DEBUG_LOCKS_WARN_ON triggered by bcm43xx-SoftMAC

DEBUG_LOCKS_WARN_ON triggered by bcm43xx-SoftMAC

[Larry Finger]

Using vanilla wireless-2.6 from Linville's tree as updated today, I
get a warning generated by the statement
         if (DEBUG_LOCKS_WARN_ON(!irqs_disabled()))
in kernel.lockdep.c. This warning only occurs when my wireless
interface is inserted and bcm43xx-softmac is activated. For that
reason I have included the SoftMAC and bcm43xx messages.

kernel: SoftMAC: sent association request!
kernel: SoftMAC: associated!
kernel: bcm43xx: set security called, .active_key = 0, .level = 2,
.enabled = 1, .encrypt = 1
kernel: bcm43xx: set security called, .enabled = 1, .encrypt = 1
kernel: BUG: warning at kernel/lockdep.c:1801/trace_hardirqs_on()
kernel:  [<c0103da7>] show_trace_log_lvl+0x197/0x1c0
kernel:  [<c01052db>] show_trace+0x1b/0x20
kernel:  [<c0105306>] dump_stack+0x26/0x30
kernel:  [<c013596f>] trace_hardirqs_on+0x17f/0x190
kernel:  [<c0104a15>] do_general_protection+0xc5/0x230
kernel:  [<c01034b1>] error_code+0x39/0x40
kernel:  [<000038c7>] 0x38c7
kernel:  [<c01052db>] show_trace+0x1b/0x20
kernel:  [<c0105306>] dump_stack+0x26/0x30
kernel:  [<c013596f>] trace_hardirqs_on+0x17f/0x190
kernel:  [<c0104a15>] do_general_protection+0xc5/0x230
kernel:  [<c01034b1>] error_code+0x39/0x40

I'm at a loss here. Can anyone explain how to interpret this dump? I
think I see a general protection fault, but what to do from there is a
mystery.

Thanks,

Larry

Re: DEBUG_LOCKS_WARN_ON triggered by bcm43xx-SoftMAC

[Michael Buesch]

On Wednesday 16 August 2006 05:54, Larry Finger wrote:
> Using vanilla wireless-2.6 from Linville's tree as updated today, I
> get a warning generated by the statement
>          if (DEBUG_LOCKS_WARN_ON(!irqs_disabled()))
> in kernel.lockdep.c. This warning only occurs when my wireless
> interface is inserted and bcm43xx-softmac is activated. For that
> reason I have included the SoftMAC and bcm43xx messages.
> 
> kernel: SoftMAC: sent association request!
> kernel: SoftMAC: associated!
> kernel: bcm43xx: set security called, .active_key = 0, .level = 2,
> .enabled = 1, .encrypt = 1
> kernel: bcm43xx: set security called, .enabled = 1, .encrypt = 1
> kernel: BUG: warning at kernel/lockdep.c:1801/trace_hardirqs_on()
> kernel:  [<c0103da7>] show_trace_log_lvl+0x197/0x1c0
> kernel:  [<c01052db>] show_trace+0x1b/0x20
> kernel:  [<c0105306>] dump_stack+0x26/0x30
> kernel:  [<c013596f>] trace_hardirqs_on+0x17f/0x190
> kernel:  [<c0104a15>] do_general_protection+0xc5/0x230
> kernel:  [<c01034b1>] error_code+0x39/0x40
> kernel:  [<000038c7>] 0x38c7
> kernel:  [<c01052db>] show_trace+0x1b/0x20
> kernel:  [<c0105306>] dump_stack+0x26/0x30
> kernel:  [<c013596f>] trace_hardirqs_on+0x17f/0x190
> kernel:  [<c0104a15>] do_general_protection+0xc5/0x230
> kernel:  [<c01034b1>] error_code+0x39/0x40
> 
> I'm at a loss here. Can anyone explain how to interpret this dump? I
> think I see a general protection fault, but what to do from there is a
> mystery.

Hm, weird bug.
I can't reproduce this on i386 or PPC.
Could it be a bug in the lockdep code?

-- 
Greetings Michael.

Re: DEBUG_LOCKS_WARN_ON triggered by bcm43xx-SoftMAC

[Arjan van de Ven]

On Wed, 2006-08-16 at 18:06 +0200, Michael Buesch wrote:
> On Wednesday 16 August 2006 05:54, Larry Finger wrote:
> > Using vanilla wireless-2.6 from Linville's tree as updated today, I
> > get a warning generated by the statement
> >          if (DEBUG_LOCKS_WARN_ON(!irqs_disabled()))
> > in kernel.lockdep.c. This warning only occurs when my wireless
> > interface is inserted and bcm43xx-softmac is activated. For that
> > reason I have included the SoftMAC and bcm43xx messages.
> > 
> > kernel: SoftMAC: sent association request!
> > kernel: SoftMAC: associated!
> > kernel: bcm43xx: set security called, .active_key = 0, .level = 2,
> > .enabled = 1, .encrypt = 1
> > kernel: bcm43xx: set security called, .enabled = 1, .encrypt = 1
> > kernel: BUG: warning at kernel/lockdep.c:1801/trace_hardirqs_on()
> > kernel:  [<c0103da7>] show_trace_log_lvl+0x197/0x1c0
> > kernel:  [<c01052db>] show_trace+0x1b/0x20
> > kernel:  [<c0105306>] dump_stack+0x26/0x30
> > kernel:  [<c013596f>] trace_hardirqs_on+0x17f/0x190
> > kernel:  [<c0104a15>] do_general_protection+0xc5/0x230
> > kernel:  [<c01034b1>] error_code+0x39/0x40
> > kernel:  [<000038c7>] 0x38c7
> > kernel:  [<c01052db>] show_trace+0x1b/0x20
> > kernel:  [<c0105306>] dump_stack+0x26/0x30
> > kernel:  [<c013596f>] trace_hardirqs_on+0x17f/0x190
> > kernel:  [<c0104a15>] do_general_protection+0xc5/0x230
> > kernel:  [<c01034b1>] error_code+0x39/0x40
> > 
> > I'm at a loss here. Can anyone explain how to interpret this dump? I
> > think I see a general protection fault, but what to do from there is a
> > mystery.
> 
> Hm, weird bug.
> I can't reproduce this on i386 or PPC.
> Could it be a bug in the lockdep code?

is there anything more in the trace???

Re: DEBUG_LOCKS_WARN_ON triggered by bcm43xx-SoftMAC

[Larry Finger]

Arjan van de Ven wrote:
> On Wed, 2006-08-16 at 18:06 +0200, Michael Buesch wrote:
>> On Wednesday 16 August 2006 05:54, Larry Finger wrote:
>>>
>>> I'm at a loss here. Can anyone explain how to interpret this dump? I
>>> think I see a general protection fault, but what to do from there is a
>>> mystery.
>> Hm, weird bug.
>> I can't reproduce this on i386 or PPC.
>> Could it be a bug in the lockdep code?
> 
> is there anything more in the trace???

No - it picks up with regular log messages at that point.

Larry

Re: DEBUG_LOCKS_WARN_ON triggered by bcm43xx-SoftMAC

[Larry Finger]

Michael Buesch wrote:
> 
> Hm, weird bug.
> I can't reproduce this on i386 or PPC.
> Could it be a bug in the lockdep code?
> 

It could be. I'll send it on to LKML. Perhaps one of the experts
there can tell us. It doesn't seem to cause any trouble, but I get
one of these when bcm43xx starts.

Are you running WPA? The message seems to occur just after
wpa_supplicant finishes the connection and sets the security flags.

Larry