graphics_Drm: pass all tests and move them to bvt-cq |
||||||||||||
Issue descriptiongraphics_Drm is running in the lab, but the binaries are not on the test image. Also the path is not set. https://wmatrix.googleplex.com/unfiltered?hide_missing=True&tests=graphics_Drm&days_back=10&releases=55 File "/usr/local/autotest/tests/graphics_Drm/graphics_Drm.py", line 34, in run_once stderr_tee=utils.TEE_TO_LOGS) File "/usr/local/autotest/common_lib/base_utils.py", line 740, in run "Command returned non-zero exit status") CmdError: Command <drm_cursor_test> failed, rc=127, Command returned non-zero exit status * Command: drm_cursor_test Exit status: 127 Duration: 0.00690698623657 stderr: /bin/bash: drm_cursor_test: command not found
,
Nov 9 2016
Lets get all these tests passing soonish and move them to bvt-cq where they will close the tree. Here is my contribution: https://chromium-review.googlesource.com/#/c/409036/
,
Nov 9 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/2f7624cbc201b83a56a941da939a4c105998d220 commit 2f7624cbc201b83a56a941da939a4c105998d220 Author: Ilja H. Friedel <ihf@chromium.org> Date: Wed Nov 09 00:28:40 2016 graphics_Drm: refactor test status. 1) Check failures in more ways and try to provide a good string to wmatrix. 2) Run all tests as part of bvt-perbuild. 3) Tag all tests after all. This will make it easier to diff across boards on wmatrix. BUG= chromium:654131 TEST=test_that $DUT graphics_Drm.atomictest Change-Id: I203e49d0a1765fe2bc77169c38191f2d3be23bef Reviewed-on: https://chromium-review.googlesource.com/409036 Commit-Ready: Ilja H. Friedel <ihf@chromium.org> Tested-by: Ilja H. Friedel <ihf@chromium.org> Reviewed-by: Zach Reizner <zachr@chromium.org> [modify] https://crrev.com/2f7624cbc201b83a56a941da939a4c105998d220/client/site_tests/graphics_Drm/control.vgem_test [modify] https://crrev.com/2f7624cbc201b83a56a941da939a4c105998d220/client/site_tests/graphics_Drm/control.linear_bo_test [modify] https://crrev.com/2f7624cbc201b83a56a941da939a4c105998d220/client/site_tests/graphics_Drm/control.vgem_fb_test [modify] https://crrev.com/2f7624cbc201b83a56a941da939a4c105998d220/client/site_tests/graphics_Drm/control.atomictest [modify] https://crrev.com/2f7624cbc201b83a56a941da939a4c105998d220/client/site_tests/graphics_Drm/graphics_Drm.py [modify] https://crrev.com/2f7624cbc201b83a56a941da939a4c105998d220/client/site_tests/graphics_Drm/control.swrast_test [modify] https://crrev.com/2f7624cbc201b83a56a941da939a4c105998d220/client/site_tests/graphics_Drm/control.null_platform_test [modify] https://crrev.com/2f7624cbc201b83a56a941da939a4c105998d220/client/site_tests/graphics_Drm/control.gamma_test [modify] https://crrev.com/2f7624cbc201b83a56a941da939a4c105998d220/client/site_tests/graphics_Drm/control.drm_cursor_test
,
Nov 11 2016
All boards fail https://wmatrix.googleplex.com/unfiltered?hide_missing=True&releases=tot&tests=graphics_Drm.atomictest https://wmatrix.googleplex.com/unfiltered?hide_missing=True&releases=tot&tests=graphics_Drm.swrast_test https://wmatrix.googleplex.com/unfiltered?hide_missing=True&releases=tot&tests=graphics_Drm.vgem_test Some boards pass, some boards fail https://wmatrix.googleplex.com/unfiltered?hide_missing=True&releases=tot&tests=graphics_Drm.drm_cursor_test https://wmatrix.googleplex.com/unfiltered?hide_missing=True&releases=tot&tests=graphics_Drm.gamma_test https://wmatrix.googleplex.com/unfiltered?hide_missing=True&releases=tot&tests=graphics_Drm.linear_bo_test https://wmatrix.googleplex.com/unfiltered?hide_missing=True&releases=tot&tests=graphics_Drm.null_platform_test https://wmatrix.googleplex.com/unfiltered?hide_missing=True&releases=tot&tests=graphics_Drm.vgem_fb_test Concrete failure reasons on the bottom of https://wmatrix.googleplex.com/unfiltered?hide_missing=True&releases=tot&tests=graphics_Drm.*
,
Nov 11 2016
Failed: atomictest (/dev/dri/card1 failed to get r) Failed: atomictest (/dev/dri/renderD129 failed to get r) Failed: atomictest (/dev/dri/renderD130 failed to get r) Failed: atomictest (failed to add properties to the set)
,
Nov 11 2016
Failed: vgem_fb_test (ERROR:bs_dumb_mmap():dumb_mmap.c:31:mmap returned MAP_FAILED: 22) Failed: vgem_fb_test (ERROR:main():vgem_fb_test.c:170:failed to open card for display) Failed: vgem_fb_test (ERROR:main():vgem_fb_test.c:176:failed to open vgem card)
,
Nov 11 2016
Failed: vgem_test ([31mfailed[0m to open vgem device: No such file or directory) Failed: vgem_test ([32mopened[0m vgem device)
,
Nov 11 2016
Failed: swrast_test (/source/core-private/drivers/nvrm/gpu/nvrm_gpu_nvgpu_linux.cpp(484) NvRmGpuNvGpu_IoctlGetCharacteristics: Using GPU capability defaults for ChromeOS kernel.) Failed: swrast_test (EGL 1.4) Failed: swrast_test (libEGL warning: DRI2: failed to open vgem (search paths /usr/lib/dri))
,
Nov 11 2016
Failed: gamma_test (ERROR:bs_dma_buf_mmap_plane():dma_buf.c:30:mmap returned MAP_FAILED: 22) Failed: gamma_test (ERROR:main():gamma_test.c:282:failed to make pipe) Failed: gamma_test (ERROR:main():gamma_test.c:315:failed to make pipe with crtc mask: 1) Failed: gamma_test (ERROR:main():gamma_test.c:315:failed to make pipe with crtc mask: 4) Failed: gamma_test (ERROR:main():gamma_test.c:329:CRTC 25 has no gamma table) Failed: gamma_test (ERROR:main():gamma_test.c:329:CRTC 26 has no gamma table) Failed: gamma_test (ERROR:main():gamma_test.c:329:CRTC 27 has no gamma table) Failed: gamma_test (ERROR:main():gamma_test.c:329:CRTC 30 has no gamma table) Failed: gamma_test (ERROR:main():gamma_test.c:329:CRTC 31 has no gamma table)
,
Nov 11 2016
Failed: linear_bo_test (/source/core-private/drivers/nvrm/gpu/nvrm_gpu_nvgpu_linux.cpp(484) NvRmGpuNvGpu_IoctlGetCharacteristics: Using GPU capability defaults for ChromeOS kernel.) Failed: linear_bo_test (ERROR:main():linear_bo_test.c:283:failed to open card for display) Failed: linear_bo_test (ERROR:main():linear_bo_test.c:365:failed to set crtc: -22) Failed: linear_bo_test (libEGL warning: DRI2: failed to open vgem (search paths /usr/lib/dri))
,
Nov 11 2016
Failed: null_platform_test (/source/core-private/drivers/nvrm/gpu/nvrm_gpu_nvgpu_linux.cpp(484) NvRmGpuNvGpu_IoctlGetCharacteristics: Using GPU capability defaults for ChromeOS kernel.) Failed: null_platform_test (ERROR:main():null_platform_test.c:111:failed to open card for display) Failed: null_platform_test (libEGL warning: DRI2: failed to open vgem (search paths /usr/lib/dri))
,
Nov 11 2016
Failed: drm_cursor_test (ERROR:bs_dma_buf_mmap_plane():dma_buf.c:30:mmap returned MAP_FAILED: 22) Failed: drm_cursor_test (ERROR:main():drm_cursor_test.c:27:failed to open card for display)
,
Nov 11 2016
And now as they say: back to work! (And please bookmark the direct links I gave in #4.)
