[PATCH v2] sched_ext: Avoid flooding the log with deprecation warnings

[PATCH v2] sched_ext: Avoid flooding the log with deprecation warnings

[Breno Leitao]

The deprecation notices for direct p->scx.slice/dsq_vtime writes and for
ops->cpu_acquire/release() use plain pr_warn(), so they repeat on every
scheduler (re)load and can flood the kernel log.

The slice/dsq_vtime notice is the worst offender: it is emitted from the
BPF verifier's btf_struct_access callback, which is re-evaluated as the
verifier explores program paths, so a single scheduler load can print it
many times -- hundreds of lines on some hosts, dozens within the same
second.

Switch both notices to pr_warn_ratelimited() so each deprecation is still
reported but bursts no longer spam the log, and add the missing newline
to the slice/dsq_vtime message.

Signed-off-by: Breno Leitao <leitao@debian.org>
---
Changes in v2:
- Switch both deprecation notices from pr_warn_once() to
  pr_warn_ratelimited() (Christian, Andrea)
- Link to v1: https://lore.kernel.org/r/20260623-scx_warning-v1-1-14bf218f4bd9@debian.org
---
 kernel/sched/ext/ext.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/sched/ext/ext.c b/kernel/sched/ext/ext.c
index 691d53fe0f648..d62b93f48a854 100644
--- a/kernel/sched/ext/ext.c
+++ b/kernel/sched/ext/ext.c
@@ -6988,7 +6988,7 @@ static int validate_ops(struct scx_sched *sch, const struct sched_ext_ops *ops)
 	 * run past the BPF allocation. Skip for cid-form.
 	 */
 	if (!sch->is_cid_type && (ops->cpu_acquire || ops->cpu_release))
-		pr_warn("ops->cpu_acquire/release() are deprecated, use sched_switch TP instead\n");
+		pr_warn_ratelimited("ops->cpu_acquire/release() are deprecated, use sched_switch TP instead\n");
 
 	/*
 	 * Sub-scheduler support is tied to the cid-form struct_ops. A sub-sched
@@ -7806,7 +7806,7 @@ static int bpf_scx_btf_struct_access(struct bpf_verifier_log *log,
 		     off + size <= offsetofend(struct task_struct, scx.slice)) ||
 		    (off >= offsetof(struct task_struct, scx.dsq_vtime) &&
 		     off + size <= offsetofend(struct task_struct, scx.dsq_vtime))) {
-			pr_warn("sched_ext: Writing directly to p->scx.slice/dsq_vtime is deprecated, use scx_bpf_task_set_slice/dsq_vtime()");
+			pr_warn_ratelimited("sched_ext: Writing directly to p->scx.slice/dsq_vtime is deprecated, use scx_bpf_task_set_slice/dsq_vtime()\n");
 			return SCALAR_VALUE;
 		}
 

---
base-commit: 4e5dfb7c84012007c3c7061126491bbc92d71bf1
change-id: 20260623-scx_warning-0df49c442afb

Best regards,
-- 
Breno Leitao <leitao@debian.org>

Re: [PATCH v2] sched_ext: Avoid flooding the log with deprecation warnings

[Andrea Righi]

Hi Breno,

On Wed, Jun 24, 2026 at 02:27:43AM -0700, Breno Leitao wrote:
> The deprecation notices for direct p->scx.slice/dsq_vtime writes and for
> ops->cpu_acquire/release() use plain pr_warn(), so they repeat on every
> scheduler (re)load and can flood the kernel log.
> 
> The slice/dsq_vtime notice is the worst offender: it is emitted from the
> BPF verifier's btf_struct_access callback, which is re-evaluated as the
> verifier explores program paths, so a single scheduler load can print it
> many times -- hundreds of lines on some hosts, dozens within the same
> second.
> 
> Switch both notices to pr_warn_ratelimited() so each deprecation is still
> reported but bursts no longer spam the log, and add the missing newline
> to the slice/dsq_vtime message.
> 
> Signed-off-by: Breno Leitao <leitao@debian.org>

Looks good to me.

Reviewed-by: Andrea Righi <arighi@nvidia.com>

Thanks,
-Andrea

> ---
> Changes in v2:
> - Switch both deprecation notices from pr_warn_once() to
>   pr_warn_ratelimited() (Christian, Andrea)
> - Link to v1: https://lore.kernel.org/r/20260623-scx_warning-v1-1-14bf218f4bd9@debian.org
> ---
>  kernel/sched/ext/ext.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/kernel/sched/ext/ext.c b/kernel/sched/ext/ext.c
> index 691d53fe0f648..d62b93f48a854 100644
> --- a/kernel/sched/ext/ext.c
> +++ b/kernel/sched/ext/ext.c
> @@ -6988,7 +6988,7 @@ static int validate_ops(struct scx_sched *sch, const struct sched_ext_ops *ops)
>  	 * run past the BPF allocation. Skip for cid-form.
>  	 */
>  	if (!sch->is_cid_type && (ops->cpu_acquire || ops->cpu_release))
> -		pr_warn("ops->cpu_acquire/release() are deprecated, use sched_switch TP instead\n");
> +		pr_warn_ratelimited("ops->cpu_acquire/release() are deprecated, use sched_switch TP instead\n");
>  
>  	/*
>  	 * Sub-scheduler support is tied to the cid-form struct_ops. A sub-sched
> @@ -7806,7 +7806,7 @@ static int bpf_scx_btf_struct_access(struct bpf_verifier_log *log,
>  		     off + size <= offsetofend(struct task_struct, scx.slice)) ||
>  		    (off >= offsetof(struct task_struct, scx.dsq_vtime) &&
>  		     off + size <= offsetofend(struct task_struct, scx.dsq_vtime))) {
> -			pr_warn("sched_ext: Writing directly to p->scx.slice/dsq_vtime is deprecated, use scx_bpf_task_set_slice/dsq_vtime()");
> +			pr_warn_ratelimited("sched_ext: Writing directly to p->scx.slice/dsq_vtime is deprecated, use scx_bpf_task_set_slice/dsq_vtime()\n");
>  			return SCALAR_VALUE;
>  		}
>  
> 
> ---
> base-commit: 4e5dfb7c84012007c3c7061126491bbc92d71bf1
> change-id: 20260623-scx_warning-0df49c442afb
> 
> Best regards,
> -- 
> Breno Leitao <leitao@debian.org>
>