On 3/11/25 17:59, Uladzislau Rezki (Sony) wrote:
The first line of the changelog needs to say:
commit dfd3df31c9db752234d7d2e09bef2aeabb643ce4 upstream.
I think Greg prefers if you resend with that fixed rather than fixing up locally. If the same backport applies to both 6.12 and 6.13 (it seems to me it does?) I guess a single mail with [PATCH 6.12.y 6.13.y] could be enough.
Apart of that, since kvfree_rcu() does reclaim memory it is worth to go with WQ_MEM_RECLAIM type of wq because it is designed for this purpose.
Fixes: 6c6c47b063b5 ("mm, slab: call kvfree_rcu_barrier() from kmem_cache_destroy()"), Reported-by: Keith Busch kbusch@kernel.org Closes: https://lore.kernel.org/all/Z7iqJtCjHKfo8Kho@kbusch-mbp/ Cc: stable@vger.kernel.org Signed-off-by: Uladzislau Rezki (Sony) urezki@gmail.com Reviewed-by: Joel Fernandes joelagnelf@nvidia.com Signed-off-by: Vlastimil Babka vbabka@suse.cz
I don't know if you need to add another Signed-off-by: Uladzislau Rezki (Sony) urezki@gmail.com
opinions on that differ and not sure where stable stands... (does "git commit -s" add it or detects your previous one?)
Thanks!
kernel/rcu/tree.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-)
diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c index ff98233d4aa5..4703b08fb882 100644 --- a/kernel/rcu/tree.c +++ b/kernel/rcu/tree.c @@ -3191,6 +3191,8 @@ void call_rcu(struct rcu_head *head, rcu_callback_t func) } EXPORT_SYMBOL_GPL(call_rcu); +static struct workqueue_struct *rcu_reclaim_wq;
/* Maximum number of jiffies to wait before draining a batch. */ #define KFREE_DRAIN_JIFFIES (5 * HZ) #define KFREE_N_BATCHES 2 @@ -3519,10 +3521,10 @@ __schedule_delayed_monitor_work(struct kfree_rcu_cpu *krcp) if (delayed_work_pending(&krcp->monitor_work)) { delay_left = krcp->monitor_work.timer.expires - jiffies; if (delay < delay_left)
mod_delayed_work(system_unbound_wq, &krcp->monitor_work, delay);
mod_delayed_work(rcu_reclaim_wq, &krcp->monitor_work, delay);
- queue_delayed_work(system_unbound_wq, &krcp->monitor_work, delay);
- queue_delayed_work(rcu_reclaim_wq, &krcp->monitor_work, delay);
} static void @@ -3620,7 +3622,7 @@ kvfree_rcu_queue_batch(struct kfree_rcu_cpu *krcp) // "free channels", the batch can handle. Break // the loop since it is done with this CPU thus // queuing an RCU work is _always_ success here.
queued = queue_rcu_work(system_unbound_wq, &krwp->rcu_work);
queued = queue_rcu_work(rcu_reclaim_wq, &krwp->rcu_work); WARN_ON_ONCE(!queued); break;@@ -3708,7 +3710,7 @@ run_page_cache_worker(struct kfree_rcu_cpu *krcp) if (rcu_scheduler_active == RCU_SCHEDULER_RUNNING && !atomic_xchg(&krcp->work_in_progress, 1)) { if (atomic_read(&krcp->backoff_page_cache_fill)) {
queue_delayed_work(system_unbound_wq,
queue_delayed_work(rcu_reclaim_wq, &krcp->page_cache_work, msecs_to_jiffies(rcu_delay_page_cache_fill_msec));@@ -5654,6 +5656,10 @@ static void __init kfree_rcu_batch_init(void) int i, j; struct shrinker *kfree_rcu_shrinker;
- rcu_reclaim_wq = alloc_workqueue("kvfree_rcu_reclaim",
WQ_UNBOUND | WQ_MEM_RECLAIM, 0);- WARN_ON(!rcu_reclaim_wq);
- /* Clamp it to [0:100] seconds interval. */ if (rcu_delay_page_cache_fill_msec < 0 || rcu_delay_page_cache_fill_msec > 100 * MSEC_PER_SEC) {