On 05.12.2010 16:02, Michal Sojka wrote:
> On Sun, 05 Dec 2010, Oliver Hartkopp wrote:
>> can you give some more details about you setup, that leads to the crash?
>>
>> 1. Are you running the setup using this patch:
>>
>>
>> http://rtime.felk.cvut.cz/gitweb/shark/linux.git/commitdiff/92487e4f349cd7518cc3202662f42fea7d42ba73
>
> Yes, but I've just tested that the panic appears even without this patch
> applied.
>
>>
>> 2. Are you using the '-e' option that creates the local echo of sent
>> CAN frames?
>
> No. I use a simple cangw -A -s can0 -d can1. It crashes immediately with
> the first routed message. I tried different messages ids and lengths but
> the result is still the same.
Yes - i don't expect this to be CAN message specific.
As i don't have all my CAN hardware here, i checked it with vcan0 & vcan1 :
At the time when i invoked
cangw -A -s vcan0 -d vcan1 -e
i get this in dmesg:
[10716.319344] BUG: sleeping function called from invalid context at
mm/slub.c:793
[10716.319357] in_atomic(): 1, irqs_disabled(): 0, pid: 7963, name: cangw
[10716.319365] 2 locks held by cangw/7963:
[10716.319371] #0: (rtnl_mutex){+.+.+.}, at: [<c13863af>] rtnl_lock+0xf/0x20
[10716.319396] #1: (cgw_list_lock){+.+.+.}, at: [<f9afbd9c>]
cgw_create_job+0x15c/0x1d0 [can_gw]
[10716.319422] Pid: 7963, comm: cangw Not tainted 2.6.37-rc4-00012-g22a5b56 #115
[10716.319429] Call Trace:
[10716.319443] [<c102c159>] __might_sleep+0x109/0x140
[10716.319456] [<c10b73de>] kmem_cache_alloc+0x3e/0xc0
[10716.319471] [<f9a4c41f>] ? can_rx_register+0x4f/0x190 [can]
[10716.319484] [<f9a4c41f>] can_rx_register+0x4f/0x190 [can]
[10716.319496] [<f9afbdc9>] cgw_create_job+0x189/0x1d0 [can_gw]
[10716.319507] [<f9afb530>] ? can_can_gw_rcv+0x0/0x140 [can_gw]
[10716.319518] [<c1386515>] rtnetlink_rcv_msg+0x135/0x220
[10716.319529] [<f9afbc40>] ? cgw_create_job+0x0/0x1d0 [can_gw]
[10716.319539] [<c13863e0>] ? rtnetlink_rcv_msg+0x0/0x220
[10716.319550] [<c138fb16>] netlink_rcv_skb+0x86/0xb0
[10716.319560] [<c13863d7>] rtnetlink_rcv+0x17/0x20
[10716.319569] [<c138f7e3>] netlink_unicast+0x243/0x290
[10716.319579] [<c13903c2>] netlink_sendmsg+0x1d2/0x2e0
[10716.319591] [<c13686b7>] sock_sendmsg+0xa7/0xd0
[10716.319603] [<c109fe2a>] ? might_fault+0x5a/0xb0
[10716.319614] [<c109fe2a>] ? might_fault+0x5a/0xb0
[10716.319624] [<c109fe70>] ? might_fault+0xa0/0xb0
[10716.319637] [<c1196478>] ? _copy_from_user+0x38/0x130
[10716.319646] [<c1367fe8>] ? copy_from_user+0x8/0x10
[10716.319656] [<c1369452>] sys_sendto+0xb2/0xe0
[10716.319668] [<c1064e49>] ? lock_release_non_nested+0x59/0x2e0
[10716.319678] [<c10a2b58>] ? handle_mm_fault+0xd8/0x650
[10716.319689] [<c109fe2a>] ? might_fault+0x5a/0xb0
[10716.319699] [<c109fe2a>] ? might_fault+0x5a/0xb0
[10716.319710] [<c136a651>] sys_socketcall+0x171/0x270
[10716.319721] [<c1002e90>] sysenter_do_call+0x12/0x36
I think, i need to check, if there was any change in netlink / rtnl_mutex in
the kernel recently. I get this message independently from CAN traffic, when
creating the routing entry.
My setup here is a recent 2.6.37-rc4 and i compiled the SocketCAN SVN source
any only insmod'ded the can{|-raw|-gw}.ko files created in the SVN.
Regards,
Oliver
_______________________________________________
Socketcan-users mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/socketcan-users
