On 12/02/16 10:14, Donald Buczek wrote:

On 11/30/16 12:43, Donald Buczek wrote:

quoted

On 11/30/16 12:09, Michal Hocko wrote:

quoted

[CCing Paul]

On Wed 30-11-16 11:28:34, Donald Buczek wrote:
[...]

quoted

shrink_active_list gets and releases the spinlock and calls 
cond_resched().
This should give other tasks a chance to run. Just as an 
experiment, I'm
trying

--- a/mm/vmscan.c
+++ b/mm/vmscan.c

@@ -1921,7 +1921,7 @@ static void shrink_active_list(unsigned long

nr_to_scan,
         spin_unlock_irq(&pgdat->lru_lock);

         while (!list_empty(&l_hold)) {
-               cond_resched();
+               cond_resched_rcu_qs();
                 page = lru_to_page(&l_hold);
                 list_del(&page->lru);

and didn't hit a rcu_sched warning for >21 hours uptime now. We'll 
see.

This is really interesting! Is it possible that the RCU stall detector
is somehow confused?

Wait... 21 hours is not yet a test result.

For the records: We didn't have any stall warnings after 2 days and 20 
hours now and so I'm quite confident, that my above patch fixed the 
problem for v4.8.0. On previous boots the rcu warnings started after 
37,0.2,1,2,0.8 hours uptime.

Now I've applied this patch to stable latest (v4.8.11) on another 
backup machine which suffered even more rcu stalls.

Donald

quoted

[...]

For the records: After 3 days and 21 hours we've got a rcu stall warning 
again [1]. So my patch didn't fix it.

Trying "[PATCH] mm, vmscan: add cond_resched into shrink_node_memcg" 
from Michal Hocko [2] on top of v4.8.12 on both servers now.

[1] https://owww.molgen.mpg.de/~buczek/321322/2016-12-06.dmesg.txt
[2] https://marc.info/?i=20161202095841.16648-1-mhocko%40kernel.org

-- 
Donald Buczek
buczek@molgen.mpg.de
Tel: +49 30 8413 1433

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>