summaryrefslogtreecommitdiffstats
path: root/gpu
Commit message (Collapse)AuthorAgeFilesLines
* Modify the base::JSONReader interface to take a set of options rather than a ↵rsesek@chromium.org2012-04-111-1/+1
| | | | | | | | | | | | | | boolean flag. This is just prep work for a large CL that rewrites the JSON parser. Splitting out this interface change will make it easier to review the other CL. BUG=121469 TEST=No functional change. Review URL: https://chromiumcodereview.appspot.com/9960077 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@131680 0039d316-1c4b-4281-b951-d872f2087c98
* Add GL_EXT_unpack_subimage support to command buffer client code.bsalomon@google.com2012-04-109-225/+512
| | | | | | | BUG=121780 Review URL: https://chromiumcodereview.appspot.com/10012057 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@131579 0039d316-1c4b-4281-b951-d872f2087c98
* Revert 131177 - aura: Change shared context to be offscreen for armmazda@chromium.org2012-04-106-35/+28
| | | | | | | | | | | | | | | | | | | | | | | | | This change causes Chrome OS to get stuck in the login screen on Lumpy and Stumpy, so it is reverted for now. BUG=chrome-os-partner:8831 TEST=Manually checked that I can log in successfully ARM drivers don't like to have 2 EGLSurface on the same Window. So we need to make the UI shared context offscren. This also requires the image transport surface to be an "offscreen" GLSurface (a pbuffer), however the GLES2Decoder needs to be "onscreen" (i.e. it calls SwapBuffers on the surface instead managing the offscreen buffers itself). So this decouples the 2 notions of "offscreen". BUG=chrome-os-partner:8284 TEST=gpu_unittests. Aura chrome --ui-use-gpu-process with poster circle and webgl on tegra, lumpy, desktop GL. Desktop chrome with poster circle and webgl. Review URL: http://codereview.chromium.org/10007034 TBR=piman@chromium.org Review URL: https://chromiumcodereview.appspot.com/10031036 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@131559 0039d316-1c4b-4281-b951-d872f2087c98
* Always write data to new buffer in SimulateAttrib0gman@chromium.org2012-04-101-9/+10
| | | | | | | | | | | | This is to work around linux nvidia driver bug. TEST=asan BUG=118970 Review URL: http://codereview.chromium.org/10019003 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@131538 0039d316-1c4b-4281-b951-d872f2087c98
* Optimize GetBucketContentsgman@chromium.org2012-04-096-109/+285
| | | | | | | | | | TEST=unit tests BUG=122642 Review URL: http://codereview.chromium.org/9999003 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@131470 0039d316-1c4b-4281-b951-d872f2087c98
* Build fix: change the code, forget to change the test.zmo@chromium.org2012-04-091-3/+3
| | | | | | | | | BUG= TEST=gpu_unittests TBR=gman Review URL: https://chromiumcodereview.appspot.com/10031017 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@131469 0039d316-1c4b-4281-b951-d872f2087c98
* Change the glGetShaderPrecisionFormat emulated values.zmo@chromium.org2012-04-091-5/+7
| | | | | | | | | | | The current returned values are wrong, so they fail the WebGL conformance test. BUG= TEST=shader-precision-format.html R=gman Review URL: https://chromiumcodereview.appspot.com/10024040 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@131465 0039d316-1c4b-4281-b951-d872f2087c98
* Hanlde long attribute name mapping in bindAttribLocation.zmo@chromium.org2012-04-075-2/+58
| | | | | | | | | BUG=122466 TEST=gl-bind-attrib-location-long-names-test.html in webgl conformance tests R=gman Review URL: https://chromiumcodereview.appspot.com/10018001 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@131216 0039d316-1c4b-4281-b951-d872f2087c98
* Delete renderbuffers when no longer referencedgman@chromium.org2012-04-064-16/+43
| | | | | | | | | | | | TEST=unit tests BUG=122007 R=apatrick@chromium.org Review URL: http://codereview.chromium.org/10000001 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@131197 0039d316-1c4b-4281-b951-d872f2087c98
* aura: Change shared context to be offscreen for armpiman@chromium.org2012-04-066-28/+35
