From: Josef Bacik <josef@redhat.com>
To: Li Zefan <lizf@cn.fujitsu.com>
Cc: Chris Mason <chris.mason@oracle.com>,
Josef Bacik <josef@redhat.com>,
linux-btrfs@vger.kernel.org, LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 1/6] btrfs: fix threshold calculation for block groups smaller than 1GB
Date: Mon, 23 Aug 2010 09:02:25 -0400 [thread overview]
Message-ID: <20100823130224.GA2404@localhost.localdomain> (raw)
In-Reply-To: <4C71CD60.5050507@cn.fujitsu.com>
On Mon, Aug 23, 2010 at 09:22:40AM +0800, Li Zefan wrote:
> If a block group is smaller than 1GB, the extent entry threadhold
> calculation will always set the threshold to 0.
>
> So as free space gets fragmented, btrfs will switch to use bitmap
> to manage free space, but then will never switch back to extents
> due to this bug.
>
> Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
> ---
> fs/btrfs/free-space-cache.c | 8 ++++++--
> 1 files changed, 6 insertions(+), 2 deletions(-)
>
> diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
> index f488fac..7edbef6 100644
> --- a/fs/btrfs/free-space-cache.c
> +++ b/fs/btrfs/free-space-cache.c
> @@ -263,14 +263,18 @@ static void recalculate_thresholds(struct btrfs_block_group_cache *block_group)
> u64 max_bytes;
> u64 bitmap_bytes;
> u64 extent_bytes;
> + u64 size = block_group->key.offset;
>
> /*
> * The goal is to keep the total amount of memory used per 1gb of space
> * at or below 32k, so we need to adjust how much memory we allow to be
> * used by extent based free space tracking
> */
> - max_bytes = MAX_CACHE_BYTES_PER_GIG *
> - (div64_u64(block_group->key.offset, 1024 * 1024 * 1024));
> + if (size < 1024 * 1024 * 1024)
> + max_bytes = MAX_CACHE_BYTES_PER_GIG;
> + else
> + max_bytes = MAX_CACHE_BYTES_PER_GIG *
> + div64_u64(size, 1024 * 1024 * 1024);
>
> /*
> * we want to account for 1 more bitmap than what we have so we can make
> --
> 1.7.0.1
>
Reviewed-by: Josef Bacik <josef@redhat.com>
Thanks,
Josef
next prev parent reply other threads:[~2010-08-23 13:02 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-08-23 1:22 [PATCH 0/6] btrfs: some bug fixes for free space cache Li Zefan
2010-08-23 1:22 ` [PATCH 1/6] btrfs: fix threshold calculation for block groups smaller than 1GB Li Zefan
2010-08-23 13:02 ` Josef Bacik [this message]
2010-08-23 1:23 ` [PATCH 2/6] btrfs: add helper function free_bitmap() Li Zefan
2010-08-23 13:03 ` Josef Bacik
2010-08-23 1:23 ` [PATCH 3/6] btrfs: free fully occupied bitmap in cluster Li Zefan
2010-08-23 13:04 ` Josef Bacik
2010-08-23 1:23 ` [PATCH 4/6] btrfs: update stats when allocating from a cluster Li Zefan
2010-08-23 13:09 ` Josef Bacik
2010-08-24 0:50 ` Li Zefan
2010-08-24 2:50 ` Josef Bacik
2010-08-23 1:23 ` [PATCH 5/6] btrfs: add a helper try_merge_free_space() Li Zefan
2010-08-23 13:15 ` Josef Bacik
2010-08-23 1:24 ` [PATCH 6/6] btrfs: check mergeable free space when removing a cluster Li Zefan
2010-08-23 13:15 ` Josef Bacik
2010-08-24 1:04 ` Li Zefan
2010-08-24 2:58 ` Josef Bacik
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=20100823130224.GA2404@localhost.localdomain \
--to=josef@redhat.com \
--cc=chris.mason@oracle.com \
--cc=linux-btrfs@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=lizf@cn.fujitsu.com \
/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;
as well as URLs for NNTP newsgroup(s).