From: Hao Jia <jiahao.kernel@gmail.com>
To: Nhat Pham <nphamcs@gmail.com>
Cc: Yosry Ahmed <yosry@kernel.org>,
akpm@linux-foundation.org, tj@kernel.org, hannes@cmpxchg.org,
shakeel.butt@linux.dev, mhocko@kernel.org, mkoutny@suse.com,
chengming.zhou@linux.dev, muchun.song@linux.dev,
roman.gushchin@linux.dev, cgroups@vger.kernel.org,
linux-mm@kvack.org, linux-kernel@vger.kernel.org,
linux-doc@vger.kernel.org, Hao Jia <jiahao1@lixiang.com>
Subject: Re: [PATCH v3 1/4] mm/zswap: Make shrink_worker writeback cursor per-memcg
Date: Tue, 2 Jun 2026 19:32:35 +0800 [thread overview]
Message-ID: <22473c1e-8448-fc18-e454-7bfb38e89561@gmail.com> (raw)
In-Reply-To: <CAKEwX=NoQNXOMDD0uTSOPWHQX-CMNU1dw=zEuFj=eLcS3fB-ow@mail.gmail.com>
On 2026/6/2 01:08, Nhat Pham wrote:
> On Mon, Jun 1, 2026 at 4:07 AM Hao Jia <jiahao.kernel@gmail.com> wrote:
>>
>>
>>
>> On 2026/5/30 09:24, Yosry Ahmed wrote:
>>> On Tue, May 26, 2026 at 07:45:58PM +0800, Hao Jia wrote:
>>>> From: Hao Jia <jiahao1@lixiang.com>
>>>>
>>>> The zswap background writeback worker shrink_worker() uses a global
>>>> cursor zswap_next_shrink, protected by zswap_shrink_lock, to round-robin
>>>> across the online memcgs under root_mem_cgroup.
>>>>
>>>> Proactive writeback also wants a similar per-memcg cursor that is
>>>> scoped to the specified memcg, so that repeated invocations against
>>>> the same memcg make forward progress across its descendant memcgs
>>>> instead of restarting from the first child memcg each time.
>>>
>>> Is this a problem in practice?
>>>
>>> Is the concern the overhead of scanning memcgs repeatedly, or lack of
>>> fairness? I wonder if we should just do writeback in batches from all
>>> memcgs, similar to how reclaim does it, then evaluate at the end if we
>>> need to start over?
>>>
>>
>> Not using a per-cgroup cursor will cause issues for "repeated
>> small-budget calls" cases. For example, repeatedly triggering a 2MB
>> writeback might result in only writing back pages from the first few
>> child memcgs every time. In the worst-case scenario (where the writeback
>> amount is less than WB_BATCH), it might only ever write back from the
>> first child memcg.
>>
>> Similar to how memory reclaim uses mem_cgroup_iter() (via struct
>> mem_cgroup_reclaim_iter) and the old shrink_worker() used
>> zswap_next_shrink, we need a shared cursor here.
>
> I think each proactive reclaim invocation just walk the entire subtree
> for page reclaim right (see shrink_node_memcgs())? Would that be
> acceptable for you?
Our current approach is very similar to how proactive memory reclaim
works in shrink_node_memcgs().
shrink_node_memcgs() first calls memcg = mem_cgroup_iter(target_memcg,
NULL, partial);. By doing this, it uses
target_memcg->nodeinfo[nid]->iter->position to retrieve the child memcg
where the last reclaim left off, and then resumes the iteration.
The catch is that zswap can't just reuse
memcg->nodeinfo[nid]->iter->position, as that would mess up the cursor
used by the memory reclaim.
>
> I also wonder if we can at least make this structure dynamically
> allocated... In a system, you only really invoke proactive reclaim
> against a few target cgroups, no?
It is possible to allocate it dynamically, but I am concerned that it
might introduce a slight performance overhead. We would need to add a
**check** like if (READ_ONCE(memcg->zswap_wb_iter)) every time
zswap_mem_cgroup_iter() is called. Furthermore, to handle concurrent
allocations, we might also need to introduce cmpxchg() to resolve race
conditions.
The additional code would look something like this:
static struct zswap_wb_iter *get_zswap_wb_iter(struct mem_cgroup *memcg)
{
struct zswap_wb_iter *iter, *new_iter;
iter = READ_ONCE(memcg->zswap_wb_iter);
if (likely(iter))
return iter;
new_iter = kzalloc(sizeof(*new_iter), GFP_KERNEL);
if (!new_iter)
return NULL;
spin_lock_init(&new_iter->lock);
if (cmpxchg(&memcg->zswap_wb_iter, NULL, new_iter) != NULL) {
/* Lost the race, someone else installed first. */
kfree(new_iter);
}
return READ_ONCE(memcg->zswap_wb_iter);
}
Thanks,
Hao
next prev parent reply other threads:[~2026-06-02 11:32 UTC|newest]
Thread overview: 76+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-05-26 11:45 [PATCH v3 0/4] mm/zswap: Implement per-cgroup proactive writeback Hao Jia
2026-05-26 11:45 ` [PATCH v3 1/4] mm/zswap: Make shrink_worker writeback cursor per-memcg Hao Jia
2026-05-29 19:51 ` Nhat Pham
2026-05-30 1:24 ` Yosry Ahmed
2026-06-01 11:07 ` Hao Jia
2026-06-01 16:44 ` Nhat Pham
2026-06-01 16:47 ` Nhat Pham
2026-06-01 17:08 ` Nhat Pham
2026-06-02 11:32 ` Hao Jia [this message]
2026-06-02 0:31 ` Yosry Ahmed
2026-06-02 11:33 ` Hao Jia
2026-06-02 23:19 ` Yosry Ahmed
2026-06-03 3:02 ` Hao Jia
2026-06-03 17:53 ` Yosry Ahmed
2026-06-04 1:58 ` Hao Jia
2026-06-04 5:34 ` Yosry Ahmed
2026-06-04 13:06 ` Hao Jia
2026-06-04 16:10 ` Yosry Ahmed
2026-06-04 17:23 ` Nhat Pham
2026-06-08 12:50 ` Hao Jia
2026-06-08 16:23 ` Nhat Pham
2026-06-08 16:44 ` Yosry Ahmed
2026-06-08 16:48 ` Yosry Ahmed
2026-06-08 18:01 ` Nhat Pham
2026-06-09 3:18 ` Hao Jia
2026-06-11 17:39 ` Yosry Ahmed
2026-06-12 16:40 ` Shakeel Butt
2026-06-12 18:15 ` Yosry Ahmed
2026-06-15 2:45 ` Hao Jia
2026-05-26 11:45 ` [PATCH v3 2/4] mm/zswap: Implement proactive writeback Hao Jia
2026-05-29 19:58 ` Nhat Pham
2026-05-30 1:40 ` Yosry Ahmed
2026-06-03 11:22 ` Hao Jia
2026-06-03 17:58 ` Yosry Ahmed
2026-06-03 18:14 ` Nhat Pham
2026-06-04 2:11 ` Hao Jia
2026-06-04 5:36 ` Yosry Ahmed
2026-06-04 14:01 ` Shakeel Butt
2026-06-08 18:30 ` Shakeel Butt
2026-06-08 19:50 ` Shakeel Butt
2026-06-08 20:19 ` Yosry Ahmed
2026-06-08 22:22 ` Shakeel Butt
2026-06-08 22:27 ` Yosry Ahmed
2026-06-09 4:19 ` YoungJun Park
2026-06-11 17:45 ` Yosry Ahmed
2026-06-11 19:12 ` Shakeel Butt
2026-06-12 7:27 ` YoungJun Park
2026-06-12 17:02 ` [swap tier discussion] " Shakeel Butt
2026-06-12 21:31 ` Yosry Ahmed
2026-06-14 9:23 ` YoungJun Park
2026-06-15 19:55 ` Yosry Ahmed
2026-06-16 1:03 ` YoungJun Park
2026-06-16 3:08 ` YoungJun Park
2026-06-16 17:30 ` Yosry Ahmed
2026-06-16 18:32 ` Nhat Pham
2026-06-16 19:54 ` Yosry Ahmed
2026-06-16 20:08 ` Nhat Pham
2026-06-16 20:10 ` Yosry Ahmed
2026-06-16 20:24 ` Nhat Pham
2026-06-16 20:26 ` Yosry Ahmed
2026-06-17 17:11 ` Nhat Pham
2026-06-16 11:44 ` Shakeel Butt
2026-05-30 1:37 ` Yosry Ahmed
2026-06-03 11:27 ` Hao Jia
2026-06-03 17:55 ` Yosry Ahmed
2026-06-03 18:23 ` Nhat Pham
2026-06-03 18:26 ` Yosry Ahmed
2026-06-03 18:34 ` Nhat Pham
2026-06-03 18:43 ` Yosry Ahmed
2026-06-03 18:51 ` Nhat Pham
2026-06-03 18:54 ` Yosry Ahmed
2026-05-26 11:46 ` [PATCH v3 3/4] mm/zswap: Add per-memcg stat for " Hao Jia
2026-05-29 20:01 ` Nhat Pham
2026-06-03 11:29 ` Hao Jia
2026-05-26 11:46 ` [PATCH v3 4/4] selftests/cgroup: Add tests for zswap " Hao Jia
2026-05-29 20:02 ` Nhat Pham
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=22473c1e-8448-fc18-e454-7bfb38e89561@gmail.com \
--to=jiahao.kernel@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=cgroups@vger.kernel.org \
--cc=chengming.zhou@linux.dev \
--cc=hannes@cmpxchg.org \
--cc=jiahao1@lixiang.com \
--cc=linux-doc@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mhocko@kernel.org \
--cc=mkoutny@suse.com \
--cc=muchun.song@linux.dev \
--cc=nphamcs@gmail.com \
--cc=roman.gushchin@linux.dev \
--cc=shakeel.butt@linux.dev \
--cc=tj@kernel.org \
--cc=yosry@kernel.org \
/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
all inboxes | Powered by JetHome®