mirror of
https://github.com/izzy2lost/dolphin.git
synced 2026-06-19 01:16:48 -07:00
Merge pull request #10889 from shuffle2/spng
replace libpng with libspng
This commit is contained in:
@@ -32,3 +32,8 @@
|
||||
path = Externals/zlib-ng/zlib-ng
|
||||
url = https://github.com/zlib-ng/zlib-ng.git
|
||||
shallow = true
|
||||
[submodule "Externals/libspng/libspng"]
|
||||
path = Externals/libspng/libspng
|
||||
url = https://github.com/randy408/libspng.git
|
||||
branch = v0.7.2
|
||||
shallow = true
|
||||
|
||||
+1
-15
@@ -53,7 +53,6 @@ if(NOT ANDROID)
|
||||
endif()
|
||||
|
||||
option(USE_SHARED_ENET "Use shared libenet if found rather than Dolphin's soon-to-compatibly-diverge version" OFF)
|
||||
option(USE_SHARED_LIBPNG "Use shared libpng if found" ON)
|
||||
option(USE_UPNP "Enables UPnP port mapping support" ON)
|
||||
option(ENABLE_NOGUI "Enable NoGUI frontend" ON)
|
||||
option(ENABLE_QT "Enable Qt (Default)" ON)
|
||||
@@ -411,9 +410,6 @@ if(CMAKE_SYSTEM_NAME MATCHES "Darwin")
|
||||
find_library(IOB_LIBRARY IOBluetooth)
|
||||
find_library(IOK_LIBRARY IOKit)
|
||||
find_library(OPENGL_LIBRARY OpenGL)
|
||||
|
||||
# We don't want to use shared libpng.
|
||||
set(USE_SHARED_LIBPNG OFF)
|
||||
endif()
|
||||
|
||||
if(ENABLE_LTO)
|
||||
@@ -810,17 +806,7 @@ else()
|
||||
set(LZO lzo2)
|
||||
endif()
|
||||
|
||||
if(USE_SHARED_LIBPNG)
|
||||
check_lib(PNG libpng png png.h QUIET)
|
||||
endif()
|
||||
if (PNG_FOUND)
|
||||
message(STATUS "Using shared libpng")
|
||||
else()
|
||||
check_vendoring_approved(libpng)
|
||||
message(STATUS "Using static libpng from Externals")
|
||||
add_subdirectory(Externals/libpng)
|
||||
set(PNG png)
|
||||
endif()
|
||||
add_subdirectory(Externals/libspng)
|
||||
|
||||
# Using static FreeSurround from Externals
|
||||
# There is no system FreeSurround library.
|
||||
|
||||
+2
-2
@@ -52,8 +52,8 @@
|
||||
<ProjectReference Include="$(ExternalsDir)liblzma\liblzma.vcxproj">
|
||||
<Project>{055a775f-b4f5-4970-9240-f6cf7661f37b}</Project>
|
||||
</ProjectReference>
|
||||
<ProjectReference Include="$(ExternalsDir)libpng\png\png.vcxproj">
|
||||
<Project>{4c9f135b-a85e-430c-bad4-4c67ef5fc12c}</Project>
|
||||
<ProjectReference Include="$(ExternalsDir)libspng\spng.vcxproj">
|
||||
<Project>{447B7B1E-1D74-4AEF-B2B9-6EB41C5D5313}</Project>
|
||||
</ProjectReference>
|
||||
<ProjectReference Include="$(ExternalsDir)libusb\libusb-1.0.vcxproj">
|
||||
<Project>{349ee8f9-7d25-4909-aaf5-ff3fade72187}</Project>
|
||||
|
||||
Vendored
-204
@@ -1,204 +0,0 @@
|
||||
# selectively extracted and adapted from the libpng CMakeLists.txt, which has the following copyright notice:
|
||||
|
||||
|
||||
# Copyright (C) 2018 Cosmin Truta
|
||||
# Copyright (C) 2007,2009-2018 Glenn Randers-Pehrson
|
||||
# Written by Christian Ehrlicher, 2007
|
||||
# Revised by Roger Lowman, 2009-2010
|
||||
# Revised by Clifford Yapp, 2011-2012,2017
|
||||
# Revised by Roger Leigh, 2016
|
||||
# Revised by Andreas Franek, 2016
|
||||
# Revised by Sam Serrels, 2017
|
||||
# Revised by Vadim Barkov, 2017
|
||||
# Revised by Vicky Pfau, 2018
|
||||
# Revised by Cameron Cawley, 2018
|
||||
# Revised by Cosmin Truta, 2018
|
||||
# Revised by Kyle Bentley, 2018
|
||||
|
||||
# This code is released under the libpng license.
|
||||
# For conditions of distribution and use, see the disclaimer
|
||||
# and license in png.h
|
||||
|
||||
|
||||
|
||||
add_library(png STATIC
|
||||
png.c
|
||||
pngerror.c
|
||||
pngget.c
|
||||
pngmem.c
|
||||
pngpread.c
|
||||
pngread.c
|
||||
pngrio.c
|
||||
pngrtran.c
|
||||
pngrutil.c
|
||||
pngset.c
|
||||
pngtrans.c
|
||||
pngwio.c
|
||||
pngwrite.c
|
||||
pngwtran.c
|
||||
pngwutil.c
|
||||
)
|
||||
|
||||
dolphin_disable_warnings_msvc(png)
|
||||
|
||||
option(PNG_HARDWARE_OPTIMIZATIONS "Enable hardware optimizations for libpng" OFF)
|
||||
|
||||
if(PNG_HARDWARE_OPTIMIZATIONS)
|
||||
|
||||
message(STATUS "CMAKE_SYSTEM_PROCESSOR: ${CMAKE_SYSTEM_PROCESSOR}")
|
||||
|
||||
# set definitions and sources for arm
|
||||
if(CMAKE_SYSTEM_PROCESSOR MATCHES "^arm" OR
|
||||
CMAKE_SYSTEM_PROCESSOR MATCHES "^aarch64")
|
||||
set(PNG_ARM_NEON_POSSIBLE_VALUES check on off)
|
||||
set(PNG_ARM_NEON "check" CACHE STRING "Enable ARM NEON optimizations:
|
||||
check: (default) use internal checking code;
|
||||
off: disable the optimizations;
|
||||
on: turn on unconditionally.")
|
||||
set_property(CACHE PNG_ARM_NEON PROPERTY STRINGS
|
||||
${PNG_ARM_NEON_POSSIBLE_VALUES})
|
||||
list(FIND PNG_ARM_NEON_POSSIBLE_VALUES ${PNG_ARM_NEON} index)
|
||||
if(index EQUAL -1)
|
||||
message(FATAL_ERROR
|
||||
"PNG_ARM_NEON must be one of [${PNG_ARM_NEON_POSSIBLE_VALUES}]")
|
||||
elseif(NOT ${PNG_ARM_NEON} STREQUAL "off")
|
||||
target_sources(png PRIVATE
|
||||
arm/arm_init.c
|
||||
arm/filter_neon.S
|
||||
arm/filter_neon_intrinsics.c
|
||||
arm/palette_neon_intrinsics.c)
|
||||
|
||||
if(${PNG_ARM_NEON} STREQUAL "on")
|
||||
target_compile_definitions(png PUBLIC -DPNG_ARM_NEON_OPT=2)
|
||||
elseif(${PNG_ARM_NEON} STREQUAL "check")
|
||||
target_compile_definitions(png PUBLIC -DPNG_ARM_NEON_CHECK_SUPPORTED)
|
||||
endif()
|
||||
else()
|
||||
target_compile_definitions(png PUBLIC -DPNG_ARM_NEON_OPT=0)
|
||||
endif()
|
||||
endif()
|
||||
|
||||
# set definitions and sources for powerpc
|
||||
if(CMAKE_SYSTEM_PROCESSOR MATCHES "^powerpc*" OR
|
||||
CMAKE_SYSTEM_PROCESSOR MATCHES "^ppc64*")
|
||||
set(PNG_POWERPC_VSX_POSSIBLE_VALUES on off)
|
||||
set(PNG_POWERPC_VSX "on" CACHE STRING "Enable POWERPC VSX optimizations:
|
||||
off: disable the optimizations.")
|
||||
set_property(CACHE PNG_POWERPC_VSX PROPERTY STRINGS
|
||||
${PNG_POWERPC_VSX_POSSIBLE_VALUES})
|
||||
list(FIND PNG_POWERPC_VSX_POSSIBLE_VALUES ${PNG_POWERPC_VSX} index)
|
||||
if(index EQUAL -1)
|
||||
message(FATAL_ERROR
|
||||
"PNG_POWERPC_VSX must be one of [${PNG_POWERPC_VSX_POSSIBLE_VALUES}]")
|
||||
elseif(NOT ${PNG_POWERPC_VSX} STREQUAL "off")
|
||||
target_sources(png PRIVATE
|
||||
powerpc/powerpc_init.c
|
||||
powerpc/filter_vsx_intrinsics.c)
|
||||
if(${PNG_POWERPC_VSX} STREQUAL "on")
|
||||
target_compile_definitions(png PUBLIC -DPNG_POWERPC_VSX_OPT=2)
|
||||
endif()
|
||||
else()
|
||||
target_compile_definitions(png PUBLIC -DPNG_POWERPC_VSX_OPT=0)
|
||||
endif()
|
||||
endif()
|
||||
|
||||
# set definitions and sources for intel
|
||||
if(CMAKE_SYSTEM_PROCESSOR MATCHES "^i?86" OR
|
||||
CMAKE_SYSTEM_PROCESSOR MATCHES "^x86_64*")
|
||||
set(PNG_INTEL_SSE_POSSIBLE_VALUES on off)
|
||||
set(PNG_INTEL_SSE "on" CACHE STRING "Enable INTEL_SSE optimizations:
|
||||
off: disable the optimizations")
|
||||
set_property(CACHE PNG_INTEL_SSE PROPERTY STRINGS
|
||||
${PNG_INTEL_SSE_POSSIBLE_VALUES})
|
||||
list(FIND PNG_INTEL_SSE_POSSIBLE_VALUES ${PNG_INTEL_SSE} index)
|
||||
if(index EQUAL -1)
|
||||
message(FATAL_ERROR
|
||||
"PNG_INTEL_SSE must be one of [${PNG_INTEL_SSE_POSSIBLE_VALUES}]")
|
||||
elseif(NOT ${PNG_INTEL_SSE} STREQUAL "off")
|
||||
target_sources(png PRIVATE
|
||||
intel/intel_init.c
|
||||
intel/filter_sse2_intrinsics.c)
|
||||
if(${PNG_INTEL_SSE} STREQUAL "on")
|
||||
target_compile_definitions(png PUBLIC -DPNG_INTEL_SSE_OPT=1)
|
||||
endif()
|
||||
else()
|
||||
target_compile_definitions(png PUBLIC -DPNG_INTEL_SSE_OPT=0)
|
||||
endif()
|
||||
endif()
|
||||
|
||||
# set definitions and sources for MIPS
|
||||
if(CMAKE_SYSTEM_PROCESSOR MATCHES "mipsel*" OR
|
||||
CMAKE_SYSTEM_PROCESSOR MATCHES "mips64el*")
|
||||
set(PNG_MIPS_MSA_POSSIBLE_VALUES on off)
|
||||
set(PNG_MIPS_MSA "on" CACHE STRING "Enable MIPS_MSA optimizations:
|
||||
off: disable the optimizations")
|
||||
set_property(CACHE PNG_MIPS_MSA PROPERTY STRINGS
|
||||
${PNG_MIPS_MSA_POSSIBLE_VALUES})
|
||||
list(FIND PNG_MIPS_MSA_POSSIBLE_VALUES ${PNG_MIPS_MSA} index)
|
||||
if(index EQUAL -1)
|
||||
message(FATAL_ERROR
|
||||
"PNG_MIPS_MSA must be one of [${PNG_MIPS_MSA_POSSIBLE_VALUES}]")
|
||||
elseif(NOT ${PNG_MIPS_MSA} STREQUAL "off")
|
||||
target_sources(png PRIVATE
|
||||
mips/mips_init.c
|
||||
mips/filter_msa_intrinsics.c)
|
||||
if(${PNG_MIPS_MSA} STREQUAL "on")
|
||||
target_compile_definitions(png PUBLIC -DPNG_MIPS_MSA_OPT=2)
|
||||
endif()
|
||||
else()
|
||||
target_compile_definitions(png PUBLIC -DPNG_MIPS_MSA_OPT=0)
|
||||
endif()
|
||||
endif()
|
||||
|
||||
else(PNG_HARDWARE_OPTIMIZATIONS)
|
||||
|
||||
# set definitions and sources for arm
|
||||
if(CMAKE_SYSTEM_PROCESSOR MATCHES "^arm" OR
|
||||
CMAKE_SYSTEM_PROCESSOR MATCHES "^aarch64")
|
||||
target_compile_definitions(png PUBLIC -DPNG_ARM_NEON_OPT=0)
|
||||
endif()
|
||||
|
||||
# set definitions and sources for powerpc
|
||||
if(CMAKE_SYSTEM_PROCESSOR MATCHES "^powerpc*" OR
|
||||
CMAKE_SYSTEM_PROCESSOR MATCHES "^ppc64*")
|
||||
target_compile_definitions(png PUBLIC -DPNG_POWERPC_VSX_OPT=0)
|
||||
endif()
|
||||
|
||||
# set definitions and sources for intel
|
||||
if(CMAKE_SYSTEM_PROCESSOR MATCHES "^i?86" OR
|
||||
CMAKE_SYSTEM_PROCESSOR MATCHES "^x86_64*")
|
||||
target_compile_definitions(png PUBLIC -DPNG_INTEL_SSE_OPT=0)
|
||||
endif()
|
||||
|
||||
# set definitions and sources for MIPS
|
||||
if(CMAKE_SYSTEM_PROCESSOR MATCHES "mipsel*" OR
|
||||
CMAKE_SYSTEM_PROCESSOR MATCHES "mips64el*")
|
||||
target_compile_definitions(png PUBLIC -DPNG_MIPS_MSA_OPT=0)
|
||||
endif()
|
||||
|
||||
endif(PNG_HARDWARE_OPTIMIZATIONS)
|
||||
|
||||
target_sources(png PRIVATE
|
||||
# public headers
|
||||
png.h
|
||||
pngconf.h
|
||||
pnglibconf.h
|
||||
|
||||
# private headers
|
||||
pngpriv.h
|
||||
pngdebug.h
|
||||
pnginfo.h
|
||||
pngstruct.h
|
||||
)
|
||||
|
||||
target_include_directories(png PUBLIC
|
||||
${CMAKE_CURRENT_SOURCE_DIR}
|
||||
)
|
||||
|
||||
target_link_libraries(png PUBLIC ZLIB::ZLIB)
|
||||
|
||||
if(NOT MSVC)
|
||||
target_compile_options(png PRIVATE
|
||||
-Wno-self-assign
|
||||
)
|
||||
endif()
|
||||
Vendored
-134
@@ -1,134 +0,0 @@
|
||||
COPYRIGHT NOTICE, DISCLAIMER, and LICENSE
|
||||
=========================================
|
||||
|
||||
PNG Reference Library License version 2
|
||||
---------------------------------------
|
||||
|
||||
* Copyright (c) 1995-2019 The PNG Reference Library Authors.
