From: David Sterba <dsterba@suse.cz>
To: Mark Fasheh <mfasheh@suse.de>
Cc: dsterba@suse.cz, Josef Bacik <jbacik@fb.com>,
linux-btrfs@vger.kernel.org, Chris Mason <clm@fb.com>
Subject: Re: btrfs: qgroup: account shared subtrees during snapshot delete
Date: Mon, 23 Jun 2014 16:49:13 +0200 [thread overview]
Message-ID: <20140623144913.GI1553@twin.jikos.cz> (raw)
In-Reply-To: <20140620152958.GR5484@wotan.suse.de>
On Fri, Jun 20, 2014 at 08:29:58AM -0700, Mark Fasheh wrote:
> +static int account_shared_subtree(struct btrfs_trans_handle *trans,
> + struct btrfs_root *root,
> + struct extent_buffer *root_eb,
> + u64 root_gen,
> + int root_level)
> +{
> + int ret;
> + int level;
> + struct extent_buffer *eb = root_eb;
> + struct btrfs_path *path = NULL;
> +
> + BUG_ON(root_level < 0 || root_level > BTRFS_MAX_LEVEL);
> + BUG_ON(root_eb == NULL);
The caller is ready for errors, can you please turn these to ane EIO?
(Separate patch is fine)
> +static int qgroup_subtree_accounting(struct btrfs_trans_handle *trans,
> + struct btrfs_fs_info *fs_info,
> + struct btrfs_qgroup_operation *oper)
> +{
> + struct ulist *roots = NULL;
> + struct ulist_node *unode;
> + struct ulist_iterator uiter;
> + struct btrfs_qgroup_list *glist;
> + struct ulist *parents;
> + int ret = 0;
> + struct btrfs_qgroup *qg;
> + u64 root_obj = 0;
> + struct seq_list elem = {};
> +
> + parents = ulist_alloc(GFP_NOFS);
> + if (!parents)
> + return -ENOMEM;
> +
> + btrfs_get_tree_mod_seq(fs_info, &elem);
> + ret = btrfs_find_all_roots(trans, fs_info, oper->bytenr,
> + elem.seq, &roots);
> + btrfs_put_tree_mod_seq(fs_info, &elem);
> + if (ret < 0)
> + return ret;
> +
> + if (roots->nnodes != 1)
> + goto out;
> +
> + ULIST_ITER_INIT(&uiter);
> + unode = ulist_next(roots, &uiter); /* Only want 1 so no need to loop */
> + /*
> + * If we find our ref root then that means all refs
> + * this extent has to the root have not yet been
> + * deleted. In that case, we do nothing and let the
> + * last ref for this bytenr drive our update.
> + *
> + * This can happen for example if an extent is
> + * referenced multiple times in a snapshot (clone,
> + * etc). If we are in the middle of snapshot removal,
> + * queued updates for such an extent will find the
> + * root if we have not yet finished removing the
> + * snapshot.
> + */
> + if (unode->val == oper->ref_root)
> + goto out;
> +
> + root_obj = unode->val;
> + BUG_ON(!root_obj);
Same here, although when called through btrfs_qgroup_account, the error
code is not properly checked in btrfs_delayed_qgroup_accounting, line 1944:
1933 int btrfs_delayed_qgroup_accounting(struct btrfs_trans_handle *trans,
1934 struct btrfs_fs_info *fs_info)
1935 {
1936 struct btrfs_qgroup_operation *oper;
1937 int ret = 0;
1938
1939 while (!list_empty(&trans->qgroup_ref_list)) {
1940 oper = list_first_entry(&trans->qgroup_ref_list,
1941 struct btrfs_qgroup_operation, list);
1942 list_del_init(&oper->list);
1943 if (!ret || !trans->aborted)
1944 ret = btrfs_qgroup_account(trans, fs_info, oper);
1945 spin_lock(&fs_info->qgroup_op_lock);
1946 rb_erase(&oper->n, &fs_info->qgroup_op_tree);
1947 spin_unlock(&fs_info->qgroup_op_lock);
1948 btrfs_put_tree_mod_seq(fs_info, &oper->elem);
1949 kfree(oper);
1950 }
1951 return ret;
1952 }
next prev parent reply other threads:[~2014-06-23 14:49 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-06-19 21:49 [PATCH 0/2] btrfs: qgroup fixes for btrfs_drop_snapshot Mark Fasheh
2014-06-19 21:49 ` btrfs: add trace for qgroup accounting Mark Fasheh
2014-06-19 21:58 ` Josef Bacik
2014-06-19 21:49 ` btrfs: qgroup: account shared subtrees during snapshot delete Mark Fasheh
2014-06-19 22:25 ` Josef Bacik
2014-06-19 23:16 ` Mark Fasheh
2014-06-19 23:17 ` Josef Bacik
2014-06-20 11:25 ` David Sterba
2014-06-20 15:29 ` Mark Fasheh
2014-06-20 15:44 ` Josef Bacik
2014-06-20 17:18 ` Mark Fasheh
2014-06-23 14:49 ` David Sterba [this message]
2014-07-07 22:09 [PATCH 0/3] btrfs: qgroup fixes for btrfs_drop_snapshot V3 Mark Fasheh
2014-07-07 22:09 ` btrfs: qgroup: account shared subtrees during snapshot delete Mark Fasheh
2015-02-01 20:51 Dan Carpenter
2015-02-06 22:16 ` Mark Fasheh
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=20140623144913.GI1553@twin.jikos.cz \
--to=dsterba@suse.cz \
--cc=clm@fb.com \
--cc=jbacik@fb.com \
--cc=linux-btrfs@vger.kernel.org \
--cc=mfasheh@suse.de \
/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.