Re: 2.6.21-rc1: known regressions (part 2)

[Jens Axboe <jens.axboe@oracle.com>]

On Tue, Feb 27 2007, Jens Axboe wrote:
> On Tue, Feb 27 2007, Ingo Molnar wrote:
> > 
> > * Jens Axboe <jens.axboe@oracle.com> wrote:
> > 
> > > > > x60 doesn't resume from S2R either, it doesn't matter if 
> > > > > CONFIG_NO_HZ is set or not though. 2.6.20 worked fine.
> > > > 
> > > > It somehow works for me. As long as I do not play with bluetooth and 
> > > > suspend to disk...
> > > 
> > > It locks solid here on resume, going back to 2.6.20 makes it work 
> > > perfectly again. In between 2.6.20 and 2.6.21-rc1 some ACPI change 
> > > broke resume, but that got fixed. Some other change later snuck in 
> > > that broke it AGAIN for me, sigh.
> > > 
> > > I don't use bluetooth nor suspend to disk.
> > 
> > resume is stock on my T60 too. So you mean v2.6.21-rc1 vanilla works 
> > fine? Do you know a commit ID that works for sure? I'd like to bisect 
> 
> Nope, 2.6.21-rc1 vanilla does not work. 2.6.20 works. 2.6.20-gitX worked
> until some acpi change broke it, the below patch fixed that for me. That
> got merged in a later 2.6.20-gitY, but then some other patch broke it
> again so that 2.6.21-rc1 is broken. Not much luck there :-)
> 
> So it looks like:
> 
> - c5a7156959e89b32260ad6072bbf5077bcdfbeee broke 2.6.20-git
> - f3ccb06f3b8e0cf42b579db21f3ca7f17fcc3f38 should fix that.
> - Something later than f3ccb06f3b8e0cf42b579db21f3ca7f17fcc3f38 broke it
>   again.
> 
> > this, but this way i might just find that ACPI change that got already 
> > fixed later on (and then got re-broken).
> 
> Yeah, it gets trickier. I'll try
> f3ccb06f3b8e0cf42b579db21f3ca7f17fcc3f38 now and see if that works, then
> bisect to 2.6.21-rc1 to find the other offender. I hope the other
> offender didn't get added before
> f3ccb06f3b8e0cf42b579db21f3ca7f17fcc3f38, we'll see :-)

f3ccb06f3b8e0cf42b579db21f3ca7f17fcc3f38 works for me, starting bisect.

-- 
Jens Axboe