|
||||
* Copyright (c) 2018-2019 Cosmin Truta.
|
||||
* Copyright (c) 2000-2002, 2004, 2006-2018 Glenn Randers-Pehrson.
|
||||
* Copyright (c) 1996-1997 Andreas Dilger.
|
||||
* Copyright (c) 1995-1996 Guy Eric Schalnat, Group 42, Inc.
|
||||
|
||||
The software is supplied "as is", without warranty of any kind,
|
||||
express or implied, including, without limitation, the warranties
|
||||
of merchantability, fitness for a particular purpose, title, and
|
||||
non-infringement. In no event shall the Copyright owners, or
|
||||
anyone distributing the software, be liable for any damages or
|
||||
other liability, whether in contract, tort or otherwise, arising
|
||||
from, out of, or in connection with the software, or the use or
|
||||
other dealings in the software, even if advised of the possibility
|
||||
of such damage.
|
||||
|
||||
Permission is hereby granted to use, copy, modify, and distribute
|
||||
this software, or portions hereof, for any purpose, without fee,
|
||||
subject to the following restrictions:
|
||||
|
||||
1. The origin of this software must not be misrepresented; you
|
||||
must not claim that you wrote the original software. If you
|
||||
use this software in a product, an acknowledgment in the product
|
||||
documentation would be appreciated, but is not required.
|
||||
|
||||
2. Altered source versions must be plainly marked as such, and must
|
||||
not be misrepresented as being the original software.
|
||||
|
||||
3. This Copyright notice may not be removed or altered from any
|
||||
source or altered source distribution.
|
||||
|
||||
|
||||
PNG Reference Library License version 1 (for libpng 0.5 through 1.6.35)
|
||||
-----------------------------------------------------------------------
|
||||
|
||||
libpng versions 1.0.7, July 1, 2000, through 1.6.35, July 15, 2018 are
|
||||
Copyright (c) 2000-2002, 2004, 2006-2018 Glenn Randers-Pehrson, are
|
||||
derived from libpng-1.0.6, and are distributed according to the same
|
||||
disclaimer and license as libpng-1.0.6 with the following individuals
|
||||
added to the list of Contributing Authors:
|
||||
|
||||
Simon-Pierre Cadieux
|
||||
Eric S. Raymond
|
||||
Mans Rullgard
|
||||
Cosmin Truta
|
||||
Gilles Vollant
|
||||
James Yu
|
||||
Mandar Sahastrabuddhe
|
||||
Google Inc.
|
||||
Vadim Barkov
|
||||
|
||||
and with the following additions to the disclaimer:
|
||||
|
||||
There is no warranty against interference with your enjoyment of
|
||||
the library or against infringement. There is no warranty that our
|
||||
efforts or the library will fulfill any of your particular purposes
|
||||
or needs. This library is provided with all faults, and the entire
|
||||
risk of satisfactory quality, performance, accuracy, and effort is
|
||||
with the user.
|
||||
|
||||
Some files in the "contrib" directory and some configure-generated
|
||||
files that are distributed with libpng have other copyright owners, and
|
||||
are released under other open source licenses.
|
||||
|
||||
libpng versions 0.97, January 1998, through 1.0.6, March 20, 2000, are
|
||||
Copyright (c) 1998-2000 Glenn Randers-Pehrson, are derived from
|
||||
libpng-0.96, and are distributed according to the same disclaimer and
|
||||
license as libpng-0.96, with the following individuals added to the
|
||||
list of Contributing Authors:
|
||||
|
||||
Tom Lane
|
||||
Glenn Randers-Pehrson
|
||||
Willem van Schaik
|
||||
|
||||
libpng versions 0.89, June 1996, through 0.96, May 1997, are
|
||||
Copyright (c) 1996-1997 Andreas Dilger, are derived from libpng-0.88,
|
||||
and are distributed according to the same disclaimer and license as
|
||||
libpng-0.88, with the following individuals added to the list of
|
||||
Contributing Authors:
|
||||
|
||||
John Bowler
|
||||
Kevin Bracey
|
||||
Sam Bushell
|
||||
Magnus Holmgren
|
||||
Greg Roelofs
|
||||
Tom Tanner
|
||||
|
||||
Some files in the "scripts" directory have other copyright owners,
|
||||
but are released under this license.
|
||||
|
||||
libpng versions 0.5, May 1995, through 0.88, January 1996, are
|
||||
Copyright (c) 1995-1996 Guy Eric Schalnat, Group 42, Inc.
|
||||
|
||||
For the purposes of this copyright and license, "Contributing Authors"
|
||||
is defined as the following set of individuals:
|
||||
|
||||
Andreas Dilger
|
||||
Dave Martindale
|
||||
Guy Eric Schalnat
|
||||
Paul Schmidt
|
||||
Tim Wegner
|
||||
|
||||
The PNG Reference Library is supplied "AS IS". The Contributing
|
||||
Authors and Group 42, Inc. disclaim all warranties, expressed or
|
||||
implied, including, without limitation, the warranties of
|
||||
merchantability and of fitness for any purpose. The Contributing
|
||||
Authors and Group 42, Inc. assume no liability for direct, indirect,
|
||||
incidental, special, exemplary, or consequential damages, which may
|
||||
result from the use of the PNG Reference Library, even if advised of
|
||||
the possibility of such damage.
|
||||
|
||||
Permission is hereby granted to use, copy, modify, and distribute this
|
||||
source code, or portions hereof, for any purpose, without fee, subject
|
||||
to the following restrictions:
|
||||
|
||||
1. The origin of this source code must not be misrepresented.
|
||||
|
||||
2. Altered versions must be plainly marked as such and must not
|
||||
be misrepresented as being the original source.
|
||||
|
||||
3. This Copyright notice may not be removed or altered from any
|
||||
source or altered source distribution.
|
||||
|
||||
The Contributing Authors and Group 42, Inc. specifically permit,
|
||||
without fee, and encourage the use of this source code as a component
|
||||
to supporting the PNG file format in commercial products. If you use
|
||||
this source code in a product, acknowledgment is not required but would
|
||||
be appreciated.
|
||||
Vendored
-136
@@ -1,136 +0,0 @@
|
||||
|
||||
/* arm_init.c - NEON optimised filter functions
|
||||
*
|
||||
* Copyright (c) 2018 Cosmin Truta
|
||||
* Copyright (c) 2014,2016 Glenn Randers-Pehrson
|
||||
* Written by Mans Rullgard, 2011.
|
||||
*
|
||||
* This code is released under the libpng license.
|
||||
* For conditions of distribution and use, see the disclaimer
|
||||
* and license in png.h
|
||||
*/
|
||||
|
||||
/* Below, after checking __linux__, various non-C90 POSIX 1003.1 functions are
|
||||
* called.
|
||||
*/
|
||||
#define _POSIX_SOURCE 1
|
||||
|
||||
#include "../pngpriv.h"
|
||||
|
||||
#ifdef PNG_READ_SUPPORTED
|
||||
|
||||
#if PNG_ARM_NEON_OPT > 0
|
||||
#ifdef PNG_ARM_NEON_CHECK_SUPPORTED /* Do run-time checks */
|
||||
/* WARNING: it is strongly recommended that you do not build libpng with
|
||||
* run-time checks for CPU features if at all possible. In the case of the ARM
|
||||
* NEON instructions there is no processor-specific way of detecting the
|
||||
* presence of the required support, therefore run-time detection is extremely
|
||||
* OS specific.
|
||||
*
|
||||
* You may set the macro PNG_ARM_NEON_FILE to the file name of file containing
|
||||
* a fragment of C source code which defines the png_have_neon function. There
|
||||
* are a number of implementations in contrib/arm-neon, but the only one that
|
||||
* has partial support is contrib/arm-neon/linux.c - a generic Linux
|
||||
* implementation which reads /proc/cpufino.
|
||||
*/
|
||||
#ifndef PNG_ARM_NEON_FILE
|
||||
# ifdef __linux__
|
||||
# define PNG_ARM_NEON_FILE "contrib/arm-neon/linux.c"
|
||||
# endif
|
||||
#endif
|
||||
|
||||
#ifdef PNG_ARM_NEON_FILE
|
||||
|
||||
#include <signal.h> /* for sig_atomic_t */
|
||||
static int png_have_neon(png_structp png_ptr);
|
||||
#include PNG_ARM_NEON_FILE
|
||||
|
||||
#else /* PNG_ARM_NEON_FILE */
|
||||
# error "PNG_ARM_NEON_FILE undefined: no support for run-time ARM NEON checks"
|
||||
#endif /* PNG_ARM_NEON_FILE */
|
||||
#endif /* PNG_ARM_NEON_CHECK_SUPPORTED */
|
||||
|
||||
#ifndef PNG_ALIGNED_MEMORY_SUPPORTED
|
||||
# error "ALIGNED_MEMORY is required; set: -DPNG_ALIGNED_MEMORY_SUPPORTED"
|
||||
#endif
|
||||
|
||||
void
|
||||
png_init_filter_functions_neon(png_structp pp, unsigned int bpp)
|
||||
{
|
||||
/* The switch statement is compiled in for ARM_NEON_API, the call to
|
||||
* png_have_neon is compiled in for ARM_NEON_CHECK. If both are defined
|
||||
* the check is only performed if the API has not set the NEON option on
|
||||
* or off explicitly. In this case the check controls what happens.
|
||||
*
|
||||
* If the CHECK is not compiled in and the option is UNSET the behavior prior
|
||||
* to 1.6.7 was to use the NEON code - this was a bug caused by having the
|
||||
* wrong order of the 'ON' and 'default' cases. UNSET now defaults to OFF,
|
||||
* as documented in png.h
|
||||
*/
|
||||
png_debug(1, "in png_init_filter_functions_neon");
|
||||
#ifdef PNG_ARM_NEON_API_SUPPORTED
|
||||
switch ((pp->options >> PNG_ARM_NEON) & 3)
|
||||
{
|
||||
case PNG_OPTION_UNSET:
|
||||
/* Allow the run-time check to execute if it has been enabled -
|
||||
* thus both API and CHECK can be turned on. If it isn't supported
|
||||
* this case will fall through to the 'default' below, which just
|
||||
* returns.
|
||||
*/
|
||||
#endif /* PNG_ARM_NEON_API_SUPPORTED */
|
||||
#ifdef PNG_ARM_NEON_CHECK_SUPPORTED
|
||||
{
|
||||
static volatile sig_atomic_t no_neon = -1; /* not checked */
|
||||
|
||||
if (no_neon < 0)
|
||||
no_neon = !png_have_neon(pp);
|
||||
|
||||
if (no_neon)
|
||||
return;
|
||||
}
|
||||
#ifdef PNG_ARM_NEON_API_SUPPORTED
|
||||
break;
|
||||
#endif
|
||||
#endif /* PNG_ARM_NEON_CHECK_SUPPORTED */
|
||||
|
||||
#ifdef PNG_ARM_NEON_API_SUPPORTED
|
||||
default: /* OFF or INVALID */
|
||||
return;
|
||||
|
||||
case PNG_OPTION_ON:
|
||||
/* Option turned on */
|
||||
break;
|
||||
}
|
||||
#endif
|
||||
|
||||
/* IMPORTANT: any new external functions used here must be declared using
|
||||
* PNG_INTERNAL_FUNCTION in ../pngpriv.h. This is required so that the
|
||||
* 'prefix' option to configure works:
|
||||
*
|
||||
* ./configure --with-libpng-prefix=foobar_
|
||||
*
|
||||
* Verify you have got this right by running the above command, doing a build
|
||||
* and examining pngprefix.h; it must contain a #define for every external
|
||||
* function you add. (Notice that this happens automatically for the
|
||||
* initialization function.)
