Messages in this thread |  | | From | "Petr Vandrovec" <> | Date | Thu, 7 Dec 2000 00:14:34 MET-1 | Subject | Re: test12pre6: BUG in schedule (sched.c, 115) |
| |
On 6 Dec 00 at 14:00, Mike Kravetz wrote: > > Are you sure that was line 115? Could it have been line 515? > Also, do you have any Oops data? >
Yesterday BUG in schedule at 515 happened for me with test12-pre4. As there were no data, and it was followed by
NMI detected lockup on CPU24 ...
(without any further data... no stack trace), I throwed it away as some -pre4 memory corruption. Maybe it was not random corruption, but stack overflow? Of course, my machine does not have 25 CPUs, but only two PIII...
And no, I have no oops data. I was reviewing Debian changes (apt-listchanges) when picture scrolled up one line (bug in sched) and then, after 5 seconds, another one line, with 'NMI detected'... And that was all, time to hit reset.
It was very patched test12-pre4, it contained Al Viro & Sct dirty buffer patches, Andrew Morton's exec_usermodehelper, and couple of other patches, brutto 153KB unzipped unified diff.
Filesystem ext2, 256MB RAM, dual PIII/450, kernel compiled with Debian's gcc-2.95.2, matroxfb. On background there were Debian's XF4.0.1-9 running with Gnome and DRI, as I was tracking why that beast reprograms matrox hardware every second (or every minute, depending on format of clock in left upper corner; but that's another story) even if X are not visible... (btw that's reason, why matroxfb does not work correctly with XF4 & DRI...) Best regards, Petr Vandrovec vandrove@vc.cvut.cz
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/
|  |