[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#705168: Any multi VGA controller systems failing to boot?



On Tue, 2013-05-07 at 15:03 -0400, Barry Fishman wrote:
> On 2013-05-07 13:20:02 EDT, Ben Hutchings wrote:
> > On Tue, May 07, 2013 at 11:14:23AM -0400, Barry Fishman wrote:
> >> 1) This bug is a month old and due to a Debian patch which was rejected
> >>    by the upstream kernel team.
> >
> > Perhaps you should tell us which that is, then.
> 
> You told me.  The "fbcon locking fixes" in linux kernel mailing list.

That was not a Debian patch, nor was it rejected by the upstream kernel
developers, nor was it the cause of the bug.

> >> 3) Its very likely the same bug as #704933, although your bug tracking
> >>    system seems unable to link related bugs together.
> >
> > Instructions for using the BTS are at http://bugs.debian.org
> 
> I should have updated the "affects" and "merge" attribute of the bug.  I
> just didn't consider that I the proper one to do triage.  I'm surprised it
> did not require someone with at least some history within Debian.

So, you decided that the Debian BTS is 'unable to link related bugs
together' despite not having read any documentation or tried to make any
changes.

> >> 4) Its not listed in the release notes for Wheezy as a problem
> >>    installing on systems with multiple VGA controllers.
> >
> > We can't list every bug in the release notes.
> >
> >> 5) I had been running Wheezy for about a year, and the only issue I had
> >>    with it occurred a month before it released, without the problem
> >>    being considered significant to apply a known fix, or even tag it
> >>    correctly, although it makes some systems unbootable.
> >
> > What known fix?  You never even followed up to say whether the
> > possible fix I found actually worked.

Sorry, some confusion here: I wrote to bug #704933 and never looked at
bug #705168 before today.

> What fix was that?  If you wanted to validate a fix you should have
> pointed me at a .deb file.  I was not sure how to follow your comments
> short of building a kernel myself.  Although I have pulled and built a
> variety source packages, I have not done this with the kernel for
> decades, and did not want to relearn how - while doing bug tracking.

Here's how to test a patch against the Debian Linux kernel package:
http://kernel-handbook.alioth.debian.org/ch-common-tasks.html#s-common-official

> > (Why is this on debian-devel?  It won't get your bug fixed any
> > faster, in fact it is worse because your message wasn't logged
> > on the bug.)
> 
> The message was not directed at fixing the bug, but in understanding the
> bug resolution process.  My first message here was about trying to see
> if others had the same issue.
> 
> I thought there was some issues in the bug tracking process that could be
> fixed.  The same bug resolution went though Ubuntu without much fuss,
> even though Debian produced a resolution first, Ubuntu got it available
> right away, and saw is was available in the upcoming release.

Unfortunately, there is at least a 2 week delay between uploading a new
kernel and getting it into installer images.  There should be a kernel
update through wheezy-updates Real Soon Now, but a lot of fixes have
missed the initial release.

> The most frustrating part is that Debian overwrote a kernel with new
> kernel version.  This makes rolling back from a bad kernel overly
> difficult.  Why should kernel versions 3.2.39 and 3.2.41 be installed
> under the same name?

Because if we change the kernel package name too often, your disk will
fill up with old kernel images.  I would like to change this, but it
won't be easy.

Ben.

-- 
Ben Hutchings
For every action, there is an equal and opposite criticism. - Harrison

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: