"Fossies" - the Fresh Open Source Software Archive

Member "pulseaudio-13.0/src/pulsecore/remap_sse.c" (13 Sep 2019, 6366 Bytes) of package /linux/misc/pulseaudio-13.0.tar.xz:


As a special service "Fossies" has tried to format the requested source page into HTML format using (guessed) C and C++ source code syntax highlighting (style: standard) with prefixed line numbers and code folding option. Alternatively you can here view or download the uninterpreted source code file. For more information about "remap_sse.c" see the Fossies "Dox" file reference documentation and the latest Fossies "Diffs" side-by-side code changes report: 12.2_vs_13.0.

    1 /***
    2   This file is part of PulseAudio.
    3 
    4   Copyright 2004-2006 Lennart Poettering
    5   Copyright 2009 Wim Taymans <wim.taymans@collabora.co.uk.com>
    6 
    7   PulseAudio is free software; you can redistribute it and/or modify
    8   it under the terms of the GNU Lesser General Public License as published
    9   by the Free Software Foundation; either version 2.1 of the License,
   10   or (at your option) any later version.
   11 
   12   PulseAudio is distributed in the hope that it will be useful, but
   13   WITHOUT ANY WARRANTY; without even the implied warranty of
   14   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
   15   General Public License for more details.
   16 
   17   You should have received a copy of the GNU Lesser General Public License
   18   along with PulseAudio; if not, see <http://www.gnu.org/licenses/>.
   19 ***/
   20 
   21 #ifdef HAVE_CONFIG_H
   22 #include <config.h>
   23 #endif
   24 
   25 #include <pulse/sample.h>
   26 #include <pulse/volume.h>
   27 #include <pulsecore/log.h>
   28 #include <pulsecore/macro.h>
   29 
   30 #include "cpu-x86.h"
   31 #include "remap.h"
   32 
   33 #define LOAD_SAMPLES                                   \
   34                 " movdqu (%1), %%xmm0           \n\t"  \
   35                 " movdqu 16(%1), %%xmm2         \n\t"  \
   36                 " movdqu 32(%1), %%xmm4         \n\t"  \
   37                 " movdqu 48(%1), %%xmm6         \n\t"  \
   38                 " movdqa %%xmm0, %%xmm1         \n\t"  \
   39                 " movdqa %%xmm2, %%xmm3         \n\t"  \
   40                 " movdqa %%xmm4, %%xmm5         \n\t"  \
   41                 " movdqa %%xmm6, %%xmm7         \n\t"
   42 
   43 #define UNPACK_SAMPLES(s)                              \
   44                 " punpckl"#s" %%xmm0, %%xmm0    \n\t"  \
   45                 " punpckh"#s" %%xmm1, %%xmm1    \n\t"  \
   46                 " punpckl"#s" %%xmm2, %%xmm2    \n\t"  \
   47                 " punpckh"#s" %%xmm3, %%xmm3    \n\t"  \
   48                 " punpckl"#s" %%xmm4, %%xmm4    \n\t"  \
   49                 " punpckh"#s" %%xmm5, %%xmm5    \n\t"  \
   50                 " punpckl"#s" %%xmm6, %%xmm6    \n\t"  \
   51                 " punpckh"#s" %%xmm7, %%xmm7    \n\t"
   52 
   53 #define STORE_SAMPLES                                  \
   54                 " movdqu %%xmm0, (%0)           \n\t"  \
   55                 " movdqu %%xmm1, 16(%0)         \n\t"  \
   56                 " movdqu %%xmm2, 32(%0)         \n\t"  \
   57                 " movdqu %%xmm3, 48(%0)         \n\t"  \
   58                 " movdqu %%xmm4, 64(%0)         \n\t"  \
   59                 " movdqu %%xmm5, 80(%0)         \n\t"  \
   60                 " movdqu %%xmm6, 96(%0)         \n\t"  \
   61                 " movdqu %%xmm7, 112(%0)        \n\t"  \
   62                 " add $64, %1                   \n\t"  \
   63                 " add $128, %0                  \n\t"
   64 
   65 #define HANDLE_SINGLE_dq()                             \
   66                 " movd (%1), %%xmm0             \n\t"  \
   67                 " punpckldq %%xmm0, %%xmm0      \n\t"  \
   68                 " movq %%xmm0, (%0)             \n\t"  \
   69                 " add $4, %1                    \n\t"  \
