Re: bug report: WARNING in bonding

[Tariq Toukan <ttoukan.linux@gmail.com>]

On 2/18/2021 7:10 PM, Tariq Toukan wrote:
> 
> 
> On 11/12/2020 6:33 PM, Ido Schimmel wrote:
>> On Thu, Nov 12, 2020 at 05:54:30PM +0200, Tariq Toukan wrote:
>>>
>>>
>>> On 11/12/2020 5:46 PM, Ido Schimmel wrote:
>>>> On Thu, Nov 12, 2020 at 05:38:44PM +0200, Tariq Toukan wrote:
>>>>> Hi all,
>>>>>
>>>>> In the past ~2-3 weeks, we started seeing the following WARNING and 
>>>>> traces
>>>>> in our regression testing systems, almost every day.
>>>>>
>>>>> Reproduction is not stable, and not isolated to a specific test, so 
>>>>> it's
>>>>> hard to bisect.
>>>>>
>>>>> Any idea what could this be?
>>>>> Or what is the suspected offending patch?
>>>>
>>>> Do you have commit f8e48a3dca06 ("lockdep: Fix preemption WARN for 
>>>> spurious
>>>> IRQ-enable")? I think it fixed the issue for me
>>>>
>>>
>>> We do have it. Yet issue still exists.
>>
>> I checked my mail and apparently we stopped seeing this warning after I
>> fixed a lockdep issue (spin_lock() vs spin_lock_bh()) in a yet to be
>> submitted patch. Do you see any other lockdep warnings in the log
>> besides this one? Maybe something in mlx4/5 which is why syzbot didn't
>> hit it?
>>
> 
> Hi,
> 
> Issue still reproduces. Even in GA kernel.
> It is always preceded by some other lockdep warning.
> 
> So to get the reproduction:
> - First, have any lockdep issue.
> - Then, open bond interface.
> 
> Any idea what could it be?
> 
> We'll share any new info as soon as we have it.
> 
> Regards,
> Tariq


Bisect shows this is the offending commit:

commit 4d004099a668c41522242aa146a38cc4eb59cb1e
Author: Peter Zijlstra <peterz@infradead.org>
Date:   Fri Oct 2 11:04:21 2020 +0200

     lockdep: Fix lockdep recursion

     Steve reported that lockdep_assert*irq*(), when nested inside lockdep
     itself, will trigger a false-positive.

     One example is the stack-trace code, as called from inside lockdep,
     triggering tracing, which in turn calls RCU, which then uses
    lockdep_assert_irqs_disabled().

     Fixes: a21ee6055c30 ("lockdep: Change hardirq{s_enabled,_context} 
to per-cpu variables")
     Reported-by: Steven Rostedt <rostedt@goodmis.org>
     Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
     Signed-off-by: Ingo Molnar <mingo@kernel.org>