| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
Move SkColorData to src/core We don't want these to leak into the public API. Change-Id: Ibdfad8ffcb1764affc84696fc07ebac33c2bcdf4 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/957556 Commit-Queue: Greg Daniel <egdaniel@google.com> Auto-Submit: Kaylee Lubick <kjlubick@google.com> Reviewed-by: Greg Daniel <egdaniel@google.com> | 1 年前 | |
Reland "Reland "Restore affine bitmap procs"" This is a reland of commit d9ebf2e8f34d58b38df1560b26729c2c3d8023d7 Includes a fix for a long-standing bug (no alpha multiplication) in the non-filtering version of this function. Original change's description: > Reland "Restore affine bitmap procs" > > This is a reland of commit fa31a916cbb8a09fc7438c2b3c8fdf378073ced1 > > Guarded to rebaseline chromium with SK_IGNORE_FAST_ROTATED_BLITTER > > Original change's description: > > Restore affine bitmap procs > > > > This effectively reverts http://review.skia.org/684577 > > When decoding images with EXIF rotation, we used to hit > > this fast-path. When it was removed, some benchmarks > > regressed significantly. > > > > Bug: b/329003512 > > Change-Id: I941d7fb3e9760711cae643ea681218e1aeee4c64 > > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/841157 > > Commit-Queue: Brian Osman <brianosman@google.com> > > Reviewed-by: John Stiles <johnstiles@google.com> > > Bug: b/329003512 > Change-Id: I2eec3d71f6347c60f7dbbba3f6987cd6ab19c776 > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/844376 > Reviewed-by: John Stiles <johnstiles@google.com> > Commit-Queue: Brian Osman <brianosman@google.com> Bug: b/329003512 Change-Id: Ia664dd784e0f0b52b71cdf49287b7802a424f16e Reviewed-on: https://skia-review.googlesource.com/c/skia/+/844816 Reviewed-by: John Stiles <johnstiles@google.com> Auto-Submit: Brian Osman <brianosman@google.com> Commit-Queue: Brian Osman <brianosman@google.com> | 2 年前 | |
Make ARGB32 blitting more consistent on loongarch codepath. This patch is ported from [1] and it can fix SkARGB32Blitter_MaskAndAntiHDrawTheSame test failed. [1] https://skia-review.googlesource.com/c/skia/+/953516 Bug: b/397749778 Change-Id: Ie74e1b3e3cf5b7c47083376355296dd180a36610 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1006816 Reviewed-by: Florin Malita <fmalita@google.com> Reviewed-by: Kaylee Lubick <kjlubick@google.com> Commit-Queue: Florin Malita <fmalita@google.com> Commit-Queue: Kaylee Lubick <kjlubick@google.com> | 1 年前 | |
Remove legacy mismatched blit code This landed in Chromium in https://crrev.com/c/6328778 Change-Id: I83d8b5a62eab3db7e6ea79e09ec456e40b3595ad Bug: b/397749778 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/960516 Reviewed-by: Florin Malita <fmalita@google.com> Commit-Queue: Kaylee Lubick <kjlubick@google.com> | 1 年前 | |
Split out memset opts into separate cpp files These used to be in SkUtils_opts. I also renamed everything so that they're declared in SkMemset.h, and defined in SkMemset_opts.h. Bug: b/40045064 Bug: b/40045066 Change-Id: Idefa5b8d5f59595ddce907742f50e02174e42964 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/740638 Commit-Queue: Brian Osman <brianosman@google.com> Reviewed-by: Kevin Lubick <kjlubick@google.com> | 2 年前 | |
Split out BitmapProcState opts into separate cpp files This CL includes two new things: 1) Slicing SkOpts by functionality, in addition to architecture. 2) Enabling architectures within the translation unit, rather than at the command line. The first attempt at #2 failed, because of a clang bug. However, that bug only impacts code that passes AVX vectors to functions. That pattern only happens in RasterPipeline, so all other SkOpts can be peeled off this way. I think it's possible to go further, and fold all the architecture specializations of a particular module (eg BitmapProcState) into a single cpp file. I started down that path, but realized the skvx includes would get pretty tricky. So I backed up and got this version working first. This is still an improvement: - Clients don't need to know about special compile flags for any of the newly added "opts" cpp files. They just get added to the soruce lists like everything else. - The initialization happens automatically, because I found a good spot to trigger it when we're choosing procs. Bug: skia:13983 Bug: skia:14355 Change-Id: Ibdcff5fbef9df25416cfd90eb1680153e54e7661 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/727257 Reviewed-by: Kevin Lubick <kjlubick@google.com> Commit-Queue: Brian Osman <brianosman@google.com> | 2 年前 | |