| | | | | | | | | | | | | | | | | ARM drivers don't like to have 2 EGLSurface on the same Window. So we need to make the UI shared context offscren. This also requires the image transport surface to be an "offscreen" GLSurface (a pbuffer), however the GLES2Decoder needs to be "onscreen" (i.e. it calls SwapBuffers on the surface instead managing the offscreen buffers itself). So this decouples the 2 notions of "offscreen". BUG=chrome-os-partner:8284 TEST=gpu_unittests. Aura chrome --ui-use-gpu-process with poster circle and webgl on tegra, lumpy, desktop GL. Desktop chrome with poster circle and webgl. Review URL: http://codereview.chromium.org/10007034 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@131177 0039d316-1c4b-4281-b951-d872f2087c98
* Allow |msg| to be NULLtbreisacher@chromium.org2012-04-061-1/+2
| | | | | | | | | | | CID=103608 BUG=none TEST=none Review URL: http://codereview.chromium.org/9836135 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@131079 0039d316-1c4b-4281-b951-d872f2087c98
* Delete framebuffers when no longer referenced.gman@chromium.org2012-04-064-21/+56
| | | | | | | | | | TEST=unit tests BUG=122007 Review URL: http://codereview.chromium.org/9995003 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@131051 0039d316-1c4b-4281-b951-d872f2087c98
* Delete Programs when no longer referencedgman@chromium.org2012-04-059-32/+82
| | | | | | | | | | | TEST=unit tests BUG=122007 R=apatrick@chromium.org Review URL: http://codereview.chromium.org/10008019 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@131016 0039d316-1c4b-4281-b951-d872f2087c98
* Don't delete textures until they're no longer referenced.gman@chromium.org2012-04-054-18/+46
| | | | | | | | | | | | TEST=unit tests BUG=122007 R=apatrick@chromium.org Review URL: http://codereview.chromium.org/9994003 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@130996 0039d316-1c4b-4281-b951-d872f2087c98
* Prevent Mipmap generation for compressed textures.gman@chromium.org2012-04-052-2/+5
| | | | | | | | | | TEST=ran WebGL conformance tests BUG=121476 Review URL: http://codereview.chromium.org/9988001 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@130993 0039d316-1c4b-4281-b951-d872f2087c98
* Make buffer not get deleted until the last reference is released.gman@chromium.org2012-04-054-19/+40
| | | | | | | | | | | | TEST=unit tests BUG=122007 R=apatrick@chromium.org Review URL: http://codereview.chromium.org/9982021 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@130865 0039d316-1c4b-4281-b951-d872f2087c98
* Remove TOOLKIT_USES_GTK and consolidate it into TOOLKIT_GTKoshima@chromium.org2012-04-052-6/+6
| | | | | | | | | | | Add gtk dependency to base.gyp for android host_os="linux" case BUG=none TEST=none Review URL: https://chromiumcodereview.appspot.com/9969080 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@130784 0039d316-1c4b-4281-b951-d872f2087c98
* Revert "Revert 130737 - Make contexts actually use ShareGroup for shared ↵gman@chromium.org2012-04-047-228/+152
| | | | | | | | | | | | | | contexts." This reverts commit 429d0c2826fb423769611fc122d2fd1ad49879e9. TEST=ran gpu_unittests BUG=120297 TBR=apatrick@chromium.org Review URL: https://chromiumcodereview.appspot.com/9961022 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@130754 0039d316-1c4b-4281-b951-d872f2087c98
* Revert 130737 - Make contexts actually use ShareGroup for shared contexts.binji@chromium.org2012-04-046-149/+143
| | | | | | | | | | | | | | | | | | | | | | | | | | | The service ALWAYS has bind_generates_resource set to true. In fact we can just remove that flag on the service in another CL. It doesn't need to be enforced on the service. It was only there to help the client find bugs. Now the client and find them itself in a Debug build. I can remove share_resources everywhere. That flag is no longer needed. Because removing that flag touches so many files I'd prefer to do that in another CL. Note: I probably can't check this in until the bugs it uncovers are fixed. Maybe they can be marked as failing. For example the webgl context-lost tests hit an assert. TEST=ran the webgl conformance tests. BUG=120297 Review URL: https://chromiumcodereview.appspot.com/9958038 TBR=gman@chromium.org Review URL: https://chromiumcodereview.appspot.com/9979032 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@130746 0039d316-1c4b-4281-b951-d872f2087c98
* Make contexts actually use ShareGroup for shared contexts.gman@chromium.org2012-04-046-143/+149