   70                 " add $8, %0                    \n\t"
   71 
   72 #define HANDLE_SINGLE_wd()                             \
   73                 " movw (%1), %w3                \n\t"  \
   74                 " movd %3, %%xmm0               \n\t"  \
   75                 " punpcklwd %%xmm0, %%xmm0      \n\t"  \
   76                 " movd %%xmm0, (%0)             \n\t"  \
   77                 " add $2, %1                    \n\t"  \
   78                 " add $4, %0                    \n\t"
   79 
   80 #define MONO_TO_STEREO(s,shift,mask)                   \
   81                 " mov %4, %2                    \n\t"  \
   82                 " sar $"#shift", %2             \n\t"  \
   83                 " cmp $0, %2                    \n\t"  \
   84                 " je 2f                         \n\t"  \
   85                 "1:                             \n\t"  \
   86                 LOAD_SAMPLES                           \
   87                 UNPACK_SAMPLES(s)                      \
   88                 STORE_SAMPLES                          \
   89                 " dec %2                        \n\t"  \
   90                 " jne 1b                        \n\t"  \
   91                 "2:                             \n\t"  \
   92                 " mov %4, %2                    \n\t"  \
   93                 " and $"#mask", %2              \n\t"  \
   94                 " je 4f                         \n\t"  \
   95                 "3:                             \n\t"  \
   96                 HANDLE_SINGLE_##s()                    \
   97                 " dec %2                        \n\t"  \
   98                 " jne 3b                        \n\t"  \
   99                 "4:                             \n\t"
  100 
  101 #if defined (__i386__) || defined (__amd64__)
  102 static void remap_mono_to_stereo_s16ne_sse2(pa_remap_t *m, int16_t *dst, const int16_t *src, unsigned n) {
  103     pa_reg_x86 temp, temp2;
  104 
  105     __asm__ __volatile__ (
  106         MONO_TO_STEREO(wd, 5, 31) /* do words to doubles */
  107         : "+r" (dst), "+r" (src), "=&r" (temp), "=&r" (temp2)
  108         : "r" ((pa_reg_x86)n)
  109         : "cc"
  110     );
  111 }
  112 
  113 /* Works for both S32NE and FLOAT32NE */
  114 static void remap_mono_to_stereo_any32ne_sse2(pa_remap_t *m, float *dst, const float *src, unsigned n) {
  115     pa_reg_x86 temp, temp2;
  116 
  117     __asm__ __volatile__ (
  118         MONO_TO_STEREO(dq, 4, 15) /* do doubles to quads */
  119         : "+r" (dst), "+r" (src), "=&r" (temp), "=&r" (temp2)
  120         : "r" ((pa_reg_x86)n)
  121         : "cc"
  122     );
  123 }
  124 
  125 /* set the function that will execute the remapping based on the matrices */
  126 static void init_remap_sse2(pa_remap_t *m) {
  127     unsigned n_oc, n_ic;
  128 
  129     n_oc = m->o_ss.channels;
  130     n_ic = m->i_ss.channels;
  131 
  132     /* find some common channel remappings, fall back to full matrix operation. */
  133     if (n_ic == 1 && n_oc == 2 &&
  134             m->map_table_i[0][0] == 0x10000 && m->map_table_i[1][0] == 0x10000) {
  135 
  136         pa_log_info("Using SSE2 mono to stereo remapping");
  137         pa_set_remap_func(m, (pa_do_remap_func_t) remap_mono_to_stereo_s16ne_sse2,
  138             (pa_do_remap_func_t) remap_mono_to_stereo_any32ne_sse2,
  139             (pa_do_remap_func_t) remap_mono_to_stereo_any32ne_sse2);
  140     }
  141 }
  142 #endif /* defined (__i386__) || defined (__amd64__) */
  143 
  144 void pa_remap_func_init_sse(pa_cpu_x86_flag_t flags) {
  145 #if defined (__i386__) || defined (__amd64__)
  146 
  147     if (flags & PA_CPU_X86_SSE2) {
  148         pa_log_info("Initialising SSE2 optimized remappers.");
  149         pa_set_init_remap_func ((pa_init_remap_func_t) init_remap_sse2);
  150     }
  151 
  152 #endif /* defined (__i386__) || defined (__amd64__) */
  153 }