,
Jan 26 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/7f6370dade8b99e124293625253472d023ebe5b6 commit 7f6370dade8b99e124293625253472d023ebe5b6 Author: Ilja H. Friedel <ihf@chromium.org> Date: Thu Jan 26 05:54:44 2017 graphics_Drm: only run tests once a day. Most are not passing and nobody is currently actively working on making them pass. BUG= chromium:654131 TEST=None. Change-Id: Id5d0477f01e90289d59ca50e6ba9f0b7173df769 Reviewed-on: https://chromium-review.googlesource.com/433437 Tested-by: Ilja H. Friedel <ihf@chromium.org> Trybot-Ready: Ilja H. Friedel <ihf@chromium.org> Reviewed-by: Zach Reizner <zachr@chromium.org> Reviewed-by: Haixia Shi <hshi@chromium.org> [modify] https://crrev.com/7f6370dade8b99e124293625253472d023ebe5b6/client/site_tests/graphics_Drm/control.vgem_test [modify] https://crrev.com/7f6370dade8b99e124293625253472d023ebe5b6/client/site_tests/graphics_Drm/control.linear_bo_test [modify] https://crrev.com/7f6370dade8b99e124293625253472d023ebe5b6/client/site_tests/graphics_Drm/control.vgem_fb_test [modify] https://crrev.com/7f6370dade8b99e124293625253472d023ebe5b6/client/site_tests/graphics_Drm/control.atomictest [modify] https://crrev.com/7f6370dade8b99e124293625253472d023ebe5b6/client/site_tests/graphics_Drm/control.swrast_test [modify] https://crrev.com/7f6370dade8b99e124293625253472d023ebe5b6/client/site_tests/graphics_Drm/control.null_platform_test [modify] https://crrev.com/7f6370dade8b99e124293625253472d023ebe5b6/client/site_tests/graphics_Drm/control.gamma_test [modify] https://crrev.com/7f6370dade8b99e124293625253472d023ebe5b6/client/site_tests/graphics_Drm/control.drm_cursor_test
,
Feb 23 2017
Mediatek based boards seem to failing the gamma_test with: Failed: gamma_test (ERROR:main():gamma_test.c:329:CRTC -332986816 has no gamma table) This patch series looks like it could solve this problem: https://chromium-review.googlesource.com/c/381796/1 +djkurtz, +Bibby, any plans to land the patch set???
,
Feb 23 2017
,
Feb 24 2017
This patch has been merged into kernel 4.4.( https://chromium-review.googlesource.com/c/428544/) I think we can merge that into kernel 3.18. @Daniel, is it ok?
,
Feb 25 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/b1144315bc2dd8cd59a3eebb42ad11784f300df4 commit b1144315bc2dd8cd59a3eebb42ad11784f300df4 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Sat Feb 25 12:19:53 2017 graphics_Drm: back out when no display is found Our Chromeboxes are not connected to an external display in the lab, and many drm-tests require connected displays. If graphics_utils finds no connected display and the test needs one, mark the test as passed (like the gbmtest does). BUG= chromium:654131 TEST=run graphics_Drm on tricky Change-Id: I806914df0af328c018055b737b27829b62aa4c09 Reviewed-on: https://chromium-review.googlesource.com/446577 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Ilja H. Friedel <ihf@chromium.org> [modify] https://crrev.com/b1144315bc2dd8cd59a3eebb42ad11784f300df4/client/site_tests/graphics_Drm/graphics_Drm.py [modify] https://crrev.com/b1144315bc2dd8cd59a3eebb42ad11784f300df4/client/site_tests/graphics_Drm/control.vgem_test [modify] https://crrev.com/b1144315bc2dd8cd59a3eebb42ad11784f300df4/client/site_tests/graphics_Drm/control.swrast_test
,
Feb 28 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/drm-tests/+/d50b1c8371bcf03ab1c3151c9f68b3fbbd3a4b85 commit d50b1c8371bcf03ab1c3151c9f68b3fbbd3a4b85 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Tue Feb 28 19:48:12 2017 drm-tests: add bs_egl_has_extensions The swrast_test needs to query extensions, so let's not duplicate code. BUG= chromium:654131 TEST=compiles Change-Id: I65c06baf07612b20abf39aefa1a601d3caffa93f Reviewed-on: https://chromium-review.googlesource.com/446680 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Zach Reizner <zachr@chromium.org> [modify] https://crrev.com/d50b1c8371bcf03ab1c3151c9f68b3fbbd3a4b85/bsdrm/src/egl.c [modify] https://crrev.com/d50b1c8371bcf03ab1c3151c9f68b3fbbd3a4b85/bsdrm/include/bs_drm.h
,
Mar 2 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/drm-tests/+/fee57f011731f6fc5da4f2858106caf733af8388 commit fee57f011731f6fc5da4f2858106caf733af8388 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Thu Mar 02 02:42:14 2017 drm-tests: swrast_test: check for EGL_KHR_no_config_context eglCreateContext can only take a NULL config if the EGL_KHR_no_config_context extension is supported by the driver. The Mali and Imagination drivers don't support this, and passing in a null config doesn't seem critical to the test, so let's add a check here. BUG= chromium:654131 TEST=swrast_test works on mali driver Change-Id: I97fd1bf3477548f8bd746473636d09e5288b604e Reviewed-on: https://chromium-review.googlesource.com/446013 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org> [modify] https://crrev.com/fee57f011731f6fc5da4f2858106caf733af8388/swrast_test.c [modify] https://crrev.com/fee57f011731f6fc5da4f2858106caf733af8388/Makefile
,
Mar 3 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/drm-tests/+/b2d9f44c5a45f3dfae478e0d4abd4f20d30fa4ee commit b2d9f44c5a45f3dfae478e0d4abd4f20d30fa4ee Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Fri Mar 03 02:34:10 2017 drm-tests: linear_bo_test: removed failing expectations On Daisy, this test is failing with: failed test case: expect_success=false format=GBM_FORMAT_XRGB8888, usage=GBM_BO_USE_SCANOUT | GBM_BO_USE_RENDERING | GBM_BO_USE_LINEAR minigbm was recently changed to simplify usages, so we don't have a separate entry for GBM_BO_USE_RENDERING and GBM_BO_USE_LINEAR in our format supported list if we always end up allocating a linear buffer (see CL:421499). More changes are expected (see CL:444251). As such, let's remove the entries that are expected to fail when: GBM_BO_USE_SCANOUT | GBM_BO_USE_RENDERING | GBM_BO_USE_LINEAR is specified. Let's also remove expect_success field. BUG= chromium:654131 TEST=linear_bo_test on Daisy Change-Id: Ica1458f79b044ad9b889633f509970a133e3c279 Reviewed-on: https://chromium-review.googlesource.com/446533 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Haixia Shi <hshi@chromium.org> Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org> [modify] https://crrev.com/b2d9f44c5a45f3dfae478e0d4abd4f20d30fa4ee/linear_bo_test.c
,
