64: #define __AVX2__ 1
 // CHECK_CORE_AVX2_M64: #define __AVX__ 1
 // CHECK_CORE_AVX2_M64: #define __BMI2__ 1
 // CHECK_CORE_AVX2_M64: #define __BMI__ 1
 // CHECK_CORE_AVX2_M64: #define __F16C__ 1
 // CHECK_CORE_AVX2_M64: #define __FMA__ 1
+// CHECK_CORE_AVX2_M64: #define __INVPCID__ 1
 // CHECK_CORE_AVX2_M64: #define __LZCNT__ 1
 // CHECK_CORE_AVX2_M64: #define __MMX__ 1
 // CHECK_CORE_AVX2_M64: #define __PCLMUL__ 1
 // CHECK_CORE_AVX2_M64: #define __POPCNT__ 1
 // CHECK_CORE_AVX2_M64: #define __RDRND__ 1
 // CHECK_CORE_AVX2_M64: #define __SSE2_MATH__ 1
 // CHECK_CORE_AVX2_M64: #define __SSE2__ 1
 // CHECK_CORE_AVX2_M64: #define __SSE3__ 1

 // CHECK_BROADWELL_M32: #define __ADX__ 1
 // CHECK_BROADWELL_M32: #define __AES__ 1
 // CHECK_BROADWELL_M32: #define __AVX2__ 1
 // CHECK_BROADWELL_M32: #define __AVX__ 1
 // CHECK_BROADWELL_M32: #define __BMI2__ 1
 // CHECK_BROADWELL_M32: #define __BMI__ 1
 // CHECK_BROADWELL_M32: #define __F16C__ 1
 // CHECK_BROADWELL_M32: #define __FMA__ 1
+// CHECK_BROADWELL_M32: #define __INVPCID__ 1
 // CHECK_BROADWELL_M32: #define __LZCNT__ 1
 // CHECK_BROADWELL_M32: #define __MMX__ 1
 // CHECK_BROADWELL_M32: #define __PCLMUL__ 1
 // CHECK_BROADWELL_M32: #define __POPCNT__ 1
 // CHECK_BROADWELL_M32: #define __PRFCHW__ 1
 // CHECK_BROADWELL_M32: #define __RDRND__ 1
 // CHECK_BROADWELL_M32: #define __RDSEED__ 1
 // CHECK_BROADWELL_M32: #define __SSE2__ 1

 // CHECK_BROADWELL_M64: #define __ADX__ 1
 // CHECK_BROADWELL_M64: #define __AES__ 1
 // CHECK_BROADWELL_M64: #define __AVX2__ 1
 // CHECK_BROADWELL_M64: #define __AVX__ 1
 // CHECK_BROADWELL_M64: #define __BMI2__ 1
 // CHECK_BROADWELL_M64: #define __BMI__ 1
 // CHECK_BROADWELL_M64: #define __F16C__ 1
 // CHECK_BROADWELL_M64: #define __FMA__ 1
+// CHECK_BROADWELL_M64: #define __INVPCID__ 1
 // CHECK_BROADWELL_M64: #define __LZCNT__ 1
 // CHECK_BROADWELL_M64: #define __MMX__ 1
 // CHECK_BROADWELL_M64: #define __PCLMUL__ 1
 // CHECK_BROADWELL_M64: #define __POPCNT__ 1
 // CHECK_BROADWELL_M64: #define __PRFCHW__ 1
 // CHECK_BROADWELL_M64: #define __RDRND__ 1
 // CHECK_BROADWELL_M64: #define __RDSEED__ 1
 // CHECK_BROADWELL_M64: #define __SSE2_MATH__ 1

 // CHECK_SKL_M32: #define __AES__ 1
 // CHECK_SKL_M32: #define __AVX2__ 1
 // CHECK_SKL_M32: #define __AVX__ 1
 // CHECK_SKL_M32: #define __BMI2__ 1
 // CHECK_SKL_M32: #define __BMI__ 1
 // CHECK_SKL_M32: #define __CLFLUSHOPT__ 1
 // CHECK_SKL_M32: #define __F16C__ 1
 // CHECK_SKL_M32: #define __FMA__ 1
+// CHECK_SKL_M32: #define __INVPCID__ 1
 // CHECK_SKL_M32: #define __LZCNT__ 1
 // CHECK_SKL_M32: #define __MMX__ 1
 // CHECK_SKL_M32: #define __MPX__ 1
 // CHECK_SKL_M32: #define __PCLMUL__ 1
 // CHECK_SKL_M32: #define __POPCNT__ 1
 // CHECK_SKL_M32: #define __PRFCHW__ 1
 // CHECK_SKL_M32: #define __RDRND__ 1
 // CHECK_SKL_M32: #define __RDSEED__ 1