| | | | | | | | | | | | | | | | | | | | | | | | The service ALWAYS has bind_generates_resource set to true. In fact we can just remove that flag on the service in another CL. It doesn't need to be enforced on the service. It was only there to help the client find bugs. Now the client and find them itself in a Debug build. I can remove share_resources everywhere. That flag is no longer needed. Because removing that flag touches so many files I'd prefer to do that in another CL. Note: I probably can't check this in until the bugs it uncovers are fixed. Maybe they can be marked as failing. For example the webgl context-lost tests hit an assert. TEST=ran the webgl conformance tests. BUG=120297 Review URL: https://chromiumcodereview.appspot.com/9958038 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@130737 0039d316-1c4b-4281-b951-d872f2087c98
* Revert "Revert 130384 - Broke a few Mac bots"gman@chromium.org2012-04-0410-309/+563
| | | | | | | | | | | | This reverts commit 7bfa1cab6da66e6923316515eaf465afb65e5481. BUG=120297 TEST=run browser_tests on linux TBR=apatrick@chromium.org Review URL: https://chromiumcodereview.appspot.com/9959107 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@130681 0039d316-1c4b-4281-b951-d872f2087c98
* Revert 130384 - Broke a few Mac botsrsleevi@chromium.org2012-04-0310-564/+309
| | | | | | | | | | | | | | | | Revert "Revert "Make ShareGroup thread safe" http://crrev.com/130034." TEST=run DRT in debug on linux BUG=120297 TBR=apatrick@chromium.org Review URL: http://codereview.chromium.org/9958087 TBR=gman@chromium.org Review URL: https://chromiumcodereview.appspot.com/9969115 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@130413 0039d316-1c4b-4281-b951-d872f2087c98
* Revert "Revert "Make ShareGroup thread safe" http://crrev.com/130034."gman@chromium.org2012-04-0310-309/+564
| | | | | | | | | | | TEST=run DRT in debug on linux BUG=120297 TBR=apatrick@chromium.org Review URL: http://codereview.chromium.org/9958087 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@130384 0039d316-1c4b-4281-b951-d872f2087c98
* Limit max texture size to 4096x4096 on intel macgman@chromium.org2012-04-031-0/+2
| | | | | | | | | | TEST=none BUG=121528 Review URL: http://codereview.chromium.org/9950072 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@130284 0039d316-1c4b-4281-b951-d872f2087c98
* PpapiCommandBufferProxy implements CommandBufferProxy.fsamuel@chromium.org2012-04-023-0/+72
| | | | | | | | | | | | | That way once the refactoring of WebGraphicsContext3DCommandBufferImpl is complete, the guest renderers will be able to use the same class. BUG=120664 TEST=manually Review URL: http://codereview.chromium.org/9904005 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@130252 0039d316-1c4b-4281-b951-d872f2087c98
* Revert "Make ShareGroup thread safe" http://crrev.com/130034.apavlov@chromium.org2012-03-3110-564/+309
| | | | | | | | | | | | Virtually all GPU-related layout tests are asserting or timing out on debug builders. This reverts commit c288ca0597d6259501104da2f804b9359cab86bb. TBR=gman@chromium.org Review URL: https://chromiumcodereview.appspot.com/9965048 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@130055 0039d316-1c4b-4281-b951-d872f2087c98
* Make ShareGroup thread safegman@chromium.org2012-03-3110-309/+564
| | | | | | | | | | | | | | In another CL I'll refactor the id handlers AND make the places sharing resources correctly set things up. TEST=none BUG=120297 Review URL: http://codereview.chromium.org/9918027 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@130034 0039d316-1c4b-4281-b951-d872f2087c98
* Revert 129879 - Revert 129819 - Add CHECK on non-generated ids in command ↵zmo@google.com2012-03-302-19/+16
| | | | | | | | | | | | | | | | | | | | | | | buffer. For Chrome and WebGL we require that all ids are generated. This is especially important for the threaded compositor as it may be using ids that are getting deleted and those ids will get auto-reused on bind. BUG=121139 TEST=ran webgl conformance tests Review URL: http://codereview.chromium.org/9903029 TBR=gman@chromium.org Review URL: https://chromiumcodereview.appspot.com/9956023 TBR=zmo@google.com Review URL: https://chromiumcodereview.appspot.com/9968018 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@129886 0039d316-1c4b-4281-b951-d872f2087c98
* Revert 129819 - Add CHECK on non-generated ids in command buffer.zmo@google.com2012-03-302-16/+19
