On Thu, Sep 16, 2021 at 04:05:05PM +0000, Deucher, Alexander wrote:
[Public]
-----Original Message----- From: Greg KH gregkh@linuxfoundation.org Sent: Thursday, September 16, 2021 11:36 AM To: Deucher, Alexander Alexander.Deucher@amd.com Cc: Michel Dänzer mdaenzer@redhat.com; Koenig, Christian Christian.Koenig@amd.com; Quan, Evan Evan.Quan@amd.com; Lazar, Lijo Lijo.Lazar@amd.com; stable@vger.kernel.org Subject: Re: FAILED: patch "[PATCH] drm/amdgpu: Cancel delayed work when GFXOFF is disabled" failed to apply to 5.14-stable tree
On Thu, Sep 16, 2021 at 02:42:42PM +0000, Deucher, Alexander wrote:
[Public]
-----Original Message----- From: Michel Dänzer mdaenzer@redhat.com Sent: Thursday, September 16, 2021 9:55 AM To: Greg KH gregkh@linuxfoundation.org Cc: Deucher, Alexander Alexander.Deucher@amd.com; Koenig, Christian Christian.Koenig@amd.com; Quan, Evan Evan.Quan@amd.com; Lazar, Lijo Lijo.Lazar@amd.com; stable@vger.kernel.org Subject: Re: FAILED: patch "[PATCH] drm/amdgpu: Cancel delayed work when GFXOFF is disabled" failed to apply to 5.14-stable tree
On 2021-09-16 15:48, Greg KH wrote:
On Thu, Sep 16, 2021 at 03:39:16PM +0200, Michel Dänzer wrote:
On 2021-09-16 15:05, gregkh@linuxfoundation.org wrote: > > The patch below does not apply to the 5.14-stable tree. > If someone wants it applied there, or to any other stable or > longterm tree, then please email the backport, including the > original git commit id to stable@vger.kernel.org.
It's already in 5.14, commit
32bc8f8373d2d6a681c96e4b25dca60d4d1c6016.
Odd, how were we supposed to know that?
Looks like the fix was merged separately for 5.14 and 5.15. I don't know how/why that happened. Alex / Christian?
The fix already landed in drm-next, but since this was before the merge window, we wanted to make sure the fix also landed in stable, so I cherry-picked it to 5.14. I'm not sure of a better way to handle these sort of cases.
You gotta give me a chance to know what happened. As the drm developers do this a lot, they have been putting the "cherry picked from" line in the patch because they can not merge between branches and keep the git id the same.
So try using the '-x' option to 'git cherry-pick' next time?
I can do that, but historically, I've gotten a lot of flack for that because the commit doesn't exist in Linus' tree yet at that point as the merge window hasn't opened yet.
I totally agree, and would argue that this is NOT the way to do it either. But it seems like this is what the i915 developers have been doing for a few years now and honestly, it's better than nothing.
But I would like a real solution here, this gets very annoying every -rc1 cycle for the DRM subsystem. You all are the only ones with this problem for some strange reason :(
greg k-h