Mar 7 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/drm-tests/+/e03544704eff6c530f7163c1927a298e84ed0fab commit e03544704eff6c530f7163c1927a298e84ed0fab Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Tue Mar 07 04:42:35 2017 drm-tests: vgem_test: don't log everything to stderr On wmatrix the vgem_test shows up failing on every platform because of this. Only log to stderr in the case of an error. BUG= chromium:654131 TEST=vgem_test passes on cyan Change-Id: Iebf885b480a7fd1530424d14fe7a91633067a091 Reviewed-on: https://chromium-review.googlesource.com/450897 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Ilja H. Friedel <ihf@chromium.org> [modify] https://crrev.com/e03544704eff6c530f7163c1927a298e84ed0fab/vgem_test.c
,
Mar 8 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/drm-tests/+/9a1cf162a93aa9bd192bcb157e261a58efe35e56 commit 9a1cf162a93aa9bd192bcb157e261a58efe35e56 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Wed Mar 08 02:07:36 2017 drm-tests: swrast_test: switch bs_egl_has_extension input The correct input order was switched. Oops. BUG= chromium:654131 TEST=none Change-Id: If60777aa3b265e1f5d4369e37ad1e0580f3f3a3c Reviewed-on: https://chromium-review.googlesource.com/450915 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org> [modify] https://crrev.com/9a1cf162a93aa9bd192bcb157e261a58efe35e56/swrast_test.c
,
Mar 8 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/4b938465a65bcd76ebfc18f62783dd4f08b8fd99 commit 4b938465a65bcd76ebfc18f62783dd4f08b8fd99 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Wed Mar 08 03:55:56 2017 graphics_Drm: rename vgem_fb_test to mmap_test With CL:339190, this test's name was changed so we should reflect that change here. BUG= chromium:654131 TEST=test_that $IP graphics_Drm.mmap_test works Change-Id: I47d2d38287f309cd779f7bef8fa6b5cb3399d802 Reviewed-on: https://chromium-review.googlesource.com/450873 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Ilja H. Friedel <ihf@chromium.org> [rename] https://crrev.com/4b938465a65bcd76ebfc18f62783dd4f08b8fd99/client/site_tests/graphics_Drm/control.mmap_test
,
Mar 10 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/drm-tests/+/5cd9c8b79222251fc8678e2cddd70ef86a63426c commit 5cd9c8b79222251fc8678e2cddd70ef86a63426c Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Fri Mar 10 21:05:15 2017 drm-tests: unbind context before eglTerminate On Oak, we have GL PVR errors due to our tests not unbinding the context before calling eglTerminate. BUG=b:36007527, chromium:654131 TEST=No PVR errors on oak Change-Id: Ia899e1a420b2b618cadc005ee59fe0c0704bdba7 Reviewed-on: https://chromium-review.googlesource.com/452543 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Haixia Shi <hshi@chromium.org> Reviewed-by: Ilja H. Friedel <ihf@chromium.org> Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org> [modify] https://crrev.com/5cd9c8b79222251fc8678e2cddd70ef86a63426c/bsdrm/src/egl.c [modify] https://crrev.com/5cd9c8b79222251fc8678e2cddd70ef86a63426c/swrast_test.c
,
Mar 17 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/302f244405270165571d72500573b4b8e094c461 commit 302f244405270165571d72500573b4b8e094c461 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Fri Mar 17 03:30:47 2017 graphics_Drm: use test result to determine failures Some GL implementations are overly verbose and print their logging information to stderr, and use words like "fail". For example, we mark the null_platform_test on Oak as failing because of this: libEGL warning: DRI2: failed to open vgem Mesa prints this to stderr even though it finds a device later. Additionally, Nyan boards print this to stderr even if a test successfully completes: NvRmGpuNvGpu_IoctlGetCharacteristics: Using GPU capability defaults for ChromeOS kernel This looks like verbose logging run amok. It's too much effort to enforce our logging standards on vendors, since some barely reply to our bugs in the first place :-). We check results after calling a vendor API implementation during a test, and fail the test if it doesn't match our expectations. Let's use that to determine success or failure. BUG= chromium:654131 TEST=null_platform_test, linear_bo_test pass on Oak, Nyan Change-Id: I8c5f1fa8644e32a77c3c81c2883f88b09e48e31c Reviewed-on: https://chromium-review.googlesource.com/455137 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Haixia Shi <hshi@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/302f244405270165571d72500573b4b8e094c461/client/site_tests/graphics_Drm/graphics_Drm.py
,
Apr 26 2017
,
Apr 27 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/drm-tests/+/fcac8009d97fc4e5776add0382cbe696c57b2a56 commit fcac8009d97fc4e5776add0382cbe696c57b2a56 Author: Gurchetan Singh <gurchetansingh@google.com> Date: Thu Apr 27 05:13:16 2017 drm-tests: swrast_test: specify surface type Makes a difference on nyan based boards. BUG= chromium:654131 TEST=Passes on nyan Change-Id: I12e43f94acb57d8ada530687bbfe93d3ca66d882 Reviewed-on: https://chromium-review.googlesource.com/487826 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Ilja H. Friedel <ihf@chromium.org> Reviewed-by: Haixia Shi <hshi@chromium.org> [modify] https://crrev.com/fcac8009d97fc4e5776add0382cbe696c57b2a56/swrast_test.c
,
May 12 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/e9b2d31d880b0755e4957c5209abc7cb6b54de29 commit e9b2d31d880b0755e4957c5209abc7cb6b54de29 Author: Gurchetan Singh <gurchetansingh@google.com> Date: Fri May 12 10:25:38 2017 graphics_Drm: move certain tests to bvt-cq These tests are passing consistently. BUG= chromium:654131 TEST=look at wmatrix https://wmatrix.googleplex.com/unfiltered?hide_missing=True&releases=tot&tests=graphics_Drm.drm_cursor_test&days_back=100 https://wmatrix.googleplex.com/unfiltered?hide_missing=True&releases=tot&tests=graphics_Drm.linear_bo_test&days_back=100 https://wmatrix.googleplex.com/unfiltered?hide_missing=True&releases=tot&tests=graphics_Drm.null_platform_test&days_back=100 https://wmatrix.googleplex.com/unfiltered?hide_missing=True&releases=tot&tests=graphics_Drm.swrast_test&days_back=100 Change-Id: I07a5e3fed70af66c94c53bc1ab3354067e89840e Reviewed-on: https://chromium-review.googlesource.com/495634 Commit-Ready: Ilja H. Friedel <ihf@chromium.org> Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Ilja H. Friedel <ihf@chromium.org> [modify] https://crrev.com/e9b2d31d880b0755e4957c5209abc7cb6b54de29/client/site_tests/graphics_Drm/control.linear_bo_test [modify] https://crrev.com/e9b2d31d880b0755e4957c5209abc7cb6b54de29/client/site_tests/graphics_Drm/control.null_platform_test [modify] https://crrev.com/e9b2d31d880b0755e4957c5209abc7cb6b54de29/client/site_tests/graphics_Drm/control.drm_cursor_test [modify] https://crrev.com/e9b2d31d880b0755e4957c5209abc7cb6b54de29/client/site_tests/graphics_Drm/control.swrast_test
,
Jun 8 2017