|
||||
*/
|
||||
pp->read_filter[PNG_FILTER_VALUE_UP-1] = png_read_filter_row_up_neon;
|
||||
|
||||
if (bpp == 3)
|
||||
{
|
||||
pp->read_filter[PNG_FILTER_VALUE_SUB-1] = png_read_filter_row_sub3_neon;
|
||||
pp->read_filter[PNG_FILTER_VALUE_AVG-1] = png_read_filter_row_avg3_neon;
|
||||
pp->read_filter[PNG_FILTER_VALUE_PAETH-1] =
|
||||
png_read_filter_row_paeth3_neon;
|
||||
}
|
||||
|
||||
else if (bpp == 4)
|
||||
{
|
||||
pp->read_filter[PNG_FILTER_VALUE_SUB-1] = png_read_filter_row_sub4_neon;
|
||||
pp->read_filter[PNG_FILTER_VALUE_AVG-1] = png_read_filter_row_avg4_neon;
|
||||
pp->read_filter[PNG_FILTER_VALUE_PAETH-1] =
|
||||
png_read_filter_row_paeth4_neon;
|
||||
}
|
||||
}
|
||||
#endif /* PNG_ARM_NEON_OPT > 0 */
|
||||
#endif /* READ */
|
||||
Vendored
-253
@@ -1,253 +0,0 @@
|
||||
|
||||
/* filter_neon.S - NEON optimised filter functions
|
||||
*
|
||||
* Copyright (c) 2018 Cosmin Truta
|
||||
* Copyright (c) 2014,2017 Glenn Randers-Pehrson
|
||||
* Written by Mans Rullgard, 2011.
|
||||
*
|
||||
* This code is released under the libpng license.
|
||||
* For conditions of distribution and use, see the disclaimer
|
||||
* and license in png.h
|
||||
*/
|
||||
|
||||
/* This is required to get the symbol renames, which are #defines, and the
|
||||
* definitions (or not) of PNG_ARM_NEON_OPT and PNG_ARM_NEON_IMPLEMENTATION.
|
||||
*/
|
||||
#define PNG_VERSION_INFO_ONLY
|
||||
#include "../pngpriv.h"
|
||||
|
||||
#if (defined(__linux__) || defined(__FreeBSD__)) && defined(__ELF__)
|
||||
.section .note.GNU-stack,"",%progbits /* mark stack as non-executable */
|
||||
#endif
|
||||
|
||||
#ifdef PNG_READ_SUPPORTED
|
||||
|
||||
/* Assembler NEON support - only works for 32-bit ARM (i.e. it does not work for
|
||||
* ARM64). The code in arm/filter_neon_intrinsics.c supports ARM64, however it
|
||||
* only works if -mfpu=neon is specified on the GCC command line. See pngpriv.h
|
||||
* for the logic which sets PNG_USE_ARM_NEON_ASM:
|
||||
*/
|
||||
#if PNG_ARM_NEON_IMPLEMENTATION == 2 /* hand-coded assembler */
|
||||
|
||||
#if PNG_ARM_NEON_OPT > 0
|
||||
|
||||
#ifdef __ELF__
|
||||
# define ELF
|
||||
#else
|
||||
# define ELF @
|
||||
#endif
|
||||
|
||||
.arch armv7-a
|
||||
.fpu neon
|
||||
|
||||
.macro func name, export=0
|
||||
.macro endfunc
|
||||
ELF .size \name, . - \name
|
||||
.endfunc
|
||||
.purgem endfunc
|
||||
.endm
|
||||
.text
|
||||
|
||||
/* Explicitly specifying alignment here because some versions of
|
||||
* GAS don't align code correctly. This is harmless in correctly
|
||||
* written versions of GAS.
|
||||
*/
|
||||
.align 2
|
||||
|
||||
.if \export
|
||||
.global \name
|
||||
.endif
|
||||
ELF .type \name, STT_FUNC
|
||||
.func \name
|
||||
\name:
|
||||
.endm
|
||||
|
||||
func png_read_filter_row_sub4_neon, export=1
|
||||
ldr r3, [r0, #4] @ rowbytes
|
||||
vmov.i8 d3, #0
|
||||
1:
|
||||
vld4.32 {d4[],d5[],d6[],d7[]}, [r1,:128]
|
||||
vadd.u8 d0, d3, d4
|
||||
vadd.u8 d1, d0, d5
|
||||
vadd.u8 d2, d1, d6
|
||||
vadd.u8 d3, d2, d7
|
||||
vst4.32 {d0[0],d1[0],d2[0],d3[0]},[r1,:128]!
|
||||
subs r3, r3, #16
|
||||
bgt 1b
|
||||
|
||||
bx lr
|
||||
endfunc
|
||||
|
||||
func png_read_filter_row_sub3_neon, export=1
|
||||
ldr r3, [r0, #4] @ rowbytes
|
||||
vmov.i8 d3, #0
|
||||
mov r0, r1
|
||||
mov r2, #3
|
||||
mov r12, #12
|
||||
vld1.8 {q11}, [r0], r12
|
||||
1:
|
||||
vext.8 d5, d22, d23, #3
|
||||
vadd.u8 d0, d3, d22
|
||||
vext.8 d6, d22, d23, #6
|
||||
vadd.u8 d1, d0, d5
|
||||
vext.8 d7, d23, d23, #1
|
||||
vld1.8 {q11}, [r0], r12
|
||||
vst1.32 {d0[0]}, [r1,:32], r2
|
||||
vadd.u8 d2, d1, d6
|
||||
vst1.32 {d1[0]}, [r1], r2
|
||||
vadd.u8 d3, d2, d7
|
||||
vst1.32 {d2[0]}, [r1], r2
|
||||
vst1.32 {d3[0]}, [r1], r2
|
||||
subs r3, r3, #12
|
||||
bgt 1b
|
||||
|
||||
bx lr
|
||||
endfunc
|
||||
|
||||
func png_read_filter_row_up_neon, export=1
|
||||
ldr r3, [r0, #4] @ rowbytes
|
||||
1:
|
||||
vld1.8 {q0}, [r1,:128]
|
||||
vld1.8 {q1}, [r2,:128]!
|
||||
vadd.u8 q0, q0, q1
|
||||
vst1.8 {q0}, [r1,:128]!
|
||||
subs r3, r3, #16
|
||||
bgt 1b
|
||||
|
||||
bx lr
|
||||
endfunc
|
||||
|
||||
func png_read_filter_row_avg4_neon, export=1
|
||||
ldr r12, [r0, #4] @ rowbytes
|
||||
vmov.i8 d3, #0
|
||||
1:
|
||||
vld4.32 {d4[],d5[],d6[],d7[]}, [r1,:128]
|
||||
vld4.32 {d16[],d17[],d18[],d19[]},[r2,:128]!
|
||||
vhadd.u8 d0, d3, d16
|
||||
vadd.u8 d0, d0, d4
|
||||
vhadd.u8 d1, d0, d17
|
||||
vadd.u8 d1, d1, d5
|
||||
vhadd.u8 d2, d1, d18
|
||||
vadd.u8 d2, d2, d6
|
||||
vhadd.u8 d3, d2, d19
|
||||
vadd.u8 d3, d3, d7
|
||||
vst4.32 {d0[0],d1[0],d2[0],d3[0]},[r1,:128]!
|
||||
subs r12, r12, #16
|
||||
bgt 1b
|
||||
|
||||
bx lr
|
||||
endfunc
|
||||
|
||||
func png_read_filter_row_avg3_neon, export=1
|
||||
push {r4,lr}
|
||||
ldr r12, [r0, #4] @ rowbytes
|
||||
vmov.i8 d3, #0
|
||||
mov r0, r1
|
||||
mov r4, #3
|
||||
mov lr, #12
|
||||
vld1.8 {q11}, [r0], lr
|
||||
1:
|
||||
vld1.8 {q10}, [r2], lr
|
||||
vext.8 d5, d22, d23, #3
|
||||
vhadd.u8 d0, d3, d20
|
||||
vext.8 d17, d20, d21, #3
|
||||
vadd.u8 d0, d0, d22
|
||||
vext.8 d6, d22, d23, #6
|
||||
vhadd.u8 d1, d0, d17
|
||||
vext.8 d18, d20, d21, #6
|
||||
vadd.u8 d1, d1, d5
|
||||
vext.8 d7, d23, d23, #1
|
||||
vld1.8 {q11}, [r0], lr
|
||||
vst1.32 {d0[0]}, [r1,:32], r4
|
||||
vhadd.u8 d2, d1, d18
|
||||
vst1.32 {d1[0]}, [r1], r4
|
||||
vext.8 d19, d21, d21, #1
|
||||
vadd.u8 d2, d2, d6
|
||||
vhadd.u8 d3, d2, d19
|
||||
vst1.32 {d2[0]}, [r1], r4
|
||||
vadd.u8 d3, d3, d7
|
||||
vst1.32 {d3[0]}, [r1], r4
|
||||
subs r12, r12, #12
|
||||
bgt 1b
|
||||
|
||||
pop {r4,pc}
|
||||
endfunc
|
||||
|
||||
.macro paeth rx, ra, rb, rc
|
||||
vaddl.u8 q12, \ra, \rb @ a + b
|
||||
vaddl.u8 q15, \rc, \rc @ 2*c
|
||||
vabdl.u8 q13, \rb, \rc @ pa
|
||||
vabdl.u8 q14, \ra, \rc @ pb
|
||||
vabd.u16 q15, q12, q15 @ pc
|
||||
vcle.u16 q12, q13, q14 @ pa <= pb
|
||||
vcle.u16 q13, q13, q15 @ pa <= pc
|
||||
vcle.u16 q14, q14, q15 @ pb <= pc
|
||||
vand q12, q12, q13 @ pa <= pb && pa <= pc
|
||||
vmovn.u16 d28, q14
|
||||
vmovn.u16 \rx, q12
|
||||
vbsl d28, \rb, \rc
|
||||
vbsl \rx, \ra, d28
|
||||
.endm
|
||||
|
||||
func png_read_filter_row_paeth4_neon, export=1
|
||||
ldr r12, [r0, #4] @ rowbytes
|
||||
vmov.i8 d3, #0
|
||||
vmov.i8 d20, #0
|
||||
1:
|
||||
vld4.32 {d4[],d5[],d6[],d7[]}, [r1,:128]
|
||||
vld4.32 {d16[],d17[],d18[],d19[]},[r2,:128]!
|
||||
paeth d0, d3, d16, d20
|
||||
vadd.u8 d0, d0, d4
|
||||
paeth d1, d0, d17, d16
|
||||
vadd.u8 d1, d1, d5
|
||||
paeth d2, d1, d18, d17
|
||||
vadd.u8 d2, d2, d6
|
||||
paeth d3, d2, d19, d18
|
||||
vmov d20, d19
|
||||
vadd.u8 d3, d3, d7
|
||||
vst4.32 {d0[0],d1[0],d2[0],d3[0]},[r1,:128]!
|
||||
subs r12, r12, #16
|
||||
bgt 1b
|
||||
|
||||
bx lr
|
||||
endfunc
|
||||
|
||||
func png_read_filter_row_paeth3_neon, export=1
|
||||
push {r4,lr}
|
||||
ldr r12, [r0, #4] @ rowbytes
|
||||
vmov.i8 d3, #0
|
||||
vmov.i8 d4, #0
|
||||
mov r0, r1
|
||||
mov r4, #3
|
||||
mov lr, #12
|
||||
vld1.8 {q11}, [r0], lr
|
||||
1:
|
||||
vld1.8 {q10}, [r2], lr
|
||||
paeth d0, d3, d20, d4
|
||||
vext.8 d5, d22, d23, #3
|
||||
vadd.u8 d0, d0, d22
|
||||
vext.8 d17, d20, d21, #3
|
||||
paeth d1, d0, d17, d20
|
||||
vst1.32 {d0[0]}, [r1,:32], r4
|
||||
vext.8 d6, d22, d23, #6
|
||||
vadd.u8 d1, d1, d5
|
||||
vext.8 d18, d20, d21, #6
|
||||
paeth d2, d1, d18, d17
|
||||
vext.8 d7, d23, d23, #1
|
||||
vld1.8 {q11}, [r0], lr
|
||||
vst1.32 {d1[0]}, [r1], r4
|
||||
vadd.u8 d2, d2, d6
|
||||
vext.8 d19, d21, d21, #1
|
||||
paeth d3, d2, d19, d18
|
||||
vst1.32 {d2[0]}, [r1], r4
|
||||
vmov d4, d19
|
||||
vadd.u8 d3, d3, d7
|
||||
vst1.32 {d3[0]}, [r1], r4
|
||||
subs r12, r12, #12
|
||||
bgt 1b
|
||||
|
||||
pop {r4,pc}
|
||||
endfunc
|
||||
#endif /* PNG_ARM_NEON_OPT > 0 */
|
||||
#endif /* PNG_ARM_NEON_IMPLEMENTATION == 2 (assembler) */
|
||||
#endif /* READ */
|
||||
-402
@@ -1,402 +0,0 @@
|
||||
|
||||
/* filter_neon_intrinsics.c - NEON optimised filter functions
|
||||
*
|
||||
* Copyright (c) 2018 Cosmin Truta
|
||||
* Copyright (c) 2014,2016 Glenn Randers-Pehrson
|
||||
* Written by James Yu <james.yu at linaro.org>, October 2013.
