Re: Kernel panic on Google Pixel devices due to regulator patch

[Greg KH]

On Wed, Dec 18, 2019 at 04:18:06PM +0000, Mark Brown wrote:
> On Wed, Dec 18, 2019 at 03:22:19PM +0100, Greg KH wrote:
> > On Wed, Dec 18, 2019 at 01:11:14PM +0000, Mark Brown wrote:
> > > On Wed, Dec 18, 2019 at 01:21:57PM +0100, Greg KH wrote:
> 
> > > > It is, but it's the latest stable kernel (well close to), and your patch
> > > > was tagged by you to be backported to here, so if there's a problem with
> > > > a stable branch, I want to know about it as I don't want to see
> > > > regressions happen in it.
> 
> > > I don't track what's in older stable kernels, it wanted to go back at
> > > least one kernel revision but the issue has been around since forever.
> 
> > Ok, you can always mark patches that way if you want to :)
> 
> That's what a tag to stable with no particular revision attached to it
> is isn't it?

No, that means "drag it as far back as Greg can easily do it" :)

> > > If you don't want to be messing with timing luck then you probably want
> > > to be having a look at what Sasha's bot is doing, it's picking up a lot
> > > of things that are *well* into this sort of territory (and the bad
> > > interactions with out of tree code territory).  I personally would not
> > > be using stable these days if I wasn't prepared to be digging into
> > > something like this.
> 
> > I watch what his bot is doing, and we have tons of testing happening as
> > well, which is reflected by the fact that THIS WAS CAUGHT HERE.  This is
> 
> You don't have anywhere near the level of testing that you'd need to
> cover what the bot is trying to pull in, the subsystem and driver
> coverage is extremely thin relative to the enthusiasm with which things
> are being picked up.  All the pushback I see in review is on me for 
> being conservative about what gets pulled into stable and worrying about
> interactions with out of tree code.
> 
> > a sign that things are working, it's just that some SoC trees are slower
> > than mainline by a few months, and that's fine.  It's worlds better than
> > the SoC trees that are no where close to mainline, and as such, totally
> > insecure :)
> 
> What you appear to have caught here is an interaction with some
> unreviewed vendor code - how much of that is going on in the vendor
> trees you're not testing?  If we want to encourage people to pull in
> stable we should be paying attention to that sort of stuff.

I get weekly merge reports from all of the major SoC vendors when they
pull these releases into their tree and run through their full suite of
tests.  So I am paying attention to this type of thing.

What I need to figure out here is what is going wrong and why the SoC's
testing did not catch this.  That's going to take a bit longer...

thanks,

greg k-h