ryan/mozjpeg
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
5e797fa69954b59e08a8fef4e696a2955b9bce5e
mozjpeg/simd/arm/aarch64/jsimd.c
DRC 607b668ff9 MSVC: Eliminate C4996 warnings in API libs 
The primary purpose of this is to encourage adoption of libjpeg-turbo in
downstream Windows projects that forbid the use of "deprecated"
functions.  libjpeg-turbo's usage of those functions was not actually
unsafe, because:

- libjpeg-turbo always checks the return value of fopen() and ensures
  that a NULL filename can never be passed to it.

- libjpeg-turbo always checks the return value of getenv() and never
  passes a NULL argument to it.

- The sprintf() calls in format_message() (jerror.c) could never
  overflow the destination string buffer or leave it unterminated as
  long as the buffer was at least JMSG_LENGTH_MAX bytes in length, as
  instructed. (Regardless, this commit replaces those calls with
  snprintf() calls.)

- libjpeg-turbo never uses sscanf() to read strings or multi-byte
  character arrays.

- Because of b7d6e84d6a, wrjpgcom
  explicitly checks the bounds of the source and destination strings
  before calling strcat() and strcpy().

- libjpeg-turbo always ensures that the destination string is
  terminated when using strncpy().
  (548490fe5e made this explicit.)

Regarding thread safety:

Technically speaking, getenv() is not thread-safe, because the returned
pointer may be invalidated if another thread sets the same environment
variable between the time that the first thread calls getenv() and the
time that that thread uses the return value.  In practice, however, this
could only occur with libjpeg-turbo if:

(1) A multithreaded calling application used the deprecated and
undocumented TJFLAG_FORCEMMX/TJFLAG_FORCESSE/TJFLAG_FORCESSE2 flags in
the TurboJPEG API or set one of the corresponding environment variables
(which are only intended for testing purposes.)  Since the TurboJPEG API
library only ever passed string constants to putenv(), the only inherent
risk (i.e. the only risk introduced by the library and not the calling
application) was that the SIMD extensions may have read an incorrect
value from one of the aforementioned environment variables.

or

(2) A multithreaded calling application modified the value of the
JPEGMEM environment variable in one thread while another thread was
reading the value of that environment variable (in the body of
jpeg_create_compress() or jpeg_create_decompress().)  Given that the
libjpeg API provides a thread-safe way for applications to modify the
default memory limit without using the JPEGMEM environment variable,
direct modification of that environment variable by calling applications
is not supported.

Microsoft's implementation of getenv_s() does not claim to be
thread-safe either, so this commit uses getenv_s() solely to mollify
Visual Studio.  New inline functions and macros (GETENV_S() and
PUTENV_S) wrap getenv_s()/_putenv_s() when building for Visual Studio
and getenv()/setenv() otherwise, but GETENV_S()/PUTENV_S() provide no
advantages over getenv()/setenv() other than parameter validation.  They
are implemented solely for convenience.

Technically speaking, strerror() is not thread-safe, because the
returned pointer may be invalidated if another thread changes the locale
and/or calls strerror() between the time that the first thread calls
strerror() and the time that that thread uses the return value.  In
practice, however, this could only occur with libjpeg-turbo if a
multithreaded calling application encountered a file I/O error in
tjLoadImage() or tjSaveImage().  Since both of those functions
immediately copy the string returned from strerror() into a thread-local
buffer, the risk is minimal, and the worst case would involve an
incorrect error string being reported to the calling application.
Regardless, this commit uses strerror_s() in the TurboJPEG API library
when building for Visual Studio.  Note that strerror_r() could have been
used on Un*x systems, but it would have been necessary to handle both
the POSIX and GNU implementations of that function and perform
widespread compatibility testing.  Such is left as an exercise for
another day.