| | | | | | | | | | | | | | | | | | For Chrome and WebGL we require that all ids are generated. This is especially important for the threaded compositor as it may be using ids that are getting deleted and those ids will get auto-reused on bind. BUG=121139 TEST=ran webgl conformance tests Review URL: http://codereview.chromium.org/9903029 TBR=gman@chromium.org Review URL: https://chromiumcodereview.appspot.com/9956023 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@129879 0039d316-1c4b-4281-b951-d872f2087c98
* Fix GL resource sharinggman@chromium.org2012-03-303-95/+111
| | | | | | | | | | | | | | | | | | | | There were 2 bugs. 1) content_gl_context.cc should have been setting bind_generates_resource to false. It should be false for everything but pepper. 2) bind_generates_resource was not getting propogated to the GPU process. TEST=ran in debugger to see that bind_generates_resource is correctly set to false in GPU process BUG=none R=apatrick@chromium.org Review URL: http://codereview.chromium.org/9932002 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@129822 0039d316-1c4b-4281-b951-d872f2087c98
* Add CHECK on non-generated ids in command buffer.gman@chromium.org2012-03-302-19/+16
| | | | | | | | | | | | | | | For Chrome and WebGL we require that all ids are generated. This is especially important for the threaded compositor as it may be using ids that are getting deleted and those ids will get auto-reused on bind. BUG=none TEST=ran webgl conformance tests Review URL: http://codereview.chromium.org/9903029 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@129819 0039d316-1c4b-4281-b951-d872f2087c98
* Add ShareGroup to GLES2Implemenationgman@chromium.org2012-03-295-0/+20
| | | | | | | | | | | | | | | | | | The idea is if you don't pass a ShareGroup it will make one. Otherwise you can do new_gles2_impl = new GLES2Implemetation( ..., some_other_gles2_impl_in_same_process->share_group(), ...); TEST=unit test BUG=120297 TBR=nfullgar@google.com,brettw@chromium.org Review URL: http://codereview.chromium.org/9837124 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@129555 0039d316-1c4b-4281-b951-d872f2087c98
* Add ShareGroup stubgman@chromium.org2012-03-286-0/+114
| | | | | | | | | | TEST=unit test BUG=120297 Review URL: http://codereview.chromium.org/9836126 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@129456 0039d316-1c4b-4281-b951-d872f2087c98
* Move gpu hash_tables and atomicops stuff gman@chromium.org2012-03-286-18/+67
| | | | | | | | | TEST=none BUG=none Review URL: http://codereview.chromium.org/9836098 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@129341 0039d316-1c4b-4281-b951-d872f2087c98
* Disable flushers on CrOS to improve GPU process saturation during bulk ↵nduca@google.com2012-03-282-1/+8
| | | | | | | | | | | | | | | texture uploads We might want to have a flame war about this. But my intuition is that this is a net-win, and that in the long run, this behavior should be up in WebGL somewhere and that the compositor and ganesh should flush manually. BUG=120548 R=jbates@chromium.org,jamesr@chromium.org,marcheu@chromium.org,piman@chromium.org Review URL: https://chromiumcodereview.appspot.com/9809013 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@129313 0039d316-1c4b-4281-b951-d872f2087c98
* make disable-glsl-translator work againgman@chromium.org2012-03-264-7/+28
| | | | | | | | | | TEST=none BUG=none Review URL: http://codereview.chromium.org/9812021 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@128849 0039d316-1c4b-4281-b951-d872f2087c98
* Fix some GPU logging printsgman@chromium.org2012-03-241-8/+10
| | | | | | | | | | | | TEST=none BUG=none TBR=apatrick@chromium.org Review URL: http://codereview.chromium.org/9836072 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@128711 0039d316-1c4b-4281-b951-d872f2087c98
* Chromium implementation of discardBackbuffer WebGraphicsContext3D extension.mmocny@chromium.org2012-03-231-0/+1
| | | | | | | | | | | Webkit side patch: https://bugs.webkit.org/show_bug.cgi?id=81383 BUG=116049 TEST=Manual Review URL: http://codereview.chromium.org/9699125 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@128501 0039d316-1c4b-4281-b951-d872f2087c98
* Modify gpu::gles2::FeatureInfo::AddFeatures to enable feature flags based on ↵bsalomon@google.com2012-03-231-3/+3
| | | | | | | | | requested features but not disable them. BUG=112165 Review URL: https://chromiumcodereview.appspot.com/9837018 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@128403 0039d316-1c4b-4281-b951-d872f2087c98
