Re: [PATCH v2] sched/fair: Prevent cfs_rq from being unthrottled with zero runtime_remaining

[Benjamin Segall <bsegall@google.com>]

Aaron Lu <ziqianlu@bytedance.com> writes:

> When a cfs_rq is to be throttled, its limbo list should be empty and
> that's why there is a warn in tg_throttle_down() for non empty
> cfs_rq->throttled_limbo_list.
>
> When running a test with the following hierarchy:
>
>           root
>         /      \
>         A*     ...
>      /  |  \   ...
>         B
>        /  \
>       C*
>
> where both A and C have quota settings, that warn on non empty limbo list
> is triggered for a cfs_rq of C, let's call it cfs_rq_c(and ignore the cpu
> part of the cfs_rq for the sake of simpler representation).
>
> Debug showed it happened like this:
> Task group C is created and quota is set, so in tg_set_cfs_bandwidth(),
> cfs_rq_c is initialized with runtime_enabled set, runtime_remaining
> equals to 0 and *unthrottled*. Before any tasks are enqueued to cfs_rq_c,
> *multiple* throttled tasks can migrate to cfs_rq_c (e.g., due to task
> group changes). When enqueue_task_fair(cfs_rq_c, throttled_task) is
> called and cfs_rq_c is in a throttled hierarchy (e.g., A is throttled),
> these throttled tasks are directly placed into cfs_rq_c's limbo list by
> enqueue_throttled_task().
>
> Later, when A is unthrottled, tg_unthrottle_up(cfs_rq_c) enqueues these
> tasks. The first enqueue triggers check_enqueue_throttle(), and with zero
> runtime_remaining, cfs_rq_c can be throttled in throttle_cfs_rq() if it
> can't get more runtime and enters tg_throttle_down(), where the warning
> is hit due to remaining tasks in the limbo list.
>
> I think it's a chaos to trigger throttle on unthrottle path, the status
> of a being unthrottled cfs_rq can be in a mixed state at the end, so fix
> this by calling throttle_cfs_rq() in tg_set_cfs_bandwidth() immediately
> after enabling bandwidth and setting runtime_remaining = 0. This ensures
> cfs_rq_c is throttled upfront and cannot enter tg_unthrottle_up() with
> zero runtime_remaining.
>
> Also, update outdated comments in tg_throttle_down() since
> unthrottle_cfs_rq() is no longer called with zero runtime_remaining.
>
> While at it, remove a redundant assignment to se in tg_throttle_down().
>
> Fixes: e1fad12dcb66("sched/fair: Switch to task based throttle model")
> Signed-off-by: Aaron Lu <ziqianlu@bytedance.com>
> Reviewed-by: K Prateek Nayak <kprateek.nayak@amd.com>
> ---
> v2: add update_rq_clock() before throttle_cfs_rq() as reported by Hao
>     Jia, or a warn on outdated rq clock is trigged in tg_throttle_down().
>     This can happen when user specified a tiny quota.
>
> Note that Hao Jia also proposed another solution by using a special flag
> when doing enqueue_task_fair() in unthrottle path to avoid doing
> check_enqueue_throttle() [0]. I think that approach is fine too and it
> also has the benefit of not needing to worry about any other potential
> cases where a cfs_rq is unthrottled with <=0 runtime_remaining. Thoughts
> on which approach to go is welcome, thanks.
> [0]: https://lore.kernel.org/lkml/c4a1bcea-fb00-6f3f-6bf6-d876393190e4@gmail.com/
>
>  kernel/sched/core.c  | 11 ++++++++++-
>  kernel/sched/fair.c  | 16 +++++++---------
>  kernel/sched/sched.h |  1 +
>  3 files changed, 18 insertions(+), 10 deletions(-)
>
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index f1ebf67b48e21..58185ec5b8efd 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -9608,7 +9608,16 @@ static int tg_set_cfs_bandwidth(struct task_group *tg,
>  		cfs_rq->runtime_enabled = runtime_enabled;
>  		cfs_rq->runtime_remaining = 0;
>  
> -		if (cfs_rq->throttled)
> +		/*
> +		 * Throttle cfs_rq now or it can be unthrottled with zero
> +		 * runtime_remaining and gets throttled on its unthrottle path.
> +		 */
> +		if (cfs_rq->runtime_enabled && !cfs_rq->throttled) {
> +			update_rq_clock(rq);
> +			throttle_cfs_rq(cfs_rq);
> +		}
> +
> +		if (!cfs_rq->runtime_enabled && cfs_rq->throttled)
>  			unthrottle_cfs_rq(cfs_rq);
>  	}
>

So if this is the only case it can come up, and it only occurs becasue
we set runtime_remaining = 0 and check in unthrottle with <= 0, then I
think we should just set runtime_remaining = 1 here. 

That seems simpler than either throttling immediately (despite likely
having plenty of cfs_b->runtime) or adding an enqueue flag. Adding NR_CPUs
nanoseconds worth of quota on configure seems fine.

unthrottle_cfs_rq/tg_unthrottle_up itself doesn't drop rq lock, so we
shouldn't be able to see cfs_rq->runtime_remaining being consumed during
it, even if it's running on a remote cpu so that threads in the cfs_rq
can be running. They should wind up stuck waiting for rq lock in order
to update runtime_remaining.

Is there anything you see missing from that approach? I think it doing =
0 in particular here is just an artifact, and while the extra check for
runtime_remaining in unthrottle isn't unreasonable, the conflict with
tg_set_cfs_bandwidth isn't a fundamental issue.