* [PATCH sched_ext/for-6.18-fixes] sched_ext: Fix scx_kick_pseqs corruption on concurrent scheduler loads
@ 2025-10-13 20:36 Andrea Righi
2025-10-14 20:29 ` Tejun Heo
0 siblings, 1 reply; 2+ messages in thread
From: Andrea Righi @ 2025-10-13 20:36 UTC (permalink / raw)
To: Tejun Heo, David Vernet, Changwoo Min
Cc: Phil Auld, Emil Tsalapatis, sched-ext, linux-kernel
If we load a BPF scheduler while another scheduler is already running,
alloc_kick_pseqs() would be called again, overwriting the previously
allocated arrays.
Fix by moving the alloc_kick_pseqs() call after the scx_enable_state()
check, ensuring that the arrays are only allocated when a scheduler can
actually be loaded.
Fixes: 14c1da3895a11 ("sched_ext: Allocate scx_kick_cpus_pnt_seqs lazily using kvzalloc()")
Signed-off-by: Andrea Righi <arighi@nvidia.com>
---
kernel/sched/ext.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
index c645d47124e72..12c9c35956926 100644
--- a/kernel/sched/ext.c
+++ b/kernel/sched/ext.c
@@ -4577,15 +4577,15 @@ static int scx_enable(struct sched_ext_ops *ops, struct bpf_link *link)
mutex_lock(&scx_enable_mutex);
- ret = alloc_kick_pseqs();
- if (ret)
- goto err_unlock;
-
if (scx_enable_state() != SCX_DISABLED) {
ret = -EBUSY;
- goto err_free_pseqs;
+ goto err_unlock;
}
+ ret = alloc_kick_pseqs();
+ if (ret)
+ goto err_unlock;
+
sch = scx_alloc_and_add_sched(ops);
if (IS_ERR(sch)) {
ret = PTR_ERR(sch);
--
2.51.0
^ permalink raw reply [flat|nested] 2+ messages in thread* Re: [PATCH sched_ext/for-6.18-fixes] sched_ext: Fix scx_kick_pseqs corruption on concurrent scheduler loads
2025-10-13 20:36 [PATCH sched_ext/for-6.18-fixes] sched_ext: Fix scx_kick_pseqs corruption on concurrent scheduler loads Andrea Righi
@ 2025-10-14 20:29 ` Tejun Heo
0 siblings, 0 replies; 2+ messages in thread
From: Tejun Heo @ 2025-10-14 20:29 UTC (permalink / raw)
To: Andrea Righi
Cc: David Vernet, Changwoo Min, Phil Auld, Emil Tsalapatis,
sched-ext, linux-kernel
On Mon, Oct 13, 2025 at 10:36:34PM +0200, Andrea Righi wrote:
> If we load a BPF scheduler while another scheduler is already running,
> alloc_kick_pseqs() would be called again, overwriting the previously
> allocated arrays.
>
> Fix by moving the alloc_kick_pseqs() call after the scx_enable_state()
> check, ensuring that the arrays are only allocated when a scheduler can
> actually be loaded.
>
> Fixes: 14c1da3895a11 ("sched_ext: Allocate scx_kick_cpus_pnt_seqs lazily using kvzalloc()")
> Signed-off-by: Andrea Righi <arighi@nvidia.com>
Applied to sched_ext/for-6.18-fixes.
Thanks.
--
tejun
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2025-10-14 20:29 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2025-10-13 20:36 [PATCH sched_ext/for-6.18-fixes] sched_ext: Fix scx_kick_pseqs corruption on concurrent scheduler loads Andrea Righi
2025-10-14 20:29 ` Tejun Heo
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®