* Fix OES_texture_float test failure.zmo@chromium.org2012-03-201-8/+38
| | | | | | | | | | | We need to map the internal_format on desktop GL. BUG=118091 TEST=webgl conformance tests R=gman Review URL: https://chromiumcodereview.appspot.com/9752013 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@127784 0039d316-1c4b-4281-b951-d872f2087c98
* Increase reschedule timeout to 1 second.jbauman@chromium.org2012-03-201-1/+1
| | | | | | | | | | | | | Due to the new per-command-buffer descheduling, this was being hit too often due to shader compiles and other expensive operations in still-scheduled command buffers. BUG= TEST= Review URL: http://codereview.chromium.org/9732023 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@127774 0039d316-1c4b-4281-b951-d872f2087c98
* Check that attachments are the correct type for the attachment typegman@chromium.org2012-03-194-19/+87
| | | | | | | | | | | | TEST=unit tests BUG=102232 R=jbauman@chromium.org Review URL: http://codereview.chromium.org/9732010 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@127559 0039d316-1c4b-4281-b951-d872f2087c98
* Remove unused code in StrictSharedIdHandlergman@chromium.org2012-03-191-10/+0
| | | | | | | | | | | | TEST=none BUG=109504 R=jbauman@chromium.org Review URL: http://codereview.chromium.org/9732004 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@127537 0039d316-1c4b-4281-b951-d872f2087c98
* Add OpenGL occlusion query support to pepper.jbauman@chromium.org2012-03-161-1/+10
| | | | | | | | | | | Should be full support for opengl es's occlusion query interface. Still only dev interfaces right now. BUG=88601 TEST= Review URL: https://chromiumcodereview.appspot.com/9601020 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@127298 0039d316-1c4b-4281-b951-d872f2087c98
* clang: Fix two -Wbool-conversion warningsthakis@chromium.org2012-03-161-2/+2
| | | | | | | | | BUG=none TEST=none Review URL: http://codereview.chromium.org/9700063 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@127128 0039d316-1c4b-4281-b951-d872f2087c98
* JSONWriter cleanup: integrate pretty print into write options.ericdingle@chromium.org2012-03-161-2/+3
| | | | | | | | | | | BUG= TEST=base_unittests TBR=abodenha@chromium.org,ajwong@chromium.org,chocobo@chromium.org,mnissler@chromium.org,akalin@chromium.org,brettw@chromium.org,arv@chromium.org Review URL: http://codereview.chromium.org/9590002 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@127080 0039d316-1c4b-4281-b951-d872f2087c98
* Make Occlusion Query Samplegman@chromium.org2012-03-155-7/+342
| | | | | | | | | | | | BUG=88601 TEST=this is the test Make Occlusion Query test Review URL: http://codereview.chromium.org/9699048 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@127009 0039d316-1c4b-4281-b951-d872f2087c98
* Plumb client side synthesized GL messages to JS Consolegman@chromium.org2012-03-152-1/+18
| | | | | | | | | | | | TEST=tested by hand BUG=none R=apatrick@chromium.org Review URL: http://codereview.chromium.org/9706042 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@126869 0039d316-1c4b-4281-b951-d872f2087c98
* Add support for GL_COMMANDS_ISSUED_CHROMIUM fence like query.gman@chromium.org2012-03-1415-442/+649
| | | | | | | | | | | | | | | | I'm not sure this is the right way to do this. I started by adding a few switch/case statements and then realized that there is no guarntee that glGenQueriesXXX exists which means I have to NOT call that for GL_COMMANDS_ISSUED_CHROMIUM queries. I ended up going down the path of a base class and this is where it lead. TEST=unit tests BUG=117768 Review URL: http://codereview.chromium.org/9694025 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@126607 0039d316-1c4b-4281-b951-d872f2087c98
* Insert No-ops into the command buffer when glQueryObjectuiv is called.gman@chromium.org2012-03-141-0/+3
| | | | | | | | | | | | | | | | This is so that if that if the GPU process dies we'll notice. Otherwise, no IPC messages will ever be sent to the GPU process and so nothing will check that it's still living. TEST=none BUG=88601 R=apatrick@chromium.org Review URL: http://codereview.chromium.org/9689083 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@126545 0039d316-1c4b-4281-b951-d872f2087c98
generated by cgit v1.1 at 2025-02-07 23:23:36 +0000