[DRBD-user] 0.7 another Oops trace

Lars Ellenberg Lars.Ellenberg at linbit.com
Mon Apr 26 17:51:47 CEST 2004

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


/ 2004-04-26 19:29:10 +0400
\ Eugene Crosser:
> On Mon, 2004-04-26 at 19:02, Lars Ellenberg wrote:
> 
> > > This I got from today's CVS version when a would be secondary booted and
> > > tried to connect.  nfsa1 was primary, nfsa2 was freshly booted.  The
> > > block device apparently stopped responding.
> 
> I do run SMP (Hyperthreading Xeon)
> "cvs -Pd" does not hint on any unupdated files.
> I build monolithic kernel.

maybe you want to
 rm -r drivers/block/drbd include/kernel/drbd*.h
and do a new make kernel-patch.

the strange thing about the oops was, that stack backtrace shows we are
*sending*, still the oops itself was reported in inet_recvmsg, not
sendmsg. I do not understand this yet.

> BTW I am now getting *lots* of these messages from the master:

still, this "NULL" is strange. no, wait, probably on your Xeon the
sector_t is u64, and %ld only uses 4 bytes of the stack, next expected
is the string pointer, but the higher 4 byte of the sector u64 get used,
which are zero :)

anyways, I'll comment out these messages and commit in a minute.

	Lars



More information about the drbd-user mailing list