Re: [PATCH 2/2] sched_ext: Fix the bypass dispatch claim for schedulers bypassed by an ancestor

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] sched_ext: Mixing goto-based cleanup and guard() in scx_bypass()
--

commit c63689b3855e65ddcc5c7f3211328556c68f2c61
Author: Tejun Heo <tj@kernel.org>

sched_ext: Fix the bypass dispatch claim for schedulers bypassed by an ancestor

This commit fixes a bypass dispatch claim leak when a sub-scheduler's depth
goes from zero to one under an ancestor. It ensures claims are properly
propagated to descendants and released when the depth returns to zero.

Cc: stable@vger.kernel.org # v7.1+

> diff --git a/kernel/sched/ext/ext.c b/kernel/sched/ext/ext.c
> index 7df47c6741009..630afee30c0c5 100644
> --- a/kernel/sched/ext/ext.c
> +++ b/kernel/sched/ext/ext.c

[ ... ]

> @@ -6197,8 +6197,12 @@ void scx_bypass(struct scx_sched *sch, bool bypass)
>  	}
>  
>  	/* disarming must come after moving all tasks out of the bypass DSQs */
> -	if (!bypass)
> -		scx_disable_bypass_dsp(sch);
> +	if (!bypass) {
> +		guard(raw_spinlock)(&scx_sched_lock);
> +		scx_for_each_descendant_pre(pos, sch)
> +			if (!pos->bypass_depth)
> +				scx_disable_bypass_dsp(pos);
> +	}
>  unlock:
>  	raw_spin_unlock_irqrestore(&scx_bypass_lock, flags);
>  }

[Severity: Low]
Does this introduce a mix of scope-based cleanup and goto-based cleanup
in scx_bypass()?

According to the cleanup subsystem guidelines (include/linux/cleanup.h),
mixing goto-based cleanup labels with guard() declarations in the same
function creates confusing ownership semantics and is explicitly prohibited.
Developers are expected to convert all resources in the function to
scope-based cleanup, or convert none of them.

This patch introduces guard(raw_spinlock)(&scx_sched_lock) while retaining
the unlock: label and goto-based cleanup that releases scx_bypass_lock.
This mixing increases the risk of future LIFO ordering mistakes or
use-after-free bugs during refactoring.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261011214028.743546-1-tj@kernel.org?part=2