From a4af89cc50f3c1035c1e0dfb50948a23107f3e95 Mon Sep 17 00:00:00 2001
From: Tejun Heo <tj@kernel.org>
Date: Thu, 14 Nov 2024 08:50:58 -1000

ops.cpu_acquire() is currently called with 0 kf_maks which is interpreted as
SCX_KF_UNLOCKED which allows all unlocked kfuncs, but ops.cpu_acquire() is
called from balance_one() under the rq lock and should only be allowed call
kfuncs that are safe under the rq lock. Update it to use SCX_KF_REST.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: David Vernet <void@manifault.com>
Cc: Zhao Mengmeng <zhaomzhao@126.com>
Link: http://lkml.kernel.org/r/ZzYvf2L3rlmjuKzh@slm.duckdns.org
Fixes: 245254f7081d ("sched_ext: Implement sched_ext_ops.cpu_acquire/release()")
---
Applied to sched_ext/for-6.12-fixes.

Thanks.

 kernel/sched/ext.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
index 19f9cb3a4190..1ae8520b63dd 100644
--- a/kernel/sched/ext.c
+++ b/kernel/sched/ext.c
@@ -2645,7 +2645,7 @@ static int balance_one(struct rq *rq, struct task_struct *prev)
 		 * emitted in scx_next_task_picked().
 		 */
 		if (SCX_HAS_OP(cpu_acquire))
-			SCX_CALL_OP(0, cpu_acquire, cpu_of(rq), NULL);
+			SCX_CALL_OP(SCX_KF_REST, cpu_acquire, cpu_of(rq), NULL);
 		rq->scx.cpu_released = false;
 	}
 
-- 
2.47.0