| Message ID | 20220126092921.298422-1-vicamo.yang@canonical.com |
|---|---|
| Headers | show |
| Series | Fix AMD Yellow Carp B0 s0i3 | expand |
Acked-by: Tim Gardner <tim.gardner@canonical.com> On 1/26/22 2:29 AM, You-Sheng Yang wrote: > BugLink: https://bugs.launchpad.net/bugs/1957026 > > [Impact] > > On AMD Yellow Carp B0 platforms, S0i3 is still not working. > > [Fix] > > This denpends on a few fixes for kernel driver that are all either in > v5.17-rc1 or in stable linux-5.1x.y, along with an already upstreamed > dmcub firmware. > > [Test Case] > > On AMD RMT CRB board, > 1. apply both kernel and firmware fixes > 2. module blacklist amd_sfh as it's still under development, > 3. trigger system suspend. Use power button to wake up the device. > 4. check /sys/kernel/debug/amd_pmc/s0ix_stats, the values shall be non-zero > > On without any of above, the board may not cut device power during suspend, > and/or fail to resume and leaving a blank screen, and/or s0ix stats remain > all zeros. > > [Where problems could occur] > > So far all the patches can be cleanly applied, and they're being > progressively proposed to linux-stable. > > [Other Info] > > For kernel driver fixes, patches for both oem-5.14 and jammy are necessary; > for firmware, that has been available in jammy branch but not yet released, > so only Focal will be nominated. > > Nicholas Kazlauskas (4): > drm/amd/display: Optimize bandwidth on following fast update > drm/amd/display: Fix surface optimization regression on Carrizo > drm/amd/display: Reset DMCUB before HW init > drm/amd/display: Don't reinitialize DMCUB on s0ix resume > > .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 35 +++++++++++++++++-- > drivers/gpu/drm/amd/display/dc/core/dc.c | 24 +++++++++---- > 2 files changed, 49 insertions(+), 10 deletions(-) >
You-Sheng Yang kirjoitti 26.1.2022 klo 11.29: > BugLink: https://bugs.launchpad.net/bugs/1957026 > > [Impact] > > On AMD Yellow Carp B0 platforms, S0i3 is still not working. > > [Fix] > > This denpends on a few fixes for kernel driver that are all either in > v5.17-rc1 or in stable linux-5.1x.y, along with an already upstreamed > dmcub firmware. > > [Test Case] > > On AMD RMT CRB board, > 1. apply both kernel and firmware fixes > 2. module blacklist amd_sfh as it's still under development, > 3. trigger system suspend. Use power button to wake up the device. > 4. check /sys/kernel/debug/amd_pmc/s0ix_stats, the values shall be non-zero > > On without any of above, the board may not cut device power during suspend, > and/or fail to resume and leaving a blank screen, and/or s0ix stats remain > all zeros. > > [Where problems could occur] > > So far all the patches can be cleanly applied, and they're being > progressively proposed to linux-stable. > > [Other Info] > > For kernel driver fixes, patches for both oem-5.14 and jammy are necessary; > for firmware, that has been available in jammy branch but not yet released, > so only Focal will be nominated. > > Nicholas Kazlauskas (4): > drm/amd/display: Optimize bandwidth on following fast update > drm/amd/display: Fix surface optimization regression on Carrizo > drm/amd/display: Reset DMCUB before HW init > drm/amd/display: Don't reinitialize DMCUB on s0ix resume > > .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 35 +++++++++++++++++-- > drivers/gpu/drm/amd/display/dc/core/dc.c | 24 +++++++++---- > 2 files changed, 49 insertions(+), 10 deletions(-) > applied to oem-5.14, thanks