Hi
(Apologies for replying in a new thread and crossposting, I'm not
subscribed to these lists and maybe this will make for useful reading.
Also please CC me in replies)
wrt Jiri Kosina's post: alpine.LNX.1.10.0809260042380.3389 () pegasus !
suse ! cz - I hit this bug at the end of August after upgrading to
Ubuntu Intrepid (which was alpha at the time).
Perhaps my timeline[0] would help provide some information to rule
potential culprits out or in by way of Ubuntu's code at the time.
Also I have some logs[2-5] that cover the period surrounding the
corruption happening.
The timeline shows a <24 hour window from my upgrading to Ubuntu
Intrepid (and thus getting a 2.6.27 kernel), and the damage actually
happening.
In conjunction with pages like
https://launchpad.net/ubuntu/+source/linux and
https://launchpad.net/ubuntu/+source/xserver-xorg-video-intel it should
be possible to figure out exactly what source I had in use.
My reading of those pages suggests I had this .config[1]
(which makes sense, it's the oldest 2.6.27 one I have installed and
there doesn't seem to have been a subsequent publishing of an Ubuntu
kernel within the failure window). It also suggests I had
xserver-xorg-video-intel version2:2.4.1-1ubuntu1 (I would not
necessarily trust the version number cf the vanilla intel driver -
tracking down the source package in Launchpad or via #ubuntu-x on
freenode would show exactly which patches were applied). Similarly,
libgl1-mesa-dri 7.1-1ubuntu1.
Apart from (obviously) any occurances where they failed to be written to
disk, I have kern.log[2] and messages[3] for the period (although not
syslog itself, that gets rotated out too quickly, unfortunately, but I
expect kern.log will have any interesting stuff there is to be found).
The mail in [0] references a kernel WARNING bug on Launchpad which ought
to be in kern.log, but just in case it isn't, that bug has it attached.
I also have logs[4][5] which show all package activity before and after
the point of ...