Add workaround for compiling with c++20 on Windows Without this change, we see the following cryptic error In file included from ../../../../../../skia/src/core/SkBitmapProcState_opts_ssse3.cpp:19: In file included from ../../../../../../skia\src/opts/SkOpts_SetTarget.h:68: In file included from C:\b\s\w\ir\clang_win\lib\clang\18\include\immintrin.h:603: C:\b\s\w\ir\clang_win\lib\clang\18\include\rtmintrin.h(36,1): error: attribute 'target' multiversioning cannot be combined with attribute 'always_inline' 36 | _xend(void) | ^ C:\b\s\w\ir\clang_win\lib\clang\18\include\intrin.h(167,6): note: function multiversioning caused by this declaration 167 | void _xend(void); | ^ This apparently [1] is because two incompatible headers are being included that have inconsistent attributes. This has been fixed in newer version of LLVM, but this should work around it on older versions. [1] https://github.com/llvm/llvm-project/issues/95133 Change-Id: I539c4d30941736d64ad1ae69e2d0ae2a6c2e155d Bug: b/311238827 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/970996 Reviewed-by: Eric Boren <borenet@google.com> Commit-Queue: Kaylee Lubick <kjlubick@google.com> | 1 年前 | |
Split out Swizzler opts into separate cpp files Bug: b/40045064 Bug: b/40045066 Change-Id: I9fd6e0818b1c3aff4176af6d5a84e8e90b9a38ec Reviewed-on: https://skia-review.googlesource.com/c/skia/+/740296 Reviewed-by: Kevin Lubick <kjlubick@google.com> Commit-Queue: Brian Osman <brianosman@google.com> | 2 年前 | |
[Loong64] Added binary compatibility support. Currently, the lasx optimization branch is used by default on the loong64 platform. Some loongarch processors(such as 2k2000) that do not support lasx instruction sets will crashed. Change-Id: I6a86e1916f1093818f117de39ada5c3be57c4d4a Reviewed-on: https://skia-review.googlesource.com/c/skia/+/873776 Reviewed-by: John Stiles <johnstiles@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Commit-Queue: Brian Osman <brianosman@google.com> | 1 年前 | |
Reland "Added AVX512 support for SkRasterPipeline_opts.h" This is a reland of commit ca883b068bad7475f5de62d1e4a8e5402fb3913f Original change's description: > Added AVX512 support for SkRasterPipeline_opts.h > > Change-Id: Ic7138faecd3887090288774373a89507da7fc87d > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/801296 > Reviewed-by: Brian Osman <brianosman@google.com> > Commit-Queue: Brian Osman <brianosman@google.com> > Reviewed-by: John Stiles <johnstiles@google.com> > Commit-Queue: John Stiles <johnstiles@google.com> Change-Id: I9a855273c4fcb0c1114261a183080a9834efef69 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/803416 Commit-Queue: Brian Osman <brianosman@google.com> Reviewed-by: John Stiles <johnstiles@google.com> | 2 年前 | |
Add kR16_unorm_SkColorType * Adding support for this color type was requested by ccameron@ while looking into accelerated RGB to YUV conversion. * This CL adds kR16_unorm_SkColorType and subsequently updates all the relevant areas (Caps format tables, testing, raster pipeline ops, etc). Bug: b/452079352 Change-Id: I2629a89c498365a1d41a52e583c0043ac74ba80d Reviewed-on: https://skia-review.googlesource.com/c/skia/+/1075756 Reviewed-by: Florin Malita <fmalita@google.com> Commit-Queue: Nicolette Prevost <nicolettep@google.com> Reviewed-by: Greg Daniel <egdaniel@google.com> | 7 个月前 | |
Move SkColorData to src/core We don't want these to leak into the public API. Change-Id: Ibdfad8ffcb1764affc84696fc07ebac33c2bcdf4 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/957556 Commit-Queue: Greg Daniel <egdaniel@google.com> Auto-Submit: Kaylee Lubick <kjlubick@google.com> Reviewed-by: Greg Daniel <egdaniel@google.com> | 1 年前 |
| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
| 1 年前 | ||
| 2 年前 | ||
| 1 年前 | ||
| 1 年前 | ||
| 2 年前 | ||
| 2 年前 | ||
| 1 年前 | ||
| 2 年前 | ||
| 1 年前 | ||
| 2 年前 | ||
| 7 个月前 | ||
| 1 年前 |