Fixes #568
2022-02-23 15:57:01 -06:00

1059 lines
24 KiB
C
Raw Blame History

/*
* jsimd_arm64.c
*
* Copyright 2009 Pierre Ossman <ossman@cendio.se> for Cendio AB
* Copyright (C) 2011, Nokia Corporation and/or its subsidiary(-ies).
* Copyright (C) 2009-2011, 2013-2014, 2016, 2018, 2020, 2022, D. R. Commander.
* Copyright (C) 2015-2016, 2018, Matthieu Darbois.
* Copyright (C) 2020, Arm Limited.
*
* Based on the x86 SIMD extension for IJG JPEG library,
* Copyright (C) 1999-2006, MIYASAKA Masaru.
* For conditions of distribution and use, see copyright notice in jsimdext.inc
*
* This file contains the interface between the "normal" portions
* of the library and the SIMD implementations when running on a
* 64-bit Arm architecture.
*/
#define JPEG_INTERNALS
#include "../../../jinclude.h"
#include "../../../jpeglib.h"
#include "../../../jsimd.h"
#include "../../../jdct.h"
#include "../../../jsimddct.h"
#include "../../jsimd.h"
#include "jconfigint.h"
#include <stdio.h>
#include <string.h>
#include <ctype.h>
#define JSIMD_FASTLD3 1
#define JSIMD_FASTST3 2
#define JSIMD_FASTTBL 4
static unsigned int simd_support = ~0;
static unsigned int simd_huffman = 1;
static unsigned int simd_features = JSIMD_FASTLD3 | JSIMD_FASTST3 |
JSIMD_FASTTBL;
#if defined(__linux__) || defined(ANDROID) || defined(__ANDROID__)
#define SOMEWHAT_SANE_PROC_CPUINFO_SIZE_LIMIT (1024 * 1024)
LOCAL(int)
check_cpuinfo(char *buffer, const char *field, char *value)
{
char *p;
if (*value == 0)
return 0;
if (strncmp(buffer, field, strlen(field)) != 0)
return 0;
buffer += strlen(field);
while (isspace(*buffer))
buffer++;
/* Check if 'value' is present in the buffer as a separate word */
while ((p = strstr(buffer, value))) {
if (p > buffer && !isspace(*(p - 1))) {
buffer++;
continue;
}
p += strlen(value);
if (*p != 0 && !isspace(*p)) {
buffer++;
continue;
}
return 1;
}
return 0;
}
LOCAL(int)
parse_proc_cpuinfo(int bufsize)
{
char *buffer = (char *)malloc(bufsize);
FILE *fd;
if (!buffer)
return 0;
fd = fopen("/proc/cpuinfo", "r");
if (fd) {
while (fgets(buffer, bufsize, fd)) {
if (!strchr(buffer, '\n') && !feof(fd)) {
/* "impossible" happened - insufficient size of the buffer! */
fclose(fd);
free(buffer);
return 0;
}
if (check_cpuinfo(buffer, "CPU part", "0xd03") ||
check_cpuinfo(buffer, "CPU part", "0xd07"))
/* The Cortex-A53 has a slow tbl implementation. We can gain a few
percent speedup by disabling the use of that instruction. The
speedup on Cortex-A57 is more subtle but still measurable. */
simd_features &= ~JSIMD_FASTTBL;
else if (check_cpuinfo(buffer, "CPU part", "0x0a1"))
/* The SIMD version of Huffman encoding is slower than the C version on
Cavium ThunderX. Also, ld3 and st3 are abyssmally slow on that
CPU. */
simd_huffman = simd_features = 0;
}
fclose(fd);
}
free(buffer);
return 1;
}
#endif
/*
* Check what SIMD accelerations are supported.
*
* FIXME: This code is racy under a multi-threaded environment.
*/
/*
* Armv8 architectures support Neon extensions by default.
* It is no longer optional as it was with Armv7.
*/
LOCAL(void)
init_simd(void)
{
#ifndef NO_GETENV
char env[2] = { 0 };
#endif
#if defined(__linux__) || defined(ANDROID) || defined(__ANDROID__)
int bufsize = 1024; /* an initial guess for the line buffer size limit */
#endif
if (simd_support != ~0U)
return;
