[Bug 241152] New: e1000 driver in thinkpad x60 fails on unsuspend with irq conflict
https://bugzilla.novell.com/show_bug.cgi?id=241152 Summary: e1000 driver in thinkpad x60 fails on unsuspend with irq conflict Product: openSUSE 10.2 Version: Final Platform: x86-64 OS/Version: Other Status: NEW Severity: Normal Priority: P5 - None Component: Kernel AssignedTo: kernel-maintainers@forge.provo.novell.com ReportedBy: wstephenson@novell.com QAContact: qa@suse.de About 50% of the time my X60 fails after coming out of suspend-to-ram, with an irq conflict. It seems to be the same problem in this lkml thread: http://lkml.org/lkml/2006/10/20/74 -- 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, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=241152 ------- Comment #1 from wstephenson@novell.com 2007-02-01 09:24 MST ------- Created an attachment (id=116870) --> (https://bugzilla.novell.com/attachment.cgi?id=116870&action=view) module debug, /proc/interrupts, uname -a, hwinfo output -- 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, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=241152 trenn@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED ------- Comment #3 from trenn@novell.com 2007-02-05 05:21 MST ------- Will, it would be great if you could search a bit on your own on that one. According to above link pci=nomsi should be the workaround you search for? If it works for you, you might want to follow that up a bit mainline and help them testing or, if there is a patch falling out that works for you, you should post it on kernel list if it's small, but I doubt it is a good idea to backport msi fixes if it's not really small (for fixing suspend to ram on X60 in 50%) -> too risky to break other machines. -- 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, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=241152 trenn@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |behlert@novell.com, bk@novell.com ------- Comment #4 from trenn@novell.com 2007-02-28 05:32 MST ------- X60 is an important machine. This probably will be a won't fix for 10.2, still we should make sure it gets fixed mainline. But according to from Andrew Morton (from link from comment #1):
hm, people report that occasionally - I don't think anyone knows what's causing it. we could waste here a lot of time for nothing...
Not sure whether this one should get closed as won't fix or whether we should wait whether it gets fixed mainline or whether we should spend the time... In the latter case I need such a machine (but do not really have the time maybe Bernhard Kaindl who fixed an mtrr problem for suspend recently and already knows a lot about the suspend process can have a look at it?). Stefan? -- 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, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=241152 bk@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO Info Provider| |wstephenson@novell.com ------- Comment #6 from bk@novell.com 2007-02-28 06:10 MST -------
From reading the last mail of the thread linked in Comment #1, this is solved in 2.6.19-rc2: http://lkml.org/lkml/2006/10/20/88
In http://lkml.org/lkml/2006/10/19/179 Eric W. Biederman says about the MSI IRQ allocator in 2.6.18:
The backwards msi irq allocator in 2.6.18 is so convoluted it may have a corner case where it fails, and that is triggering this mess.
So indeed, fixing this on 10.2 would require to replace the MSI IRQ allocator, and I don't think it's worth do do that when the simple workaround of pci=nomsi works and it's fixed in 10.3. Even for SLED10, pci=nomsi sould be acceptable and we could do this automatically for the X60, based on DMI. Will or anybody with an X60 could verify the above by adding pci=nomsi to their kernel agruments in /boot/grub/menu.lst and verify the real fix by installing the kernel from any 10.3 Alpha or the kernel of the day (KOTD). Will, can you do this? -- 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, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=241152 ------- Comment #7 from wstephenson@novell.com 2007-02-28 06:57 MST ------- I am testing pci=nomsi at the moment -- 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, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=241152 wstephenson@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED Info Provider|wstephenson@novell.com | ------- Comment #8 from wstephenson@novell.com 2007-05-24 12:06 MST ------- Haven't seen this problem in 10.3 at all. pci=nomsi helps on 10.2 -- 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, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=241152#c9
Thomas Renninger
participants (1)
-
bugzilla_noreply＠novell.com