a lot of the graphics_sanity are failing, is this related ? https://cros-goldeneye.corp.google.com/chromeos/healthmonitoring/testHealthDashboard?suite=bvt-cq&daysBack=7&milestone=&chromeOsVersion=&testNameFilter= if you look at https://cros-goldeneye.corp.google.com/chromeos/console/viewBuild?version=9592.11.0&channel=dev&type=release under rambi, the 4-5 failure bvt-cq test failures are all graphics related Here's the cautotest link https://uberchromeos-server38.corp.google.com/new_tko/#tab_id=spreadsheet_view&row=job_name%252Ctest_name&column=status&show_incomplete=true&show_only_latest=false&show_invalid=true&condition=job_name%2520LIKE%2520%27banjo-release/R60-9592.11.0/bvt-cq/%2525%27%2520AND%2520test_name%2520%253C%253E%2520%27SERVER_JOB%27%2520AND%2520test_name%2520NOT%2520LIKE%2520%27CLIENT%255C_JOB%2525%27 The failures are https://screenshot.googleplex.com/WRxf2TExUr4.png
,
Jun 8 2017
A test matrix: https://cros-goldeneye.corp.google.com/chromeos/healthmonitoring/testDetails?testName=graphics_Drm.drm_cursor_test&suite=bvt-cq&daysBack=7&version&milestone It looks the drm_cursor_test passes: 06/06 15:04:23.321 INFO | graphics_Drm:0119| -----------------[drm_cursor_test]----------------- 06/06 15:04:23.322 DEBUG| utils:0203| Running 'wflinfo -p null -a gles2' 06/06 15:04:24.338 INFO | graphics_utils:1002| Found gles3.0. 06/06 15:04:24.339 DEBUG| utils:0203| Running 'modetest -c' 06/06 15:04:24.442 DEBUG| utils:0203| Running 'modetest -c' 06/06 15:04:24.549 DEBUG| graphics_Drm:0122| Running test drm_cursor_test. 06/06 15:04:24.550 DEBUG| utils:0203| Running 'drm_cursor_test' 06/06 15:04:36.357 INFO | graphics_Drm:0079| Passed: drm_cursor_test But unfortunately, graphics_Drm marks this as a failure. pwang@, can you take a look?
,
Jun 8 2017
I am fairly sure this is issue 730272 again. It looks like an R60 control file runs the R61 python of graphics_Drm.py. We should check if it goes away tomorrow.
,
Oct 11 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/9e6101834c047c211314e66fb0537be0b22255ca commit 9e6101834c047c211314e66fb0537be0b22255ca Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Wed Oct 11 05:13:50 2017 graphics_Drm: run all atomictests It passes now on Kevin/Eve, let's start running them all. BUG= chromium:654131 TEST=atomictest -t all Change-Id: Iaf92e9c038201c326d244c22de76c46468bb7d4a Reviewed-on: https://chromium-review.googlesource.com/709555 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Po-Hsien Wang <pwang@chromium.org> [modify] https://crrev.com/9e6101834c047c211314e66fb0537be0b22255ca/client/site_tests/graphics_Drm/graphics_Drm.py
,
Nov 15 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/b1d31e84fbb73cd59311a0a38694565d35a0220c commit b1d31e84fbb73cd59311a0a38694565d35a0220c Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Wed Nov 15 08:39:52 2017 graphics_Drm: remove gamma_test There is no plan to use gamma functions using the old API (maybe as an optional atomic property), and we can't expect all vendors to pass this test. BUG= chromium:654131 TEST=None Change-Id: Ic3c817db300cc824ac77a9163e8a9aee9bb32d89 Reviewed-on: https://chromium-review.googlesource.com/770233 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/b1d31e84fbb73cd59311a0a38694565d35a0220c/client/site_tests/graphics_Drm/graphics_Drm.py [delete] https://crrev.com/b02675147f22255c68769a8226908491a216546d/client/site_tests/graphics_Drm/control.gamma_test
,
Nov 18 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/drm-tests/+/0bc1fa8ec152c725cd7233ec5fe858beb4501708 commit 0bc1fa8ec152c725cd7233ec5fe858beb4501708 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Sat Nov 18 04:18:47 2017 drm-tests: build with large file support Otherwise dumb mmap doesn't work correctly. BUG= chromium:654131 TEST=vgem_test on Kevin passes Change-Id: Ib961c51108dd39ecfb23002a008a42abd606675e Reviewed-on: https://chromium-review.googlesource.com/775783 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Ilja H. Friedel <ihf@chromium.org> [modify] https://crrev.com/0bc1fa8ec152c725cd7233ec5fe858beb4501708/Makefile
,
Nov 21 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/drm-tests/+/4268fcac40e0504a9188b345b4e768282b48804e commit 4268fcac40e0504a9188b345b4e768282b48804e Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Tue Nov 21 02:13:43 2017 drm-tests: disable vkCreateDmaBufImageINTEL vk_glow is failing in the lab. vkCreateDmaBufImageINTEL doesn't work since anv_CreateDmaBufImageINTEL closes the fd (i.e, pCreateInfo->fd) before calling anv_bo_cache_import, which requires said prime fd. However, vkCreateDmaBufImageINTEL will be deprecated very soon so there's no point in fixing this. Let's wait until VK_EXT_external_memory_dma_buf / VK_EXT_image_drm_format_modifier are stable and don't test this code path for now. Also, clang-format was run (it delivers a different result everytime the clang in the chroot changes, which happens quite frequently). BUG= chromium:654131 TEST=emerge-reef drm-tests Change-Id: If26f5ba7d5f57d7998a6f9d0509a28232a35d0ae Reviewed-on: https://chromium-review.googlesource.com/777987 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Chad Versace <chadversary@chromium.org> [modify] https://crrev.com/4268fcac40e0504a9188b345b4e768282b48804e/vk_glow.c
,
Nov 25 2017
Hi everyone! Can you help me in reviewing https://chromium-review.googlesource.com/c/chromiumos/platform/drm-tests/+/789713 ! With the above changes, vgemtest is passing in kernel v4.12 and above!
,
Nov 29 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/drm-tests/+/68da5c0142d3f6f87a02749798bab9e26b048b3d commit 68da5c0142d3f6f87a02749798bab9e26b048b3d Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Wed Nov 29 04:17:03 2017 drm-tests: cut the number of frames in the mmap_test Otherwise, it's timing out in the lab on certain devices. BUG= chromium:654131 = TEST=graphics_Drm on minnie Change-Id: I26ce3333a8cf98b2b330d1669e95d1d599a8c27e Reviewed-on: https://chromium-review.googlesource.com/777946 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Chad Versace <chadversary@chromium.org> [modify] https://crrev.com/68da5c0142d3f6f87a02749798bab9e26b048b3d/mmap_test.c
,
Nov 29 2017
Hi gurchetan, I've noticed a very peculiar behavior for non graphics_Drm.bvt tests in autotest framework! It seems like for experimental tests like 'atomictest' the test setup is not stopping ui. Case #1 - test_that -b [board name][DUT IP Address] graphics_Drm fails on atomictest with ERROR:enable_crtc():atomictest.c:586:[CRTC:37]: failed to find mode. case #2- test_that -b [board name][DUT IP Address] graphics_Drm.atomictest Pass case #3- from DUT console- $ stop ui; atomictest -t all;start ui; Pass Can you please look into it. Thanks.
