Re: kernel regression on 164UX
(You might want to turn off HTML mail for mailing list posts.)
On Tue, Aug 3, 2010 at 5:56 PM, Ian Las <ialas@yahoo.com> wrote:
> Completed initial tests.
> Last working kernel for 164Ux is 2.6.22.19
> First broken kernel is 2.6.23
> Installed Git package.
> Need some help how to continue:
> As linux-2.6.23 is older than 2.6.22.19 would it be possible to specify 2.6.22.19 as good and 2.6.23 as bad ?
> It has a higher revision number but is actually older !
> Or is it necessary to start with a good older revision which would be 2.6.22.9 ?
> Do I need the git repository from " git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git " ?
> What would be the best starting point ?
>
> Ian
Hi Ian,
Yes, the first step is `git clone
git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git`
If 2.6.22 is working, and 2.6.23 is not working, we can bisect this to
find the breaking commit. (2.6.22.19 -> 2.6.23 is not a linear change,
as they both came from 2.6.22, and not one from another.)
So from your linux-2.6 git directory, you would do
git bisect start
git bisect good v2.6.22
git bisect bad v2.6.23
and then it'll guide you through the process. If you're able to boot
an intermediate kernel, you'll run `git bisect good` and if you're not
able to boot a particular revision, you'll run `git bisect bad` until
you find the exact commit that caused the problem.
Thanks for doing this. Let me know if there's anything I can do to help.
Matt
Reply to: