From: Qi Zheng <zhengqi.arch@bytedance.com>
To: Kent Overstreet <kent.overstreet@linux.dev>
Cc: david@fromorbit.com, linux-fsdevel@vger.kernel.org,
linux-mm@kvack.org, Andrew Morton <akpm@linux-foundation.org>,
Roman Gushchin <roman.gushchin@linux.dev>
Subject: Re: [PATCH 02/10] mm: shrinker: Add a .to_text() method for shrinkers
Date: Mon, 26 Aug 2024 11:10:08 +0800 [thread overview]
Message-ID: <4733d230-0935-4bff-a17e-8c6735ad16b4@bytedance.com> (raw)
In-Reply-To: <h4i4wn5xnics2wjuwzjmx6pscfuajhrkwbz2sceiihktgzuefr@llohtbim43jg>
On 2024/8/26 11:04, Kent Overstreet wrote:
> On Mon, Aug 26, 2024 at 11:01:50AM GMT, Qi Zheng wrote:
>>
>>
>> On 2024/8/25 03:10, Kent Overstreet wrote:
>>> This adds a new callback method to shrinkers which they can use to
>>> describe anything relevant to memory reclaim about their internal state,
>>> for example object dirtyness.
>>>
>>> This patch also adds shrinkers_to_text(), which reports on the top 10
>>> shrinkers - by object count - in sorted order, to be used in OOM
>>> reporting.
>>>
>>> Cc: Andrew Morton <akpm@linux-foundation.org>
>>> Cc: Qi Zheng <zhengqi.arch@bytedance.com>
>>> Cc: Roman Gushchin <roman.gushchin@linux.dev>
>>> Cc: linux-mm@kvack.org
>>> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
>>> ---
>>> include/linux/shrinker.h | 7 +++-
>>> mm/shrinker.c | 73 +++++++++++++++++++++++++++++++++++++++-
>>> 2 files changed, 78 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/include/linux/shrinker.h b/include/linux/shrinker.h
>>> index 1a00be90d93a..6193612617a1 100644
>>> --- a/include/linux/shrinker.h
>>> +++ b/include/linux/shrinker.h
>>> @@ -24,6 +24,8 @@ struct shrinker_info {
>>> struct shrinker_info_unit *unit[];
>>> };
>>> +struct seq_buf;
>>> +
>>> /*
>>> * This struct is used to pass information from page reclaim to the shrinkers.
>>> * We consolidate the values for easier extension later.
>>> @@ -80,10 +82,12 @@ struct shrink_control {
>>> * @flags determine the shrinker abilities, like numa awareness
>>> */
>>> struct shrinker {
>>> + const char *name;
>>> unsigned long (*count_objects)(struct shrinker *,
>>> struct shrink_control *sc);
>>> unsigned long (*scan_objects)(struct shrinker *,
>>> struct shrink_control *sc);
>>> + void (*to_text)(struct seq_buf *, struct shrinker *);
>>> long batch; /* reclaim batch size, 0 = default */
>>> int seeks; /* seeks to recreate an obj */
>>> @@ -110,7 +114,6 @@ struct shrinker {
>>> #endif
>>> #ifdef CONFIG_SHRINKER_DEBUG
>>> int debugfs_id;
>>> - const char *name;
>>> struct dentry *debugfs_entry;
>>> #endif
>>> /* objs pending delete, per node */
>>> @@ -135,6 +138,8 @@ __printf(2, 3)
>>> struct shrinker *shrinker_alloc(unsigned int flags, const char *fmt, ...);
>>> void shrinker_register(struct shrinker *shrinker);
>>> void shrinker_free(struct shrinker *shrinker);
>>> +void shrinker_to_text(struct seq_buf *, struct shrinker *);
>>> +void shrinkers_to_text(struct seq_buf *);
>>> static inline bool shrinker_try_get(struct shrinker *shrinker)
>>> {
>>> diff --git a/mm/shrinker.c b/mm/shrinker.c
>>> index dc5d2a6fcfc4..ad52c269bb48 100644
>>> --- a/mm/shrinker.c
>>> +++ b/mm/shrinker.c
>>> @@ -1,8 +1,9 @@
>>> // SPDX-License-Identifier: GPL-2.0
>>> #include <linux/memcontrol.h>
>>> +#include <linux/rculist.h>
>>> #include <linux/rwsem.h>
>>> +#include <linux/seq_buf.h>
>>> #include <linux/shrinker.h>
>>> -#include <linux/rculist.h>
>>> #include <trace/events/vmscan.h>
>>> #include "internal.h"
>>> @@ -807,3 +808,73 @@ void shrinker_free(struct shrinker *shrinker)
>>> call_rcu(&shrinker->rcu, shrinker_free_rcu_cb);
>>> }
>>> EXPORT_SYMBOL_GPL(shrinker_free);
>>> +
>>> +void shrinker_to_text(struct seq_buf *out, struct shrinker *shrinker)
>>> +{
>>> + struct shrink_control sc = { .gfp_mask = GFP_KERNEL, };
>>> +
>>> + seq_buf_puts(out, shrinker->name);
>>> + seq_buf_printf(out, " objects: %lu\n", shrinker->count_objects(shrinker, &sc));
>>> +
>>> + if (shrinker->to_text) {
>>> + shrinker->to_text(out, shrinker);
>>> + seq_buf_puts(out, "\n");
>>> + }
>>> +}
>>> +
>>> +/**
>>> + * shrinkers_to_text - Report on shrinkers with highest usage
>>> + *
>>> + * This reports on the top 10 shrinkers, by object counts, in sorted order:
>>> + * intended to be used for OOM reporting.
