Re: RT and Cascade interrupts

[Trond Myklebust <trond.myklebust@fys.uio.no>]

lau den 28.05.2005 Klokka 13:48 (-0400) skreiv john cooper:

> The other possibility to fix the original problem within
> the scope of the RPC code was to replace the bit state of
> RPC_TASK_HAS_TIMER with a count so we don't obscure the
> fact the timer was requeued during the preemption window.
> This happens as rpc_run_timer() does an unconditional
> clear_bit(RPC_TASK_HAS_TIMER,..) before returning.
> 
> Another would be to check timer->base/timer_pending()
> in rpc_run_timer() and to omit doing a clear_bit()
> if the timer is found to have been requeued.

Could you please explain why you think such a scenario is possible? The
timer functions themselves should never be causing a re-queue, and every
iteration through the loop in __rpc_execute() should cause any pending
timer to be killed, as should rpc_release_task().

That's why we can use del_singleshot_timer_sync() in the first place:
because there is no recursion, and no re-queueing that can cause races.
I don't see how either preemption or RT will change that (and if they
do, then _that_ is the real bug that needs fixing).

Cheers,
  Trond