mbox series

[v6,00/10] drm: ci: fixes

Message ID 20231129121841.253223-1-vignesh.raman@collabora.com (mailing list archive)
Headers show
Series drm: ci: fixes | expand

Message

Vignesh Raman Nov. 29, 2023, 12:18 p.m. UTC
The patch series contains improvements, enabling new ci jobs which
enables testing for Mediatek MT8173, Qualcomm APQ 8016 and VirtIO GPU,
fixing issues with the ci jobs and updating the expectation files.

v2:
  - Use fdtoverlay command to merge overlay dtbo with the base dtb instead of modifying the kernel sources
  - Reworded the commit message for enabling jobs
  - Added a new patch in the series to use scripts/config to enable/disable configs

v3:
  - New patch in the series to add device tree overlay in arch/arm64/boot/dts/qcom
  - drm-ci scripts to use device tree overlay from arch/arm64/boot/dts/qcom and compile base device tree with overlay support
  - New patch in the series to enable CONFIG_REGULATOR_DA9211 in defconfig
  - Remove CONFIG_RTC_DRV_MT6397=y as it is already enabled in defconfig

v4:
  - Drop 'enable CONFIG_REGULATOR_DA9211 in defconfig' patch as it is sent upstream as a seperate patch
  - Use apq8016-sbc-usb-host.dtb which allows the USB controllers to work in host mode.
    This patch depends on https://lore.kernel.org/lkml/20230911161518.650726-1-vignesh.raman@collabora.com/

v5:
  - Added a new patch in the series to set IGT_FORCE_DRIVER to 'mediatek' for mt8173
  - Added a new patch in the series to make artifacts available for virtio jobs
  - Added a new patch in the series to add pipeline url to fails and flakes files
  - Generate fails and flakes file with the updated xfails script - https://www.spinics.net/lists/kernel/msg4959630.html
  - Drop 'virtio: Update ci variables' patch as the tests which causes the malloc issue are skipped

v6:
  - Updated commit message for enable DA9211 regulator fix 
  - Use GPU_VERSION instead of CI_JOB_NAME to check if it is mt8173 while setting IGT_FORCE_DRIVER
  - Added a new patch in the series to uprev IGT to fix memory corruption
  - Added a new patch in the series to update drm ci documentation
  - Generate fails file with drm-misc-next

Vignesh Raman (10):
  drm: ci: igt_runner: Remove todo
  drm: ci: Force db410c to host mode
  drm: ci: arm64.config: Enable DA9211 regulator
  drm: ci: Enable new jobs
  drm: ci: Use scripts/config to enable/disable configs
  drm: ci: mediatek: Set IGT_FORCE_DRIVER for mt8173
  drm: ci: virtio: Make artifacts available
  drm: ci: uprev IGT
  drm/doc: ci: Add IGT version details for flaky tests
  drm: ci: Update xfails

 Documentation/gpu/automated_testing.rst       |  7 +--
 drivers/gpu/drm/ci/arm64.config               |  1 +
 drivers/gpu/drm/ci/build.sh                   | 16 +++----
 drivers/gpu/drm/ci/gitlab-ci.yml              |  2 +-
 drivers/gpu/drm/ci/igt_runner.sh              |  5 +-
 drivers/gpu/drm/ci/test.yml                   | 13 ++----
 .../drm/ci/xfails/mediatek-mt8173-fails.txt   | 13 ++++--
 .../gpu/drm/ci/xfails/msm-apq8016-fails.txt   |  5 ++
 .../drm/ci/xfails/virtio_gpu-none-fails.txt   | 46 +++++++++++++++++++
 9 files changed, 82 insertions(+), 26 deletions(-)

Comments

Daniel Stone Nov. 29, 2023, 1:22 p.m. UTC | #1
Hi Vignesh,

On Wed, 29 Nov 2023 at 12:19, Vignesh Raman <vignesh.raman@collabora.com> wrote:
> Expected driver for mt8173 is "mediatek" and for mt8183
> it is "panfrost". Set IGT_FORCE_DRIVER to 'mediatek' as
> the expected driver for mt8173.

Actually, for mt8183 it's both. And for mt8173 it will probably be
mediatek+pvr pretty soon. Each of these SoCs (like most Arm devices)
have a separate display controller and GPU, with different drivers for
each. They'll run different tests with different xfails. So we should
figure out a way to support igt running for both devices on the one
system.

Cheers,
Daniel
Vignesh Raman Nov. 30, 2023, 1:50 p.m. UTC | #2
Hi Daniel,

On 29/11/23 18:52, Daniel Stone wrote:
> Hi Vignesh,
> 
> On Wed, 29 Nov 2023 at 12:19, Vignesh Raman <vignesh.raman@collabora.com> wrote:
>> Expected driver for mt8173 is "mediatek" and for mt8183
>> it is "panfrost". Set IGT_FORCE_DRIVER to 'mediatek' as
>> the expected driver for mt8173.
> 
> Actually, for mt8183 it's both. And for mt8173 it will probably be
> mediatek+pvr pretty soon. Each of these SoCs (like most Arm devices)
> have a separate display controller and GPU, with different drivers for
> each. They'll run different tests with different xfails. So we should
> figure out a way to support igt running for both devices on the one
> system.

Agree. Will work on separate patch to test both the drivers for these 
devices.

Regards,
Vignesh