|
||||
* Based on filter_neon.S, written by Mans Rullgard, 2011.
|
||||
*
|
||||
* This code is released under the libpng license.
|
||||
* For conditions of distribution and use, see the disclaimer
|
||||
* and license in png.h
|
||||
*/
|
||||
|
||||
#include "../pngpriv.h"
|
||||
|
||||
#ifdef PNG_READ_SUPPORTED
|
||||
|
||||
/* This code requires -mfpu=neon on the command line: */
|
||||
#if PNG_ARM_NEON_IMPLEMENTATION == 1 /* intrinsics code from pngpriv.h */
|
||||
|
||||
#if defined(_MSC_VER) && defined(_M_ARM64)
|
||||
# include <arm64_neon.h>
|
||||
#else
|
||||
# include <arm_neon.h>
|
||||
#endif
|
||||
|
||||
/* libpng row pointers are not necessarily aligned to any particular boundary,
|
||||
* however this code will only work with appropriate alignment. arm/arm_init.c
|
||||
* checks for this (and will not compile unless it is done). This code uses
|
||||
* variants of png_aligncast to avoid compiler warnings.
|
||||
*/
|
||||
#define png_ptr(type,pointer) png_aligncast(type *,pointer)
|
||||
#define png_ptrc(type,pointer) png_aligncastconst(const type *,pointer)
|
||||
|
||||
/* The following relies on a variable 'temp_pointer' being declared with type
|
||||
* 'type'. This is written this way just to hide the GCC strict aliasing
|
||||
* warning; note that the code is safe because there never is an alias between
|
||||
* the input and output pointers.
|
||||
*
|
||||
* When compiling with MSVC ARM64, the png_ldr macro can't be passed directly
|
||||
* to vst4_lane_u32, because of an internal compiler error inside MSVC.
|
||||
* To avoid this compiler bug, we use a temporary variable (vdest_val) to store
|
||||
* the result of png_ldr.
|
||||
*/
|
||||
#define png_ldr(type,pointer)\
|
||||
(temp_pointer = png_ptr(type,pointer), *temp_pointer)
|
||||
|
||||
#if PNG_ARM_NEON_OPT > 0
|
||||
|
||||
void
|
||||
png_read_filter_row_up_neon(png_row_infop row_info, png_bytep row,
|
||||
png_const_bytep prev_row)
|
||||
{
|
||||
png_bytep rp = row;
|
||||
png_bytep rp_stop = row + row_info->rowbytes;
|
||||
png_const_bytep pp = prev_row;
|
||||
|
||||
png_debug(1, "in png_read_filter_row_up_neon");
|
||||
|
||||
for (; rp < rp_stop; rp += 16, pp += 16)
|
||||
{
|
||||
uint8x16_t qrp, qpp;
|
||||
|
||||
qrp = vld1q_u8(rp);
|
||||
qpp = vld1q_u8(pp);
|
||||
qrp = vaddq_u8(qrp, qpp);
|
||||
vst1q_u8(rp, qrp);
|
||||
}
|
||||
}
|
||||
|
||||
void
|
||||
png_read_filter_row_sub3_neon(png_row_infop row_info, png_bytep row,
|
||||
png_const_bytep prev_row)
|
||||
{
|
||||
png_bytep rp = row;
|
||||
png_bytep rp_stop = row + row_info->rowbytes;
|
||||
|
||||
uint8x16_t vtmp = vld1q_u8(rp);
|
||||
uint8x8x2_t *vrpt = png_ptr(uint8x8x2_t, &vtmp);
|
||||
uint8x8x2_t vrp = *vrpt;
|
||||
|
||||
uint8x8x4_t vdest;
|
||||
vdest.val[3] = vdup_n_u8(0);
|
||||
|
||||
png_debug(1, "in png_read_filter_row_sub3_neon");
|
||||
|
||||
for (; rp < rp_stop;)
|
||||
{
|
||||
uint8x8_t vtmp1, vtmp2;
|
||||
uint32x2_t *temp_pointer;
|
||||
|
||||
vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 3);
|
||||
vdest.val[0] = vadd_u8(vdest.val[3], vrp.val[0]);
|
||||
vtmp2 = vext_u8(vrp.val[0], vrp.val[1], 6);
|
||||
vdest.val[1] = vadd_u8(vdest.val[0], vtmp1);
|
||||
|
||||
vtmp1 = vext_u8(vrp.val[1], vrp.val[1], 1);
|
||||
vdest.val[2] = vadd_u8(vdest.val[1], vtmp2);
|
||||
vdest.val[3] = vadd_u8(vdest.val[2], vtmp1);
|
||||
|
||||
vtmp = vld1q_u8(rp + 12);
|
||||
vrpt = png_ptr(uint8x8x2_t, &vtmp);
|
||||
vrp = *vrpt;
|
||||
|
||||
vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[0]), 0);
|
||||
rp += 3;
|
||||
vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[1]), 0);
|
||||
rp += 3;
|
||||
vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[2]), 0);
|
||||
rp += 3;
|
||||
vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[3]), 0);
|
||||
rp += 3;
|
||||
}
|
||||
|
||||
PNG_UNUSED(prev_row)
|
||||
}
|
||||
|
||||
void
|
||||
png_read_filter_row_sub4_neon(png_row_infop row_info, png_bytep row,
|
||||
png_const_bytep prev_row)
|
||||
{
|
||||
png_bytep rp = row;
|
||||
png_bytep rp_stop = row + row_info->rowbytes;
|
||||
|
||||
uint8x8x4_t vdest;
|
||||
vdest.val[3] = vdup_n_u8(0);
|
||||
|
||||
png_debug(1, "in png_read_filter_row_sub4_neon");
|
||||
|
||||
for (; rp < rp_stop; rp += 16)
|
||||
{
|
||||
uint32x2x4_t vtmp = vld4_u32(png_ptr(uint32_t,rp));
|
||||
uint8x8x4_t *vrpt = png_ptr(uint8x8x4_t,&vtmp);
|
||||
uint8x8x4_t vrp = *vrpt;
|
||||
uint32x2x4_t *temp_pointer;
|
||||
uint32x2x4_t vdest_val;
|
||||
|
||||
vdest.val[0] = vadd_u8(vdest.val[3], vrp.val[0]);
|
||||
vdest.val[1] = vadd_u8(vdest.val[0], vrp.val[1]);
|
||||
vdest.val[2] = vadd_u8(vdest.val[1], vrp.val[2]);
|
||||
vdest.val[3] = vadd_u8(vdest.val[2], vrp.val[3]);
|
||||
|
||||
vdest_val = png_ldr(uint32x2x4_t, &vdest);
|
||||
vst4_lane_u32(png_ptr(uint32_t,rp), vdest_val, 0);
|
||||
}
|
||||
|
||||
PNG_UNUSED(prev_row)
|
||||
}
|
||||
|
||||
void
|
||||
png_read_filter_row_avg3_neon(png_row_infop row_info, png_bytep row,
|
||||
png_const_bytep prev_row)
|
||||
{
|
||||
png_bytep rp = row;
|
||||
png_const_bytep pp = prev_row;
|
||||
png_bytep rp_stop = row + row_info->rowbytes;
|
||||
|
||||
uint8x16_t vtmp;
|
||||
uint8x8x2_t *vrpt;
|
||||
uint8x8x2_t vrp;
|
||||
uint8x8x4_t vdest;
|
||||
vdest.val[3] = vdup_n_u8(0);
|
||||
|
||||
vtmp = vld1q_u8(rp);
|
||||
vrpt = png_ptr(uint8x8x2_t,&vtmp);
|
||||
vrp = *vrpt;
|
||||
|
||||
png_debug(1, "in png_read_filter_row_avg3_neon");
|
||||
|
||||
for (; rp < rp_stop; pp += 12)
|
||||
{
|
||||
uint8x8_t vtmp1, vtmp2, vtmp3;
|
||||
|
||||
uint8x8x2_t *vppt;
|
||||
uint8x8x2_t vpp;
|
||||
|
||||
uint32x2_t *temp_pointer;
|
||||
|
||||
vtmp = vld1q_u8(pp);
|
||||
vppt = png_ptr(uint8x8x2_t,&vtmp);
|
||||
vpp = *vppt;
|
||||
|
||||
vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 3);
|
||||
vdest.val[0] = vhadd_u8(vdest.val[3], vpp.val[0]);
|
||||
vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]);
|
||||
|
||||
vtmp2 = vext_u8(vpp.val[0], vpp.val[1], 3);
|
||||
vtmp3 = vext_u8(vrp.val[0], vrp.val[1], 6);
|
||||
vdest.val[1] = vhadd_u8(vdest.val[0], vtmp2);
|
||||
vdest.val[1] = vadd_u8(vdest.val[1], vtmp1);
|
||||
|
||||
vtmp2 = vext_u8(vpp.val[0], vpp.val[1], 6);
|
||||
vtmp1 = vext_u8(vrp.val[1], vrp.val[1], 1);
|
||||
|
||||
vtmp = vld1q_u8(rp + 12);
|
||||
vrpt = png_ptr(uint8x8x2_t,&vtmp);
|
||||
vrp = *vrpt;
|
||||
|
||||
vdest.val[2] = vhadd_u8(vdest.val[1], vtmp2);
|
||||
vdest.val[2] = vadd_u8(vdest.val[2], vtmp3);
|
||||
|
||||
vtmp2 = vext_u8(vpp.val[1], vpp.val[1], 1);
|
||||
|
||||
vdest.val[3] = vhadd_u8(vdest.val[2], vtmp2);
|
||||
vdest.val[3] = vadd_u8(vdest.val[3], vtmp1);
|
||||
|
||||
vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[0]), 0);
|
||||
rp += 3;
|
||||
vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[1]), 0);
|
||||
rp += 3;
|
||||
vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[2]), 0);
|
||||
rp += 3;
|
||||
vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[3]), 0);
|
||||
rp += 3;
|
||||
}
|
||||
}
|
||||
|
||||
void
|
||||
png_read_filter_row_avg4_neon(png_row_infop row_info, png_bytep row,
|
||||
png_const_bytep prev_row)
|
||||
{
|
||||
png_bytep rp = row;
|
||||
png_bytep rp_stop = row + row_info->rowbytes;
|
||||
png_const_bytep pp = prev_row;
|
||||
|
||||
uint8x8x4_t vdest;
|
||||
vdest.val[3] = vdup_n_u8(0);
|
||||
|
||||
png_debug(1, "in png_read_filter_row_avg4_neon");
|
||||
|
||||
for (; rp < rp_stop; rp += 16, pp += 16)
|
||||
{
|
||||
uint32x2x4_t vtmp;
|
||||
uint8x8x4_t *vrpt, *vppt;
|
||||
uint8x8x4_t vrp, vpp;
|
||||
uint32x2x4_t *temp_pointer;
|
||||
uint32x2x4_t vdest_val;
|
||||
|
||||
vtmp = vld4_u32(png_ptr(uint32_t,rp));
|
||||
vrpt = png_ptr(uint8x8x4_t,&vtmp);
|
||||
vrp = *vrpt;
|
||||
vtmp = vld4_u32(png_ptrc(uint32_t,pp));
|
||||
vppt = png_ptr(uint8x8x4_t,&vtmp);
|
||||
vpp = *vppt;
|
||||
|
||||
vdest.val[0] = vhadd_u8(vdest.val[3], vpp.val[0]);
|
||||
vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]);
|
||||
vdest.val[1] = vhadd_u8(vdest.val[0], vpp.val[1]);
|
||||
vdest.val[1] = vadd_u8(vdest.val[1], vrp.val[1]);
|
||||
vdest.val[2] = vhadd_u8(vdest.val[1], vpp.val[2]);
|
||||
vdest.val[2] = vadd_u8(vdest.val[2], vrp.val[2]);
|
||||
vdest.val[3] = vhadd_u8(vdest.val[2], vpp.val[3]);
|
||||
vdest.val[3] = vadd_u8(vdest.val[3], vrp.val[3]);
|
||||
|
||||
vdest_val = png_ldr(uint32x2x4_t, &vdest);
|
||||
vst4_lane_u32(png_ptr(uint32_t,rp), vdest_val, 0);
|
||||
}
|
||||
}
|
||||
|
||||
static uint8x8_t
|
||||
paeth(uint8x8_t a, uint8x8_t b, uint8x8_t c)
|
||||
{
|
||||
uint8x8_t d, e;
|
||||
uint16x8_t p1, pa, pb, pc;
|
||||
|
||||
p1 = vaddl_u8(a, b); /* a + b */
|
||||
pc = vaddl_u8(c, c); /* c * 2 */
|
||||
pa = vabdl_u8(b, c); /* pa */
|
||||
pb = vabdl_u8(a, c); /* pb */
|
||||
pc = vabdq_u16(p1, pc); /* pc */
|
||||
|
||||
p1 = vcleq_u16(pa, pb); /* pa <= pb */
|
||||
pa = vcleq_u16(pa, pc); /* pa <= pc */
|
||||
pb = vcleq_u16(pb, pc); /* pb <= pc */
|
||||
|
||||
p1 = vandq_u16(p1, pa); /* pa <= pb && pa <= pc */
|
||||
|
||||
d = vmovn_u16(pb);
|
||||
e = vmovn_u16(p1);
|
||||
|
||||
d = vbsl_u8(d, b, c);
|
||||
e = vbsl_u8(e, a, d);
|
||||
|
||||
return e;
|
||||
}
|
||||
|
||||
void
|
||||
png_read_filter_row_paeth3_neon(png_row_infop row_info, png_bytep row,
|
||||
png_const_bytep prev_row)
|
||||
{
|
||||
png_bytep rp = row;
|
||||
