#10969 NORM 11.2.0-: rtnl deadlock related to libertas reset?

[Zarro Boogs per Child]

#10969: rtnl deadlock related to libertas reset?
---------------------------------+------------------------------------------
           Reporter:  greenfeld  |       Owner:  dsd                              
               Type:  defect     |      Status:  new                              
           Priority:  normal     |   Milestone:  11.2.0-final                     
          Component:  kernel     |     Version:  Development build as of this date
         Resolution:             |    Keywords:                                   
        Next_action:  diagnose   |    Verified:  0                                
Deployment_affected:             |   Blockedby:                                   
           Blocking:             |  
---------------------------------+------------------------------------------

Comment(by dsd):

 Debugged on IRC: NetworkManager is holding the rtnl, and the trace of that
 process (not included in the attachment due to truncation) is:

 {{{
 [60074.960014] ea737d9c 0000000b ea8703e0 00000000 ea7340b0 b043583d
 ea8fc678 ea8fc678
 [60074.960014] Call Trace:
 [60074.960014] [<ed13acdf>] __lbs_cmd+0xe8/0x17e [libertas]
 [60074.960014] [<b054323c>] ? __debug_check_no_obj_freed+0x4f/0x137
 [60074.960014] [<b043583d>] ? autoremove_wake_function+0x0/0x34
 [60074.960014] [<ed147dc3>] lbs_get_wireless_stats+0x1c0/0x31d [libertas]
 [60074.960014] [<ed138eb1>] ? lbs_cmd_copyback+0x0/0x26 [libertas]
 [60074.960014] [<b0706fdb>] get_wireless_stats+0x42/0x44
 [60074.960014] [<b0706fef>] iw_handler_get_iwstats+0x12/0x41
 [60074.960014] [<b0707259>] ioctl_standard_call+0x207/0x296
 [60074.960014] [<b071933e>] ? sub_preempt_count+0x81/0x8e
 [60074.960014] [<b0715bc7>] ? mutex_lock_nested+0x223/0x23d
 [60074.960014] [<b0669b9a>] ? __dev_get_by_name+0x6e/0x7f
 [60074.960014] [<b0707395>] wext_handle_ioctl+0xad/0x18c
 [60074.960014] [<b0706fdd>] ? iw_handler_get_iwstats+0x0/0x41
 [60074.960014] [<b066c9c5>] dev_ioctl+0x600/0x614
 [60074.960014] [<b0496ee6>] ? fd_install+0x26/0x4a
 [60074.960014] [<b065cd21>] sock_ioctl+0x1fc/0x208
 [60074.960014] [<b04a4513>] vfs_ioctl+0x33/0x91
 [60074.960014] [<b065cb25>] ? sock_ioctl+0x0/0x208
 [60074.960014] [<b04a4ad7>] do_vfs_ioctl+0x4a1/0x4cc
 [60074.960014] [<b0499684>] ? fget_light+0x3f/0xd6
 [60074.960014] [<b041b42c>] ? get_parent_ip+0xb/0x31
 [60074.960014] [<b071933e>] ? sub_preempt_count+0x81/0x8e
 [60074.960014] [<b04a4b48>] sys_ioctl+0x46/0x66
 [60074.960014] [<b04026d0>] sysenter_do_call+0x12/0x31
 }}}

 i.e. smells very much like what I fixed yesterday, bad command timeout
 handling...

-- 
Ticket URL: <http://dev.laptop.org/ticket/10969#comment:5>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system