Repeated kernel hangs on 2.6.32-220.2.1.el6: umount stuck in synchronize_sched_expedited for 9 hours

Issue

• NFS umount stuck in synchronize_sched_expedited() for 9 hours.
• Using sysrq to trigger a core dump and examining the core shows the the NFS umount process stuck with a backtrace similar to the following

PID: 31208  TASK: ffff88063b78cac0  CPU: 0   COMMAND: "umount"
 #0 [ffff8805f4313b48] schedule at ffffffff814eca93
 #1 [ffff8805f4313c10] schedule_timeout at ffffffff814ed985
 #2 [ffff8805f4313cc0] wait_for_common at ffffffff814ed603
 #3 [ffff8805f4313d50] wait_for_completion at ffffffff814ed71d
 #4 [ffff8805f4313d60] synchronize_sched_expedited at ffffffff81062f51
 #5 [ffff8805f4313dc0] synchronize_rcu_expedited at ffffffff810ddc7e
 #6 [ffff8805f4313dd0] bdi_remove_from_list at ffffffff811340ea
 #7 [ffff8805f4313df0] bdi_unregister at ffffffff811341b1
 #8 [ffff8805f4313e20] nfs_put_super at ffffffffa04899e9 [nfs]
 #9 [ffff8805f4313e30] generic_shutdown_super at ffffffff8117864b
#10 [ffff8805f4313e50] kill_anon_super at ffffffff81178736
#11 [ffff8805f4313e70] nfs_kill_super at ffffffffa048c415 [nfs]
#12 [ffff8805f4313e90] deactivate_super at ffffffff811796c0
#13 [ffff8805f4313eb0] mntput_no_expire at ffffffff8119563f
#14 [ffff8805f4313ee0] sys_umount at ffffffff811960db