x86: AVX2 high bit-depth pixel_ssd

[x264] / common / cpu.h

diff --git a/common/cpu.h b/common/cpu.h
index e78148ade8b83baa60af48bbfd792695a6f6fe19..7f76e435b050e6b357b1fda0acb569f1f6ee4869 100644 (file)
--- a/common/cpu.h
+++ b/common/cpu.h
@@ -1,10 +1,9 @@
 /*****************************************************************************
- * cpu.h: h264 encoder library
+ * cpu.h: cpu detection
  *****************************************************************************
- * Copyright (C) 2003 Laurent Aimar
- * $Id: cpu.h,v 1.1 2004/06/03 19:27:06 fenrir Exp $
+ * Copyright (C) 2004-2013 x264 project
  *
- * Authors: Laurent Aimar <fenrir@via.ecp.fr>
+ * Authors: Loren Merritt <lorenm@u.washington.edu>
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -18,7 +17,10 @@
  *
  * You should have received a copy of the GNU General Public License
  * along with this program; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111, USA.
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02111, USA.
+ *
+ * This program is also available under a commercial proprietary license.
+ * For more information, contact us at licensing@x264.com.
  *****************************************************************************/
 
 #ifndef X264_CPU_H
@@ -26,9 +28,25 @@
 
 uint32_t x264_cpu_detect( void );
 int      x264_cpu_num_processors( void );
-
-/* probably MMX(EXT) centric but .... */
-void     x264_cpu_restore( uint32_t cpu );
+void     x264_cpu_emms( void );
+void     x264_cpu_sfence( void );
+#if HAVE_MMX
+/* There is no way to forbid the compiler from using float instructions
+ * before the emms so miscompilation could theoretically occur in the
+ * unlikely event that the compiler reorders emms and float instructions. */
+#if HAVE_X86_INLINE_ASM
+/* Clobbering memory makes the compiler less likely to reorder code. */
+#define x264_emms() asm volatile( "emms":::"memory","st","st(1)","st(2)", \
+                                  "st(3)","st(4)","st(5)","st(6)","st(7)" )
+#else
+#define x264_emms() x264_cpu_emms()
+#endif
+#else
+#define x264_emms()
+#endif
+#define x264_sfence x264_cpu_sfence
+void     x264_cpu_mask_misalign_sse( void );
+void     x264_safe_intel_cpu_indicator_init( void );
 
 /* kluge:
  * gcc can't give variables any greater alignment than the stack frame has.
@@ -37,16 +55,19 @@ void     x264_cpu_restore( uint32_t cpu );
  * gcc 4.2 introduced __attribute__((force_align_arg_pointer)) to fix this
  * problem, but I don't want to require such a new version.
  * This applies only to x86_32, since other architectures that need alignment
- * also have ABIs that ensure aligned stack. */
-#if defined(ARCH_X86) && defined(HAVE_MMX)
-void x264_stack_align( void (*func)(x264_t*), x264_t *arg );
+ * either have ABIs that ensure aligned stack, or don't support it at all. */
+#if ARCH_X86 && HAVE_MMX
+int x264_stack_align( void (*func)(), ... );
+#define x264_stack_align(func,...) x264_stack_align((void (*)())func, __VA_ARGS__)
 #else
-#define x264_stack_align(func,arg) func(arg)
+#define x264_stack_align(func,...) func(__VA_ARGS__)
 #endif
 
-extern const struct {
-    const char name[8];
-    int flags;
-} x264_cpu_names[];
+typedef struct
+{
+    const char name[16];
+    uint32_t flags;
+} x264_cpu_name_t;
+extern const x264_cpu_name_t x264_cpu_names[];
 
 #endif