 // CHECK_SKL_M64: #define __AES__ 1
 // CHECK_SKL_M64: #define __AVX2__ 1
 // CHECK_SKL_M64: #define __AVX__ 1
 // CHECK_SKL_M64: #define __BMI2__ 1
 // CHECK_SKL_M64: #define __BMI__ 1
 // CHECK_SKL_M64: #define __CLFLUSHOPT__ 1
 // CHECK_SKL_M64: #define __F16C__ 1
 // CHECK_SKL_M64: #define __FMA__ 1
+// CHECK_SKL_M64: #define __INVPCID__ 1
 // CHECK_SKL_M64: #define __LZCNT__ 1
 // CHECK_SKL_M64: #define __MMX__ 1
 // CHECK_SKL_M64: #define __MPX__ 1
 // CHECK_SKL_M64: #define __PCLMUL__ 1
 // CHECK_SKL_M64: #define __POPCNT__ 1
 // CHECK_SKL_M64: #define __PRFCHW__ 1
 // CHECK_SKL_M64: #define __RDRND__ 1
 // CHECK_SKL_M64: #define __RDSEED__ 1

 // CHECK_SKX_M32: #define __AVX512VL__ 1
 // CHECK_SKX_M32: #define __AVX__ 1
 // CHECK_SKX_M32: #define __BMI2__ 1
 // CHECK_SKX_M32: #define __BMI__ 1
 // CHECK_SKX_M32: #define __CLFLUSHOPT__ 1
 // CHECK_SKX_M32: #define __CLWB__ 1
 // CHECK_SKX_M32: #define __F16C__ 1
 // CHECK_SKX_M32: #define __FMA__ 1
+// CHECK_SKX_M32: #define __INVPCID__ 1
 // CHECK_SKX_M32: #define __LZCNT__ 1
 // CHECK_SKX_M32: #define __MMX__ 1
 // CHECK_SKX_M32: #define __MPX__ 1
 // CHECK_SKX_M32: #define __PCLMUL__ 1
 // CHECK_SKX_M32: #define __PKU__ 1
 // CHECK_SKX_M32: #define __POPCNT__ 1
 // CHECK_SKX_M32: #define __PRFCHW__ 1
 // CHECK_SKX_M32: #define __RDRND__ 1

 // CHECK_SKX_M64: #define __AVX512VL__ 1
 // CHECK_SKX_M64: #define __AVX__ 1
 // CHECK_SKX_M64: #define __BMI2__ 1
 // CHECK_SKX_M64: #define __BMI__ 1
 // CHECK_SKX_M64: #define __CLFLUSHOPT__ 1
 // CHECK_SKX_M64: #define __CLWB__ 1
 // CHECK_SKX_M64: #define __F16C__ 1
 // CHECK_SKX_M64: #define __FMA__ 1
+// CHECK_SKX_M64: #define __INVPCID__ 1
 // CHECK_SKX_M64: #define __LZCNT__ 1
 // CHECK_SKX_M64: #define __MMX__ 1
 // CHECK_SKX_M64: #define __MPX__ 1
 // CHECK_SKX_M64: #define __PCLMUL__ 1
 // CHECK_SKX_M64: #define __PKU__ 1
 // CHECK_SKX_M64: #define __POPCNT__ 1
 // CHECK_SKX_M64: #define __PRFCHW__ 1
 // CHECK_SKX_M64: #define __RDRND__ 1

 // CHECK_CNL_M32: #define __AVX512VL__ 1
 // CHECK_CNL_M32: #define __AVX__ 1
 // CHECK_CNL_M32: #define __BMI2__ 1
 // CHECK_CNL_M32: #define __BMI__ 1
 // CHECK_CNL_M32: #define __CLFLUSHOPT__ 1
 // CHECK_CNL_M32-NOT: #define __CLWB__ 1
 // CHECK_CNL_M32: #define __F16C__ 1
 // CHECK_CNL_M32: #define __FMA__ 1
+// CHECK_CNL_M32: #define __INVPCID__ 1
 // CHECK_CNL_M32: #define __LZCNT__ 1
 // CHECK_CNL_M32: #define __MMX__ 1
 // CHECK_CNL_M32: #define __MPX__ 1
 // CHECK_CNL_M32: #define __PCLMUL__ 1
 // CHECK_CNL_M32: #define __PKU__ 1
 // CHECK_CNL_M32: #define __POPCNT__ 1
 // CHECK_CNL_M32: #define __PRFCHW__ 1
 // CHECK_CNL_M32: #define __RDRND__ 1

 // CHECK_CNL_M64: #define __AVX512VL__ 1
 // CHECK_CNL_M64: #define __AVX__ 1
 // CHECK_CNL_M64: #define __BMI2__ 1
 // CHECK_CNL_M64: #define __BMI__ 1
 // CHECK_CNL_M64: #define __CLFLUSHOPT__ 1
 // CHECK_CNL_M64-NOT: #define __CLWB__ 1
 // CHECK_CNL_M64: #define __F16C__ 1
 // CHECK_CNL_M64: #define __FMA__ 1
+// CHECK_CNL_M64: #define __INVPCID__ 1
 // CHECK_CNL_M64: #define __LZCNT__ 1
 // CHECK_CNL_M64: #define __MMX__ 1
 // CHECK_CNL_M64: #define __MPX__ 1
 // CHECK_CNL_M64: #define __PCLMUL__ 1
 // CHECK_CNL_M64: #define __PKU__ 1
 // CHECK_CNL_M64: #define __POPCNT__ 1
 // CHECK_CNL_M64: #define __PRFCHW__ 1
 // CHECK_CNL_M64: #define __RDRND__ 1

