On Fri, 2009-06-12 at 12:24 -0400, Arnout Boelens wrote: > Package: linux-image-2.6.26-2-amd64 > Version: 2.6.26-15lenny3 > Severity: normal > > > Every once I a while I get kernel oops on one of my computers and find the > following messages back in my syslog: > > Jun 12 10:42:03 yama kernel: [848178.269002] BUG: unable to handle kernel > paging request at 000000002f467780 > Jun 12 10:42:03 yama kernel: [848178.269002] IP: [<000000002f467780>] [...] > Jun 12 10:42:03 yama kernel: [848178.269002] Call Trace: > Jun 12 10:42:03 yama kernel: [848178.269002] [<ffffffffa031f9ab>] ? > :sunrpc:rpcauth_lookup_credcache+0xae/0x1e9 [...]
That corresponds to this line in rpcauth_lookup_credcache(): if (!entry->cr_ops->crmatch(acred, entry, flags)) The function pointer entry->cr_ops->crmatch has turned out not to be a valid function address, and since entry->cr_ops should be pointing to a static constant structure which cannot itself be corrupted this suggests that *entry has somehow been corrupted. There is an upstream change that might deal with this issue, but I'm not convinced that it's related: commit 5f707eb429e2c98dfd564ffbbd9f536bf493d869 Author: Trond Myklebust <trond.mykleb...@netapp.com> Date: Tue Oct 28 15:21:42 2008 -0400 SUNRPC: Fix potential race in put_rpccred() Ben. -- Ben Hutchings Never attribute to conspiracy what can adequately be explained by stupidity.
signature.asc
Description: This is a digitally signed message part