Frank Barchard cf101116c9 Remove initialize to zero on output variables for inline.
Inline that uses temporary variables is currently initializing them
to 0 and passing in as output "+r".
This CL replaces the output constraint to "=&r" for most meaning an
output with early write (before inputs).  This allows the initialize
to zero step to be removed, saving 1 instruction.

BUG=libyuv:580
TESTED=local libyuv build on gcc/linux and try bots
R=harryjin@google.com

Review URL: https://codereview.chromium.org/1895743008 .
2016-04-18 16:24:26 -07:00
..
compare_common.cc xmmword cast for clang 2015-08-18 11:13:12 -07:00
compare_gcc.cc nolint removed 2015-08-31 10:52:13 -07:00
compare_neon64.cc xmmword cast for clang 2015-08-18 11:13:12 -07:00
compare_neon.cc xmmword cast for clang 2015-08-18 11:13:12 -07:00
compare_win.cc xmmword cast for clang 2015-08-18 11:13:12 -07:00
compare.cc xmmword cast for clang 2015-08-18 11:13:12 -07:00
convert_argb.cc rename MIPS_DSPR2 to DSPR2 for consistency 2016-02-05 14:49:54 -08:00
convert_from_argb.cc refactor ARGBToI422 using ARGBToI420 internally 2016-01-12 17:05:49 -08:00
convert_from.cc rename MIPS_DSPR2 to DSPR2 for consistency 2016-02-05 14:49:54 -08:00
convert_jpeg.cc libyuv::MJPGToI420() and libyuv::MJPGToARGB() return failure if callback to JPeg fails. 2014-01-28 03:08:59 +00:00
convert_to_argb.cc Fix temporary stride for ConvertToARGB with rotation. 2016-04-11 15:21:04 -07:00
convert_to_i420.cc Remove Q420 fourcc support. 2015-02-11 18:20:54 +00:00
convert.cc rename MIPS_DSPR2 to DSPR2 for consistency 2016-02-05 14:49:54 -08:00
cpu_id.cc add if defined(_MSC_FULL_VER) for NaCL 2016-04-01 17:48:23 -07:00
mjpeg_decoder.cc nolint removed 2015-08-31 10:52:13 -07:00
mjpeg_validate.cc validate scan EOI from end for better coverage 2015-09-14 10:58:51 -07:00
planar_functions.cc Fix stride bug for msan on I420Interpolate. 2016-04-08 15:58:53 -07:00
rotate_any.cc rename MIPS_DSPR2 to DSPR2 for consistency 2016-02-05 14:49:54 -08:00
rotate_argb.cc rotate include and proto cleanup 2015-07-22 18:09:04 -07:00
rotate_common.cc rotate include and proto cleanup 2015-07-22 18:09:04 -07:00
rotate_gcc.cc use visual c 32 bit code for clangcl 2015-08-11 10:10:45 -07:00
rotate_mips.cc rename MIPS_DSPR2 to DSPR2 for consistency 2016-02-05 14:49:54 -08:00
rotate_neon64.cc Remove initialize to zero on output variables for inline. 2016-04-18 16:24:26 -07:00
rotate_neon.cc Remove initialize to zero on output variables for inline. 2016-04-18 16:24:26 -07:00
rotate_win.cc use visual c 32 bit code for clangcl 2015-08-11 10:10:45 -07:00
rotate.cc rename MIPS_DSPR2 to DSPR2 for consistency 2016-02-05 14:49:54 -08:00
row_any.cc add perf data files to ignores 2016-02-17 21:37:09 -08:00
row_common.cc Fix stride bug for msan on I420Interpolate. 2016-04-08 15:58:53 -07:00
row_gcc.cc Remove initialize to zero on output variables for inline. 2016-04-18 16:24:26 -07:00
row_mips.cc Remove initialize to zero on output variables for inline. 2016-04-18 16:24:26 -07:00
row_neon64.cc Port ARGBToRGB565 from aarch64 neon to 32 bit 2016-02-29 12:22:25 -08:00
row_neon.cc Port ARGBToRGB565 from aarch64 neon to 32 bit 2016-02-29 12:22:25 -08:00
row_win.cc Port ARGBToJ420 to AVX2 2016-02-17 23:16:39 -08:00
scale_any.cc Odd width variation of scale down by 2 for subsampling 2016-01-06 15:12:17 -08:00
scale_argb.cc rename MIPS_DSPR2 to DSPR2 for consistency 2016-02-05 14:49:54 -08:00
scale_common.cc rename MIPS_DSPR2 to DSPR2 for consistency 2016-02-05 14:49:54 -08:00
scale_gcc.cc Remove initialize to zero on output variables for inline. 2016-04-18 16:24:26 -07:00
scale_mips.cc rename MIPS_DSPR2 to DSPR2 for consistency 2016-02-05 14:49:54 -08:00
scale_neon64.cc Remove initialize to zero on output variables for inline. 2016-04-18 16:24:26 -07:00
scale_neon.cc Remove initialize to zero on output variables for inline. 2016-04-18 16:24:26 -07:00
scale_win.cc Remove use_sysroot=0 2016-01-11 14:57:50 -08:00
scale.cc rename MIPS_DSPR2 to DSPR2 for consistency 2016-02-05 14:49:54 -08:00
video_common.cc Remove bayer format support from libyuv. This format is very rare and used on legacy hardware. Its not well optimized and has bugs related to odd widths. Removing the format will allow tests to pass under more circumstances, run faster and allow focus on higher priority quality and performance issues. 2015-02-09 19:58:19 +00:00