simd_support = 0;
simd_support |= JSIMD_NEON;
#if defined(__linux__) || defined(ANDROID) || defined(__ANDROID__)
while (!parse_proc_cpuinfo(bufsize)) {
bufsize *= 2;
if (bufsize > SOMEWHAT_SANE_PROC_CPUINFO_SIZE_LIMIT)
break;
}
#endif
#ifndef NO_GETENV
/* Force different settings through environment variables */
if (!GETENV_S(env, 2, "JSIMD_FORCENEON") && !strcmp(env, "1"))
simd_support = JSIMD_NEON;
if (!GETENV_S(env, 2, "JSIMD_FORCENONE") && !strcmp(env, "1"))
simd_support = 0;
if (!GETENV_S(env, 2, "JSIMD_NOHUFFENC") && !strcmp(env, "1"))
simd_huffman = 0;
if (!GETENV_S(env, 2, "JSIMD_FASTLD3") && !strcmp(env, "1"))
simd_features |= JSIMD_FASTLD3;
if (!GETENV_S(env, 2, "JSIMD_FASTLD3") && !strcmp(env, "0"))
simd_features &= ~JSIMD_FASTLD3;
if (!GETENV_S(env, 2, "JSIMD_FASTST3") && !strcmp(env, "1"))
simd_features |= JSIMD_FASTST3;
if (!GETENV_S(env, 2, "JSIMD_FASTST3") && !strcmp(env, "0"))
simd_features &= ~JSIMD_FASTST3;
#endif
}
GLOBAL(int)
jsimd_can_rgb_ycc(void)
{
init_simd();
/* The code is optimised for these values only */
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if ((RGB_PIXELSIZE != 3) && (RGB_PIXELSIZE != 4))
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_can_rgb_gray(void)
{
init_simd();
/* The code is optimised for these values only */
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if ((RGB_PIXELSIZE != 3) && (RGB_PIXELSIZE != 4))
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_can_ycc_rgb(void)
{
init_simd();
/* The code is optimised for these values only */
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if ((RGB_PIXELSIZE != 3) && (RGB_PIXELSIZE != 4))
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_can_ycc_rgb565(void)
{
init_simd();
/* The code is optimised for these values only */
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(void)
jsimd_rgb_ycc_convert(j_compress_ptr cinfo, JSAMPARRAY input_buf,
JSAMPIMAGE output_buf, JDIMENSION output_row,
int num_rows)
{
void (*neonfct) (JDIMENSION, JSAMPARRAY, JSAMPIMAGE, JDIMENSION, int);
switch (cinfo->in_color_space) {
case JCS_EXT_RGB:
#ifndef NEON_INTRINSICS
if (simd_features & JSIMD_FASTLD3)
#endif
neonfct = jsimd_extrgb_ycc_convert_neon;
#ifndef NEON_INTRINSICS
else
neonfct = jsimd_extrgb_ycc_convert_neon_slowld3;
#endif
break;
case JCS_EXT_RGBX:
case JCS_EXT_RGBA:
neonfct = jsimd_extrgbx_ycc_convert_neon;
break;
case JCS_EXT_BGR:
#ifndef NEON_INTRINSICS
if (simd_features & JSIMD_FASTLD3)
#endif
neonfct = jsimd_extbgr_ycc_convert_neon;
#ifndef NEON_INTRINSICS
else
neonfct = jsimd_extbgr_ycc_convert_neon_slowld3;
#endif
break;
case JCS_EXT_BGRX:
case JCS_EXT_BGRA:
neonfct = jsimd_extbgrx_ycc_convert_neon;
break;
case JCS_EXT_XBGR:
case JCS_EXT_ABGR:
neonfct = jsimd_extxbgr_ycc_convert_neon;
break;
case JCS_EXT_XRGB:
case JCS_EXT_ARGB:
neonfct = jsimd_extxrgb_ycc_convert_neon;
break;
default:
#ifndef NEON_INTRINSICS
if (simd_features & JSIMD_FASTLD3)
#endif
neonfct = jsimd_extrgb_ycc_convert_neon;
#ifndef NEON_INTRINSICS
else
neonfct = jsimd_extrgb_ycc_convert_neon_slowld3;
#endif
break;
}
neonfct(cinfo->image_width, input_buf, output_buf, output_row, num_rows);
}
GLOBAL(void)
jsimd_rgb_gray_convert(j_compress_ptr cinfo, JSAMPARRAY input_buf,
JSAMPIMAGE output_buf, JDIMENSION output_row,
int num_rows)
{
