cpu.h 2.8 KB

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667686970717273
  1. /*****************************************************************************
  2. * cpu.h: cpu detection
  3. *****************************************************************************
  4. * Copyright (C) 2004-2018 x264 project
  5. *
  6. * Authors: Loren Merritt <lorenm@u.washington.edu>
  7. *
  8. * This program is free software; you can redistribute it and/or modify
  9. * it under the terms of the GNU General Public License as published by
  10. * the Free Software Foundation; either version 2 of the License, or
  11. * (at your option) any later version.
  12. *
  13. * This program is distributed in the hope that it will be useful,
  14. * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15. * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  16. * GNU General Public License for more details.
  17. *
  18. * You should have received a copy of the GNU General Public License
  19. * along with this program; if not, write to the Free Software
  20. * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA.
  21. *
  22. * This program is also available under a commercial proprietary license.
  23. * For more information, contact us at licensing@x264.com.
  24. *****************************************************************************/
  25. #ifndef X264_CPU_H
  26. #define X264_CPU_H
  27. uint32_t x264_cpu_detect( void );
  28. int x264_cpu_num_processors( void );
  29. void x264_cpu_emms( void );
  30. void x264_cpu_sfence( void );
  31. #if HAVE_MMX
  32. /* There is no way to forbid the compiler from using float instructions
  33. * before the emms so miscompilation could theoretically occur in the
  34. * unlikely event that the compiler reorders emms and float instructions. */
  35. #if HAVE_X86_INLINE_ASM
  36. /* Clobbering memory makes the compiler less likely to reorder code. */
  37. #define x264_emms() asm volatile( "emms":::"memory","st","st(1)","st(2)", \
  38. "st(3)","st(4)","st(5)","st(6)","st(7)" )
  39. #else
  40. #define x264_emms() x264_cpu_emms()
  41. #endif
  42. #else
  43. #define x264_emms()
  44. #endif
  45. #define x264_sfence x264_cpu_sfence
  46. /* kludge:
  47. * gcc can't give variables any greater alignment than the stack frame has.
  48. * We need 32 byte alignment for AVX2, so here we make sure that the stack is
  49. * aligned to 32 bytes.
  50. * gcc 4.2 introduced __attribute__((force_align_arg_pointer)) to fix this
  51. * problem, but I don't want to require such a new version.
  52. * aligning to 32 bytes only works if the compiler supports keeping that
  53. * alignment between functions (osdep.h handles manual alignment of arrays
  54. * if it doesn't).
  55. */
  56. #if HAVE_MMX && (STACK_ALIGNMENT > 16 || (ARCH_X86 && STACK_ALIGNMENT > 4))
  57. intptr_t x264_stack_align( void (*func)(), ... );
  58. #define x264_stack_align(func,...) x264_stack_align((void (*)())func, __VA_ARGS__)
  59. #else
  60. #define x264_stack_align(func,...) func(__VA_ARGS__)
  61. #endif
  62. typedef struct
  63. {
  64. const char *name;
  65. uint32_t flags;
  66. } x264_cpu_name_t;
  67. extern const x264_cpu_name_t x264_cpu_names[];
  68. #endif