* Chaskiel M Grundman [2005-04-25 14:14:25 -0400]:
> --On Monday, April 25, 2005 13:11:53 -0400 Derrick J Brashear
> <[EMAIL PROTECTED]> wrote:
>
> >I wonder if this is the inode semaphore the configure test was
> >(incompletely) added for in 1.3.81, maybe...
>
> The inotify stuff does not appear in stock 2.6.11.6 or 2.6.11.7. It does
> appear in genpatches-2.6-11.09-extras.tar.bz2 (4800_inotify-0.22-2.patch)
The kernel I saw this on does indeed include that patch. I'll try to
back it out (probably not before tomorrow evening, though).
My original description was inaccurate in a few respects: mounting /afs
works fine, it's the first attempt to open the directory afterwards that
oopses. And I hadn't noticed a third lock in crref()/crfree(). Here's
the relevant log excerpt:
Unable to handle kernel NULL pointer dereference at 0000000000000000
RIP:<ffffffff802f2ba7>{__down+103}
[...]
Call Trace:<ffffffff8861d413>{:libafs:crfree+3}
<ffffffff8012bb90>{default_wake_function+0}
<ffffffff88621497>{:libafs:afs_linux_open+55}
<ffffffff802f4073>{__down_failed+53} <ffffffff8018e388>{.text.lock.inotify+5}
<ffffffff80168c9d>{sys_open+109} <ffffffff8010d15a>{system_call+126}
inotify, indeed. You were correct. Thanks.
_______________________________________________
OpenAFS-info mailing list
[email protected]
https://lists.openafs.org/mailman/listinfo/openafs-info