,
Nov 30 2017
#39, which machine/build are you testing on? According to our Eve results, graphics_Drm hasn't had the issue you described: https://stainless.corp.google.com/search?exclude_cts=true&exclude_non_release=true&board=%5Eeve%24&test=%5Egraphics_&suite=%5Egraphics%5C_per%5C-day%24&view=matrix&col=build&row=test&first_date=20171103&last_date=20171130 It's possible that 'stop ui' is racy with the start of the test. Can you test putting in a sleep after self._services.stop_services() in graphics_Drm.py to see if that helps your case?
,
Dec 2 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/3bee138615b16ca0d291dc4b7c04058aaa17c275 commit 3bee138615b16ca0d291dc4b7c04058aaa17c275 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Sat Dec 02 06:45:35 2017 CHROMIUM: add internal cursor fb to fb list The legacy drm_mode_cursor / drm_mode_cursor2 ioctls end up creating an interal framebuffer. However, where this framebuffer should be freed is not immediately obvious, since the ioctls don't give userspace an handle to this internal fb. Let's add it to fb list, so drm_fb_release (called by drm_release) will close it. BUG= chromium:654131 TEST=After drm_cursor_test, internal fb is no longer seen via modetest This should help with graphics_Drm flakiness. Change-Id: Ie754f6f518f663bf7e308dd83ebf7f18f49a26bb Reviewed-on: https://chromium-review.googlesource.com/804665 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/3bee138615b16ca0d291dc4b7c04058aaa17c275/drivers/gpu/drm/drm_plane.c
,
Dec 5 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/be6f5553bacddc4fad6eeebc9013aa5ecc76c4da commit be6f5553bacddc4fad6eeebc9013aa5ecc76c4da Author: Hirokazu Honda <hiroh@chromium.org> Date: Tue Dec 05 04:02:55 2017 Revert "CHROMIUM: add internal cursor fb to fb list" This reverts commit 3bee138615b16ca0d291dc4b7c04058aaa17c275. Reason for revert: This CL causes a crush on eve in few seconds after logging in. Original change's description: > CHROMIUM: add internal cursor fb to fb list > > The legacy drm_mode_cursor / drm_mode_cursor2 ioctls end up creating > an interal framebuffer. However, where this framebuffer should be freed > is not immediately obvious, since the ioctls don't give userspace an > handle to this internal fb. Let's add it to fb list, so drm_fb_release > (called by drm_release) will close it. > > BUG= chromium:654131 > TEST=After drm_cursor_test, internal fb is no longer seen via modetest > This should help with graphics_Drm flakiness. > > Change-Id: Ie754f6f518f663bf7e308dd83ebf7f18f49a26bb > Reviewed-on: https://chromium-review.googlesource.com/804665 > Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> > Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> > Reviewed-by: Stphane Marchesin <marcheu@chromium.org> Bug: chromium:654131 Change-Id: Ia207d024467b6b5b5aa865f36e47345ec8b953b0 Reviewed-on: https://chromium-review.googlesource.com/804610 Commit-Ready: Pawel Osciak <posciak@chromium.org> Tested-by: Ricky Liang <jcliang@chromium.org> Reviewed-by: Ricky Liang <jcliang@chromium.org> [modify] https://crrev.com/be6f5553bacddc4fad6eeebc9013aa5ecc76c4da/drivers/gpu/drm/drm_plane.c
,
Dec 7 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/drm-tests/+/fb2ddcf4acf95ff6b5b559d9548f483e5e91269d commit fb2ddcf4acf95ff6b5b559d9548f483e5e91269d Author: Abhijeet Kumar <abhijeet.kumar@intel.corp-partner.google.com> Date: Thu Dec 07 13:09:28 2017 drm-tests: changing sys_card_path for vgem device From 4.12 kernel version and above, vgem_init() is registering the device as a purely virtual device. Thus drm device is not available at "/sys/devices/platform/devices/vgem". And hence test is failing with "failed to open vgem device: No such file or directory" BUG= chromium:654131 TEST=emerge-{$board} drm-tests Change-Id: I204f2bd33f983ee4b75ba0cbeb15718013240d54 Signed-off-by: Abhijeet Kumar<abhijeet.kumar@intel.corp-partner.google.com> Reviewed-on: https://chromium-review.googlesource.com/789713 Commit-Ready: Abhijeet Kumar <abhijeet.kumar@intel.corp-partner.google.com> Tested-by: Anand Naik <anand.naik@intel.corp-partner.google.com> Reviewed-by: Zach Reizner <zachr@chromium.org> [modify] https://crrev.com/fb2ddcf4acf95ff6b5b559d9548f483e5e91269d/vgem_test.c [modify] https://crrev.com/fb2ddcf4acf95ff6b5b559d9548f483e5e91269d/Makefile
,
Dec 8 2017
,
Dec 8 2017
We decided not to move all of the tests to bvt-cq since some of them take too long. But we should track them on cros-perf ( crbug.com/779727 ). There are 3 issues remaining: 1) vgem_test failing on peach_pit (I think pwang@ is looking into that). 2) vgem_test/vk_glow is not running on Elm/Hana (I'll take a look into that). 3) atomictest is flaky sometimes (fixed with crrev.com/c/818321).
