From: Chao Yu <chao@kernel.org>
To: Daeho Jeong <daeho43@gmail.com>
Cc: chao@kernel.org, jaegeuk@kernel.org,
linux-kernel@vger.kernel.org,
linux-f2fs-devel@lists.sourceforge.net
Subject: Re: [f2fs-dev] [PATCH v3 01/12] f2fs: cache: implement metadata cache
Date: Fri, 28 Aug 2026 09:36:19 +0800 [thread overview]
Message-ID: <e223e7e8-92cb-4264-a39e-08ca1dfb4cce@kernel.org> (raw)
In-Reply-To: <CACOAw_z-QjubNTcodONJoa07qQvvNufBE7sDVegyonJoFPoARQ@mail.gmail.com>
On 8/28/26 00:56, Daeho Jeong wrote:
> The idea is to drop list_lock completely from f2fs_find_cache() by
> deferring LRU rotation to the shrinker (like page cache's
> PG_referenced / clock algorithm):
Ah, I see, let me implement the logic as you suggested, thanks for the idea!
>
> In f2fs_find_cache() (no list_lock at all):
>
> if (entry) {
> f2fs_cache_get(entry);
> set_bit(F2FS_BLOCK_REFERENCED, &entry->state);
> }
>
> In f2fs_do_shrink_cache() Phase 1 (under list_lock):
>
> list_for_each_entry_safe(entry, next, &cache->lru_list, list) {
> /* If accessed, give it a second chance and rotate to tail */
> if (test_and_clear_bit(F2FS_BLOCK_REFERENCED, &entry->state)) {
> list_move_tail(&entry->list, &cache->lru_list);
> continue;
> }
> ...
> /* Reclaim cold entry */
> }
>
> This completely eliminates list_lock overhead from the read lookup
> path while keeping active entries protected from eviction.
Thanks,
next prev parent reply other threads:[~2026-08-28 1:36 UTC|newest]
Thread overview: 42+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-08-25 13:01 [PATCH v3 00/12] f2fs: introduce " Chao Yu
2026-08-25 13:01 ` [PATCH v3 01/12] f2fs: cache: implement " Chao Yu
2026-08-26 19:14 ` [f2fs-dev] " Daeho Jeong
2026-08-27 1:16 ` Chao Yu
2026-08-27 16:56 ` Daeho Jeong
2026-08-28 1:36 ` Chao Yu [this message]
2026-08-25 13:01 ` [PATCH v3 02/12] f2fs: cache: initialize meta cache Chao Yu
2026-08-25 13:01 ` [PATCH v3 03/12] f2fs: cache: introduce shrinker Chao Yu
2026-08-26 19:19 ` [f2fs-dev] " Daeho Jeong
2026-08-27 1:19 ` Chao Yu
2026-08-27 17:00 ` Daeho Jeong
2026-08-28 1:59 ` Chao Yu
2026-08-28 2:56 ` Chao Yu
2026-08-28 18:18 ` Daeho Jeong
2026-08-25 13:01 ` [PATCH v3 04/12] f2fs: cache: introduce writeback thread Chao Yu
2026-08-26 19:37 ` [f2fs-dev] " Daeho Jeong
2026-08-27 1:28 ` Chao Yu
2026-08-27 17:03 ` Daeho Jeong
2026-08-25 13:01 ` [PATCH v3 05/12] f2fs: cache: use meta cache Chao Yu
2026-08-26 20:22 ` [f2fs-dev] " Daeho Jeong
2026-08-27 2:26 ` Chao Yu
2026-08-27 6:46 ` Chao Yu
2026-08-25 13:01 ` [PATCH v3 06/12] f2fs: cache: initialize node cache Chao Yu
2026-08-25 13:01 ` [PATCH v3 07/12] f2fs: cache: use " Chao Yu
2026-08-25 13:01 ` [PATCH v3 08/12] f2fs: cache: initialize compress cache Chao Yu
2026-08-25 13:01 ` [PATCH v3 09/12] f2fs: cache: use " Chao Yu
2026-08-26 20:04 ` [f2fs-dev] " Daeho Jeong
2026-08-26 20:23 ` Daeho Jeong
2026-08-27 2:27 ` Chao Yu
2026-08-27 2:22 ` Chao Yu
2026-08-27 17:12 ` Daeho Jeong
2026-08-27 3:08 ` Chao Yu
2026-08-27 17:13 ` Daeho Jeong
2026-08-27 3:24 ` Chao Yu
2026-08-27 17:20 ` Daeho Jeong
2026-08-28 3:32 ` Chao Yu
2026-08-28 12:00 ` Chao Yu
2026-08-28 12:25 ` Chao Yu
2026-08-28 18:22 ` Daeho Jeong
2026-08-25 13:01 ` [PATCH v3 10/12] f2fs: cache: support fault injection Chao Yu
2026-08-25 13:01 ` [PATCH v3 11/12] f2fs: cache: introduce tracepoints Chao Yu
2026-08-25 13:01 ` [PATCH v3 12/12] f2fs: cache: show per-cache usage in debugfs Chao Yu
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=e223e7e8-92cb-4264-a39e-08ca1dfb4cce@kernel.org \
--to=chao@kernel.org \
--cc=daeho43@gmail.com \
--cc=jaegeuk@kernel.org \
--cc=linux-f2fs-devel@lists.sourceforge.net \
--cc=linux-kernel@vger.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®