png_const_bytep pp = prev_row;
|
||||
png_bytep rp_stop = row + row_info->rowbytes;
|
||||
|
||||
uint8x16_t vtmp;
|
||||
uint8x8x2_t *vrpt;
|
||||
uint8x8x2_t vrp;
|
||||
uint8x8_t vlast = vdup_n_u8(0);
|
||||
uint8x8x4_t vdest;
|
||||
vdest.val[3] = vdup_n_u8(0);
|
||||
|
||||
vtmp = vld1q_u8(rp);
|
||||
vrpt = png_ptr(uint8x8x2_t,&vtmp);
|
||||
vrp = *vrpt;
|
||||
|
||||
png_debug(1, "in png_read_filter_row_paeth3_neon");
|
||||
|
||||
for (; rp < rp_stop; pp += 12)
|
||||
{
|
||||
uint8x8x2_t *vppt;
|
||||
uint8x8x2_t vpp;
|
||||
uint8x8_t vtmp1, vtmp2, vtmp3;
|
||||
uint32x2_t *temp_pointer;
|
||||
|
||||
vtmp = vld1q_u8(pp);
|
||||
vppt = png_ptr(uint8x8x2_t,&vtmp);
|
||||
vpp = *vppt;
|
||||
|
||||
vdest.val[0] = paeth(vdest.val[3], vpp.val[0], vlast);
|
||||
vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]);
|
||||
|
||||
vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 3);
|
||||
vtmp2 = vext_u8(vpp.val[0], vpp.val[1], 3);
|
||||
vdest.val[1] = paeth(vdest.val[0], vtmp2, vpp.val[0]);
|
||||
vdest.val[1] = vadd_u8(vdest.val[1], vtmp1);
|
||||
|
||||
vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 6);
|
||||
vtmp3 = vext_u8(vpp.val[0], vpp.val[1], 6);
|
||||
vdest.val[2] = paeth(vdest.val[1], vtmp3, vtmp2);
|
||||
vdest.val[2] = vadd_u8(vdest.val[2], vtmp1);
|
||||
|
||||
vtmp1 = vext_u8(vrp.val[1], vrp.val[1], 1);
|
||||
vtmp2 = vext_u8(vpp.val[1], vpp.val[1], 1);
|
||||
|
||||
vtmp = vld1q_u8(rp + 12);
|
||||
vrpt = png_ptr(uint8x8x2_t,&vtmp);
|
||||
vrp = *vrpt;
|
||||
|
||||
vdest.val[3] = paeth(vdest.val[2], vtmp2, vtmp3);
|
||||
vdest.val[3] = vadd_u8(vdest.val[3], vtmp1);
|
||||
|
||||
vlast = vtmp2;
|
||||
|
||||
vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[0]), 0);
|
||||
rp += 3;
|
||||
vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[1]), 0);
|
||||
rp += 3;
|
||||
vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[2]), 0);
|
||||
rp += 3;
|
||||
vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[3]), 0);
|
||||
rp += 3;
|
||||
}
|
||||
}
|
||||
|
||||
void
|
||||
png_read_filter_row_paeth4_neon(png_row_infop row_info, png_bytep row,
|
||||
png_const_bytep prev_row)
|
||||
{
|
||||
png_bytep rp = row;
|
||||
png_bytep rp_stop = row + row_info->rowbytes;
|
||||
png_const_bytep pp = prev_row;
|
||||
|
||||
uint8x8_t vlast = vdup_n_u8(0);
|
||||
uint8x8x4_t vdest;
|
||||
vdest.val[3] = vdup_n_u8(0);
|
||||
|
||||
png_debug(1, "in png_read_filter_row_paeth4_neon");
|
||||
|
||||
for (; rp < rp_stop; rp += 16, pp += 16)
|
||||
{
|
||||
uint32x2x4_t vtmp;
|
||||
uint8x8x4_t *vrpt, *vppt;
|
||||
uint8x8x4_t vrp, vpp;
|
||||
uint32x2x4_t *temp_pointer;
|
||||
uint32x2x4_t vdest_val;
|
||||
|
||||
vtmp = vld4_u32(png_ptr(uint32_t,rp));
|
||||
vrpt = png_ptr(uint8x8x4_t,&vtmp);
|
||||
vrp = *vrpt;
|
||||
vtmp = vld4_u32(png_ptrc(uint32_t,pp));
|
||||
vppt = png_ptr(uint8x8x4_t,&vtmp);
|
||||
vpp = *vppt;
|
||||
|
||||
vdest.val[0] = paeth(vdest.val[3], vpp.val[0], vlast);
|
||||
vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]);
|
||||
vdest.val[1] = paeth(vdest.val[0], vpp.val[1], vpp.val[0]);
|
||||
vdest.val[1] = vadd_u8(vdest.val[1], vrp.val[1]);
|
||||
vdest.val[2] = paeth(vdest.val[1], vpp.val[2], vpp.val[1]);
|
||||
vdest.val[2] = vadd_u8(vdest.val[2], vrp.val[2]);
|
||||
vdest.val[3] = paeth(vdest.val[2], vpp.val[3], vpp.val[2]);
|
||||
vdest.val[3] = vadd_u8(vdest.val[3], vrp.val[3]);
|
||||
|
||||
vlast = vpp.val[3];
|
||||
|
||||
vdest_val = png_ldr(uint32x2x4_t, &vdest);
|
||||
vst4_lane_u32(png_ptr(uint32_t,rp), vdest_val, 0);
|
||||
}
|
||||
}
|
||||
|
||||
#endif /* PNG_ARM_NEON_OPT > 0 */
|
||||
#endif /* PNG_ARM_NEON_IMPLEMENTATION == 1 (intrinsics) */
|
||||
#endif /* READ */
|
||||
-149
@@ -1,149 +0,0 @@
|
||||
|
||||
/* palette_neon_intrinsics.c - NEON optimised palette expansion functions
|
||||
*
|
||||
* Copyright (c) 2018-2019 Cosmin Truta
|
||||
* Copyright (c) 2017-2018 Arm Holdings. All rights reserved.
|
||||
* Written by Richard Townsend <Richard.Townsend@arm.com>, February 2017.
|
||||
*
|
||||
* This code is released under the libpng license.
|
||||
* For conditions of distribution and use, see the disclaimer
|
||||
* and license in png.h
|
||||
*/
|
||||
|
||||
#include "../pngpriv.h"
|
||||
|
||||
#if PNG_ARM_NEON_IMPLEMENTATION == 1
|
||||
|
||||
#if defined(_MSC_VER) && defined(_M_ARM64)
|
||||
# include <arm64_neon.h>
|
||||
#else
|
||||
# include <arm_neon.h>
|
||||
#endif
|
||||
|
||||
/* Build an RGBA8 palette from the separate RGB and alpha palettes. */
|
||||
void
|
||||
png_riffle_palette_neon(png_structrp png_ptr)
|
||||
{
|
||||
png_const_colorp palette = png_ptr->palette;
|
||||
png_bytep riffled_palette = png_ptr->riffled_palette;
|
||||
png_const_bytep trans_alpha = png_ptr->trans_alpha;
|
||||
int num_trans = png_ptr->num_trans;
|
||||
int i;
|
||||
|
||||
png_debug(1, "in png_riffle_palette_neon");
|
||||
|
||||
/* Initially black, opaque. */
|
||||
uint8x16x4_t w = {{
|
||||
vdupq_n_u8(0x00),
|
||||
vdupq_n_u8(0x00),
|
||||
vdupq_n_u8(0x00),
|
||||
vdupq_n_u8(0xff),
|
||||
}};
|
||||
|
||||
/* First, riffle the RGB colours into an RGBA8 palette.
|
||||
* The alpha component is set to opaque for now.
|
||||
*/
|
||||
for (i = 0; i < 256; i += 16)
|
||||
{
|
||||
uint8x16x3_t v = vld3q_u8((png_const_bytep)(palette + i));
|
||||
w.val[0] = v.val[0];
|
||||
w.val[1] = v.val[1];
|
||||
w.val[2] = v.val[2];
|
||||
vst4q_u8(riffled_palette + (i << 2), w);
|
||||
}
|
||||
|
||||
/* Fix up the missing transparency values. */
|
||||
for (i = 0; i < num_trans; i++)
|
||||
riffled_palette[(i << 2) + 3] = trans_alpha[i];
|
||||
}
|
||||
|
||||
/* Expands a palettized row into RGBA8. */
|
||||
int
|
||||
png_do_expand_palette_rgba8_neon(png_structrp png_ptr, png_row_infop row_info,
|
||||
png_const_bytep row, png_bytepp ssp, png_bytepp ddp)
|
||||
{
|
||||
png_uint_32 row_width = row_info->width;
|
||||
const png_uint_32 *riffled_palette =
|
||||
(const png_uint_32 *)png_ptr->riffled_palette;
|
||||
const png_int_32 pixels_per_chunk = 4;
|
||||
int i;
|
||||
|
||||
png_debug(1, "in png_do_expand_palette_rgba8_neon");
|
||||
|
||||
if (row_width < pixels_per_chunk)
|
||||
return 0;
|
||||
|
||||
/* This function originally gets the last byte of the output row.
|
||||
* The NEON part writes forward from a given position, so we have
|
||||
* to seek this back by 4 pixels x 4 bytes.
|
||||
*/
|
||||
*ddp = *ddp - ((pixels_per_chunk * sizeof(png_uint_32)) - 1);
|
||||
|
||||
for (i = 0; i < row_width; i += pixels_per_chunk)
|
||||
{
|
||||
uint32x4_t cur;
|
||||
png_bytep sp = *ssp - i, dp = *ddp - (i << 2);
|
||||
cur = vld1q_dup_u32 (riffled_palette + *(sp - 3));
|
||||
cur = vld1q_lane_u32(riffled_palette + *(sp - 2), cur, 1);
|
||||
cur = vld1q_lane_u32(riffled_palette + *(sp - 1), cur, 2);
|
||||
cur = vld1q_lane_u32(riffled_palette + *(sp - 0), cur, 3);
|
||||
vst1q_u32((void *)dp, cur);
|
||||
}
|
||||
if (i != row_width)
|
||||
{
|
||||
/* Remove the amount that wasn't processed. */
|
||||
i -= pixels_per_chunk;
|
||||
}
|
||||
|
||||
/* Decrement output pointers. */
|
||||
*ssp = *ssp - i;
|
||||
*ddp = *ddp - (i << 2);
|
||||
return i;
|
||||
}
|
||||
|
||||
/* Expands a palettized row into RGB8. */
|
||||
int
|
||||
png_do_expand_palette_rgb8_neon(png_structrp png_ptr, png_row_infop row_info,
|
||||
png_const_bytep row, png_bytepp ssp, png_bytepp ddp)
|
||||
{
|
||||
png_uint_32 row_width = row_info->width;
|
||||
png_const_bytep palette = (png_const_bytep)png_ptr->palette;
|
||||
const png_uint_32 pixels_per_chunk = 8;
|
||||
int i;
|
||||
|
||||
png_debug(1, "in png_do_expand_palette_rgb8_neon");
|
||||
|
||||
if (row_width <= pixels_per_chunk)
|
||||
return 0;
|
||||
|
||||
/* Seeking this back by 8 pixels x 3 bytes. */
|
||||
*ddp = *ddp - ((pixels_per_chunk * sizeof(png_color)) - 1);
|
||||
|
||||
for (i = 0; i < row_width; i += pixels_per_chunk)
|
||||
{
|
||||
uint8x8x3_t cur;
|
||||
png_bytep sp = *ssp - i, dp = *ddp - ((i << 1) + i);
|
||||
cur = vld3_dup_u8(palette + sizeof(png_color) * (*(sp - 7)));
|
||||
cur = vld3_lane_u8(palette + sizeof(png_color) * (*(sp - 6)), cur, 1);
|
||||
cur = vld3_lane_u8(palette + sizeof(png_color) * (*(sp - 5)), cur, 2);
|
||||
cur = vld3_lane_u8(palette + sizeof(png_color) * (*(sp - 4)), cur, 3);
|
||||
cur = vld3_lane_u8(palette + sizeof(png_color) * (*(sp - 3)), cur, 4);
|
||||
cur = vld3_lane_u8(palette + sizeof(png_color) * (*(sp - 2)), cur, 5);
|
||||
cur = vld3_lane_u8(palette + sizeof(png_color) * (*(sp - 1)), cur, 6);
|
||||
cur = vld3_lane_u8(palette + sizeof(png_color) * (*(sp - 0)), cur, 7);
|
||||
vst3_u8((void *)dp, cur);
|
||||
}
|
||||
|
||||
if (i != row_width)
|
||||
{
|
||||
/* Remove the amount that wasn't processed. */
|
||||
i -= pixels_per_chunk;
|
||||
}
|
||||
|
||||
/* Decrement output pointers. */
|
||||
*ssp = *ssp - i;
|
||||
*ddp = *ddp - ((i << 1) + i);
|
||||
return i;
|
||||
}
|
||||
|
||||
#endif /* PNG_ARM_NEON_IMPLEMENTATION */
|
||||
-391
@@ -1,391 +0,0 @@
|
||||
|
||||
/* filter_sse2_intrinsics.c - SSE2 optimized filter functions
|
||||
*
|
||||
* Copyright (c) 2018 Cosmin Truta
|
||||
* Copyright (c) 2016-2017 Glenn Randers-Pehrson
|
||||
* Written by Mike Klein and Matt Sarett
|
||||
* Derived from arm/filter_neon_intrinsics.c
|
||||
*
|
||||
* This code is released under the libpng license.
