xfs
[Top] [All Lists]

Re: [PATCH v3 4/4] xfs: fold xfs_create_tmpfile() into xfs_create()

To: Brian Foster <bfoster@xxxxxxxxxx>
Subject: Re: [PATCH v3 4/4] xfs: fold xfs_create_tmpfile() into xfs_create()
From: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Tue, 15 Apr 2014 10:51:30 -0700
Cc: xfs@xxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <1397578706-5385-5-git-send-email-bfoster@xxxxxxxxxx>
References: <1397578706-5385-1-git-send-email-bfoster@xxxxxxxxxx> <1397578706-5385-5-git-send-email-bfoster@xxxxxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Tue, Apr 15, 2014 at 12:18:26PM -0400, Brian Foster wrote:
> xfs_create_tmpfile() duplicates most of xfs_create() with the exception
> of a couple minor differences. We use a unique transaction reservation
> and place the allocated inode on the unlinked list rather than create a
> directory entry for it.
> 
> Fold xfs_create_tmpfile() into xfs_create() to reduce some of this
> duplication. The name parameter that represents the name of the
> directory entry to create is now conditional and its existence dictates
> whether a directory entry is created for the inode or not.
> 
> Signed-off-by: Brian Foster <bfoster@xxxxxxxxxx>

Looks correct although I still can't get excited about it..

Reviewed-by: Christoph Hellwig <hch@xxxxxx>

<Prev in Thread] Current Thread [Next in Thread>