mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Tejun Heo <tj@kernel.org>
To: sched-ext@lists.linux.dev
Cc: David Vernet <void@manifault.com>,
	Andrea Righi <arighi@nvidia.com>,
	Changwoo Min <changwoo@igalia.com>,
	Emil Tsalapatis <emil@etsalapatis.com>,
	David Dai <david.dai@linux.dev>,
	linux-kernel@vger.kernel.org, Tejun Heo <tj@kernel.org>,
	Sashiko <sashiko-bot@kernel.org>
Subject: [PATCH 18/17] sched_ext: Fix use-after-free of a NONE rescuee leaving sched_ext
Date: Fri, 23 Oct 2026 10:06:39 -1000	[thread overview]
Message-ID: <20261023200640.272834-1-tj@kernel.org> (raw)
In-Reply-To: <20261023175608.4120910-1-tj@kernel.org>

An earlier patch, "sched_ext: Fix bypassed tasks stalling when the host
lacks SCX_CAP_BASE", sends a bypassed task to rescue when its host lacks
BASE on the cpu. That reaches a task scx_fail_parent() punted to its
bypassing parent at NONE, so such a task can be a rescuee.

switched_from_scx() returned early for a NONE task and skipped the slice end
scx_disable_task() does for the other states. A NONE rescuee leaving
sched_ext while running, through a setscheduler or a priority boost with
slice left, left rq->scx.rescue.curr pointing at it, a use-after-free once
the task exits.

End the slice for a NONE task there, which ends the rescue with it.

Reported-by: Sashiko <sashiko-bot@kernel.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
---
 kernel/sched/ext/ext.c | 15 +++++++--------
 1 file changed, 7 insertions(+), 8 deletions(-)

diff --git a/kernel/sched/ext/ext.c b/kernel/sched/ext/ext.c
index a274210007fa..8241cf374ef9 100644
--- a/kernel/sched/ext/ext.c
+++ b/kernel/sched/ext/ext.c
@@ -4960,16 +4960,15 @@ static void switched_from_scx(struct rq *rq, struct task_struct *p)
 		return;
 
 	/*
-	 * %NONE means SCX is no longer tracking @p at the task level (e.g.
-	 * scx_fail_parent() handed @p back to the parent at NONE pending the
-	 * parent's own teardown). There is nothing to disable; calling
-	 * scx_disable_task() would WARN on the non-%ENABLED state and trigger a
-	 * NONE -> READY validation failure.
+	 * A %NONE task, one scx_fail_parent() punted to its bypassing parent,
+	 * has nothing to disable and scx_disable_task() would WARN on it. It
+	 * can still be a rescuee under the parent, so end its slice here as
+	 * scx_disable_task() does for the other states.
 	 */
 	if (scx_get_task_state(p) == SCX_TASK_NONE)
-		return;
-
-	scx_disable_task(scx_task_sched(p), p);
+		scx_task_slice_ended(rq, p);
+	else
+		scx_disable_task(scx_task_sched(p), p);
 }
 
 static void switched_to_scx(struct rq *rq, struct task_struct *p) {}
-- 
2.56.0


  parent reply	other threads:[~2026-10-23 20:06 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-10-23 17:55 [PATCHSET v3 sched_ext/for-7.4] sched_ext: Fix bypassed sub-scheduler tasks stalling when the host lacks caps Tejun Heo
2026-10-23 17:55 ` [PATCH 01/17] sched_ext: Pack scx_ent_flags' multi-bit fields at the top Tejun Heo
2026-10-23 17:55 ` [PATCH 02/17] sched_ext: Clear holding_cpu where the claim is confirmed Tejun Heo
2026-10-23 17:55 ` [PATCH 03/17] sched_ext: Fix bypassed tasks stranded by the reenqueue limit Tejun Heo
2026-10-23 17:55 ` [PATCH 04/17] sched_ext: Fix bypassed tasks looping on a host without SCX_CAP_ENQ Tejun Heo
2026-10-23 17:55 ` [PATCH 05/17] sched_ext: Move refill_task_slice_dfl() below apply_slice_vtime() Tejun Heo
2026-10-23 17:55 ` [PATCH 06/17] sched_ext: Fix rescue served time disagreeing with the slice under proxy execution Tejun Heo
2026-10-23 17:55 ` [PATCH 07/17] sched_ext: Restore a rescuee's slice after ops.stopping() Tejun Heo
2026-10-23 17:55 ` [PATCH 08/17] sched_ext: Fix a cpu idling with a runnable rescuee Tejun Heo
2026-10-23 17:56 ` [PATCH 09/17] sched_ext: Mark the rescue grant as a kernel slice with SCX_TASK_KERNEL_SLICE Tejun Heo
2026-10-23 17:56 ` [PATCH 10/17] sched_ext: Refuse extending a kernel slice Tejun Heo
2026-10-23 17:56 ` [PATCH 11/17] sched_ext: Fix a cpu idling after put_prev_task_scx() queues its task on the local DSQ Tejun Heo
2026-10-23 17:56 ` [PATCH 12/17] sched_ext: Expire prev's slice in dispatch_one() under bypass Tejun Heo
2026-10-23 17:56 ` [PATCH 13/17] sched_ext: Fix bypassed tasks stalling when the host lacks SCX_CAP_BASE Tejun Heo
2026-10-23 17:56 ` [PATCH 14/17] sched_ext: Fix the bypass load balancer stranding tasks on cpus where " Tejun Heo
2026-10-23 17:56 ` [PATCH 15/17] sched_ext: Factor out bypass_lb_move() and insert with the donor DSQ unlocked Tejun Heo
2026-10-23 17:56 ` [PATCH 16/17] sched_ext: Let the bypass load balancer move tasks homed on other cpus Tejun Heo
2026-10-23 17:56 ` [PATCH 17/17] sched_ext: Fix queued bypassed tasks stalling when the host loses SCX_CAP_BASE Tejun Heo
2026-10-23 20:06 ` Tejun Heo [this message]
2026-10-23 20:06 ` [PATCH 19/17] sched_ext: Fix the stall watchdog blaming a bypassing owner for a stuck task Tejun Heo
2026-10-23 20:10 ` [PATCH 20/17] sched_ext: Fix a lone IMMED task waiting behind a kept prev Tejun Heo

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20261023200640.272834-1-tj@kernel.org \
    --to=tj@kernel.org \
    --cc=arighi@nvidia.com \
    --cc=changwoo@igalia.com \
    --cc=david.dai@linux.dev \
    --cc=emil@etsalapatis.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=sashiko-bot@kernel.org \
    --cc=sched-ext@lists.linux.dev \
    --cc=void@manifault.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®