,
Dec 9 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/drm-tests/+/04e756b3349fcd711c828cb09b5c7be9927977b4 commit 04e756b3349fcd711c828cb09b5c7be9927977b4 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Sat Dec 09 07:13:57 2017 drm-tests: clean up state before atomictest The DRM UAPI states that cursor and overlay framebuffers may be present after a CRTC disable, so zero this out so we can get a clean slate. BUG= chromium:654131 TEST=graphics_Drm no longer flaky on Kevin/Eve Change-Id: I4a495fb5999b8e8ad95196e70714986786dc8506 Reviewed-on: https://chromium-review.googlesource.com/818321 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/04e756b3349fcd711c828cb09b5c7be9927977b4/atomictest.c
,
Dec 11 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/1e7e6488889c77fb7c9178eafb039e2d0641289c commit 1e7e6488889c77fb7c9178eafb039e2d0641289c Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Mon Dec 11 20:47:49 2017 graphics_Drm: don't run vgem on baseboard oak It's too hard to fix and not worth the effort. We should focus on making the test pass on newer kernels (which it does). It should, however, work on peach_pi and peach_pit (v3.8) once we enable vgem on it. BUG= chromium:654131 TEST=run graphics_Drm.vgem_test on Elm Change-Id: I0a52cd5e407491f73136b834ccddefcbf63b3172 Reviewed-on: https://chromium-review.googlesource.com/818444 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/1e7e6488889c77fb7c9178eafb039e2d0641289c/client/site_tests/graphics_Drm/graphics_Drm.py
,
Dec 11 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/overlays/board-overlays/+/b579cabd3a54abffcfa70c6b38200979ccfc36bc commit b579cabd3a54abffcfa70c6b38200979ccfc36bc Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Mon Dec 11 20:48:52 2017 baseboard-oak: Enable USE=vulkan We build the vk_glow test when USE=vulkan is enabled. We attempt to run vk_glow when libvulkan is present. libvulkan is present on Elm/Hana, but we don't build the vk_glow test. We then attempt to run the vk_glow test. Let's fix this. BUG= chromium:654131 TEST=emerge-elm drm-tests. vk_glow doesn't work on Elm due the fact it can't support XB24 for scanout. However, vk_glow doesn't render anything on Eve either so let's build it anyways until we figure what to do regard this test. CQ-DEPEND=CL:818530 Change-Id: I2400ea0041da7c8e80a78d3337058f3603de42cc Reviewed-on: https://chromium-review.googlesource.com/818383 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Ilja H. Friedel <ihf@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> Reviewed-by: Daniel Kurtz <djkurtz@chromium.org> [modify] https://crrev.com/b579cabd3a54abffcfa70c6b38200979ccfc36bc/baseboard-oak/profiles/base/make.defaults
,
Dec 11 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/drm-tests/+/5f45e43e2be8060a02ce029b5a0351553fd32192 commit 5f45e43e2be8060a02ce029b5a0351553fd32192 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Mon Dec 11 20:47:42 2017 drm-tests: use undef instead #define blah 0 This helps build vk_glow for platforms other than Intel. BUG= chromium:654131 TEST=emerge-oak drm-tests Change-Id: I145db3d70bf6e1110f00336bb8efed743a5aa093 Reviewed-on: https://chromium-review.googlesource.com/818530 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/5f45e43e2be8060a02ce029b5a0351553fd32192/vk_glow.c
,
Dec 12 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/overlays/board-overlays/+/3b2030ab267cf8e226bed82d5803e34a499c24d2 commit 3b2030ab267cf8e226bed82d5803e34a499c24d2 Author: Daniel Kurtz <djkurtz@chromium.org> Date: Tue Dec 12 06:44:23 2017 chipset-mt8173: Remove deqp vulkan USE flag This USE flag is already set globally in baseboard-oak overlay [0]. [0] http://crrev.com/c/818383 BUG= chromium:654131 TEST=emerge-elm -pvq deqp => Shows "vulkan" USE is selected CQ-DEPEND=818383 Change-Id: I96ff2e41ac44e24819f8fec8db275c0c95c50724 Reviewed-on: https://chromium-review.googlesource.com/820095 Commit-Ready: Ilja H. Friedel <ihf@chromium.org> Tested-by: Ilja H. Friedel <ihf@chromium.org> Reviewed-by: Ilja H. Friedel <ihf@chromium.org> Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org> [modify] https://crrev.com/3b2030ab267cf8e226bed82d5803e34a499c24d2/chipset-mt8173/profiles/base/package.use
,
Dec 12 2017
The following revision refers to this bug: https://chrome-internal.googlesource.com/chromeos/overlays/chipset-mt8173-private/+/7addc8065cf9d5febd4c1be8704472e4498e9601 commit 7addc8065cf9d5febd4c1be8704472e4498e9601 Author: Daniel Kurtz <djkurtz@chromium.org> Date: Tue Dec 12 21:50:59 2017
,
Jan 11 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/1774bbd8a58a97a073e574959d0b153a33a9e8d6 commit 1774bbd8a58a97a073e574959d0b153a33a9e8d6 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Thu Jan 11 20:02:14 2018 graphics_Drm: run vgem_test on ARM boards We've figured out the caching issues, so let's run this everywhere. BUG= chromium:654131 TEST=vgem_test works on ARM Change-Id: Ib919be4fb14765bb973a4083cbc071877c086d23 Reviewed-on: https://chromium-review.googlesource.com/860739 Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Ilja H. Friedel <ihf@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/1774bbd8a58a97a073e574959d0b153a33a9e8d6/client/site_tests/graphics_Drm/graphics_Drm.py
,
Jan 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/3b8f1516ba2c66b01b0331f310383e172fc58fa1 commit 3b8f1516ba2c66b01b0331f310383e172fc58fa1 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Fri Jan 12 06:22:33 2018 CHROMIUM: add case of CONFIG_ARM64 in drm_clflush_sg The dma API apparent works for ARM64. We thought this might not be the case, since we were under the impression there's a swiotlb backend for the API (like Intel has). Atleast for v3.18, this works. Let's add a case for it in drm_cache file. BUG= chromium:654131 TEST=vgem_test works on Elm Change-Id: I2b385204dd901ef3261e3dc263e3402285e18abf Reviewed-on: https://chromium-review.googlesource.com/860737 Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/3b8f1516ba2c66b01b0331f310383e172fc58fa1/drivers/gpu/drm/i915/i915_gem.c [modify] https://crrev.com/3b8f1516ba2c66b01b0331f310383e172fc58fa1/include/drm/drmP.h [modify] https://crrev.com/3b8f1516ba2c66b01b0331f310383e172fc58fa1/drivers/gpu/drm/drm_cache.c
,
Jan 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/629384ae5c0e81607f98a278edbff7b10b5ebe37 commit 629384ae5c0e81607f98a278edbff7b10b5ebe37 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Fri Jan 12 06:22:34 2018 CHROMIUM: drm/vgem: flush after getting pages The pages that are fetched in vgem_get_pages() may be still be in the cache. We're experiencing a case where eviction of the cache races with the writes we do. To fix this, flush the cache during vgem_get_pages(). BUG= chromium:654131 TEST=vgem_test works on Elm Change-Id: I00c99f63e10aa9001043ce74912297e2bffb6478 Reviewed-on: https://chromium-review.googlesource.com/860738 Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/629384ae5c0e81607f98a278edbff7b10b5ebe37/drivers/gpu/drm/vgem/vgem_drv.h [modify] https://crrev.com/629384ae5c0e81607f98a278edbff7b10b5ebe37/drivers/gpu/drm/vgem/vgem_drv.c
