From: Hans Wennborg <hans@hanshq.net>
Date: Tue, 17 Nov 2015 18:46:48 +0000 (+0000)
Subject: bmiintrin.h: Allow using the tzcnt intrinsics for non-BMI targets
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=b016519fe108cb4bf1e2f0981ee6f1892d16c0e7;p=clang

bmiintrin.h: Allow using the tzcnt intrinsics for non-BMI targets

The tzcnt intrinsics are used non non-BMI targets by code (e.g. ffmpeg)
that uses it as a potentially faster BSF.

The TZCNT instruction is special in that it's encoded in a
backward-compatible way and behaves as BSF on non-BMI targets.

Differential Revision: http://reviews.llvm.org/D14748

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@253358 91177308-0d34-0410-b5e6-96231b3b80d8
---

diff --git a/lib/Headers/bmiintrin.h b/lib/Headers/bmiintrin.h
index dc2f83f3e2..da98792d83 100644
--- a/lib/Headers/bmiintrin.h
+++ b/lib/Headers/bmiintrin.h
@@ -39,7 +39,12 @@
 /* Define the default attributes for the functions in this file. */
 #define __DEFAULT_FN_ATTRS __attribute__((__always_inline__, __nodebug__, __target__("bmi")))
 
-static __inline__ unsigned short __DEFAULT_FN_ATTRS
+/* Allow using the tzcnt intrinsics even for non-BMI targets. Since the TZCNT
+   instruction behaves as BSF on non-BMI targets, there is code that expects
+   to use it as a potentially faster version of BSF. */
+#define __RELAXED_FN_ATTRS __attribute__((__always_inline__, __nodebug__))
+
+static __inline__ unsigned short __RELAXED_FN_ATTRS
 __tzcnt_u16(unsigned short __X)
 {
   return __X ? __builtin_ctzs(__X) : 16;
@@ -83,7 +88,7 @@ __blsr_u32(unsigned int __X)
   return __X & (__X - 1);
 }
 
-static __inline__ unsigned int __DEFAULT_FN_ATTRS
+static __inline__ unsigned int __RELAXED_FN_ATTRS
 __tzcnt_u32(unsigned int __X)
 {
   return __X ? __builtin_ctz(__X) : 32;
@@ -136,7 +141,7 @@ __blsr_u64(unsigned long long __X)
   return __X & (__X - 1);
 }
 
-static __inline__ unsigned long long __DEFAULT_FN_ATTRS
+static __inline__ unsigned long long __RELAXED_FN_ATTRS
 __tzcnt_u64(unsigned long long __X)
 {
   return __X ? __builtin_ctzll(__X) : 64;
@@ -145,5 +150,6 @@ __tzcnt_u64(unsigned long long __X)
 #endif /* __x86_64__ */
 
 #undef __DEFAULT_FN_ATTRS
+#undef __RELAXED_FN_ATTRS
 
 #endif /* __BMIINTRIN_H */