https://bugzilla.novell.com/show_bug.cgi?id=440891
User teheo@novell.com added comment
https://bugzilla.novell.com/show_bug.cgi?id=440891#c30
Tejun Heo changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|REOPENED |NEEDINFO
Info Provider| |dave.plater@yahoo.co.uk
--- Comment #30 from Tejun Heo 2009-01-16 01:32:18 MST ---
Hello,
Okay, I think now I understand the problem. Your optical drive is experiencing
a lot of timeouts and L-EC uncorrectable media errors, which basically means
that the drive is having a lot of problem reading data off the media. As the
error rate is very high, this causes various interesting problems.
1. As the PATA channel is shared by the harddrive and the cdrom, while the
cdrom is timing out and failing the drive isn't accessible, so the system will
be very unresponsive in general.
2. The frequent IO errors trigger various error conditions on upper layers.
isofs will complain that the media is gone while the filesystem is mounted and
so on.
3. The kernel had problem handling timeout conditions which was fixed by the
patch posted here. The frequent timeouts triggered this condition and that's
why you saw the WARNING message. This bug is fixed but the underlying hardware
problem still remains.
4. ata_piix is generally good at recovering from such error conditions but it's
not uncommon for IDE controllers to lock the whole machine up after timeout or
other failures especially when its FIFO status isn't matched by what the driver
thinks it is. This usually manifests as the IDE controller hanging while
holding the PCI bus. When the machine falls into this state, nothing can
really help. CPU simply cannot access the memory or IO devices. So, only
hardreset would work. This is why you saw the hard lock up. Maybe there's
some room for improvement regarding how libata EH tries to recover the
controller after such timeouts but in certain chipsets (especially the old
ones) sometimes it's just how the hardware is built. Some of them simply can't
recover from certain type of failures. Intel ones are one of the best behaving
ones tho. Can you please post the output of "lspci -nn"?
At any rate, the baseline is that your optical drive is having a LOT of problem
doing things that it's supposed to. It could be that the drive is failing or
it simply doesn't like the media it's being fed. The failure pattern differs
according to the type of media, right?
--
Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.