Re: [patch] del_timer_sync scalability patch

[Ingo Molnar]

* Christoph Lameter <christoph@xxxxxxxxxxx> wrote:

> The following patch removes the magic in the timer_list structure
> (Andrew suggested that we may not need it anymore) and replaces it
> with two u8 variables that give us some additional state of the timer

The 'remove the magic' observation is not a 'backdoor' to introduce new
fields 'for free'. So please dont mix the two things into one patch.

> +       u8 running;             /* function is currently executing */
> +       u8 shutdown;            /* do not schedule this timer */

it may as well be cleaner to do the timer->base_running thing after all,
and to do this in __run_timers():

	timer->base = NULL;
	timer->base_running = base;

note that we cannot clear ->base_running in __run_timers() [we dont own
any access to the timer anymore, it may be kfree()d, etc.], so
del_timer_sync() has to make sure that the timer->base_running info is
not stale - it is a hint only. But it looks doable, and it should solve
the NUMA scalability problem.

	Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/