Oliver Hartkopp wrote:
> Peter Zijlstra wrote:
>> On Tue, 2008-11-25 at 12:43 +0100, Peter Zijlstra wrote:
>>  
>>> Hi,
>>>
>>> This is an attempt at removing some of the hrtimer complexity by
>>> reducing the number of callback modes to 1.
>>>
>>> This means that all hrtimer callback functions will be ran from 
>>> HARD-irq
>>> context.
>>>
>>> I went through all the 30 odd hrtimer callback functions in the kernel
>>> and saw only one that I'm not quite sure of, which is the one in
>>> net/can/bcm.c - hence I'm CC-ing the folks responsible for that code.
>>>     
>
> Thanks Peter.
>
> Indeed i assumed my hrtimer callbacks to run in soft-irq. I tried the 
> can-bcm protocol with Ingos current linux-2.6-sched-devel.git 
> including your patches and i did not see any issues so far. And i do 
> not expect any (recursion) problems with hrtimer_forward() in my code 
> either.
>
> But i'm not that familiar with the timer context's stuff, that i would 
> like guaranty that the functions i use in bcm_send_user() and in 
> bcm_can_tx() are always safe to be used in hard-irq context.
>
> It would be nice, if you could give me some support by double checking 
> the correctness of the hard-irq context in the given functions.

Hi Peter,

i ran a heavy load test, which get's (reproducible) the attached outputs ...

Maybe it's not that good to define the hrtimer context to be always 
hard-irq.

Any idea?

The Kernel is Ingos linux-2.6-sched-devel Linux version 
2.6.28-rc7-tip-01309-g1e7292b (hartko@vwagwolkf320) (gcc version 4.3.2 
(Debian 4.3.2-1) ) #1 SMP ...

If you need some additional information to build up the test environment 
your own, please let me know.

Best regards,
Oliver