Red Hat Bugzilla – Bug 156636
[IT 54907] System oopsing in __get_lease()
Last modified: 2013-08-05 21:14:03 EDT
Kernel has oopsed several times. Netdump has capture the oops but not a complete vmcore. The oops below is representative. The changes documented in the following link are not present in the RH kernel. I feel the race identified by Stephen Rothwell involving the checking of whether inode->i_flock is not NULL and the subseqeunt use of that pointer is a likely explanation for these oopses. Kernel 2.4.9-e.49summit CPU: 0 EIP: 0010:[<c015a483>] Not tainted EFLAGS: 00010246 EIP is at __get_lease [kernel] 0x43 eax: 00000000 ebx: 00000000 ecx: 00000000 edx: c0692000 esi: 00000000 edi: 00000000 ebp: f4129080 esp: c0693efc ds: 0018 es: 0018 ss: 0018 Process cluster_smbd (pid: 21235, stackpage=c0693000) Stack: e23499a0 dfda3006 c01291fe 00002710 c01528b9 f4129080 ffffffff 00008001 c0693f7c c01540de f4129080 00008001 00000000 00000004 f1edd900 00000000 d8069180 bffec968 d8069180 bffec968 c0693f88 c015b291 f6f4decc 00008000 Call Trace: [<c01291fe>] in_group_p [kernel] 0x1e (0xc0693f04) pg. 7 [<c01528b9>] vfs_permission [kernel] 0x79 (0xc0693f0c) [<c01540de>] open_namei [kernel] 0x3de (0xc0693f20) [<c015b291>] fcntl_setlk64 [kernel] 0x1c1 (0xc0693f50) [<c01470d6>] filp_open [kernel] 0x36 (0xc0693f68) [<c01527fe>] getname [kernel] 0x5e (0xc0693f88) [<c01473d6>] sys_open [kernel] 0x36 (0xc0693fa0) [<c01073e3>] system_call [kernel] 0x33 (0xc0693fc0)
Created attachment 113943 [details] Patch to fix This is the final patch agreed upon in the IT. It's the least invasive fix.
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on the solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHSA-2005-529.html