void (*neonfct) (JDIMENSION, JSAMPARRAY, JSAMPIMAGE, JDIMENSION, int);
switch (cinfo->in_color_space) {
case JCS_EXT_RGB:
neonfct = jsimd_extrgb_gray_convert_neon;
break;
case JCS_EXT_RGBX:
case JCS_EXT_RGBA:
neonfct = jsimd_extrgbx_gray_convert_neon;
break;
case JCS_EXT_BGR:
neonfct = jsimd_extbgr_gray_convert_neon;
break;
case JCS_EXT_BGRX:
case JCS_EXT_BGRA:
neonfct = jsimd_extbgrx_gray_convert_neon;
break;
case JCS_EXT_XBGR:
case JCS_EXT_ABGR:
neonfct = jsimd_extxbgr_gray_convert_neon;
break;
case JCS_EXT_XRGB:
case JCS_EXT_ARGB:
neonfct = jsimd_extxrgb_gray_convert_neon;
break;
default:
neonfct = jsimd_extrgb_gray_convert_neon;
break;
}
neonfct(cinfo->image_width, input_buf, output_buf, output_row, num_rows);
}
GLOBAL(void)
jsimd_ycc_rgb_convert(j_decompress_ptr cinfo, JSAMPIMAGE input_buf,
JDIMENSION input_row, JSAMPARRAY output_buf,
int num_rows)
{
void (*neonfct) (JDIMENSION, JSAMPIMAGE, JDIMENSION, JSAMPARRAY, int);
switch (cinfo->out_color_space) {
case JCS_EXT_RGB:
#ifndef NEON_INTRINSICS
if (simd_features & JSIMD_FASTST3)
#endif
neonfct = jsimd_ycc_extrgb_convert_neon;
#ifndef NEON_INTRINSICS
else
neonfct = jsimd_ycc_extrgb_convert_neon_slowst3;
#endif
break;
case JCS_EXT_RGBX:
case JCS_EXT_RGBA:
neonfct = jsimd_ycc_extrgbx_convert_neon;
break;
case JCS_EXT_BGR:
#ifndef NEON_INTRINSICS
if (simd_features & JSIMD_FASTST3)
#endif
neonfct = jsimd_ycc_extbgr_convert_neon;
#ifndef NEON_INTRINSICS
else
neonfct = jsimd_ycc_extbgr_convert_neon_slowst3;
#endif
break;
case JCS_EXT_BGRX:
case JCS_EXT_BGRA:
neonfct = jsimd_ycc_extbgrx_convert_neon;
break;
case JCS_EXT_XBGR:
case JCS_EXT_ABGR:
neonfct = jsimd_ycc_extxbgr_convert_neon;
break;
case JCS_EXT_XRGB:
case JCS_EXT_ARGB:
neonfct = jsimd_ycc_extxrgb_convert_neon;
break;
default:
#ifndef NEON_INTRINSICS
if (simd_features & JSIMD_FASTST3)
#endif
neonfct = jsimd_ycc_extrgb_convert_neon;
#ifndef NEON_INTRINSICS
else
neonfct = jsimd_ycc_extrgb_convert_neon_slowst3;
#endif
break;
}
neonfct(cinfo->output_width, input_buf, input_row, output_buf, num_rows);
}
GLOBAL(void)
jsimd_ycc_rgb565_convert(j_decompress_ptr cinfo, JSAMPIMAGE input_buf,
JDIMENSION input_row, JSAMPARRAY output_buf,
int num_rows)
{
jsimd_ycc_rgb565_convert_neon(cinfo->output_width, input_buf, input_row,
output_buf, num_rows);
}
GLOBAL(int)
jsimd_can_h2v2_downsample(void)
{
init_simd();
/* The code is optimised for these values only */
if (BITS_IN_JSAMPLE != 8)
return 0;
if (DCTSIZE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_can_h2v1_downsample(void)
{
init_simd();
/* The code is optimised for these values only */
if (BITS_IN_JSAMPLE != 8)
return 0;
if (DCTSIZE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(void)
jsimd_h2v2_downsample(j_compress_ptr cinfo, jpeg_component_info *compptr,
JSAMPARRAY input_data, JSAMPARRAY output_data)
{
jsimd_h2v2_downsample_neon(cinfo->image_width, cinfo->max_v_samp_factor,
compptr->v_samp_factor, compptr->width_in_blocks,
input_data, output_data);
}
GLOBAL(void)
jsimd_h2v1_downsample(j_compress_ptr cinfo, jpeg_component_info *compptr,
JSAMPARRAY input_data, JSAMPARRAY output_data)
{
jsimd_h2v1_downsample_neon(cinfo->image_width, cinfo->max_v_samp_factor,
compptr->v_samp_factor, compptr->width_in_blocks,
input_data, output_data);
}
GLOBAL(int)
jsimd_can_h2v2_upsample(void)
{
init_simd();
