All of lore.kernel.org
 help / color / mirror / Atom feed
From: Joel Becker <Joel.Becker@oracle.com>
To: Louis Rilling <louis.rilling@kerlabs.com>
Cc: linux-kernel@vger.kernel.org, ocfs2-devel@oss.oracle.com
Subject: [Ocfs2-devel] [BUGFIX][PATCH 3/3] configfs: Fix failing symlink() making	rmdir() fail
Date: Thu, 19 Jun 2008 15:16:05 -0700	[thread overview]
Message-ID: <20080619221605.GE10888@mail.oracle.com> (raw)
In-Reply-To: <1213724243-29183-4-git-send-email-louis.rilling@kerlabs.com>

On Tue, Jun 17, 2008 at 07:37:23PM +0200, Louis Rilling wrote:

	Ok, I can see some of why I hated this.

> +
>  		spin_lock(&configfs_dirent_lock);
> -		if (target_sd->s_type & CONFIGFS_USET_DROPPING) {
> -			spin_unlock(&configfs_dirent_lock);
> -			config_item_put(item);
> -			kfree(sl);
> -			return -EPERM;
> -		}
> -		list_add(&sl->sl_list, &target_sd->s_links);
> +		/*
> +		 * Force rmdir() of parent_item to wait until we know if we
> +		 * succeed.
> +		 */
> +		parent_sd->s_type |= CONFIGFS_USET_ATTACHING;
>  		spin_unlock(&configfs_dirent_lock);
> +
>  		ret = configfs_create_link(sl, parent_item->ci_dentry,
>  					   dentry);
> -		if (ret) {
> -			spin_lock(&configfs_dirent_lock);
> -			list_del_init(&sl->sl_list);
> +
> +		spin_lock(&configfs_dirent_lock);
> +		parent_sd->s_type &= ~CONFIGFS_USET_ATTACHING;
> +
> +		if (ret || target_sd->s_type & CONFIGFS_USET_DROPPING) {

	Use parenthesis.  Actually, separate out the error cases.

> +			struct configfs_dirent *sd = NULL;
> +
> +			if (!ret) {
> +				sd = dentry->d_fsdata;
> +				list_del_init(&sd->s_sibling);
> +			}
>  			spin_unlock(&configfs_dirent_lock);
> +
> +			if (!ret) {
> +				configfs_drop_dentry(sd, dentry->d_parent);
> +				d_delete(dentry);
> +				configfs_put(sd);
> +			}

	This open-code of the VFS munging is ripe to break when the VFS
changes or other configfs changes happen.  The real issue is that you
are reimplementing the core of configfs_unlink().  Note how the core VFS
munging of configfs_rmdir() is separated out so that configfs_mkdir()
can also use it in the failure case?  Do the same with unlink() and it
will read much better ("if (DROPPING) configfs_delete_link()").  Call it
configfs_remove_link() or configfs_delete_link().

Joel

-- 

Life's Little Instruction Book #337

	"Reread your favorite book."

Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127

WARNING: multiple messages have this Message-ID (diff)
From: Joel Becker <Joel.Becker@oracle.com>
To: Louis Rilling <louis.rilling@kerlabs.com>
Cc: linux-kernel@vger.kernel.org, ocfs2-devel@oss.oracle.com
Subject: Re: [BUGFIX][PATCH 3/3] configfs: Fix failing symlink() making rmdir() fail
Date: Thu, 19 Jun 2008 15:16:05 -0700	[thread overview]
Message-ID: <20080619221605.GE10888@mail.oracle.com> (raw)
In-Reply-To: <1213724243-29183-4-git-send-email-louis.rilling@kerlabs.com>

On Tue, Jun 17, 2008 at 07:37:23PM +0200, Louis Rilling wrote:

	Ok, I can see some of why I hated this.

