Droid 4 regression in 5.5-rc1, armsoc-soc tree
tony at atomide.com
Thu Jan 2 11:25:40 PST 2020
* Pavel Machek <pavel at ucw.cz> [200102 19:18]:
> > * Arnd Bergmann <arnd at arndb.de> [200102 14:48]:
> > > On Thu, Jan 2, 2020 at 3:06 PM Pavel Machek <pavel at ucw.cz> wrote:
> > >
> > > > # good: [ab818f0999dc73af3f966194d087e9f6650f939f] Merge tag
> > > > # 'omap-for-v5.5/maintainers-signed' of
> > > > # git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap into
> > > > # arm/soc
> > > > git bisect good ab818f0999dc73af3f966194d087e9f6650f939f
> > > > # first bad commit: [38206c24ab09b4f4c2a57de5c1af0bb2e69cf5b6] Merge tag
> > > > # 'armsoc-soc' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
> > > >
> > > > ...and am not sure how to interpret the result, or what to do next :(.
> > >
> > > This means that either the bisection is unreliable (it might be 100%
> > > reproducible or you mixed up good/bad once), or that the arm-soc
> > > branch was indeed good by itself but had a conflict with some other
> > > change that torvalds merged earlier.
> > >
> > > I would suggest to go back to the list I had extracted in my previous
> > > email and revert all the commits that I identify there, on top of
> > > linux-5.5-rc4. If some commit is not easily reverted, skip that one
> > > at first.
> > >
> > > Ideally this gets you a working kernel based on -rc4, and then
> > > you can bisect that to find which revert fixed it.
> > Also please check your dtb is up to date. We still have a lot of
> > stuff missing from the device tree meaning an old dtb will likely
> > be incomplete for many devices.
> Aha, so it works ok when using matching kernel + dtb.
OK good to hear.
> Is 5.5 kernel with 5.3 dtb situation worth debugging?
Well producing a warning based on something would be good. Not sure
what the logic would be though, checking for some minimum dtb version
required would have to be bumped until we have complete dts data.
More information about the linux-arm-kernel