[PATCH v2 00/12][SRU][OEM-5.14]Fix i915 TypeC disconnect problems for Intel ADL-P
Andrea Righi
andrea.righi at canonical.com
Fri Nov 26 08:08:06 UTC 2021
On Fri, Nov 26, 2021 at 04:04:31PM +0800, Chris Chiu wrote:
> On Fri, Nov 26, 2021 at 3:55 PM Andrea Righi <andrea.righi at canonical.com> wrote:
> >
> > On Thu, Nov 25, 2021 at 06:10:13PM +0800, Chris Chiu wrote:
> > > BugLink: https://bugs.launchpad.net/bugs/1952041
> > >
> > > [Impact]
> > > When the ADL-P system connects the external display via TypeC port, it will hang after unplugging the TypeC connector. The system will never come back until reboot.
> > >
> > > [Fix]
> > > Intel has released a patch set to fix the TypeC PHY connect/disconnect logic. The shift for ownership of PHY and power domain will be handled correctly for ADL-P.
> > >
> > > [Test]
> > > The ADL-P system will no longer freeze and the ownership will shift correctly after disconnecting the external display connects via TypeC port.
> > >
> > > [Where problem could occur]
> > > It's kind of a big refactor for the i915 TypeC PHY handling logic. Don't know if there's any problems on older platforms. Targeting only on Jammy/Unstable and latest OEM kernel for lower risk.
> > >
> > > v2: remove SRU for U since #4 of this series should be cleanly cherry-picked to U and J.
> >
> > I was considering to apply this also to jammy:linux, but "big refactor"
> > is a little worrying in terms of maintainability of the i915 code, that
> > always gets problematic due to the inevitable custom patches that we
> > usually apply...
> >
> > Do you think there's any chance that we get this via the regular updates
> > from stable? Maybe if it wasn't not done already we should try to ask to
> > apply this to stable.
> >
> > Thanks,
> > -Andrea
>
> The patch set is in 5.15.0-rc4 now so the regular update would also get
> these patches applied. My major concern is these patches are too new
> for older platforms (ex. TGL) which we haven't yet time for verification and
> I don't want to incur huge impact on older kernels for this.
Ok, in that case I'd suggest to hold on a little bit for jammy:linux
(since we're pushing to get a kernel promoted to the release pocket). If
we get these patches with the regular updates fine, otherwise we can
apply these patches once we have a "more stable" jammy:linux.
Thanks,
-Andrea
More information about the kernel-team
mailing list