|
||||
* For conditions of distribution and use, see the disclaimer
|
||||
* and license in png.h
|
||||
*/
|
||||
|
||||
#include "../pngpriv.h"
|
||||
|
||||
#ifdef PNG_READ_SUPPORTED
|
||||
|
||||
#if PNG_INTEL_SSE_IMPLEMENTATION > 0
|
||||
|
||||
#include <immintrin.h>
|
||||
|
||||
/* Functions in this file look at most 3 pixels (a,b,c) to predict the 4th (d).
|
||||
* They're positioned like this:
|
||||
* prev: c b
|
||||
* row: a d
|
||||
* The Sub filter predicts d=a, Avg d=(a+b)/2, and Paeth predicts d to be
|
||||
* whichever of a, b, or c is closest to p=a+b-c.
|
||||
*/
|
||||
|
||||
static __m128i load4(const void* p) {
|
||||
int tmp;
|
||||
memcpy(&tmp, p, sizeof(tmp));
|
||||
return _mm_cvtsi32_si128(tmp);
|
||||
}
|
||||
|
||||
static void store4(void* p, __m128i v) {
|
||||
int tmp = _mm_cvtsi128_si32(v);
|
||||
memcpy(p, &tmp, sizeof(int));
|
||||
}
|
||||
|
||||
static __m128i load3(const void* p) {
|
||||
png_uint_32 tmp = 0;
|
||||
memcpy(&tmp, p, 3);
|
||||
return _mm_cvtsi32_si128(tmp);
|
||||
}
|
||||
|
||||
static void store3(void* p, __m128i v) {
|
||||
int tmp = _mm_cvtsi128_si32(v);
|
||||
memcpy(p, &tmp, 3);
|
||||
}
|
||||
|
||||
void png_read_filter_row_sub3_sse2(png_row_infop row_info, png_bytep row,
|
||||
png_const_bytep prev)
|
||||
{
|
||||
/* The Sub filter predicts each pixel as the previous pixel, a.
|
||||
* There is no pixel to the left of the first pixel. It's encoded directly.
|
||||
* That works with our main loop if we just say that left pixel was zero.
|
||||
*/
|
||||
size_t rb;
|
||||
|
||||
__m128i a, d = _mm_setzero_si128();
|
||||
|
||||
png_debug(1, "in png_read_filter_row_sub3_sse2");
|
||||
|
||||
rb = row_info->rowbytes;
|
||||
while (rb >= 4) {
|
||||
a = d; d = load4(row);
|
||||
d = _mm_add_epi8(d, a);
|
||||
store3(row, d);
|
||||
|
||||
row += 3;
|
||||
rb -= 3;
|
||||
}
|
||||
if (rb > 0) {
|
||||
a = d; d = load3(row);
|
||||
d = _mm_add_epi8(d, a);
|
||||
store3(row, d);
|
||||
|
||||
row += 3;
|
||||
rb -= 3;
|
||||
}
|
||||
PNG_UNUSED(prev)
|
||||
}
|
||||
|
||||
void png_read_filter_row_sub4_sse2(png_row_infop row_info, png_bytep row,
|
||||
png_const_bytep prev)
|
||||
{
|
||||
/* The Sub filter predicts each pixel as the previous pixel, a.
|
||||
* There is no pixel to the left of the first pixel. It's encoded directly.
|
||||
* That works with our main loop if we just say that left pixel was zero.
|
||||
*/
|
||||
size_t rb;
|
||||
|
||||
__m128i a, d = _mm_setzero_si128();
|
||||
|
||||
png_debug(1, "in png_read_filter_row_sub4_sse2");
|
||||
|
||||
rb = row_info->rowbytes+4;
|
||||
while (rb > 4) {
|
||||
a = d; d = load4(row);
|
||||
d = _mm_add_epi8(d, a);
|
||||
store4(row, d);
|
||||
|
||||
row += 4;
|
||||
rb -= 4;
|
||||
}
|
||||
PNG_UNUSED(prev)
|
||||
}
|
||||
|
||||
void png_read_filter_row_avg3_sse2(png_row_infop row_info, png_bytep row,
|
||||
png_const_bytep prev)
|
||||
{
|
||||
/* The Avg filter predicts each pixel as the (truncated) average of a and b.
|
||||
* There's no pixel to the left of the first pixel. Luckily, it's
|
||||
* predicted to be half of the pixel above it. So again, this works
|
||||
* perfectly with our loop if we make sure a starts at zero.
|
||||
*/
|
||||
|
||||
size_t rb;
|
||||
|
||||
const __m128i zero = _mm_setzero_si128();
|
||||
|
||||
__m128i b;
|
||||
__m128i a, d = zero;
|
||||
|
||||
png_debug(1, "in png_read_filter_row_avg3_sse2");
|
||||
rb = row_info->rowbytes;
|
||||
while (rb >= 4) {
|
||||
__m128i avg;
|
||||
b = load4(prev);
|
||||
a = d; d = load4(row );
|
||||
|
||||
/* PNG requires a truncating average, so we can't just use _mm_avg_epu8 */
|
||||
avg = _mm_avg_epu8(a,b);
|
||||
/* ...but we can fix it up by subtracting off 1 if it rounded up. */
|
||||
avg = _mm_sub_epi8(avg, _mm_and_si128(_mm_xor_si128(a,b),
|
||||
_mm_set1_epi8(1)));
|
||||
d = _mm_add_epi8(d, avg);
|
||||
store3(row, d);
|
||||
|
||||
prev += 3;
|
||||
row += 3;
|
||||
rb -= 3;
|
||||
}
|
||||
if (rb > 0) {
|
||||
__m128i avg;
|
||||
b = load3(prev);
|
||||
a = d; d = load3(row );
|
||||
|
||||
/* PNG requires a truncating average, so we can't just use _mm_avg_epu8 */
|
||||
avg = _mm_avg_epu8(a,b);
|
||||
/* ...but we can fix it up by subtracting off 1 if it rounded up. */
|
||||
avg = _mm_sub_epi8(avg, _mm_and_si128(_mm_xor_si128(a,b),
|
||||
_mm_set1_epi8(1)));
|
||||
|
||||
d = _mm_add_epi8(d, avg);
|
||||
store3(row, d);
|
||||
|
||||
prev += 3;
|
||||
row += 3;
|
||||
rb -= 3;
|
||||
}
|
||||
}
|
||||
|
||||
void png_read_filter_row_avg4_sse2(png_row_infop row_info, png_bytep row,
|
||||
png_const_bytep prev)
|
||||
{
|
||||
/* The Avg filter predicts each pixel as the (truncated) average of a and b.
|
||||
* There's no pixel to the left of the first pixel. Luckily, it's
|
||||
* predicted to be half of the pixel above it. So again, this works
|
||||
* perfectly with our loop if we make sure a starts at zero.
|
||||
*/
|
||||
size_t rb;
|
||||
const __m128i zero = _mm_setzero_si128();
|
||||
__m128i b;
|
||||
__m128i a, d = zero;
|
||||
|
||||
png_debug(1, "in png_read_filter_row_avg4_sse2");
|
||||
|
||||
rb = row_info->rowbytes+4;
|
||||
while (rb > 4) {
|
||||
__m128i avg;
|
||||
b = load4(prev);
|
||||
a = d; d = load4(row );
|
||||
|
||||
/* PNG requires a truncating average, so we can't just use _mm_avg_epu8 */
|
||||
avg = _mm_avg_epu8(a,b);
|
||||
/* ...but we can fix it up by subtracting off 1 if it rounded up. */
|
||||
avg = _mm_sub_epi8(avg, _mm_and_si128(_mm_xor_si128(a,b),
|
||||
_mm_set1_epi8(1)));
|
||||
|
||||
d = _mm_add_epi8(d, avg);
|
||||
store4(row, d);
|
||||
|
||||
prev += 4;
|
||||
row += 4;
|
||||
rb -= 4;
|
||||
}
|
||||
}
|
||||
|
||||
/* Returns |x| for 16-bit lanes. */
|
||||
static __m128i abs_i16(__m128i x) {
|
||||
#if PNG_INTEL_SSE_IMPLEMENTATION >= 2
|
||||
return _mm_abs_epi16(x);
|
||||
#else
|
||||
/* Read this all as, return x<0 ? -x : x.
|
||||
* To negate two's complement, you flip all the bits then add 1.
|
||||
*/
|
||||
__m128i is_negative = _mm_cmplt_epi16(x, _mm_setzero_si128());
|
||||
|
||||
/* Flip negative lanes. */
|
||||
x = _mm_xor_si128(x, is_negative);
|
||||
|
||||
/* +1 to negative lanes, else +0. */
|
||||
x = _mm_sub_epi16(x, is_negative);
|
||||
return x;
|
||||
#endif
|
||||
}
|
||||
|
||||
/* Bytewise c ? t : e. */
|
||||
static __m128i if_then_else(__m128i c, __m128i t, __m128i e) {
|
||||
#if PNG_INTEL_SSE_IMPLEMENTATION >= 3
|
||||
return _mm_blendv_epi8(e,t,c);
|
||||
#else
|
||||
return _mm_or_si128(_mm_and_si128(c, t), _mm_andnot_si128(c, e));
|
||||
#endif
|
||||
}
|
||||
|
||||
void png_read_filter_row_paeth3_sse2(png_row_infop row_info, png_bytep row,
|
||||
png_const_bytep prev)
|
||||
{
|
||||
/* Paeth tries to predict pixel d using the pixel to the left of it, a,
|
||||
* and two pixels from the previous row, b and c:
|
||||
* prev: c b
|
||||
* row: a d
|
||||
* The Paeth function predicts d to be whichever of a, b, or c is nearest to
|
||||
* p=a+b-c.
|
||||
*
|
||||
* The first pixel has no left context, and so uses an Up filter, p = b.
|
||||
* This works naturally with our main loop's p = a+b-c if we force a and c
|
||||
* to zero.
|
||||
* Here we zero b and d, which become c and a respectively at the start of
|
||||
* the loop.
|
||||
*/
|
||||
size_t rb;
|
||||
const __m128i zero = _mm_setzero_si128();
|
||||
__m128i c, b = zero,
|
||||
a, d = zero;
|
||||
|
||||
png_debug(1, "in png_read_filter_row_paeth3_sse2");
|
||||
|
||||
rb = row_info->rowbytes;
|
||||
while (rb >= 4) {
|
||||
/* It's easiest to do this math (particularly, deal with pc) with 16-bit
|
||||
* intermediates.
|
||||
*/
|
||||
__m128i pa,pb,pc,smallest,nearest;
|
||||
c = b; b = _mm_unpacklo_epi8(load4(prev), zero);
|
||||
a = d; d = _mm_unpacklo_epi8(load4(row ), zero);
|
||||
|
||||
/* (p-a) == (a+b-c - a) == (b-c) */
|
||||
|
||||
pa = _mm_sub_epi16(b,c);
|
||||
|
||||
/* (p-b) == (a+b-c - b) == (a-c) */
|
||||
pb = _mm_sub_epi16(a,c);
|
||||
|
||||
/* (p-c) == (a+b-c - c) == (a+b-c-c) == (b-c)+(a-c) */
|
||||
pc = _mm_add_epi16(pa,pb);
|
||||
|
||||
pa = abs_i16(pa); /* |p-a| */
|
||||
pb = abs_i16(pb); /* |p-b| */
|
||||
pc = abs_i16(pc); /* |p-c| */
|
||||
|
||||
smallest = _mm_min_epi16(pc, _mm_min_epi16(pa, pb));
|
||||
|
||||
/* Paeth breaks ties favoring a over b over c. */
|
||||
nearest = if_then_else(_mm_cmpeq_epi16(smallest, pa), a,
|
||||
if_then_else(_mm_cmpeq_epi16(smallest, pb), b,
|
||||
c));
|
||||
|
||||
/* Note `_epi8`: we need addition to wrap modulo 255. */
|
||||
d = _mm_add_epi8(d, nearest);
|
||||
store3(row, _mm_packus_epi16(d,d));
|
||||
|
||||
prev += 3;
|
||||
row += 3;
|
||||
rb -= 3;
|
||||
}
|
||||
if (rb > 0) {
|
||||
/* It's easiest to do this math (particularly, deal with pc) with 16-bit
|
||||
* intermediates.
