On Fri, Dec 31, 2021 at 11:30:34AM +0100, Christian Brauner wrote: > On Thu, Dec 30, 2021 at 03:14:33PM -0800, Linus Torvalds wrote: > > On Thu, Dec 30, 2021 at 11:23 AM Christian Brauner > > <christian.brauner@xxxxxxxxxx> wrote: > > > > > > Would you be ok with applying this fix directly? I > > > > Done. > > > > That said, I would have liked a "Fixes:" tag, or some indication of > > how far back the stable people should take this.. > > Ugh, I missed to add that. > >From a pure upstream stable perspective the only relevant and still > supported kernel that should get this fix is 5.15. > > I can make it a custom to mark all patches that should go to stable with > the first kernel version where a given fix should be applied. In this > case this whould've meant I'd given it: > > Cc: <stable@xxxxxxxxxxxxxxx> # v5.12+ > > For upstream stable maintainers it should be clear that since the only > supported stable version within the range is v5.15. > For downstream users/distros it should help to identify whether they > still run/maintain a kernel that falls within the range of kernels that > would technically be eligible for this fix. > > I haven't seen whether we prefer the Cc: with # v*.**+ syntax to a > simple Cc: without it nowadays. If you do the # v.** syntax, or the Fixes: tag, then I know exactly how far back to backport things. If a failure occurs in backporting to a listed place, then I will send you a FAILED email. If there is no such marking, then I just have to guess myself and if I start to get conflicts on older kernels, I just stop and do not send a FAILED email if it does not look obviously relevant to me. I'll just queue this up for 5.15, thanks. greg k-h
