On Wed, 25 Sep 2019, Balasubramani Vivekanandan wrote:
>
> Since it is now allowed to start the hrtimer from the callback, there is
Is now allowed?
> no need for the try to cancel logic. All that is now removed.
Sure I can see that it is removed from the patch, but why and why is it
correct?
When a cpu requests broadcasting, before starting the tick broadcast
hrtimer, bc_set_next() checks if the timer callback (bc_handler) is
active using hrtimer_try_to_cancel(). But hrtimer_try_to_cancel() does
not provide the required synchronization when the callback is active on
other core.
The cal
Sorry, I missed to fix few checkpatch warnings. Corrected it now
+ Peter Zijlstra, linux-renesas-soc
When a cpu requests broadcasting, before starting the tick broadcast
hrtimer, bc_set_next() checks if the timer callback (bc_handler) is
active using hrtimer_try_to_cancel(). But hrtimer_try_to_cancel() does
not provide the required synchronization when the callback is active on
other core.
The cal
> Find a completely untested patch below
Thanks for the patch. I tested your patch and it is working fine. I do
not notice the rcu stall warnings anymore. I have just updated the
formatting of brackets. Please check and apply
6 matches
Mail list logo
