Re: [PATCH v3] cgroup: Add pids controller event when fork fails because of pid limit
From: Kenny Yu
Date: Tue Jun 21 2016 - 14:33:58 EST
Thanks for the feedback Tejun!
On 6/21/16, 1:12 PM, "Tejun Heo" <htejun@xxxxxxxxx on behalf of tj@xxxxxxxxxx> wrote:
>Hello,
>
>Just a couple nits.
>
>On Tue, Jun 21, 2016 at 09:56:38AM -0700, Kenny Yu wrote:
>> Summary:
>
>No need for "Summary:" tag.
>
>> This patch adds more visibility into the pids controller when the controller
>> rejects a fork request. Whenever fork fails because the limit on the number of
>> pids in the cgroup is reached, the controller will log this and also notify the
>> newly added cgroups events file. The `max` key in the events file represents
>> the number of times fork failed because of the pids controller.
>>
>> This change also adds an atomic boolean to prevent logging too much (e.g. a fork
>> bomb). The message is logged once per cgroup until the next time the pids limit
>> changes.
>
>The above paragraph isn't uptodate anymore.
Thanks! Will change.
>
>> @@ -213,10 +220,23 @@ static int pids_can_fork(struct task_struct *task)
>> {
>> struct cgroup_subsys_state *css;
>> struct pids_cgroup *pids;
>> + int err;
>> + int events_limit;
>>
>> css = task_css_check(current, pids_cgrp_id, true);
>> pids = css_pids(css);
>> - return pids_try_charge(pids, 1);
>> + err = pids_try_charge(pids, 1);
>> + if (err) {
>> + events_limit = atomic64_inc_return(&pids->events_limit);
>> + cgroup_file_notify(&pids->events_file);
>> + /* Only log the first time events_limit is incremented. */
>> + if (events_limit == 1) {
>> + pr_info("cgroup: fork rejected by pids controller in ");
>> + pr_cont_cgroup_path(task_cgroup(current, pids_cgrp_id));
>> + pr_cont("\n");
>> + }
>> + }
>> + return err;
>> }
>
>It'd be better to use atomic64_inc_and_test() instead.
>
> if (err) {
> if (atomic64_inc_and_test()) {
> pr_xxx...;
> }
> cgroup_file_notify(&pids->events_file);
> }
>
According to the docs https://www.kernel.org/doc/Documentation/atomic_ops.txt ,
it looks like atomic_inc_and_test returns "a boolean indicating whether the resulting
counter value was zero or not", which will only happen when the counter goes from
negative to 0. I'll keep it as atomic_inc_return and get rid of the temp variable.
>Thanks.
>
>--
>tejun
Thanks,
Kenny