[Fwd: moradin 2006-08-02 11:02 System Events]
Subsystems (2)
Hello!
Rebooting one of our NFS file servers with 2.6.17.7, I just got the following OOPS:
BUG: unable to handle kernel NULL pointer dereference at virtual address
00000041
printing eip:c01fe571
*pde = 00000000
Oops: 0000 [#1] SMP Modules linked in: parport_pc lp parport nfsd exportfs binfmt_misc autofs4 thermal ipv6 8021q ide_cd 8250_pnp 8250 serial_core floppy ohci_hcd usbcore i2c_piix4 i2c_core ipmi_si ipmi_watchdog ipmi_devintf ipmi_msghandler CPU: 1 EIP: 0060:[<c01fe571>] Not tainted VLI EFLAGS: 00010246 (2.6.17.7-station #1)
EIP is at _atomic_dec_and_lock+0x9/0x48
eax: 00000041 ebx: 00000041 ecx: 00000001 edx: c041b280
esi: c041b280 edi: f6f675c0 ebp: f7636dc0 esp: f7636db8
ds: 007b es: 007b ss: 0068
Process nfsd (pid: 1964, threadinfo=f7636000 task=f7e22a70)
Stack: 00000041 f8b8d958 f7636dcc c0351d77 f6f675c0 f7636dd8 f8b8d992 f6f675cc f7636de8 c01ff767 f8bb1500 f8baf260 f7636e08 c0353df6 f6c42340 f8bb1500 f7636e2c 00000000 00000080 00000008 f7636e24 f8b8def2 00000008 f7636e2c Call Trace: <c01034a0> show_stack_log_lvl+0x88/0x90 <c01035fc> show_registers+0x115/0x17d <c01037ed> die+0x115/0x1ee <c0112644> do_page_fault+0x466/0x555 <c0103147> error_code+0x4f/0x54 <c0351d77> auth_domain_put+0x10/0x3f <f8b8d992> expkey_put+0x3a/0x44 [nfsd] <c01ff767> kref_put+0x66/0x77 <c0353df6> sunrpc_cache_lookup+0xbb/0xf5 <f8b8def2> svc_expkey_lookup+0xa1/0xaf [nfsd] <f8b8e635> exp_find_key+0x5c/0x81 [nfsd <f8b8efd1> exp_find+0x11/0x75 [nfsd] <f8b89e6e> fh_verify+0x15d/0x470 [nfsd] <f8b8b57c> nfsd_open+0x21/0x14d [nfsd] <f8b8be1b> nfsd_read+0x6f/0xa9 [nfsd] <f8b91c58> nfsd3_proc_read+0xdb/0x137 [nfsd] <f8b8865a> nfsd_dispatch+0xca/0x190 [nfsd] <c034f93f> svc_process+0x393/0x5c4 <f8b88437> nfsd+0x189/0x2e2 [nfsd] <c0100e31> kernel_thread_helper+0x5/0xb Code: 0f a3 0d 78 ae 48 c0 19 c0 85 c0 75 10 89 c8 89 da e8 8a ff ff ff
83 f8 03 89 c1 7e e3 5b 89 c8 5d c3 55 89 e5 56 89 d6 53 89 c3 <8b> 10 83 fa 01 74 11 8d 4a ff 89 d0 f0 0f b1 0b 39 d0 74 04 89
EIP: [<c01fe571>] _atomic_dec_and_lock+0x9/0x48 SS:ESP 0068:f7636db8
Doing a quick search did't reveal any similar reported bug. I'm not subsribed to the nfs-ml, so please cc: me on replies.
BYtE
Philipp
Signature · 14 lines
Dipl.-Inform. [email protected] Abteilung Systemsoftware und verteilte Systeme, Fk. II Carl von Ossietzky Universitaet Oldenburg, 26111 Oldenburg, Germany http://www.svs.informatik.uni-oldenburg.de/contact/pmhahn/ Telefon: +49 441 798-2866 Telefax: +49 441 798-2756 ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ NFS maillist - [email protected] https://lists.sourceforge.net/lists/listinfo/nfs
Quoted 4 lines · On Wednesday August 2, [email protected] wrote:
Hello!
Rebooting one of our NFS file servers with 2.6.17.7, I just got the following OOPS:
Thanks for the report.
The bug was fairly easy to find and fix.
I think this would be appropriate for the next 2.6.17.x stable kernel, and definitely for 2.6.18. (hence 'akpm' and 'stable' cc:ed).
It is not relevant for earlier kernels (e.g. 2.6.16).
Patch was made against 2.6.18-rc2-mm1, but applies equally to 2.6.17.7.
Thanks again,
NeilBrown
---------------------------------------------
Fix race related problem when adding items to and svcrpc auth cache.
If we don't find the item we are lookng for, we allocate a new one, and then grab the lock again and search to see if it has been added while we did the alloc.
If it had been added we need to 'cache_put' the newly created item that we are never going to use. But as it hasn't been initialised properly, putting it can cause an oops.
So move the ->init call earlier to that it will always be fully initilised if we have to put it.
Thanks to Philipp Matthias Hahn <[email protected]> for reporting the problem.
### Diffstat output
./net/sunrpc/cache.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)
net/sunrpc/cache.c +5 -1
@@ -71,7 +71,12 @@ struct cache_head *sunrpc_cache_lookup(s new = detail->alloc ();if (!new)return NULL ; +/* must fully initialise 'new', else +* we might get lose if we need to +* cache_put it soon. +*/ cache_init (new); + detail->init (new, key);write_lock (&detail->hash_lock);
@@ -85,7 +90,6 @@ struct cache_head *sunrpc_cache_lookup(sreturn tmp; } } - detail->init (new, key); new->next = *head; *head = new; detail->entries++;
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
Greg KH
AppliedQuoted 11 lines · Neil Brown
Quoted 4 lines · On Wednesday August 2, [email protected] wrote:
Hello!
Rebooting one of our NFS file servers with 2.6.17.7, I just got the following OOPS:
Thanks for the report.
The bug was fairly easy to find and fix.
I think this would be appropriate for the next 2.6.17.x stable kernel, and definitely for 2.6.18. (hence 'akpm' and 'stable' cc:ed).
Quoted 4 lines
It is not relevant for earlier kernels (e.g. 2.6.16).
Patch was made against 2.6.18-rc2-mm1, but applies equally to 2.6.17.7.
Queued to -stable.
thanks,
greg k-h