From: Lukas Czerner <lczerner@redhat.com>
To: Jinke Han <hanjinke.666@bytedance.com>
Cc: tytso@mit.edu, adilger.kernel@dilger.ca,
linux-ext4@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] ext4: fix trim range leak
Date: Wed, 15 Jun 2022 10:40:17 +0200 [thread overview]
Message-ID: <20220615084017.xwexup5ckrrpevhe@fedora> (raw)
In-Reply-To: <20220614044647.21846-1-hanjinke.666@bytedance.com>
On Tue, Jun 14, 2022 at 12:46:47PM +0800, Jinke Han wrote:
> From: hanjinke <hanjinke.666@bytedance.com>
>
> When release group lock, a large number of blocks may be alloc from
> the group(e.g. not from the rest of target trim range). This may
> lead end of the loop and leave the rest of trim range unprocessed.
Hi,
you're correct. Indeed it's possible to miss some of the blocks this
way.
But I wonder how much of a problem this actually is? I'd think that the
optimization you just took out is very usefull, especially with larger
minlen and more fragmented free space it'll save us a lot of cycles.
Do you have any performance numbers for this change?
Perhaps we don't have to remove it completely, rather zero the
free_count every time bb_free changes? Would that be worth it?
-Lukas
>
> Signed-off-by: hanjinke <hanjinke.666@bytedance.com>
> ---
> fs/ext4/mballoc.c | 6 +-----
> 1 file changed, 1 insertion(+), 5 deletions(-)
>
> diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
> index 9f12f29bc346..45eb9ee20947 100644
> --- a/fs/ext4/mballoc.c
> +++ b/fs/ext4/mballoc.c
> @@ -6345,14 +6345,13 @@ static int ext4_try_to_trim_range(struct super_block *sb,
> __acquires(ext4_group_lock_ptr(sb, e4b->bd_group))
> __releases(ext4_group_lock_ptr(sb, e4b->bd_group))
> {
> - ext4_grpblk_t next, count, free_count;
> + ext4_grpblk_t next, count;
> void *bitmap;
>
> bitmap = e4b->bd_bitmap;
> start = (e4b->bd_info->bb_first_free > start) ?
> e4b->bd_info->bb_first_free : start;
> count = 0;
> - free_count = 0;
>
> while (start <= max) {
> start = mb_find_next_zero_bit(bitmap, max + 1, start);
> @@ -6367,7 +6366,6 @@ __releases(ext4_group_lock_ptr(sb, e4b->bd_group))
> break;
> count += next - start;
> }
> - free_count += next - start;
> start = next + 1;
>
> if (fatal_signal_pending(current)) {
> @@ -6381,8 +6379,6 @@ __releases(ext4_group_lock_ptr(sb, e4b->bd_group))
> ext4_lock_group(sb, e4b->bd_group);
> }
>
> - if ((e4b->bd_info->bb_free - free_count) < minblocks)
> - break;
> }
>
> return count;
> --
> 2.20.1
>
next prev parent reply other threads:[~2022-06-15 8:40 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-06-14 4:46 [PATCH] ext4: fix trim range leak Jinke Han
2022-06-15 8:40 ` Lukas Czerner [this message]
2022-06-16 6:09 ` [External] " hanjinke
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=20220615084017.xwexup5ckrrpevhe@fedora \
--to=lczerner@redhat.com \
--cc=adilger.kernel@dilger.ca \
--cc=hanjinke.666@bytedance.com \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=tytso@mit.edu \
/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