CentOS - Oct 2011 - Centos 5.7 - bridging firewall crash (only since the update) - dmesg output included - Trying to vfree() nonexistent vm area (ffffffff885d9000) ....

Hi.

We have several Centos 5 firewalls in our company  - they are all in
bridging mode.

We updated them to the latest Centos 5.x last night.

4 hours later there was a crash..  (previously it had never crashed - and
had been running for at least 1.5 yrs)

The crash occurred when someone added a rule, after rebooting it was fine  -
the thing is it can't have been a bad rule as the first thing the server
does is load the firewall (same) rules..

In /var/log/messages I can see:-

--------------------------------------------------------------------------------------------
Oct 13 09:27:46 new kernel: Trying to vfree() nonexistent vm area
(ffffffff885d9000)
Oct 13 09:27:46 new kernel: WARNING: at mm/vmalloc.c:329 __vunmap()
Oct 13 09:27:46 new kernel:
Oct 13 09:27:46 new kernel: Call Trace:
Oct 13 09:27:46 new kernel:  [<ffffffff800a8b6b>]
sys_delete_module+0x1b3/0x1d4
Oct 13 09:27:46 new kernel:  [<ffffffff8005d28d>] tracesys+0xd5/0xe0
Oct 13 09:27:46 new kernel:

--------------------------------------------------------------------------------------------

Followed by (a few times)

--------------------------------------------------------------------------------------------
Oct 13 09:27:46 new kernel: WARNING: at mm/vmalloc.c:97 vmap_pte_range()
Oct 13 09:27:46 new kernel:
Oct 13 09:27:46 new kernel: Call Trace:
Oct 13 09:27:46 new kernel:  [<ffffffff884e7000>]
:iw_cm:iw_cm_init_qp_attr+0x0/0xb1
Oct 13 09:27:46 new kernel:  [<ffffffff800d3e04>] map_vm_area+0x216/0x306
Oct 13 09:27:46 new kernel:  [<ffffffff80001ff8>] _stext+0xff8/0x1000
Oct 13 09:27:46 new kernel:  [<ffffffff800d4219>]
__vmalloc_area_node+0x134/0x15c
Oct 13 09:27:46 new kernel:  [<ffffffff800403a4>] load_module+0x961/0x19f5
Oct 13 09:27:46 new kernel:  [<ffffffff800a2dfd>]
autoremove_wake_function+0x0/0x2e
Oct 13 09:27:46 new kernel:  [<ffffffff8000e0aa>]
do_mmap_pgoff+0x3d7/0x780
Oct 13 09:27:46 new kernel:  [<ffffffff8000e2e8>]
do_mmap_pgoff+0x615/0x780
Oct 13 09:27:46 new kernel:  [<ffffffff800b9c60>]
audit_syscall_entry+0x1a8/0x1d3
Oct 13 09:27:46 new kernel:  [<ffffffff800a92f0>]
sys_init_module+0x5b/0x206
Oct 13 09:27:46 new kernel:  [<ffffffff8005d28d>] tracesys+0xd5/0xe0
Oct 13 09:27:46 new kernel:
Oct 13 09:27:46 new kernel: WARNING: at mm/vmalloc.c:97 vmap_pte_range()
Oct 13 09:27:46 new kernel:
Oct 13 09:27:46 new kernel: Call Trace:
Oct 13 09:27:46 new kernel:  [<ffffffff800d3e04>] map_vm_area+0x216/0x306
Oct 13 09:27:46 new kernel:  [<ffffffff80001ff8>] _stext+0xff8/0x1000
Oct 13 09:27:46 new kernel:  [<ffffffff800d4219>]
__vmalloc_area_node+0x134/0x15c
Oct 13 09:27:46 new kernel:  [<ffffffff800403a4>] load_module+0x961/0x19f5
Oct 13 09:27:46 new kernel:  [<ffffffff800a2dfd>]
autoremove_wake_function+0x0/0x2e
Oct 13 09:27:46 new kernel:  [<ffffffff8000e0aa>]
do_mmap_pgoff+0x3d7/0x780
Oct 13 09:27:46 new kernel:  [<ffffffff8000e2e8>]
do_mmap_pgoff+0x615/0x780
Oct 13 09:27:46 new kernel:  [<ffffffff800b9c60>]
audit_syscall_entry+0x1a8/0x1d3
Oct 13 09:27:46 new kernel:  [<ffffffff800a92f0>]
sys_init_module+0x5b/0x206
Oct 13 09:27:46 new kernel:  [<ffffffff8005d28d>] tracesys+0xd5/0xe0
Oct 13 09:27:46 new kernel:

--------------------------------------------------------------------------------------------

Any ideas what caused this ?

Or how to investigate further ?

Regards