Hard lockup on netconsole - netconsole_netdev_event() with use of bonding driver
Issue
Kernel panic - vmcore generated with significant stack:
#0 [ffff880c2da259f8] schedule at ffffffff814fd7e2
#1 [ffff880c2da25ac0] __mutex_lock_slowpath at ffffffff814fee8e
#2 [ffff880c2da25b30] mutex_lock at ffffffff814fed2b
#3 [ffff880c2da25b50] rtnl_lock at ffffffff81448675 **<--- netlink mutex lock try => deadlock**
#4 [ffff880c2da25b60] netpoll_cleanup at ffffffff8144f1dc
#5 [ffff880c2da25b80] netconsole_netdev_event at ffffffffa024c137 [netconsole]
#6 [ffff880c2da25bc0] notifier_call_chain at ffffffff81503325
#7 [ffff880c2da25c00] raw_notifier_call_chain at ffffffff81098076
#8 [ffff880c2da25c10] call_netdevice_notifiers at ffffffff8143b5ab
#9 [ffff880c2da25c20] bond_enslave at ffffffffa0269158 [bonding]
#10 [ffff880c2da25cc0] bond_do_ioctl at ffffffffa026a175 [bonding]
#11 [ffff880c2da25d40] dev_ifsioc at ffffffff8143ddde
#12 [ffff880c2da25d70] dev_ioctl at ffffffff8143e2d3 **<--- netlink mutex acquired**
#13 [ffff880c2da25e30] sock_ioctl at ffffffff81426c0d
#14 [ffff880c2da25e60] vfs_ioctl at ffffffff8118dec2
#15 [ffff880c2da25ea0] do_vfs_ioctl at ffffffff8118e064
#16 [ffff880c2da25f30] sys_ioctl at ffffffff8118e5e1
#17 [ffff880c2da25f80] system_call_fastpath at ffffffff8100b0f2
Netconsole has been configured to use a NIC (ethX) for logging over network and the same NIC was assumed to be slave for some current bond master.
Environment
This is relevant only for the kernel-2.6.32-279.el6.x86_64
with netconsole and bonding in use over the same NIC
Red Hat Enterprise Linux (RHEL)
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.