/* The code is optimised for these values only */
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_can_h2v1_upsample(void)
{
init_simd();
/* The code is optimised for these values only */
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(void)
jsimd_h2v2_upsample(j_decompress_ptr cinfo, jpeg_component_info *compptr,
JSAMPARRAY input_data, JSAMPARRAY *output_data_ptr)
{
jsimd_h2v2_upsample_neon(cinfo->max_v_samp_factor, cinfo->output_width,
input_data, output_data_ptr);
}
GLOBAL(void)
jsimd_h2v1_upsample(j_decompress_ptr cinfo, jpeg_component_info *compptr,
JSAMPARRAY input_data, JSAMPARRAY *output_data_ptr)
{
jsimd_h2v1_upsample_neon(cinfo->max_v_samp_factor, cinfo->output_width,
input_data, output_data_ptr);
}
GLOBAL(int)
jsimd_can_h2v2_fancy_upsample(void)
{
init_simd();
/* The code is optimised for these values only */
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_can_h2v1_fancy_upsample(void)
{
init_simd();
/* The code is optimised for these values only */
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_can_h1v2_fancy_upsample(void)
{
init_simd();
/* The code is optimised for these values only */
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(void)
jsimd_h2v2_fancy_upsample(j_decompress_ptr cinfo, jpeg_component_info *compptr,
JSAMPARRAY input_data, JSAMPARRAY *output_data_ptr)
{
jsimd_h2v2_fancy_upsample_neon(cinfo->max_v_samp_factor,
compptr->downsampled_width, input_data,
output_data_ptr);
}
GLOBAL(void)
jsimd_h2v1_fancy_upsample(j_decompress_ptr cinfo, jpeg_component_info *compptr,
JSAMPARRAY input_data, JSAMPARRAY *output_data_ptr)
{
jsimd_h2v1_fancy_upsample_neon(cinfo->max_v_samp_factor,
compptr->downsampled_width, input_data,
output_data_ptr);
}
GLOBAL(void)
jsimd_h1v2_fancy_upsample(j_decompress_ptr cinfo, jpeg_component_info *compptr,
JSAMPARRAY input_data, JSAMPARRAY *output_data_ptr)
{
jsimd_h1v2_fancy_upsample_neon(cinfo->max_v_samp_factor,
compptr->downsampled_width, input_data,
output_data_ptr);
}
GLOBAL(int)
jsimd_can_h2v2_merged_upsample(void)
{
init_simd();
/* The code is optimised for these values only */
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_can_h2v1_merged_upsample(void)
{
init_simd();
/* The code is optimised for these values only */
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(void)
jsimd_h2v2_merged_upsample(j_decompress_ptr cinfo, JSAMPIMAGE input_buf,
JDIMENSION in_row_group_ctr, JSAMPARRAY output_buf)
{
void (*neonfct) (JDIMENSION, JSAMPIMAGE, JDIMENSION, JSAMPARRAY);
switch (cinfo->out_color_space) {
case JCS_EXT_RGB:
neonfct = jsimd_h2v2_extrgb_merged_upsample_neon;
break;
case JCS_EXT_RGBX:
case JCS_EXT_RGBA:
neonfct = jsimd_h2v2_extrgbx_merged_upsample_neon;
break;
case JCS_EXT_BGR:
neonfct = jsimd_h2v2_extbgr_merged_upsample_neon;
break;
case JCS_EXT_BGRX:
case JCS_EXT_BGRA:
neonfct = jsimd_h2v2_extbgrx_merged_upsample_neon;
break;
case JCS_EXT_XBGR:
case JCS_EXT_ABGR:
neonfct = jsimd_h2v2_extxbgr_merged_upsample_neon;
break;
case JCS_EXT_XRGB:
case JCS_EXT_ARGB:
neonfct = jsimd_h2v2_extxrgb_merged_upsample_neon;
break;
default:
neonfct = jsimd_h2v2_extrgb_merged_upsample_neon;
break;
}
neonfct(cinfo->output_width, input_buf, in_row_group_ctr, output_buf);
}
GLOBAL(void)
jsimd_h2v1_merged_upsample(j_decompress_ptr cinfo, JSAMPIMAGE input_buf,
