i.MX6 cpufreq : rcu_sched kthread starved

[Jean-Michel Hautbois]

Hi Russell,

2015-06-11 11:24 GMT+02:00 Russell King - ARM Linux <linux at arm.linux.org.uk>:
> On Thu, Jun 11, 2015 at 10:48:31AM +0200, Jean-Michel Hautbois wrote:
>> [  388.902571] INFO: rcu_sched detected stalls on CPUs/tasks: { 1}
>> (detected by 0, t=2102 jiffies, g=7088, c=7087, q=34)
>> [  388.913334] Task dump for CPU 1:
>> [  388.916588] irq/298-IC PP   R running      0    46      2 0x00000002
>> [  388.923016] Backtrace:
>> [  388.925541] [<8000cbc0>] (vfp_double_ftoui) from [<807fd2c8>]
>> (schedule+0x4c/0xa4)
>
> This does not look very good.  If vfp_double_ftoui really is being called
> from this path, then that means you're trying to use VFP inside the
> kernel, and that's not permitted.

I am pretty sure I didn't write anything related to floating point
unit... Thanks for pointing it out !

>> [  392.127701] CPU: 0 PID: 25 Comm: kworker/0:1 Tainted: G    B    C
>>    4.0.5 #49
>
> This also tells me that your kernel has thrown a bad_page() exception,
> and you have modules from drivers/staging loaded - so it's possible
> that the problem you're seeing here is caused by either an unstable
> staging module, or whatever caused the bad_page() exception.
>
> Always investigate the first error first...

My bad, I have not read deeply enough...
Well, let's find who is doing this...
Thank you !
JM