> +
>  		spin_lock(&configfs_dirent_lock);
> -		if (target_sd->s_type & CONFIGFS_USET_DROPPING) {
> -			spin_unlock(&configfs_dirent_lock);
> -			config_item_put(item);
> -			kfree(sl);
> -			return -EPERM;
> -		}
> -		list_add(&sl->sl_list, &target_sd->s_links);
> +		/*
> +		 * Force rmdir() of parent_item to wait until we know if we
> +		 * succeed.
> +		 */
> +		parent_sd->s_type |= CONFIGFS_USET_ATTACHING;
>  		spin_unlock(&configfs_dirent_lock);
> +
>  		ret = configfs_create_link(sl, parent_item->ci_dentry,
>  					   dentry);
> -		if (ret) {
> -			spin_lock(&configfs_dirent_lock);
> -			list_del_init(&sl->sl_list);
> +
> +		spin_lock(&configfs_dirent_lock);
> +		parent_sd->s_type &= ~CONFIGFS_USET_ATTACHING;
> +
> +		if (ret || target_sd->s_type & CONFIGFS_USET_DROPPING) {

	Use parenthesis.  Actually, separate out the error cases.

> +			struct configfs_dirent *sd = NULL;
> +
> +			if (!ret) {
> +				sd = dentry->d_fsdata;
> +				list_del_init(&sd->s_sibling);
> +			}
>  			spin_unlock(&configfs_dirent_lock);
> +
> +			if (!ret) {
> +				configfs_drop_dentry(sd, dentry->d_parent);
> +				d_delete(dentry);
> +				configfs_put(sd);
> +			}

	This open-code of the VFS munging is ripe to break when the VFS
changes or other configfs changes happen.  The real issue is that you
are reimplementing the core of configfs_unlink().  Note how the core VFS
munging of configfs_rmdir() is separated out so that configfs_mkdir()
can also use it in the failure case?  Do the same with unlink() and it
will read much better ("if (DROPPING) configfs_delete_link()").  Call it
configfs_remove_link() or configfs_delete_link().

Joel

-- 

Life's Little Instruction Book #337

	"Reread your favorite book."

Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker@oracle.com
Phone: (650) 506-8127

  parent reply	other threads:[~2008-06-19 22:16 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-06-17 17:37 [Ocfs2-devel] [BUGFIX][PATCH 0/3] configfs: symlink() fixes Louis Rilling
2008-06-17 17:37 ` Louis Rilling
2008-06-17 17:37 ` [Ocfs2-devel] [BUGFIX][PATCH 1/3] configfs: Fix symlink() to a removing item Louis Rilling
2008-06-17 17:37   ` Louis Rilling
2008-06-17 22:17   ` [Ocfs2-devel] " Joel Becker
2008-06-17 22:17     ` Joel Becker
2008-06-17 17:37 ` [Ocfs2-devel] [BUGFIX][PATCH 2/3] configfs: Rename CONFIGFS_USET_IN_MKDIR to CONFIGFS_USET_ATTACHING Louis Rilling
2008-06-17 17:37   ` Louis Rilling
2008-06-17 17:37 ` [Ocfs2-devel] [BUGFIX][PATCH 3/3] configfs: Fix failing symlink() making rmdir() fail Louis Rilling
2008-06-17 17:37   ` Louis Rilling
2008-06-17 22:15   ` [Ocfs2-devel] " Joel Becker
2008-06-17 22:15     ` Joel Becker
2008-06-18 11:40     ` [Ocfs2-devel] " Louis Rilling
2008-06-18 11:40       ` Louis Rilling
2008-06-18 20:11       ` [Ocfs2-devel] " Joel Becker
2008-06-18 20:11         ` Joel Becker
2008-06-19  9:28         ` [Ocfs2-devel] " Louis Rilling
2008-06-19  9:28           ` Louis Rilling
2008-06-19 22:03           ` [Ocfs2-devel] " Joel Becker
2008-06-19 22:03             ` Joel Becker
2008-06-19 22:16   ` Joel Becker [this message]
2008-06-19 22:16     ` Joel Becker
2008-06-20 12:09     ` [Ocfs2-devel] [PATCH] " Louis Rilling
2008-06-20 12:09       ` Louis Rilling
2008-06-20 22:42       ` [Ocfs2-devel] " Joel Becker
2008-06-20 22:42         ` Joel Becker
2008-06-20 22:44       ` [Ocfs2-devel] " Joel Becker
2008-06-20 22:44         ` Joel Becker
2008-06-23 12:05         ` [Ocfs2-devel] " Louis Rilling
2008-06-23 12:05           ` Louis Rilling

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=20080619221605.GE10888@mail.oracle.com \
    --to=joel.becker@oracle.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=louis.rilling@kerlabs.com \
    --cc=ocfs2-devel@oss.oracle.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 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.