public inbox for linux-btrfs@vger.kernel.org
 help / color / mirror / Atom feed
From: David Woodhouse <dwmw2@infradead.org>
To: Chris Mason <chris.mason@oracle.com>
Cc: Balaji Rao <balajirrao@gmail.com>, linux-btrfs@vger.kernel.org
Subject: Re: [PATCH] NFS support for btrfs - v2
Date: Mon, 18 Aug 2008 21:20:39 +0100	[thread overview]
Message-ID: <1219090839.3184.436.camel@pmac.infradead.org> (raw)
In-Reply-To: <1219088859.14063.35.camel@think.oraclecorp.com>

On Mon, 2008-08-18 at 15:47 -0400, Chris Mason wrote:
> Lets pretend I had put in commments something like the code below.
> The important part is that directories have only one link, so they
> have only one backref.

OK. Now can I rip that code out anyway? The VFS will never call
btrfs_lookup() for ".." -- not since the 2.2 kernel :)

I'm still a little confused about precisely what btrfs_search_slot()
returns when it doesn't find a match -- that's probably where the
documentation would be more useful.

What I have in btrfs_get_parent() is correct, I think -- but could
probably be simplified to your version, searching for (u64)-1 and then
not needing to call btrfs_next_leaf().

Something like this?

(And am I fixing a use-after-free bug by moving that btrfs_free_path()
down by a line, at the end?)

diff --git a/export.c b/export.c
index 797b4cb..830e87a 100644
--- a/export.c
+++ b/export.c
@@ -156,27 +156,27 @@ static struct dentry *btrfs_get_parent(struct dentry *child)
 
 	key.objectid = dir->i_ino;
 	btrfs_set_key_type(&key, BTRFS_INODE_REF_KEY);
-	key.offset = 0;
+	key.offset = (u64)-1;
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
-	BUG_ON(ret == 0);
-	ret = 0;
 
 	leaf = path->nodes[0];
 	slot = path->slots[0];
-	nritems = btrfs_header_nritems(leaf);
-	if (slot >= nritems) {
-		ret = btrfs_next_leaf(root, path);
-		if (ret) {
-			btrfs_free_path(path);
-			goto out;
-		}
-		leaf = path->nodes[0];
-		slot = path->slots[0];
+	if (ret < 0 && slot == 0) {
+		btrfs_free_path(path);
+		goto out;
 	}
+	/*
+	  This will be in the slot _before_ the one that btrfs_search_slot()
+	   returns. And for some reason which dwmw2 doesn't quite understand,
+	   that'll definitely be in the same leaf that btrfs_search_slot()
+	   returned -- even if btrfs_search_slot() had to look in the _next_
+	   leaf to find the first key which is greater than what we asked for 
+	*/
+	slot--;
 
+	btrfs_item_key_to_cpu(leaf, &key, slot);
 	btrfs_free_path(path);
 
-	btrfs_item_key_to_cpu(leaf, &key, slot);
 	if (key.objectid != dir->i_ino || key.type != BTRFS_INODE_REF_KEY)
 		goto out;
 


-- 
David Woodhouse                            Open Source Technology Centre
David.Woodhouse@intel.com                              Intel Corporation




  reply	other threads:[~2008-08-18 20:20 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-07-20 20:31 [PATCH] NFS support for btrfs - v2 Balaji Rao
2008-08-17 11:53 ` David Woodhouse
2008-08-17 12:51   ` Balaji Rao
2008-08-17 12:56     ` David Woodhouse
2008-08-17 13:24       ` Balaji Rao
2008-08-17 13:30         ` David Woodhouse
2008-08-17 14:17           ` David Woodhouse
2008-08-17 16:10             ` [PATCH] rewrite btrfs_readdir() David Woodhouse
2008-08-18 18:46               ` Chris Mason
2008-08-18 19:08                 ` David Woodhouse
2008-08-18 19:24                   ` Chris Mason
2008-08-18 19:32                     ` David Woodhouse
2008-08-17 13:40         ` [PATCH] NFS support for btrfs - v2 David Woodhouse
2008-08-18 19:23           ` Chris Mason
2008-08-18 19:33             ` David Woodhouse
2008-08-18 19:47               ` Chris Mason
2008-08-18 20:20                 ` David Woodhouse [this message]
2008-08-18 20:32                   ` Chris Mason
2008-08-18 21:52                     ` David Woodhouse
2008-08-19 11:54                       ` Chris Mason
2008-08-19 14:49                         ` David Woodhouse
2008-08-19 21:34                           ` David Woodhouse
2008-08-19  0:16                   ` Christoph Hellwig
2008-08-19  0:21                     ` David Woodhouse
2008-08-18 11:51     ` David Woodhouse
2008-08-18 12:10       ` David Woodhouse
2008-08-18 19:15         ` Chris Mason

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=1219090839.3184.436.camel@pmac.infradead.org \
    --to=dwmw2@infradead.org \
    --cc=balajirrao@gmail.com \
    --cc=chris.mason@oracle.com \
    --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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox