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: <bab60fe4-964c-43a6-ecce-4cbd4981d875@ya.ru>
Date:   Thu, 1 Jun 2023 03:57:10 +0300
From:   Kirill Tkhai <tkhai@...ru>
To:     Qi Zheng <qi.zheng@...ux.dev>, paulmck@...nel.org
Cc:     RCU <rcu@...r.kernel.org>, Yujie Liu <yujie.liu@...el.com>,
        oe-lkp@...ts.linux.dev, lkp@...el.com,
        linux-kernel@...r.kernel.org,
        Andrew Morton <akpm@...ux-foundation.org>,
        Vlastimil Babka <vbabka@...e.cz>,
        Roman Gushchin <roman.gushchin@...ux.dev>,
        Christian König <christian.koenig@....com>,
        David Hildenbrand <david@...hat.com>,
        Davidlohr Bueso <dave@...olabs.net>,
        Johannes Weiner <hannes@...xchg.org>,
        Michal Hocko <mhocko@...nel.org>,
        Muchun Song <muchun.song@...ux.dev>,
        Shakeel Butt <shakeelb@...gle.com>,
        Yang Shi <shy828301@...il.com>, linux-mm@...ck.org,
        ying.huang@...el.com, feng.tang@...el.com, fengwei.yin@...el.com
Subject: Re: [linus:master] [mm] f95bdb700b: stress-ng.ramfs.ops_per_sec
 -88.8% regression

Hi,

On 30.05.2023 06:07, Qi Zheng wrote:
> Hi,
> 
> On 2023/5/29 20:51, Paul E. McKenney wrote:
>> On Mon, May 29, 2023 at 10:39:21AM +0800, Qi Zheng wrote:
> 
> [...]
> 
>>>
>>> Thanks for such a detailed explanation.
>>>
>>> Now I think we can continue to try to complete the idea[1] from
>>> Kirill Tkhai. The patch moves heavy synchronize_srcu() to delayed
>>> work, so it doesn't affect on user-visible unregistration speed.
>>>
>>> [1]. https://lore.kernel.org/lkml/153365636747.19074.12610817307548583381.stgit@localhost.localdomain/
>>
>> A blast from the past!  ;-)
>>
>> But yes, moving the long-latency synchronize_srcu() off the user-visible
>> critical code path can be even better.
> 
> Yeah, I applied these patches  ([PATCH RFC 04/10]~[PATCH RFC 10/10],
> with few conflicts), the ops/s does get back to the previous levels.
> 
> I'll continue updating this patchset after doing more testing.

You may also fix the issue using the below generic solution.

In addition to this we need patch, which calls unregister_shrinker_delayed_initiate()
instead of unregister_shrinker() in deactivate_locked_super(), and calls
unregister_shrinker_delayed_finalize() from destroy_super_work(). Compilation tested only.

---
 include/linux/shrinker.h |  2 ++
 mm/vmscan.c              | 38 ++++++++++++++++++++++++++++++--------
 2 files changed, 32 insertions(+), 8 deletions(-)
diff --git a/include/linux/shrinker.h b/include/linux/shrinker.h
index 224293b2dd06..4ba2986716d3 100644
--- a/include/linux/shrinker.h
+++ b/include/linux/shrinker.h
@@ -4,6 +4,7 @@
 
 #include <linux/atomic.h>
 #include <linux/types.h>
+#include <linux/rwsem.h>
 
 /*
  * This struct is used to pass information from page reclaim to the shrinkers.
@@ -83,6 +84,7 @@ struct shrinker {
 #endif
 	/* objs pending delete, per node */
 	atomic_long_t *nr_deferred;
+	struct rw_semaphore rwsem;
 };
 #define DEFAULT_SEEKS 2 /* A good number if you don't know better. */
 
diff --git a/mm/vmscan.c b/mm/vmscan.c
index eeca83e28c9b..19fc129771ce 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -706,6 +706,7 @@ static int __prealloc_shrinker(struct shrinker *shrinker)
 	if (!shrinker->nr_deferred)
 		return -ENOMEM;
 
