Stable kernels 2.6.32.10 and 2.6.33.1 [LWN.net]

Stable kernels 2.6.32.10 and 2.6.33.1

Posted Mar 16, 2010 1:49 UTC (Tue) by maks (subscriber, #32426) [Link]

the sync is relative, but better then it used to be.
Indeed stable@k.o is pretty easy to feed and many distro devs use that,
but it can be improved. XFS is known fucked in 2.6.32. EXT4 quota afaik also still stumbles around. iwlagn can allocate mem up to order 4, which sucks on many laptops..

Fedora is already shipping DRM 2.6.34-rc1.
Debian and Ubuntu already pulled DRM up to 2.6.33.1.

Fedora adds some dvb and other patches like lirc see:
http://cvs.fedoraproject.org/viewvc/rpms/kernel/F-12/

Ubuntu has a mechanism to add newer modules.
afair they shipp seperate newer iwlwifi, similar wireless drivers and .. separate to their linux-2.6 tree.

Despite those criticism all in all 2.6.32 is shaping up nicely. :)

Stable kernels 2.6.32.10 and 2.6.33.1

Posted Mar 16, 2010 13:33 UTC (Tue) by lwkejrlej (guest, #64237) [Link]

Despite those criticism all in all 2.6.32 is shaping up nicely. :)

Indeed, and I'm grateful (looking forward to RHEL 6). However, seeing as we're at the 2.6.34 cycle now, it's quite amazing at the amount of bugs being picked up in 2.6.32. This leads to two interesting questions:

1. Why weren't the bugs picked up during the "stabilisation" phase of 2.6.32?
2. Given the amount of bugs in 2.6.32.0, how much faith should we have in 2.6.34 (when it's released), or in fact any other following kernel ?

Stable kernels 2.6.32.10 and 2.6.33.1

Posted Mar 16, 2010 16:11 UTC (Tue) by proski (subscriber, #104) [Link]

That's mostly bugs that affect few users. Also, most fixes are not for regressions. That is, they fix something that never worked before. Many patches are a side effect of the ongoing development. That is, simple self-contained changes that may affect some users are committed to the stable kernels in addition to the bleeding edge. Just look at the changelog to see what I mean.

Actually, some fixes are not for bugs and I don't think they should be part of the stable series. For instance:

USB: ftdi_sio: sort PID/VID entries in new ftdi_sio_ids.h header

No users should suffer in any way because PID/VID entries are not sorted.

Stable kernels 2.6.32.10 and 2.6.33.1

Posted Mar 18, 2010 23:04 UTC (Thu) by nix (subscriber, #2304) [Link]

I suspect that sort of thing goes in because it's needed for later patches
to apply unchanged.

Stable kernels 2.6.32.10 and 2.6.33.1

Posted Mar 16, 2010 15:11 UTC (Tue) by jengelh (subscriber, #33263) [Link]

>XFS is known fucked in 2.6.32.

Link?

Stable kernels 2.6.32.10 and 2.6.33.1

Posted Mar 16, 2010 15:20 UTC (Tue) by maks (subscriber, #32426) [Link]

conversation on stable@kernel.org about xfs state and possible fs corruptions plus first band aid fedora patch:
http://cvs.fedoraproject.org/viewvc/rpms/kernel/F-12/xfs_...

afais current git stable-queue is not yet fed with patches by them.