[Ksummit-2012-discuss] [ATTEND] stable kernel stuff and grumpy maintainers

Dave Jones davej at redhat.com
Tue Jun 19 21:21:42 UTC 2012


On Tue, Jun 19, 2012 at 10:49:24AM -0700, Greg Kroah-Hartman wrote:

 > > Is it a case that the distro tree fixes are not core patches or are they
 > > backports that are not getting resubmitted for -stable?
 > 
 > It's a case that distros are putting patches into their trees that are
 > backports of bugfixes that are upstream, and are not letting me know
 > that they should be included in the stable tree.
 > 
 > Debian is the exception, they do a wonderful job, thanks to Ben and
 > others.  Fedora and Canonical are getting better, but could still
 > improve, a bunch.

It's the usual story of just being utterly buried in bugs.

For the most part, when we identify an upstream patch that needs backporting,
we'll poke the relevant upstream maintainer to send it along.
Like with everything, there are exceptions, notably where we're not sure
if a particular backport actually fixes a problem until we have confirmation etc.

There's also a case where stuff falls through the cracks when we rebase too.
Once we've moved F16 to 3.4 for example, we immediately stop caring about 3.3 and earlier.
If we had 3.3 specific patches in our tree at the point of the rebase that are now
in 3.4, we just drop them.  A release or two ago, one of the Debian guys picked
up a bunch of those iirc. Some of those should have been pushed to stable sooner for sure.

And then of course, sometimes.. we just suck. We're trying to fix that part.
You have our email addresses. Holler if something doesn't look right.

	Dave



More information about the Ksummit-2012-discuss mailing list