JDIMENSION in_row_group_ctr, JSAMPARRAY output_buf)
{
void (*neonfct) (JDIMENSION, JSAMPIMAGE, JDIMENSION, JSAMPARRAY);
switch (cinfo->out_color_space) {
case JCS_EXT_RGB:
neonfct = jsimd_h2v1_extrgb_merged_upsample_neon;
break;
case JCS_EXT_RGBX:
case JCS_EXT_RGBA:
neonfct = jsimd_h2v1_extrgbx_merged_upsample_neon;
break;
case JCS_EXT_BGR:
neonfct = jsimd_h2v1_extbgr_merged_upsample_neon;
break;
case JCS_EXT_BGRX:
case JCS_EXT_BGRA:
neonfct = jsimd_h2v1_extbgrx_merged_upsample_neon;
break;
case JCS_EXT_XBGR:
case JCS_EXT_ABGR:
neonfct = jsimd_h2v1_extxbgr_merged_upsample_neon;
break;
case JCS_EXT_XRGB:
case JCS_EXT_ARGB:
neonfct = jsimd_h2v1_extxrgb_merged_upsample_neon;
break;
default:
neonfct = jsimd_h2v1_extrgb_merged_upsample_neon;
break;
}
neonfct(cinfo->output_width, input_buf, in_row_group_ctr, output_buf);
}
GLOBAL(int)
jsimd_can_convsamp(void)
{
init_simd();
/* The code is optimised for these values only */
if (DCTSIZE != 8)
return 0;
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if (sizeof(DCTELEM) != 2)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_can_convsamp_float(void)
{
return 0;
}
GLOBAL(void)
jsimd_convsamp(JSAMPARRAY sample_data, JDIMENSION start_col,
DCTELEM *workspace)
{
jsimd_convsamp_neon(sample_data, start_col, workspace);
}
GLOBAL(void)
jsimd_convsamp_float(JSAMPARRAY sample_data, JDIMENSION start_col,
FAST_FLOAT *workspace)
{
}
GLOBAL(int)
jsimd_can_fdct_islow(void)
{
init_simd();
/* The code is optimised for these values only */
if (DCTSIZE != 8)
return 0;
if (sizeof(DCTELEM) != 2)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_can_fdct_ifast(void)
{
init_simd();
/* The code is optimised for these values only */
if (DCTSIZE != 8)
return 0;
if (sizeof(DCTELEM) != 2)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_can_fdct_float(void)
{
return 0;
}
GLOBAL(void)
jsimd_fdct_islow(DCTELEM *data)
{
jsimd_fdct_islow_neon(data);
}
GLOBAL(void)
jsimd_fdct_ifast(DCTELEM *data)
{
jsimd_fdct_ifast_neon(data);
}
GLOBAL(void)
jsimd_fdct_float(FAST_FLOAT *data)
{
}
GLOBAL(int)
jsimd_can_quantize(void)
{
init_simd();
/* The code is optimised for these values only */
if (DCTSIZE != 8)
return 0;
if (sizeof(JCOEF) != 2)
return 0;
if (sizeof(DCTELEM) != 2)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_can_quantize_float(void)
{
return 0;
}
GLOBAL(void)
jsimd_quantize(JCOEFPTR coef_block, DCTELEM *divisors, DCTELEM *workspace)
{
jsimd_quantize_neon(coef_block, divisors, workspace);
}
GLOBAL(void)
jsimd_quantize_float(JCOEFPTR coef_block, FAST_FLOAT *divisors,
FAST_FLOAT *workspace)
{
}
GLOBAL(int)
jsimd_can_idct_2x2(void)
{
init_simd();
/* The code is optimised for these values only */
if (DCTSIZE != 8)
return 0;
if (sizeof(JCOEF) != 2)
return 0;
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if (sizeof(ISLOW_MULT_TYPE) != 2)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_can_idct_4x4(void)
{
init_simd();
/* The code is optimised for these values only */
if (DCTSIZE != 8)
return 0;
if (sizeof(JCOEF) != 2)
return 0;
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if (sizeof(ISLOW_MULT_TYPE) != 2)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(void)
jsimd_idct_2x2(j_decompress_ptr cinfo, jpeg_component_info *compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf,
