ACK: [PATCH 0/2][SRU][Lunar] Fix gfxoff workqueue to recover S0i3 on AMD Phoenix platform

Tim Gardner tim.gardner at canonical.com
Tue May 30 13:13:03 UTC 2023


On 5/30/23 3:25 AM, You-Sheng Yang wrote:
> BugLink: https://bugs.launchpad.net/bugs/2020685
> 
> [Impact]
> 
> On AMD Phoenix platforms, system may hang, or reboot while resuming from
> suspend.
> 
> [Fix]
> 
> Two patches in v6.4-rc3 for amdgpu. The first one is pulled for
> dependency.
> 
> [Test Case]
> 
> $ checkbox-cli run com.canonical.certification::suspend/suspend_advanced_auto
> $ checkbox-cli run com.canonical.certification::stress/s2idle_pm-graph_30
> $ checkbox-cli run com.canonical.certification::power-management/suspend_30_cycles_with_reboots
> 
> [Where problems could occur]
> 
> The fix is to safely guard register access as it should be. No potential
> problem identified so far.
> 
> [Other Info]
> 
> This should affects v6.1 or above as which AMD Phoenix are supported,
> and has been included in mantic-6.3 through stable-6.3.4 and in oem-6.1
> through stable-6.1.30, leaving lunar-6.2 unfixed only.
> 
> Alex Deucher (1):
>    drm/amdgpu/gfx11: update gpu_clock_counter logic
> 
> Tong Liu01 (1):
>    drm/amdgpu: refine get gpu clock counter method
> 
>   drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c | 28 ++++++++++++++++++++------
>   1 file changed, 22 insertions(+), 6 deletions(-)
> 
Acked-by: Tim Gardner <tim.gardner at canonical.com>
-- 
-----------
Tim Gardner
Canonical, Inc




More information about the kernel-team mailing list