|
||||
*/
|
||||
__m128i pa,pb,pc,smallest,nearest;
|
||||
c = b; b = _mm_unpacklo_epi8(load3(prev), zero);
|
||||
a = d; d = _mm_unpacklo_epi8(load3(row ), zero);
|
||||
|
||||
/* (p-a) == (a+b-c - a) == (b-c) */
|
||||
pa = _mm_sub_epi16(b,c);
|
||||
|
||||
/* (p-b) == (a+b-c - b) == (a-c) */
|
||||
pb = _mm_sub_epi16(a,c);
|
||||
|
||||
/* (p-c) == (a+b-c - c) == (a+b-c-c) == (b-c)+(a-c) */
|
||||
pc = _mm_add_epi16(pa,pb);
|
||||
|
||||
pa = abs_i16(pa); /* |p-a| */
|
||||
pb = abs_i16(pb); /* |p-b| */
|
||||
pc = abs_i16(pc); /* |p-c| */
|
||||
|
||||
smallest = _mm_min_epi16(pc, _mm_min_epi16(pa, pb));
|
||||
|
||||
/* Paeth breaks ties favoring a over b over c. */
|
||||
nearest = if_then_else(_mm_cmpeq_epi16(smallest, pa), a,
|
||||
if_then_else(_mm_cmpeq_epi16(smallest, pb), b,
|
||||
c));
|
||||
|
||||
/* Note `_epi8`: we need addition to wrap modulo 255. */
|
||||
d = _mm_add_epi8(d, nearest);
|
||||
store3(row, _mm_packus_epi16(d,d));
|
||||
|
||||
prev += 3;
|
||||
row += 3;
|
||||
rb -= 3;
|
||||
}
|
||||
}
|
||||
|
||||
void png_read_filter_row_paeth4_sse2(png_row_infop row_info, png_bytep row,
|
||||
png_const_bytep prev)
|
||||
{
|
||||
/* Paeth tries to predict pixel d using the pixel to the left of it, a,
|
||||
* and two pixels from the previous row, b and c:
|
||||
* prev: c b
|
||||
* row: a d
|
||||
* The Paeth function predicts d to be whichever of a, b, or c is nearest to
|
||||
* p=a+b-c.
|
||||
*
|
||||
* The first pixel has no left context, and so uses an Up filter, p = b.
|
||||
* This works naturally with our main loop's p = a+b-c if we force a and c
|
||||
* to zero.
|
||||
* Here we zero b and d, which become c and a respectively at the start of
|
||||
* the loop.
|
||||
*/
|
||||
size_t rb;
|
||||
const __m128i zero = _mm_setzero_si128();
|
||||
__m128i pa,pb,pc,smallest,nearest;
|
||||
__m128i c, b = zero,
|
||||
a, d = zero;
|
||||
|
||||
png_debug(1, "in png_read_filter_row_paeth4_sse2");
|
||||
|
||||
rb = row_info->rowbytes+4;
|
||||
while (rb > 4) {
|
||||
/* It's easiest to do this math (particularly, deal with pc) with 16-bit
|
||||
* intermediates.
|
||||
*/
|
||||
c = b; b = _mm_unpacklo_epi8(load4(prev), zero);
|
||||
a = d; d = _mm_unpacklo_epi8(load4(row ), zero);
|
||||
|
||||
/* (p-a) == (a+b-c - a) == (b-c) */
|
||||
pa = _mm_sub_epi16(b,c);
|
||||
|
||||
/* (p-b) == (a+b-c - b) == (a-c) */
|
||||
pb = _mm_sub_epi16(a,c);
|
||||
|
||||
/* (p-c) == (a+b-c - c) == (a+b-c-c) == (b-c)+(a-c) */
|
||||
pc = _mm_add_epi16(pa,pb);
|
||||
|
||||
pa = abs_i16(pa); /* |p-a| */
|
||||
pb = abs_i16(pb); /* |p-b| */
|
||||
pc = abs_i16(pc); /* |p-c| */
|
||||
|
||||
smallest = _mm_min_epi16(pc, _mm_min_epi16(pa, pb));
|
||||
|
||||
/* Paeth breaks ties favoring a over b over c. */
|
||||
nearest = if_then_else(_mm_cmpeq_epi16(smallest, pa), a,
|
||||
if_then_else(_mm_cmpeq_epi16(smallest, pb), b,
|
||||
c));
|
||||
|
||||
/* Note `_epi8`: we need addition to wrap modulo 255. */
|
||||
d = _mm_add_epi8(d, nearest);
|
||||
store4(row, _mm_packus_epi16(d,d));
|
||||
|
||||
prev += 4;
|
||||
row += 4;
|
||||
rb -= 4;
|
||||
}
|
||||
}
|
||||
|
||||
#endif /* PNG_INTEL_SSE_IMPLEMENTATION > 0 */
|
||||
#endif /* READ */
|
||||
Vendored
-52
@@ -1,52 +0,0 @@
|
||||
|
||||
/* intel_init.c - SSE2 optimized filter functions
|
||||
*
|
||||
* Copyright (c) 2018 Cosmin Truta
|
||||
* Copyright (c) 2016-2017 Glenn Randers-Pehrson
|
||||
* Written by Mike Klein and Matt Sarett, Google, Inc.
|
||||
* Derived from arm/arm_init.c
|
||||
*
|
||||
* This code is released under the libpng license.
|
||||
* For conditions of distribution and use, see the disclaimer
|
||||
* and license in png.h
|
||||
*/
|
||||
|
||||
#include "../pngpriv.h"
|
||||
|
||||
#ifdef PNG_READ_SUPPORTED
|
||||
#if PNG_INTEL_SSE_IMPLEMENTATION > 0
|
||||
|
||||
void
|
||||
png_init_filter_functions_sse2(png_structp pp, unsigned int bpp)
|
||||
{
|
||||
/* The techniques used to implement each of these filters in SSE operate on
|
||||
* one pixel at a time.
|
||||
* So they generally speed up 3bpp images about 3x, 4bpp images about 4x.
|
||||
* They can scale up to 6 and 8 bpp images and down to 2 bpp images,
|
||||
* but they'd not likely have any benefit for 1bpp images.
|
||||
* Most of these can be implemented using only MMX and 64-bit registers,
|
||||
* but they end up a bit slower than using the equally-ubiquitous SSE2.
|
||||
*/
|
||||
png_debug(1, "in png_init_filter_functions_sse2");
|
||||
if (bpp == 3)
|
||||
{
|
||||
pp->read_filter[PNG_FILTER_VALUE_SUB-1] = png_read_filter_row_sub3_sse2;
|
||||
pp->read_filter[PNG_FILTER_VALUE_AVG-1] = png_read_filter_row_avg3_sse2;
|
||||
pp->read_filter[PNG_FILTER_VALUE_PAETH-1] =
|
||||
png_read_filter_row_paeth3_sse2;
|
||||
}
|
||||
else if (bpp == 4)
|
||||
{
|
||||
pp->read_filter[PNG_FILTER_VALUE_SUB-1] = png_read_filter_row_sub4_sse2;
|
||||
pp->read_filter[PNG_FILTER_VALUE_AVG-1] = png_read_filter_row_avg4_sse2;
|
||||
pp->read_filter[PNG_FILTER_VALUE_PAETH-1] =
|
||||
png_read_filter_row_paeth4_sse2;
|
||||
}
|
||||
|
||||
/* No need optimize PNG_FILTER_VALUE_UP. The compiler should
|
||||
* autovectorize.
|
||||
*/
|
||||
}
|
||||
|
||||
#endif /* PNG_INTEL_SSE_IMPLEMENTATION > 0 */
|
||||
#endif /* PNG_READ_SUPPORTED */
|
||||
-808
File diff suppressed because it is too large
Load Diff
Vendored
-130
@@ -1,130 +0,0 @@
|
||||
|
||||
/* mips_init.c - MSA optimised filter functions
|
||||
*
|
||||
* Copyright (c) 2018 Cosmin Truta
|
||||
* Copyright (c) 2016 Glenn Randers-Pehrson
|
||||
* Written by Mandar Sahastrabuddhe, 2016.
|
||||
*
|
||||
* This code is released under the libpng license.
|
||||
* For conditions of distribution and use, see the disclaimer
|
||||
* and license in png.h
|
||||
*/
|
||||
|
||||
/* Below, after checking __linux__, various non-C90 POSIX 1003.1 functions are
|
||||
* called.
|
||||
*/
|
||||
#define _POSIX_SOURCE 1
|
||||
|
||||
#include <stdio.h>
|
||||
#include "../pngpriv.h"
|
||||
|
||||
#ifdef PNG_READ_SUPPORTED
|
||||
|
||||
#if PNG_MIPS_MSA_OPT > 0
|
||||
#ifdef PNG_MIPS_MSA_CHECK_SUPPORTED /* Do run-time checks */
|
||||
/* WARNING: it is strongly recommended that you do not build libpng with
|
||||
* run-time checks for CPU features if at all possible. In the case of the MIPS
|
||||
* MSA instructions there is no processor-specific way of detecting the
|
||||
* presence of the required support, therefore run-time detection is extremely
|
||||
* OS specific.
|
||||
*
|
||||
* You may set the macro PNG_MIPS_MSA_FILE to the file name of file containing
|
||||
* a fragment of C source code which defines the png_have_msa function. There
|
||||
* are a number of implementations in contrib/mips-msa, but the only one that
|
||||
* has partial support is contrib/mips-msa/linux.c - a generic Linux
|
||||
* implementation which reads /proc/cpufino.
|
||||
*/
|
||||
#ifndef PNG_MIPS_MSA_FILE
|
||||
# ifdef __linux__
|
||||
# define PNG_MIPS_MSA_FILE "contrib/mips-msa/linux.c"
|
||||
# endif
|
||||
#endif
|
||||
|
||||
#ifdef PNG_MIPS_MSA_FILE
|
||||
|
||||
#include <signal.h> /* for sig_atomic_t */
|
||||
static int png_have_msa(png_structp png_ptr);
|
||||
#include PNG_MIPS_MSA_FILE
|
||||
|
||||
#else /* PNG_MIPS_MSA_FILE */
|
||||
# error "PNG_MIPS_MSA_FILE undefined: no support for run-time MIPS MSA checks"
|
||||
#endif /* PNG_MIPS_MSA_FILE */
|
||||
#endif /* PNG_MIPS_MSA_CHECK_SUPPORTED */
|
||||
|
||||
#ifndef PNG_ALIGNED_MEMORY_SUPPORTED
|
||||
# error "ALIGNED_MEMORY is required; set: -DPNG_ALIGNED_MEMORY_SUPPORTED"
|
||||
#endif
|
||||
|
||||
void
|
||||
png_init_filter_functions_msa(png_structp pp, unsigned int bpp)
|
||||
{
|
||||
/* The switch statement is compiled in for MIPS_MSA_API, the call to
|
||||
* png_have_msa is compiled in for MIPS_MSA_CHECK. If both are defined
|
||||
* the check is only performed if the API has not set the MSA option on
|
||||
* or off explicitly. In this case the check controls what happens.
|
||||
*/
|
||||
|
||||
#ifdef PNG_MIPS_MSA_API_SUPPORTED
|
||||
switch ((pp->options >> PNG_MIPS_MSA) & 3)
|
||||
{
|
||||
case PNG_OPTION_UNSET:
|
||||
/* Allow the run-time check to execute if it has been enabled -
|
||||
* thus both API and CHECK can be turned on. If it isn't supported
|
||||
* this case will fall through to the 'default' below, which just
|
||||
* returns.
|
||||
*/
|
||||
#endif /* PNG_MIPS_MSA_API_SUPPORTED */
|
||||
#ifdef PNG_MIPS_MSA_CHECK_SUPPORTED
|
||||
{
|
||||
static volatile sig_atomic_t no_msa = -1; /* not checked */
|
||||
|
||||
if (no_msa < 0)
|
||||
no_msa = !png_have_msa(pp);
|
||||
|
||||
if (no_msa)
|
||||
return;
|
||||
}
|
||||
#ifdef PNG_MIPS_MSA_API_SUPPORTED
|
||||
break;
|
||||
#endif
|
||||
#endif /* PNG_MIPS_MSA_CHECK_SUPPORTED */
|
||||
|
||||
#ifdef PNG_MIPS_MSA_API_SUPPORTED
|
||||
default: /* OFF or INVALID */
|
||||
return;
|
||||
|
||||
case PNG_OPTION_ON:
|
||||
/* Option turned on */
|
||||
break;
|
||||
}
|
||||
#endif
|
||||
|
||||
/* IMPORTANT: any new external functions used here must be declared using
|
||||
* PNG_INTERNAL_FUNCTION in ../pngpriv.h. This is required so that the
|
||||
* 'prefix' option to configure works:
|
||||
*
|
||||
* ./configure --with-libpng-prefix=foobar_
|
||||
*
|
||||
* Verify you have got this right by running the above command, doing a build
|
||||
* and examining pngprefix.h; it must contain a #define for every external
|
||||
* function you add. (Notice that this happens automatically for the
|
||||
* initialization function.)