JDIMENSION output_col)
{
jsimd_idct_2x2_neon(compptr->dct_table, coef_block, output_buf, output_col);
}
GLOBAL(void)
jsimd_idct_4x4(j_decompress_ptr cinfo, jpeg_component_info *compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf,
JDIMENSION output_col)
{
jsimd_idct_4x4_neon(compptr->dct_table, coef_block, output_buf, output_col);
}
GLOBAL(int)
jsimd_can_idct_islow(void)
{
init_simd();
/* The code is optimised for these values only */
if (DCTSIZE != 8)
return 0;
if (sizeof(JCOEF) != 2)
return 0;
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if (sizeof(ISLOW_MULT_TYPE) != 2)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_can_idct_ifast(void)
{
init_simd();
/* The code is optimised for these values only */
if (DCTSIZE != 8)
return 0;
if (sizeof(JCOEF) != 2)
return 0;
if (BITS_IN_JSAMPLE != 8)
return 0;
if (sizeof(JDIMENSION) != 4)
return 0;
if (sizeof(IFAST_MULT_TYPE) != 2)
return 0;
if (IFAST_SCALE_BITS != 2)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_can_idct_float(void)
{
return 0;
}
GLOBAL(void)
jsimd_idct_islow(j_decompress_ptr cinfo, jpeg_component_info *compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf,
JDIMENSION output_col)
{
jsimd_idct_islow_neon(compptr->dct_table, coef_block, output_buf,
output_col);
}
GLOBAL(void)
jsimd_idct_ifast(j_decompress_ptr cinfo, jpeg_component_info *compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf,
JDIMENSION output_col)
{
jsimd_idct_ifast_neon(compptr->dct_table, coef_block, output_buf,
output_col);
}
GLOBAL(void)
jsimd_idct_float(j_decompress_ptr cinfo, jpeg_component_info *compptr,
JCOEFPTR coef_block, JSAMPARRAY output_buf,
JDIMENSION output_col)
{
}
GLOBAL(int)
jsimd_can_huff_encode_one_block(void)
{
init_simd();
if (DCTSIZE != 8)
return 0;
if (sizeof(JCOEF) != 2)
return 0;
if (simd_support & JSIMD_NEON && simd_huffman)
return 1;
return 0;
}
GLOBAL(JOCTET *)
jsimd_huff_encode_one_block(void *state, JOCTET *buffer, JCOEFPTR block,
int last_dc_val, c_derived_tbl *dctbl,
c_derived_tbl *actbl)
{
#ifndef NEON_INTRINSICS
if (simd_features & JSIMD_FASTTBL)
#endif
return jsimd_huff_encode_one_block_neon(state, buffer, block, last_dc_val,
dctbl, actbl);
#ifndef NEON_INTRINSICS
else
return jsimd_huff_encode_one_block_neon_slowtbl(state, buffer, block,
last_dc_val, dctbl, actbl);
#endif
}
GLOBAL(int)
jsimd_can_encode_mcu_AC_first_prepare(void)
{
init_simd();
if (DCTSIZE != 8)
return 0;
if (sizeof(JCOEF) != 2)
return 0;
if (SIZEOF_SIZE_T != 8)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(void)
jsimd_encode_mcu_AC_first_prepare(const JCOEF *block,
const int *jpeg_natural_order_start, int Sl,
int Al, JCOEF *values, size_t *zerobits)
{
jsimd_encode_mcu_AC_first_prepare_neon(block, jpeg_natural_order_start,
Sl, Al, values, zerobits);
}
GLOBAL(int)
jsimd_can_encode_mcu_AC_refine_prepare(void)
{
init_simd();
if (DCTSIZE != 8)
return 0;
if (sizeof(JCOEF) != 2)
return 0;
if (SIZEOF_SIZE_T != 8)
return 0;
if (simd_support & JSIMD_NEON)
return 1;
return 0;
}
GLOBAL(int)
jsimd_encode_mcu_AC_refine_prepare(const JCOEF *block,
const int *jpeg_natural_order_start, int Sl,
int Al, JCOEF *absvalues, size_t *bits)
{
return jsimd_encode_mcu_AC_refine_prepare_neon(block,
jpeg_natural_order_start,
Sl, Al, absvalues, bits);
}
Reference in New Issue View Git Blame Copy Permalink