 // CHECK_ICL_M32: #define __AVX__ 1
 // CHECK_ICL_M32: #define __BMI2__ 1
 // CHECK_ICL_M32: #define __BMI__ 1
 // CHECK_ICL_M32: #define __CLFLUSHOPT__ 1
 // CHECK_ICL_M32: #define __CLWB__ 1
 // CHECK_ICL_M32: #define __F16C__ 1
 // CHECK_ICL_M32: #define __FMA__ 1
 // CHECK_ICL_M32: #define __GFNI__ 1
+// CHECK_ICL_M32: #define __INVPCID__ 1
 // CHECK_ICL_M32: #define __LZCNT__ 1
 // CHECK_ICL_M32: #define __MMX__ 1
 // CHECK_ICL_M32: #define __MPX__ 1
 // CHECK_ICL_M32: #define __PCLMUL__ 1
 // CHECK_ICL_M32: #define __PKU__ 1
 // CHECK_ICL_M32: #define __POPCNT__ 1
 // CHECK_ICL_M32: #define __PRFCHW__ 1
 // CHECK_ICL_M32: #define __RDPID__ 1

 // CHECK_ICL_M64: #define __AVX__ 1
 // CHECK_ICL_M64: #define __BMI2__ 1
 // CHECK_ICL_M64: #define __BMI__ 1
 // CHECK_ICL_M64: #define __CLFLUSHOPT__ 1
 // CHECK_ICL_M64: #define __CLWB__ 1
 // CHECK_ICL_M64: #define __F16C__ 1
 // CHECK_ICL_M64: #define __FMA__ 1
 // CHECK_ICL_M64: #define __GFNI__ 1
+// CHECK_ICL_M64: #define __INVPCID__ 1
 // CHECK_ICL_M64: #define __LZCNT__ 1
 // CHECK_ICL_M64: #define __MMX__ 1
 // CHECK_ICL_M64: #define __MPX__ 1
 // CHECK_ICL_M64: #define __PCLMUL__ 1
 // CHECK_ICL_M64: #define __PKU__ 1
 // CHECK_ICL_M64: #define __POPCNT__ 1
 // CHECK_ICL_M64: #define __PRFCHW__ 1
 // CHECK_ICL_M64: #define __RDPID__ 1

 // CHECK_ICX_M32: #define __AVX__ 1
 // CHECK_ICX_M32: #define __BMI2__ 1
 // CHECK_ICX_M32: #define __BMI__ 1
 // CHECK_ICX_M32: #define __CLFLUSHOPT__ 1
 // CHECK_ICX_M32: #define __CLWB__ 1
 // CHECK_ICX_M32: #define __F16C__ 1
 // CHECK_ICX_M32: #define __FMA__ 1
 // CHECK_ICX_M32: #define __GFNI__ 1
+// CHECK_ICX_M32: #define __INVPCID__ 1
 // CHECK_ICX_M32: #define __LZCNT__ 1
 // CHECK_ICX_M32: #define __MMX__ 1
 // CHECK_ICX_M32: #define __MPX__ 1
 // CHECK_ICX_M32: #define __PCLMUL__ 1
 // CHECK_ICX_M32: #define __PCONFIG__ 1
 // CHECK_ICX_M32: #define __PKU__ 1
 // CHECK_ICX_M32: #define __POPCNT__ 1
 // CHECK_ICX_M32: #define __PRFCHW__ 1

 // CHECK_ICX_M64: #define __AVX__ 1
 // CHECK_ICX_M64: #define __BMI2__ 1
 // CHECK_ICX_M64: #define __BMI__ 1
 // CHECK_ICX_M64: #define __CLFLUSHOPT__ 1
 // CHECK_ICX_M64: #define __CLWB__ 1
 // CHECK_ICX_M64: #define __F16C__ 1
 // CHECK_ICX_M64: #define __FMA__ 1
 // CHECK_ICX_M64: #define __GFNI__ 1
+// CHECK_ICX_M64: #define __INVPCID__ 1
 // CHECK_ICX_M64: #define __LZCNT__ 1
 // CHECK_ICX_M64: #define __MMX__ 1
 // CHECK_ICX_M64: #define __MPX__ 1
 // CHECK_ICX_M64: #define __PCLMUL__ 1
 // CHECK_ICX_M64: #define __PCONFIG__ 1
 // CHECK_ICX_M64: #define __PKU__ 1
 // CHECK_ICX_M64: #define __POPCNT__ 1
 // CHECK_ICX_M64: #define __PRFCHW__ 1