|
||||
*/
|
||||
pp->read_filter[PNG_FILTER_VALUE_UP-1] = png_read_filter_row_up_msa;
|
||||
|
||||
if (bpp == 3)
|
||||
{
|
||||
pp->read_filter[PNG_FILTER_VALUE_SUB-1] = png_read_filter_row_sub3_msa;
|
||||
pp->read_filter[PNG_FILTER_VALUE_AVG-1] = png_read_filter_row_avg3_msa;
|
||||
pp->read_filter[PNG_FILTER_VALUE_PAETH-1] = png_read_filter_row_paeth3_msa;
|
||||
}
|
||||
|
||||
else if (bpp == 4)
|
||||
{
|
||||
pp->read_filter[PNG_FILTER_VALUE_SUB-1] = png_read_filter_row_sub4_msa;
|
||||
pp->read_filter[PNG_FILTER_VALUE_AVG-1] = png_read_filter_row_avg4_msa;
|
||||
pp->read_filter[PNG_FILTER_VALUE_PAETH-1] = png_read_filter_row_paeth4_msa;
|
||||
}
|
||||
}
|
||||
#endif /* PNG_MIPS_MSA_OPT > 0 */
|
||||
#endif /* READ */
|
||||
Vendored
-4607
File diff suppressed because it is too large
Load Diff
Vendored
-3247
File diff suppressed because it is too large
Load Diff
Vendored
-50
@@ -1,50 +0,0 @@
|
||||
<?xml version="1.0" encoding="utf-8"?>
|
||||
<Project>
|
||||
<Import Project="..\..\..\Source\VSProps\Base.Macros.props" />
|
||||
<Import Project="$(VSPropsDir)Base.Targets.props" />
|
||||
<PropertyGroup Label="Globals">
|
||||
<ProjectGuid>{4C9F135B-A85E-430C-BAD4-4C67EF5FC12C}</ProjectGuid>
|
||||
</PropertyGroup>
|
||||
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.Default.props" />
|
||||
<Import Project="$(VSPropsDir)Configuration.StaticLibrary.props" />
|
||||
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
|
||||
<ImportGroup Label="ExtensionSettings" />
|
||||
<ImportGroup Label="PropertySheets">
|
||||
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
|
||||
<Import Project="$(VSPropsDir)Base.props" />
|
||||
<Import Project="$(VSPropsDir)ClDisableAllWarnings.props" />
|
||||
</ImportGroup>
|
||||
<PropertyGroup Label="UserMacros" />
|
||||
<ItemGroup>
|
||||
<Text Include="..\CMakeLists.txt" />
|
||||
</ItemGroup>
|
||||
<ItemGroup>
|
||||
<ClCompile Include="..\png.c" />
|
||||
<ClCompile Include="..\pngerror.c" />
|
||||
<ClCompile Include="..\pngget.c" />
|
||||
<ClCompile Include="..\pngmem.c" />
|
||||
<ClCompile Include="..\pngpread.c" />
|
||||
<ClCompile Include="..\pngread.c" />
|
||||
<ClCompile Include="..\pngrio.c" />
|
||||
<ClCompile Include="..\pngrtran.c" />
|
||||
<ClCompile Include="..\pngrutil.c" />
|
||||
<ClCompile Include="..\pngset.c" />
|
||||
<ClCompile Include="..\pngtrans.c" />
|
||||
<ClCompile Include="..\pngwio.c" />
|
||||
<ClCompile Include="..\pngwrite.c" />
|
||||
<ClCompile Include="..\pngwtran.c" />
|
||||
<ClCompile Include="..\pngwutil.c" />
|
||||
</ItemGroup>
|
||||
<ItemGroup>
|
||||
<ClInclude Include="..\png.h" />
|
||||
<ClInclude Include="..\pngconf.h" />
|
||||
<ClInclude Include="..\pngdebug.h" />
|
||||
<ClInclude Include="..\pnginfo.h" />
|
||||
<ClInclude Include="..\pnglibconf.h" />
|
||||
<ClInclude Include="..\pngpriv.h" />
|
||||
<ClInclude Include="..\pngstruct.h" />
|
||||
</ItemGroup>
|
||||
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
|
||||
<ImportGroup Label="ExtensionTargets">
|
||||
</ImportGroup>
|
||||
</Project>
|
||||
Vendored
-623
File diff suppressed because it is too large
Load Diff
Vendored
-153
@@ -1,153 +0,0 @@
|
||||
|
||||
/* pngdebug.h - Debugging macros for libpng, also used in pngtest.c
|
||||
*
|
||||
* Copyright (c) 2018 Cosmin Truta
|
||||
* Copyright (c) 1998-2002,2004,2006-2013 Glenn Randers-Pehrson
|
||||
* Copyright (c) 1996-1997 Andreas Dilger
|
||||
* Copyright (c) 1995-1996 Guy Eric Schalnat, Group 42, Inc.
|
||||
*
|
||||
* This code is released under the libpng license.
|
||||
* For conditions of distribution and use, see the disclaimer
|
||||
* and license in png.h
|
||||
*/
|
||||
|
||||
/* Define PNG_DEBUG at compile time for debugging information. Higher
|
||||
* numbers for PNG_DEBUG mean more debugging information. This has
|
||||
* only been added since version 0.95 so it is not implemented throughout
|
||||
* libpng yet, but more support will be added as needed.
|
||||
*
|
||||
* png_debug[1-2]?(level, message ,arg{0-2})
|
||||
* Expands to a statement (either a simple expression or a compound
|
||||
* do..while(0) statement) that outputs a message with parameter
|
||||
* substitution if PNG_DEBUG is defined to 2 or more. If PNG_DEBUG
|
||||
* is undefined, 0 or 1 every png_debug expands to a simple expression
|
||||
* (actually ((void)0)).
|
||||
*
|
||||
* level: level of detail of message, starting at 0. A level 'n'
|
||||
* message is preceded by 'n' 3-space indentations (not implemented
|
||||
* on Microsoft compilers unless PNG_DEBUG_FILE is also
|
||||
* defined, to allow debug DLL compilation with no standard IO).
|
||||
* message: a printf(3) style text string. A trailing '\n' is added
|
||||
* to the message.
|
||||
* arg: 0 to 2 arguments for printf(3) style substitution in message.
|
||||
*/
|
||||
#ifndef PNGDEBUG_H
|
||||
#define PNGDEBUG_H
|
||||
/* These settings control the formatting of messages in png.c and pngerror.c */
|
||||
/* Moved to pngdebug.h at 1.5.0 */
|
||||
# ifndef PNG_LITERAL_SHARP
|
||||
# define PNG_LITERAL_SHARP 0x23
|
||||
# endif
|
||||
# ifndef PNG_LITERAL_LEFT_SQUARE_BRACKET
|
||||
# define PNG_LITERAL_LEFT_SQUARE_BRACKET 0x5b
|
||||
# endif
|
||||
# ifndef PNG_LITERAL_RIGHT_SQUARE_BRACKET
|
||||
# define PNG_LITERAL_RIGHT_SQUARE_BRACKET 0x5d
|
||||
# endif
|
||||
# ifndef PNG_STRING_NEWLINE
|
||||
# define PNG_STRING_NEWLINE "\n"
|
||||
# endif
|
||||
|
||||
#ifdef PNG_DEBUG
|
||||
# if (PNG_DEBUG > 0)
|
||||
# if !defined(PNG_DEBUG_FILE) && defined(_MSC_VER)
|
||||
# include <crtdbg.h>
|
||||
# if (PNG_DEBUG > 1)
|
||||
# ifndef _DEBUG
|
||||
# define _DEBUG
|
||||
# endif
|
||||
# ifndef png_debug
|
||||
# define png_debug(l,m) _RPT0(_CRT_WARN,m PNG_STRING_NEWLINE)
|
||||
# endif
|
||||
# ifndef png_debug1
|
||||
# define png_debug1(l,m,p1) _RPT1(_CRT_WARN,m PNG_STRING_NEWLINE,p1)
|
||||
# endif
|
||||
# ifndef png_debug2
|
||||
# define png_debug2(l,m,p1,p2) \
|
||||
_RPT2(_CRT_WARN,m PNG_STRING_NEWLINE,p1,p2)
|
||||
# endif
|
||||
# endif
|
||||
# else /* PNG_DEBUG_FILE || !_MSC_VER */
|
||||
# ifndef PNG_STDIO_SUPPORTED
|
||||
# include <stdio.h> /* not included yet */
|
||||
# endif
|
||||
# ifndef PNG_DEBUG_FILE
|
||||
# define PNG_DEBUG_FILE stderr
|
||||
# endif /* PNG_DEBUG_FILE */
|
||||
|
||||
# if (PNG_DEBUG > 1)
|
||||
# ifdef __STDC__
|
||||
# ifndef png_debug
|
||||
# define png_debug(l,m) \
|
||||
do { \
|
||||
int num_tabs=l; \
|
||||
fprintf(PNG_DEBUG_FILE,"%s" m PNG_STRING_NEWLINE,(num_tabs==1 ? " " : \
|
||||
(num_tabs==2 ? " " : (num_tabs>2 ? " " : "")))); \
|
||||
} while (0)
|
||||
# endif
|
||||
# ifndef png_debug1
|
||||
# define png_debug1(l,m,p1) \
|
||||
do { \
|
||||
int num_tabs=l; \
|
||||
fprintf(PNG_DEBUG_FILE,"%s" m PNG_STRING_NEWLINE,(num_tabs==1 ? " " : \
|
||||
(num_tabs==2 ? " " : (num_tabs>2 ? " " : ""))),p1); \
|
||||
} while (0)
|
||||
# endif
|
||||
# ifndef png_debug2
|
||||
# define png_debug2(l,m,p1,p2) \
|
||||
do { \
|
||||
int num_tabs=l; \
|
||||
fprintf(PNG_DEBUG_FILE,"%s" m PNG_STRING_NEWLINE,(num_tabs==1 ? " " : \
|
||||
(num_tabs==2 ? " " : (num_tabs>2 ? " " : ""))),p1,p2);\
|
||||
} while (0)
|
||||
# endif
|
||||
# else /* __STDC __ */
|
||||
# ifndef png_debug
|
||||
# define png_debug(l,m) \
|
||||
do { \
|
||||
int num_tabs=l; \
|
||||
char format[256]; \
|
||||
snprintf(format,256,"%s%s%s",(num_tabs==1 ? "\t" : \
|
||||
(num_tabs==2 ? "\t\t":(num_tabs>2 ? "\t\t\t":""))), \
|
||||
m,PNG_STRING_NEWLINE); \
|
||||
fprintf(PNG_DEBUG_FILE,format); \
|
||||
} while (0)
|
||||
# endif
|
||||
# ifndef png_debug1
|
||||
# define png_debug1(l,m,p1) \
|
||||
do { \
|
||||
int num_tabs=l; \
|
||||
char format[256]; \
|
||||
snprintf(format,256,"%s%s%s",(num_tabs==1 ? "\t" : \
|
||||
(num_tabs==2 ? "\t\t":(num_tabs>2 ? "\t\t\t":""))), \
|
||||
m,PNG_STRING_NEWLINE); \
|
||||
fprintf(PNG_DEBUG_FILE,format,p1); \
|
||||
} while (0)
|
||||
# endif
|
||||
# ifndef png_debug2
|
||||
# define png_debug2(l,m,p1,p2) \
|
||||
do { \
|
||||
int num_tabs=l; \
|
||||
char format[256]; \
|
||||
snprintf(format,256,"%s%s%s",(num_tabs==1 ? "\t" : \
|
||||
(num_tabs==2 ? "\t\t":(num_tabs>2 ? "\t\t\t":""))), \
|
||||
m,PNG_STRING_NEWLINE); \
|
||||
fprintf(PNG_DEBUG_FILE,format,p1,p2); \
|
||||
} while (0)
|
||||
# endif
|
||||
# endif /* __STDC __ */
|
||||
# endif /* (PNG_DEBUG > 1) */
|
||||
|
||||
# endif /* _MSC_VER */
|
||||
# endif /* (PNG_DEBUG > 0) */
|
||||
#endif /* PNG_DEBUG */
|
||||
#ifndef png_debug
|
||||
# define png_debug(l, m) ((void)0)
|
||||
#endif
|
||||
#ifndef png_debug1
|
||||
# define png_debug1(l, m, p1) ((void)0)
|
||||
#endif
|
||||
#ifndef png_debug2
|
||||
# define png_debug2(l, m, p1, p2) ((void)0)
|
||||
#endif
|
||||
#endif /* PNGDEBUG_H */
|
||||
Vendored
-963
File diff suppressed because it is too large
Load Diff
Vendored
-1249
File diff suppressed because it is too large
Load Diff
Some files were not shown because too many files have changed in this diff Show More
Reference in New Issue
Block a user