,
Jan 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/95ea570a90f93b682779e0cd77ef9f553d374b20 commit 95ea570a90f93b682779e0cd77ef9f553d374b20 Author: Tomasz Figa <tfiga@chromium.org> Date: Fri Jan 12 22:30:54 2018 BACKPORT: CHROMIUM: drm/vgem: Add short circuit for reimporting our GEMs This fixes support for mmap after reimporting buffers previously exported in another process. BUG=b:27243707 TEST=mmap on reimported buffer seems to work Signed-off-by: Tomasz Figa <tfiga@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/340869 (cherry picked from commit 5f60803b96f62f8f37d87164f2f251e7cc6a8f91) Signed-off-by: Gurchetan Singh <gurchetansingh@chromium.org> BUG= chromium:654131 TEST=vgem_test works on Peach Pi CQ-DEPEND=CL:862970 Change-Id: I44121bff5429ad3e4349eb102ef5a8447dab3eb8 Reviewed-on: https://chromium-review.googlesource.com/858599 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/95ea570a90f93b682779e0cd77ef9f553d374b20/drivers/gpu/drm/vgem/vgem_drv.c [modify] https://crrev.com/95ea570a90f93b682779e0cd77ef9f553d374b20/drivers/gpu/drm/vgem/vgem_drv.h [modify] https://crrev.com/95ea570a90f93b682779e0cd77ef9f553d374b20/drivers/gpu/drm/vgem/vgem_dma_buf.c
,
Jan 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/5af5ced65aa79505196109dc9f5b429a6892805b commit 5af5ced65aa79505196109dc9f5b429a6892805b Author: Tomasz Figa <tfiga@chromium.org> Date: Fri Jan 12 22:30:55 2018 BACKPORT: CHROMIUM: drm/vgem: Add support for PRIME mmap Since we can now simply call mmap() on a prime FD in userspace, let's wire this properly in the driver, so that such attempts can succeed. BUG=b:28803368 TEST=Boot cyan-cheets image in qemu Signed-off-by: Tomasz Figa <tfiga@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/356270 Reviewed-by: Zach Reizner <zachr@chromium.org> (cherry picked from commit 183c84ec39916b48b23a94433eb77084c869c91e) Signed-off-by: Gurchetan Singh <gurchetansingh@chromium.org> BUG= chromium:654131 TEST=vgem_test on Tegra Change-Id: Iac8448c0b2b07cc523147e6982d6acdb5564e127 Reviewed-on: https://chromium-review.googlesource.com/858600 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/5af5ced65aa79505196109dc9f5b429a6892805b/drivers/gpu/drm/vgem/vgem_drv.c [modify] https://crrev.com/5af5ced65aa79505196109dc9f5b429a6892805b/drivers/gpu/drm/vgem/vgem_drv.h [modify] https://crrev.com/5af5ced65aa79505196109dc9f5b429a6892805b/drivers/gpu/drm/vgem/vgem_dma_buf.c
,
Jan 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/f986ce9da38d6a6fc3eb47f3f89d3c9495c045b9 commit f986ce9da38d6a6fc3eb47f3f89d3c9495c045b9 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Fri Jan 12 22:30:57 2018 CHROMIUM: add case of CONFIG_ARM in drm_clflush_sg The dma API works for flushing the cache on ARM32 (though not necessarily for ARM64). Let's add a case for it in drm_cache file. BUG= chromium:654131 TEST=vgem_test works on Peach Pi Change-Id: I54387efada0189dc5f13c8dfbc7eb71ba7809326 Reviewed-on: https://chromium-review.googlesource.com/858601 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/f986ce9da38d6a6fc3eb47f3f89d3c9495c045b9/drivers/gpu/drm/i915/i915_gem.c [modify] https://crrev.com/f986ce9da38d6a6fc3eb47f3f89d3c9495c045b9/include/drm/drmP.h [modify] https://crrev.com/f986ce9da38d6a6fc3eb47f3f89d3c9495c045b9/drivers/gpu/drm/drm_cache.c
,
Jan 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/e86d2925023291100a221c408b76c289c2bbf795 commit e86d2925023291100a221c408b76c289c2bbf795 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Fri Jan 12 22:30:58 2018 CHROMIUM: drm/vgem: flush after getting pages The pages that are fetched in vgem_get_pages() may be still be in the cache. We're experiencing a case where eviction of the cache races with the writes we do. To fix this, flush the cache during vgem_get_pages(). BUG= chromium:654131 TEST=vgem_test works on Tegra Change-Id: I6d65793ff179e209a380c167b1a3003e5d14dc3b Reviewed-on: https://chromium-review.googlesource.com/858602 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/e86d2925023291100a221c408b76c289c2bbf795/drivers/gpu/drm/vgem/vgem_drv.h [modify] https://crrev.com/e86d2925023291100a221c408b76c289c2bbf795/drivers/gpu/drm/vgem/vgem_drv.c
,
Jan 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/32ce012b023e79fac13bc59bf1918fdc0dec6f5a commit 32ce012b023e79fac13bc59bf1918fdc0dec6f5a Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Fri Jan 12 22:30:59 2018 CHROMIUM: drm/vgem: implement everything required by drm_gem_prime_dmabuf_ops, remove vgem_dmabuf_ops This CL is a backport of CL:243369, but no cherry-picks since copy and pasting was easier. BUG= chromium:654131 TEST=vgem_test works on Peach Pi CQ-DEPEND=CL:858599 Change-Id: I31baf22abee1fb32b749590519f6f79fe4068992 Reviewed-on: https://chromium-review.googlesource.com/862970 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/32ce012b023e79fac13bc59bf1918fdc0dec6f5a/drivers/gpu/drm/vgem/vgem_drv.c [modify] https://crrev.com/32ce012b023e79fac13bc59bf1918fdc0dec6f5a/drivers/gpu/drm/vgem/vgem_drv.h [modify] https://crrev.com/32ce012b023e79fac13bc59bf1918fdc0dec6f5a/drivers/gpu/drm/vgem/vgem_dma_buf.c
,
Jan 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/b5e38f9c094071b7240ed51142d1610f59a98b6e commit b5e38f9c094071b7240ed51142d1610f59a98b6e Author: Tomasz Figa <tfiga@chromium.org> Date: Fri Jan 12 22:31:01 2018 BACKPORT: CHROMIUM: drm/vgem: Add short circuit for reimporting our GEMs This fixes support for mmap after reimporting buffers previously exported in another process. BUG=b:27243707 TEST=mmap on reimported buffer seems to work Signed-off-by: Tomasz Figa <tfiga@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/340869 (cherry picked from commit 5f60803b96f62f8f37d87164f2f251e7cc6a8f91) Signed-off-by: Gurchetan Singh <gurchetansingh@chromium.org> BUG= chromium:654131 TEST=vgem_test and gbmtest on ARM v3.14 works Change-Id: I57a1807806bbc9292058ff01b6fe38484a97478f Reviewed-on: https://chromium-review.googlesource.com/858263 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/b5e38f9c094071b7240ed51142d1610f59a98b6e/drivers/gpu/drm/vgem/vgem_drv.c [modify] https://crrev.com/b5e38f9c094071b7240ed51142d1610f59a98b6e/drivers/gpu/drm/vgem/vgem_drv.h [modify] https://crrev.com/b5e38f9c094071b7240ed51142d1610f59a98b6e/drivers/gpu/drm/vgem/vgem_dma_buf.c
,