+	init_rwsem(&shrinker->rwsem);
 	return 0;
 }
 
@@ -757,7 +758,9 @@ void register_shrinker_prepared(struct shrinker *shrinker)
 {
 	mutex_lock(&shrinker_mutex);
 	list_add_tail_rcu(&shrinker->list, &shrinker_list);
+	down_write(&shrinker->rwsem);
 	shrinker->flags |= SHRINKER_REGISTERED;
+	up_write(&shrinker->rwsem);
 	shrinker_debugfs_add(shrinker);
 	mutex_unlock(&shrinker_mutex);
 }
@@ -802,7 +805,7 @@ EXPORT_SYMBOL(register_shrinker);
 /*
  * Remove one
  */
-void unregister_shrinker(struct shrinker *shrinker)
+void unregister_shrinker_delayed_initiate(struct shrinker *shrinker)
 {
 	struct dentry *debugfs_entry;
 	int debugfs_id;
@@ -812,20 +815,33 @@ void unregister_shrinker(struct shrinker *shrinker)
 
 	mutex_lock(&shrinker_mutex);
 	list_del_rcu(&shrinker->list);
+	down_write(&shrinker->rwsem);
 	shrinker->flags &= ~SHRINKER_REGISTERED;
+	up_write(&shrinker->rwsem);
 	if (shrinker->flags & SHRINKER_MEMCG_AWARE)
 		unregister_memcg_shrinker(shrinker);
 	debugfs_entry = shrinker_debugfs_detach(shrinker, &debugfs_id);
 	mutex_unlock(&shrinker_mutex);
 
+	shrinker_debugfs_remove(debugfs_entry, debugfs_id); // This is moved in your patch
+}
+EXPORT_SYMBOL(unregister_shrinker_delayed_initiate);
+
+void unregister_shrinker_delayed_finalize(struct shrinker *shrinker)
+{
 	atomic_inc(&shrinker_srcu_generation);
 	synchronize_srcu(&shrinker_srcu);
 
-	shrinker_debugfs_remove(debugfs_entry, debugfs_id);
-
 	kfree(shrinker->nr_deferred);
 	shrinker->nr_deferred = NULL;
 }
+EXPORT_SYMBOL(unregister_shrinker_delayed_finalize);
+
+void unregister_shrinker(struct shrinker *shrinker)
+{
+	unregister_shrinker_delayed_initiate(shrinker);
+	unregister_shrinker_delayed_finalize(shrinker);
+}
 EXPORT_SYMBOL(unregister_shrinker);
 
 /**
@@ -856,9 +872,14 @@ static unsigned long do_shrink_slab(struct shrink_control *shrinkctl,
 					  : SHRINK_BATCH;
 	long scanned = 0, next_deferred;
 
+	down_read(&shrinker->rwsem);
+	if (!(shrinker->flags & SHRINKER_REGISTERED))
+		goto unlock;
 	freeable = shrinker->count_objects(shrinker, shrinkctl);
-	if (freeable == 0 || freeable == SHRINK_EMPTY)
-		return freeable;
+	if (freeable == 0 || freeable == SHRINK_EMPTY) {
+		freed = freeable;
+		goto unlock;
+	}
 
 	/*
 	 * copy the current shrinker scan count into a local variable
@@ -935,6 +956,8 @@ static unsigned long do_shrink_slab(struct shrink_control *shrinkctl,
 	 * manner that handles concurrent updates.
 	 */
 	new_nr = add_nr_deferred(next_deferred, shrinker, shrinkctl);
+unlock:
+	up_read(&shrinker->rwsem);
 
 	trace_mm_shrink_slab_end(shrinker, shrinkctl->nid, freed, nr, new_nr, total_scan);
 	return freed;
@@ -968,9 +991,8 @@ static unsigned long shrink_slab_memcg(gfp_t gfp_mask, int nid,
 		struct shrinker *shrinker;
 
 		shrinker = idr_find(&shrinker_idr, i);
-		if (unlikely(!shrinker || !(shrinker->flags & SHRINKER_REGISTERED))) {
-			if (!shrinker)
-				clear_bit(i, info->map);
+		if (unlikely(!shrinker)) {
+			clear_bit(i, info->map);
 			continue;
 		}
 


Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