From: Miao Xie <miaox@cn.fujitsu.com>
To: Itaru Kitayama <kitayama@cl.bb4u.ne.jp>
Cc: linux-btrfs@vger.kernel.org
Subject: Re: [PATCH V3] btrfs: implement delayed inode items operation
Date: Wed, 16 Mar 2011 22:24:31 +0800 [thread overview]
Message-ID: <4D80C81F.60507@cn.fujitsu.com> (raw)
In-Reply-To: <20110316184850.f25551a6.kitayama@cl.bb4u.ne.jp>
On Wed, 16 Mar 2011 18:48:50 +0900, Itaru Kitayama wrote:
> Hi Miao,
>
> The V4 still hangs. Does the new function, btrfs_update_inode_nodelayed() you introduced to the V4
> need take care of a delayed node (commit delayed items of the inode and set the delayed note to NULL)?
I think there no delayed item because we commit all the delayed items when we start committing
transaction.
>
> Below is my experimental patch on top of V4 trying to avoid taking trans_mutex in btrfs_remove_delayed_node(),
> but still hangs. Could you take a look at it? Thanks.
It is so strange I can not reproduce it on my machine. But by reviewing the code, I think some inodes
may be updated when committing the transaction, it makes the filesystem hang.
>
> Itaru
>
> diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c
> index 495a6af..b6ea082 100644
> --- a/fs/btrfs/delayed-inode.c
> +++ b/fs/btrfs/delayed-inode.c
> @@ -1006,9 +1006,9 @@ int btrfs_run_delayed_items(struct btrfs_trans_handle *trans,
> return ret;
> }
>
> -int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
> - struct btrfs_root *root,
> - struct inode *inode)
> +int __btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
> + struct btrfs_root *root,
> + struct inode *inode, int delayed_update)
> {
> struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode);
> struct btrfs_path *path;
> @@ -1036,7 +1036,7 @@ int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
> if (!ret)
> ret = btrfs_delete_delayed_items(trans, path, root,
> delayed_node);
> - if (!ret)
> + if (!ret && delayed_update)
> ret = btrfs_update_delayed_inode(trans, root, path,
> delayed_node);
> btrfs_free_path(path);
> @@ -1044,13 +1044,27 @@ int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
> return ret;
> }
>
> +int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
> + struct btrfs_root *root,
> + struct inode *inode)
> +{
> + return __btrfs_commit_inode_delayed_items(trans, root, inode, 1);
> +}
> +
> +int btrfs_commit_inode_delayed_items_no_delayed_update(struct btrfs_trans_handle *trans,
> + struct btrfs_root *root,
> + struct inode *inode)
> +{
> + return __btrfs_commit_inode_delayed_items(trans, root, inode, 0);
> +}
> +
> int btrfs_remove_delayed_node(struct inode *inode)
> {
> struct btrfs_trans_handle *trans;
> struct btrfs_root *root = BTRFS_I(inode)->root;
> struct btrfs_delayed_node *delayed_node = BTRFS_I(inode)->delayed_node;
> int ret;
> -
> +
> if (!delayed_node)
> return 0;
>
> diff --git a/fs/btrfs/delayed-inode.h b/fs/btrfs/delayed-inode.h
> index d4134bc..b6986df 100644
> --- a/fs/btrfs/delayed-inode.h
> +++ b/fs/btrfs/delayed-inode.h
> @@ -114,6 +114,11 @@ void btrfs_balance_delayed_items(struct btrfs_root *root);
> int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
> struct btrfs_root *root,
> struct inode *inode);
> +int btrfs_commit_inode_delayed_items_no_delayed_update(
> + struct btrfs_trans_handle *trans,
> + struct btrfs_root *root,
> + struct inode *inode);
> +
> /* Used for evicting the inode. */
> int btrfs_remove_delayed_node(struct inode *inode);
> int btrfs_kill_delayed_inode_items(struct btrfs_trans_handle *trans,
> diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
> index 5299d9c..e198051 100644
> --- a/fs/btrfs/inode.c
> +++ b/fs/btrfs/inode.c
> @@ -2611,8 +2611,12 @@ noinline int btrfs_update_inode_nodelayed(struct btrfs_trans_handle *trans,
> struct btrfs_inode_item *inode_item;
> struct btrfs_path *path;
> struct extent_buffer *leaf;
> + struct btrfs_delayed_node *delayed_node;
> int ret;
>
> + btrfs_commit_inode_delayed_items_no_delayed_update(trans, root, inode);
> + delayed_node = NULL;
> +
> path = btrfs_alloc_path();
> if (!path)
> return -ENOMEM;
>
next prev parent reply other threads:[~2011-03-16 14:24 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-03-09 15:03 [PATCH V3] btrfs: implement delayed inode items operation Itaru Kitayama
2011-03-10 6:21 ` Miao Xie
2011-03-16 9:48 ` Itaru Kitayama
2011-03-16 14:24 ` Miao Xie [this message]
2011-03-16 17:49 ` Chris Mason
2011-03-17 1:37 ` Itaru Kitayama
2011-03-18 3:37 ` Miao Xie
2011-03-18 8:59 ` Itaru Kitayama
2011-03-18 9:18 ` Miao Xie
-- strict thread matches above, loose matches on Subject: below --
2011-03-03 6:06 Miao Xie
2011-03-05 7:01 ` Itaru Kitayama
2011-03-06 23:25 ` Tsutomu Itoh
2011-03-07 1:57 ` Miao Xie
2011-03-10 3:07 ` Itaru Kitayama
2011-03-08 16:56 ` David Sterba
2011-03-09 5:40 ` Miao Xie
2011-03-11 15:24 ` David Sterba
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=4D80C81F.60507@cn.fujitsu.com \
--to=miaox@cn.fujitsu.com \
--cc=kitayama@cl.bb4u.ne.jp \
--cc=linux-btrfs@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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.