Jan 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/cb04462f273e3a7af95cda600ff7370918775494 commit cb04462f273e3a7af95cda600ff7370918775494 Author: Tomasz Figa <tfiga@chromium.org> Date: Fri Jan 12 22:31:02 2018 BACKPORT: CHROMIUM: drm/vgem: Add support for PRIME mmap Since we can now simply call mmap() on a prime FD in userspace, let's wire this properly in the driver, so that such attempts can succeed. BUG=b:28803368 TEST=Boot cyan-cheets image in qemu Signed-off-by: Tomasz Figa <tfiga@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/356270 Reviewed-by: Zach Reizner <zachr@chromium.org> (cherry picked from commit 183c84ec39916b48b23a94433eb77084c869c91e) Signed-off-by: Gurchetan Singh <gurchetansingh@chromium.org> BUG= chromium:654131 TEST=vgem_test on ARM v3.14 works Change-Id: Ic030b9ac2a9d310de9dde78b7f77b1edb42d9276 Reviewed-on: https://chromium-review.googlesource.com/858264 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/cb04462f273e3a7af95cda600ff7370918775494/drivers/gpu/drm/vgem/vgem_drv.c [modify] https://crrev.com/cb04462f273e3a7af95cda600ff7370918775494/drivers/gpu/drm/vgem/vgem_drv.h [modify] https://crrev.com/cb04462f273e3a7af95cda600ff7370918775494/drivers/gpu/drm/vgem/vgem_dma_buf.c
,
Jan 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/5526cbe148db37ba7d6745e4149336cadb057cef commit 5526cbe148db37ba7d6745e4149336cadb057cef Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Fri Jan 12 22:31:03 2018 CHROMIUM: add case of CONFIG_ARM in drm_clflush_sg The dma API works for flushing the cache on ARM32 (though not necessarily for ARM64). Let's add a case for it in drm_cache file. BUG= chromium:654131 TEST=vgem_test works on ARM v3.14 Change-Id: I100fcf2accc46f44fdcfc295b264fbfc929bab05 Reviewed-on: https://chromium-review.googlesource.com/858265 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/5526cbe148db37ba7d6745e4149336cadb057cef/drivers/gpu/drm/i915/i915_gem.c [modify] https://crrev.com/5526cbe148db37ba7d6745e4149336cadb057cef/include/drm/drmP.h [modify] https://crrev.com/5526cbe148db37ba7d6745e4149336cadb057cef/drivers/gpu/drm/drm_cache.c
,
Jan 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/16c04b4eedcfe1869ba0ff20df0e74eecba31457 commit 16c04b4eedcfe1869ba0ff20df0e74eecba31457 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Fri Jan 12 22:31:05 2018 CHROMIUM: drm/vgem: flush after getting pages The pages that are fetched in vgem_get_pages() may be still be in the cache. We're experiencing a case where eviction of the cache races with the writes we do. To fix this, flush the cache during vgem_get_pages(). BUG= chromium:654131 TEST=vgem_test works on ARM v3.14 Change-Id: I9a36a01d86859b4c06e5fa41199f3a284b3df871 Reviewed-on: https://chromium-review.googlesource.com/858266 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/16c04b4eedcfe1869ba0ff20df0e74eecba31457/drivers/gpu/drm/vgem/vgem_drv.h [modify] https://crrev.com/16c04b4eedcfe1869ba0ff20df0e74eecba31457/drivers/gpu/drm/vgem/vgem_drv.c
,
Jan 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/b8ef74392c7b3b61a2495c21927b2f836f807d9c commit b8ef74392c7b3b61a2495c21927b2f836f807d9c Author: Tomasz Figa <tfiga@chromium.org> Date: Fri Jan 12 22:30:48 2018 BACKPORT: CHROMIUM: drm/vgem: Add short circuit for reimporting our GEMs This fixes support for mmap after reimporting buffers previously exported in another process. BUG=b:27243707 TEST=mmap on reimported buffer seems to work Signed-off-by: Tomasz Figa <tfiga@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/340869 (cherry picked from commit 5f60803b96f62f8f37d87164f2f251e7cc6a8f91) Signed-off-by: Gurchetan Singh <gurchetansingh@chromium.org> BUG= chromium:654131 TEST=vgem_test works on Tegra CQ-DEPEND=CL:862994 Change-Id: Ib2ca651a2a4c7549024f61982c840cee61009b25 Reviewed-on: https://chromium-review.googlesource.com/858490 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/b8ef74392c7b3b61a2495c21927b2f836f807d9c/drivers/gpu/drm/vgem/vgem_drv.c [modify] https://crrev.com/b8ef74392c7b3b61a2495c21927b2f836f807d9c/drivers/gpu/drm/vgem/vgem_drv.h [modify] https://crrev.com/b8ef74392c7b3b61a2495c21927b2f836f807d9c/drivers/gpu/drm/vgem/vgem_dma_buf.c
,
Jan 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/4d7d6e1ebc531234841f02e1d0fcbd011082b1b2 commit 4d7d6e1ebc531234841f02e1d0fcbd011082b1b2 Author: Tomasz Figa <tfiga@chromium.org> Date: Fri Jan 12 22:30:50 2018 BACKPORT: CHROMIUM: drm/vgem: Add support for PRIME mmap Since we can now simply call mmap() on a prime FD in userspace, let's wire this properly in the driver, so that such attempts can succeed. BUG=b:28803368 TEST=Boot cyan-cheets image in qemu Signed-off-by: Tomasz Figa <tfiga@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/356270 Reviewed-by: Zach Reizner <zachr@chromium.org> (cherry picked from commit 183c84ec39916b48b23a94433eb77084c869c91e) Signed-off-by: Gurchetan Singh <gurchetansingh@chromium.org> BUG= chromium:654131 TEST=vgem_test on Tegra Change-Id: I757bf59e26e8c9c0174c7c1e7761d47938fb44a9 Reviewed-on: https://chromium-review.googlesource.com/858491 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/4d7d6e1ebc531234841f02e1d0fcbd011082b1b2/drivers/gpu/drm/vgem/vgem_drv.c [modify] https://crrev.com/4d7d6e1ebc531234841f02e1d0fcbd011082b1b2/drivers/gpu/drm/vgem/vgem_drv.h [modify] https://crrev.com/4d7d6e1ebc531234841f02e1d0fcbd011082b1b2/drivers/gpu/drm/vgem/vgem_dma_buf.c
,
Jan 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/6507bec21d05dd0885a12aa09e67cbb9e7f27554 commit 6507bec21d05dd0885a12aa09e67cbb9e7f27554 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Fri Jan 12 22:30:51 2018 CHROMIUM: add case of CONFIG_ARM in drm_clflush_sg The dma API works for flushing the cache on ARM32 (though not necessarily for ARM64). Let's add a case for it in drm_cache file. BUG= chromium:654131 TEST=vgem_test works on Tegra Change-Id: Ibdbb7323374d928149870e5528fd68d1418c079b Reviewed-on: https://chromium-review.googlesource.com/858492 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/6507bec21d05dd0885a12aa09e67cbb9e7f27554/drivers/gpu/drm/i915/i915_gem.c [modify] https://crrev.com/6507bec21d05dd0885a12aa09e67cbb9e7f27554/include/drm/drmP.h [modify] https://crrev.com/6507bec21d05dd0885a12aa09e67cbb9e7f27554/drivers/gpu/drm/drm_cache.c
,
Jan 12 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/kernel/+/3b9daaae6e0f9d4a09f4c7e4f226fb184a051d61 commit 3b9daaae6e0f9d4a09f4c7e4f226fb184a051d61 Author: Gurchetan Singh <gurchetansingh@chromium.org> Date: Fri Jan 12 22:30:52 2018 CHROMIUM: drm/vgem: flush after getting pages The pages that are fetched in vgem_get_pages() may be still be in the cache. We're experiencing a case where eviction of the cache races with the writes we do. To fix this, flush the cache during vgem_get_pages(). BUG= chromium:654131 TEST=vgem_test works on Tegra Change-Id: I75e3f09959b67880d3d8c14ed03b8eaf09c2f34a Reviewed-on: https://chromium-review.googlesource.com/858493 Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org> Tested-by: Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by: Stéphane Marchesin <marcheu@chromium.org> [modify] https://crrev.com/3b9daaae6e0f9d4a09f4c7e4f226fb184a051d61/drivers/gpu/drm/vgem/vgem_drv.h [modify] https://crrev.com/3b9daaae6e0f9d4a09f4c7e4f226fb184a051d61/drivers/gpu/drm/vgem/vgem_drv.c
,
Aug 24
I don't think the intention is to move the drm-tests to the CQ anymore. |
||||||||||||
►
Sign in to add a comment |
||||||||||||
Comment 1 by ihf@chromium.org
, Oct 25 2016