>>> + */
>>> +void shrinkers_to_text(struct seq_buf *out)
>>> +{
>>> + struct shrinker *shrinker;
>>> + struct shrinker_by_mem {
>>> + struct shrinker *shrinker;
>>> + unsigned long mem;
>>> + } shrinkers_by_mem[10];
>>> + int i, nr = 0;
>>> +
>>> + if (!mutex_trylock(&shrinker_mutex)) {
>>> + seq_buf_puts(out, "(couldn't take shrinker lock)");
>>> + return;
>>> + }
>>
>> I remember I pointed out that the RCU + refcount method should be used
>> here. Otherwise you will block other shrinkers from
>> registering/unregistering, etc.
>
> The more complex iteration isn't needed here - this is a slowpath
> function and we're not doing anything blocking inside it.
The shrinker list may be very long, and can you guarantee that shrinkers
other than bcachefs won't do anything blocking? ;)
next prev parent reply other threads:[~2024-08-26 3:10 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-08-24 19:10 [PATCH 00/10] shrinker debugging, .to_text() report (resend) Kent Overstreet
2024-08-24 19:10 ` [PATCH 01/10] seq_buf: seq_buf_human_readable_u64() Kent Overstreet
2024-08-24 19:10 ` [PATCH 02/10] mm: shrinker: Add a .to_text() method for shrinkers Kent Overstreet
2024-08-26 3:01 ` Qi Zheng
2024-08-26 3:04 ` Kent Overstreet
2024-08-26 3:10 ` Qi Zheng [this message]
2024-08-26 3:32 ` Kent Overstreet
2024-08-28 3:32 ` Dave Chinner
2024-08-24 19:10 ` [PATCH 03/10] mm: shrinker: Add new stats for .to_text() Kent Overstreet
2024-08-28 2:36 ` Dave Chinner
2024-08-28 2:42 ` Kent Overstreet
2024-08-24 19:10 ` [PATCH 04/10] mm: Centralize & improve oom reporting in show_mem.c Kent Overstreet
2024-08-24 19:10 ` [PATCH 05/10] mm: shrinker: Add shrinker_to_text() to debugfs interface Kent Overstreet
2024-08-24 19:10 ` [PATCH 06/10] bcachefs: shrinker.to_text() methods Kent Overstreet
2024-08-24 19:10 ` [PATCH 07/10] percpu: per_cpu_sum() Kent Overstreet
2024-08-24 19:10 ` [PATCH 08/10] fs: Add super_block->s_inodes_nr Kent Overstreet
2024-08-24 19:10 ` [PATCH 09/10] fs/dcache: Add per-sb accounting for nr dentries Kent Overstreet
2024-08-24 19:10 ` [PATCH 10/10] fs: super_cache_to_text() Kent Overstreet
2024-08-28 3:43 ` Dave Chinner
2024-08-28 1:51 ` [PATCH 00/10] shrinker debugging, .to_text() report (resend) Dave Chinner
2024-08-28 2:23 ` Kent Overstreet
[not found] <20240824180454.3160385-1-kent.overstreet@linux.dev>
2024-08-24 18:04 ` [PATCH 02/10] mm: shrinker: Add a .to_text() method for shrinkers Kent Overstreet
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=4733d230-0935-4bff-a17e-8c6735ad16b4@bytedance.com \
--to=zhengqi.arch@bytedance.com \
--cc=akpm@linux-foundation.org \
--cc=david@fromorbit.com \
--cc=kent.overstreet@linux.dev \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=roman.gushchin@linux.dev \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox