linux-kernel - Re: [PATCH -next v2 2/4] rcu/nocb: Add warning if no rcuog wake up attempt happened during overload

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <4AD1E2C8-0947-403A-B5C9-B6BE4231D576@nvidia.com>
Date: Fri, 16 Jan 2026 09:42:59 -0500
From: <joelagnelf@...dia.com>
To: <linux-kernel@...r.kernel.org>
Cc: Paul E McKenney <paulmck@...nel.org>, Boqun Feng <boqun.feng@...il.com>, <rcu@...r.kernel.org>, Frederic Weisbecker <frederic@...nel.org>
Subject: Re: [PATCH -next v2 2/4] rcu/nocb: Add warning if no rcuog wake up attempt happened during overload

> On Jan 14, 2026, at 12:32 PM, Joel Fernandes <joelagnelf@...dia.com> wrote:
> To be sure we have no rcog wake ups that were lost, add a warning
> to cover the case where the rdp is overloaded with callbacks but
> no wake up was attempted.
> 
> [applied Frederic's adjustment to clearing of nocb_gp_handling flag]

Frederic,

Is it possible for you to do a quick review of these last few? They're mostly 
simple. It would be great if we can pick it for this upcoming merge window.

Thanks!

 - Joel 




> 
> Signed-off-by: Joel Fernandes <joelagnelf@...dia.com>
> ---
> kernel/rcu/tree.c      |  4 ++++
> kernel/rcu/tree.h      |  1 +
> kernel/rcu/tree_nocb.h | 11 ++++++++++-
> 3 files changed, 15 insertions(+), 1 deletion(-)
> 
> diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
> index 2921ffb19939..958b61be87ea 100644
> --- a/kernel/rcu/tree.c
> +++ b/kernel/rcu/tree.c
> @@ -3767,6 +3767,10 @@ static void rcu_barrier_entrain(struct rcu_data *rdp)
>        debug_rcu_head_unqueue(&rdp->barrier_head);
>        rcu_barrier_trace(TPS("IRQNQ"), -1, rcu_state.barrier_sequence);
>    }
> +#ifdef CONFIG_RCU_NOCB_CPU
> +    /* wake_nocb implies all CBs queued before were bypass/lazy. */
> +    WARN_ON_ONCE(wake_nocb && !rdp->nocb_gp_handling);
> +#endif
>    rcu_nocb_unlock(rdp);
>    if (wake_nocb)
>        wake_nocb_gp(rdp);
> diff --git a/kernel/rcu/tree.h b/kernel/rcu/tree.h
> index 7dfc57e9adb1..af1d065e3215 100644
> --- a/kernel/rcu/tree.h
> +++ b/kernel/rcu/tree.h
> @@ -257,6 +257,7 @@ struct rcu_data {
>    unsigned long nocb_gp_loops;    /* # passes through wait code. */
>    struct swait_queue_head nocb_gp_wq; /* For nocb kthreads to sleep on. */
>    bool nocb_cb_sleep;        /* Is the nocb CB thread asleep? */
> +    bool nocb_gp_handling;        /* Is rcuog handling this rdp? */
>    struct task_struct *nocb_cb_kthread;
>    struct list_head nocb_head_rdp; /*
>                     * Head of rcu_data list in wakeup chain,
> diff --git a/kernel/rcu/tree_nocb.h b/kernel/rcu/tree_nocb.h
> index f525e4f7985b..acca24670a8c 100644
> --- a/kernel/rcu/tree_nocb.h
> +++ b/kernel/rcu/tree_nocb.h
> @@ -546,6 +546,7 @@ static void __call_rcu_nocb_wake(struct rcu_data *rdp, bool was_alldone,
>    lazy_len = READ_ONCE(rdp->lazy_len);
>    if (was_alldone) {
>        rdp->qlen_last_fqs_check = len;
> +        rdp->nocb_gp_handling = true;
>        rcu_nocb_unlock(rdp);
>        // Only lazy CBs in bypass list
>        if (lazy_len && bypass_len == lazy_len) {
> @@ -563,7 +564,8 @@ static void __call_rcu_nocb_wake(struct rcu_data *rdp, bool was_alldone,
> 
>        return;
>    } else if (len > rdp->qlen_last_fqs_check + qhimark) {
> -        /* ... or if many callbacks queued. */
> +        /* Callback overload condition. */
> +        WARN_ON_ONCE(!rdp->nocb_gp_handling);
>        rdp->qlen_last_fqs_check = len;
>        j = jiffies;
>        if (j != rdp->nocb_gp_adv_time &&
> @@ -732,6 +734,12 @@ static void nocb_gp_wait(struct rcu_data *my_rdp)
>            needwait_gp = true;
>            trace_rcu_nocb_wake(rcu_state.name, rdp->cpu,
>                        TPS("NeedWaitGP"));
> +        } else if (!rcu_cblist_n_cbs(&rdp->nocb_bypass)) {
> +            /*
> +             * No pending callbacks and no bypass callbacks.
> +             * The rcuog kthread is done handling this rdp.
> +             */
> +            rdp->nocb_gp_handling = false;
>        }
>        if (rcu_segcblist_ready_cbs(&rdp->cblist)) {
>            needwake = rdp->nocb_cb_sleep;
> @@ -1254,6 +1262,7 @@ lazy_rcu_shrink_scan(struct shrinker *shrink, struct shrink_control *sc)
>            continue;
>        }
>        rcu_nocb_try_flush_bypass(rdp, jiffies);
> +        rdp->nocb_gp_handling = true;
>        rcu_nocb_unlock_irqrestore(rdp, flags);
>        wake_nocb_gp(rdp);
>        sc->nr_to_scan -= _count;
> --
> 2.34.1