QuadEthernet driver `sunqe' broken
Hi,
(Note: I tried posting this message to `sparclinux' but Bogofilter at
vger.kernel.org is so zealous that all my attempts failed!)
The `sunqe' driver in `linux-image-2.6.18-4-sparc32' is broken. I.e.,
it loads fine, but then reports the following erroneous ethernet
addresses:
sunqe.c:v4.0 June 23, 2006 David S. Miller (davem@davemloft.net)
eth0: qe channel[0] 00:00:00:00:00:00
eth1: qe channel[1] 00:00:00:00:00:00
eth2: qe channel[2] 00:00:00:00:00:00
eth3: qe channel[3] 00:00:00:00:00:00
Also, `rmmod sunqe' yields the following oops:
[f001b3b0 : of_device_remove+0x2c/0x34 ]
[f0116f70 : __device_release_driver+0x78/0x8c ]
[f0117308 : driver_detach+0x90/0x274 ]
[f01164c0 : bus_remove_driver+0x7c/0xa0 ]
[f0117524 : driver_unregister+0x10/0x1c ]
[fe650504 : qec_exit+0x10/0x5c [sunqe] ]
[f004cd44 : sys_delete_module+0x1cc/0x1f4 ]
[f0011634 : syscall_is_too_hard+0x3c/0x40 ]
[00011220 : 0x11228 ]
[f013a82c : unregister_netdevice+0x184/0x248 ]
[fe64f058 : qec_sbus_remove+0x18/0x70 [sunqe] ]
[f001b3b0 : of_device_remove+0x2c/0x34 ]
[f0116f70 : __device_release_driver+0x78/0x8c ]
[f0117308 : driver_detach+0x90/0x274 ]
[f01164c0 : bus_remove_driver+0x7c/0xa0 ]
[f0117524 : driver_unregister+0x10/0x1c ]
[fe650504 : qec_exit+0x10/0x5c [sunqe] ]
[f004cd44 : sys_delete_module+0x1cc/0x1f4 ]
[f0011634 : syscall_is_too_hard+0x3c/0x40 ]
[00011220 : 0x11228 ]
[f003d9d0 : notifier_call_chain+0x1c/0x4c ]
[f013a82c : unregister_netdevice+0x184/0x248 ]
[fe64f058 : qec_sbus_remove+0x18/0x70 [sunqe] ]
[f001b3b0 : of_device_remove+0x2c/0x34 ]
[f0116f70 : __device_release_driver+0x78/0x8c ]
[f0117308 : driver_detach+0x90/0x274 ]
[f01164c0 : bus_remove_driver+0x7c/0xa0 ]
[f0117524 : driver_unregister+0x10/0x1c ]
[fe650504 : qec_exit+0x10/0x5c [sunqe] ]
[f004cd44 : sys_delete_module+0x1cc/0x1f4 ]
[f0011634 : syscall_is_too_hard+0x3c/0x40 ]
[00011220 : 0x11228 ]
kernel BUG at net/core/dev.c:3236!
\|/ ____ \|/
"@'/ ,. \`@"
/_| \__/ |_\
\__U_/
rmmod(3507): Kernel bad trap [#1]
PSR: 414000c5 PC: f01390c8 NPC: f01390cc Y: 00000000 Not tainted
PC: <free_netdev+0x48/0x54>
%%G: fc000000 414000e4 f01d15f8 f1bb1e60 0000000c f022b740 f12e8000 00000000
%%O: 00000026 f01b7038 00000ca4 00000004 fc000000 00000003 f12e9b58 f01390c0
RPC: <free_netdev+0x40/0x54>
%%L: 419000e6 00000010 f0233044 00000010 00000004 00000000 00000000 5002a000
%%I: f1ce1800 00000004 0000e000 f002f000 f0230c60 f0230c00 f12e9bc0 fe64f0a4
Caller[fe64f0a4]: qec_sbus_remove+0x64/0x70 [sunqe]
Caller[f001b3b0]: of_device_remove+0x2c/0x34
Caller[f0116f70]: __device_release_driver+0x78/0x8c
Caller[f0117308]: driver_detach+0x90/0x274
Caller[f01164c0]: bus_remove_driver+0x7c/0xa0
Caller[f0117524]: driver_unregister+0x10/0x1c
Caller[fe650504]: qec_exit+0x10/0x5c [sunqe]
Caller[f004cd44]: sys_delete_module+0x1cc/0x1f4
Caller[f0011634]: syscall_is_too_hard+0x3c/0x40
Caller[00011220]: 0x11228
Instruction DUMP: 92102ca4 7ffb658c 90122038 <91d02005> c220a174 7fff7a57 91ee21d0 01000000 9de3bf98
I couldn't find any evidence in the `sparclinux' mailing list archive
that this has been fixed in recent kernels. I did find a report stating
that the module wouldn't even load on 2.6.12, though [0].
I was previously using this driver on a 2.4.26 kernel where it worked
fine.
Thanks,
Ludovic.
[0] http://www.mail-archive.com/sparclinux@vger.kernel.org/msg00235.html
Reply to: