[clang][X86] Add __cpuidex function to cpuid.h (#97785)

MSVC has a __cpuidex function implemented to call the underlying cpuid
instruction which accepts a leaf, subleaf, and data array that the
output data is written into. This patch adds this functionality into
clang under the cpuid.h header. This also makes clang match GCC's
behavior. GCC has had __cpuidex in its cpuid.h since 2020.

This is another attempt to land https://reviews.llvm.org/D158348.

clang/lib/Headers/cpuid.h

@@ -339,4 +339,13 @@ static __inline int __get_cpuid_count (unsigned int __leaf,
     return 1;
 }
 
+// In some configurations, __cpuidex is defined as a builtin (primarily
+// -fms-extensions) which will conflict with the __cpuidex definition below.
+#if !(__has_builtin(__cpuidex))
+static __inline void __cpuidex(int __cpu_info[4], int __leaf, int __subleaf) {
+  __cpuid_count(__leaf, __subleaf, __cpu_info[0], __cpu_info[1], __cpu_info[2],
+                __cpu_info[3]);
+}
+#endif
+
 #endif /* __CPUID_H */

clang/test/Headers/__cpuidex_conflict.c

@@ -0,0 +1,22 @@
+// Make sure that __cpuidex in cpuid.h doesn't conflict with the MS
+// extensions built in by ensuring compilation succeeds:
+// RUN: %clang_cc1 %s -ffreestanding -fms-extensions -fms-compatibility \
+// RUN:  -fms-compatibility-version=19.00 -triple x86_64-pc-windows-msvc -emit-llvm -o -
+// %clang_cc1 %s -ffreestanding -triple x86_64-w64-windows-gnu -fms-extensions -emit-llvm -o -
+// RUN: %clang_cc1 %s -ffreestanding -fopenmp -fopenmp-is-target-device -aux-triple x86_64-unknown-linux-gnu
+
+typedef __SIZE_TYPE__ size_t;
+
+// We declare __cpuidex here as where the buitlin should be exposed (MSVC), the
+// declaration is in <intrin.h>, but <intrin.h> is not available from all the
+// targets that are being tested here.
+void __cpuidex (int[4], int, int);
+
+#include <cpuid.h>
+
+int cpuid_info[4];
+
+void test_cpuidex(unsigned level, unsigned count) {
+  __cpuidex(cpuid_info, level, count);
+}
+

clang/test/Headers/cpuid.c

@@ -6,14 +6,19 @@
 
 // CHECK-64: {{.*}} call { i32, i32, i32, i32 } asm "  xchgq  %rbx,${1:q}\0A cpuid\0A xchgq %rbx,${1:q}", "={ax},=r,={cx},={dx},0,~{dirflag},~{fpsr},~{flags}"(i32 %{{[a-z0-9]+}})
 // CHECK-64: {{.*}} call { i32, i32, i32, i32 } asm "  xchgq  %rbx,${1:q}\0A  cpuid\0A  xchgq  %rbx,${1:q}", "={ax},=r,={cx},={dx},0,2,~{dirflag},~{fpsr},~{flags}"(i32 %{{[a-z0-9]+}}, i32 %{{[a-z0-9]+}})
+// CHECK-64: {{.*}} call { i32, i32, i32, i32 } asm "  xchgq  %rbx,${1:q}\0A  cpuid\0A  xchgq  %rbx,${1:q}", "={ax},=r,={cx},={dx},0,2,~{dirflag},~{fpsr},~{flags}"(i32 %{{[a-z0-9]+}}, i32 %{{[a-z0-9]+}})
 
 // CHECK-32: {{.*}} call { i32, i32, i32, i32 } asm "cpuid", "={ax},={bx},={cx},={dx},0,~{dirflag},~{fpsr},~{flags}"(i32 %{{[a-z0-9]+}})
 // CHECK-32: {{.*}} call { i32, i32, i32, i32 } asm "cpuid", "={ax},={bx},={cx},={dx},0,2,~{dirflag},~{fpsr},~{flags}"(i32 %{{[a-z0-9]+}}, i32 %{{[a-z0-9]+}})
+// CHECK-32: {{.*}} call { i32, i32, i32, i32 } asm "cpuid", "={ax},={bx},={cx},={dx},0,2,~{dirflag},~{fpsr},~{flags}"(i32 %{{[a-z0-9]+}}, i32 %{{[a-z0-9]+}})
 
 unsigned eax0, ebx0, ecx0, edx0;
 unsigned eax1, ebx1, ecx1, edx1;
 
+int cpuid_info[4];
+
 void test_cpuid(unsigned level, unsigned count) {
   __cpuid(level, eax1, ebx1, ecx1, edx1);
   __cpuid_count(level, count, eax0, ebx0, ecx0, edx0);
+  __cpuidex(cpuid_info, level, count);
 }