Re: [PATCH bpf-next] sched_ext: Reset task fields on exit of an SCX_TASK_READY task

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [High] sched_ext: call to undeclared scx_reset_task_fields() breaks
  build
--

commit 64db5de46ee85990aea72e2451a03a9902ada739
Author: Rahad Bhuiya <rahadbhuiya2021@gmail.com>

sched_ext: Reset task fields on exit of an SCX_TASK_READY task

This commit calls scx_reset_task_fields() in __scx_disable_and_exit_task()
when a task is in the SCX_TASK_READY state.

> diff --git a/kernel/sched/ext/ext.c b/kernel/sched/ext/ext.c
> index 0403c46342813..1cb1532512e6a 100644
> --- a/kernel/sched/ext/ext.c
> +++ b/kernel/sched/ext/ext.c
> @@ -4014,6 +4014,7 @@ void __scx_disable_and_exit_task(struct scx_sched *sch, struct task_struct *p)
>  		args.cancelled = true;
>  		break;
>  	case SCX_TASK_READY:
> +		scx_reset_task_fields(task_rq(p), p);

[Severity: High]
Does calling scx_reset_task_fields() here break the build when
CONFIG_SCHED_CLASS_EXT is enabled?  Commit a298fd5f4ee8 ("sched_ext: Fix
use-after-free of a NONE rescuee leaving its sched while running") is not
present in this tree, so scx_disable_task() still resets task fields inline
and scx_reset_task_fields() is neither declared nor defined.

Even on top of commit a298fd5f4ee8 ("sched_ext: Fix use-after-free of a
NONE rescuee leaving its sched while running"), a task in SCX_TASK_READY is
not under a scheduler's active runtime control (any transition from
SCX_TASK_ENABLED via switched_from_scx() already ran scx_disable_task()),
and calling scx_reset_task_fields() during scx_rehome_task() would zero the
default initial p->scx.slice set by init_scx_entity() or scx_enable().

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261106142818.1475-1-rahadbhuiya2021@gmail.com?part=1