Re: [PATCH 5/5] xfs: assert that we hold the ilock for extent map access

From: Dave Chinner <david@fromorbit.com>
To: Christoph Hellwig <hch@infradead.org>
Cc: xfs@oss.sgi.com
Subject: Re: [PATCH 5/5] xfs: assert that we hold the ilock for extent map access
Date: Fri, 6 Dec 2013 08:10:47 +1100	[thread overview]
Message-ID: <20131205211047.GG29897@dastard> (raw)
In-Reply-To: <20131205155951.874279041@bombadil.infradead.org>

On Thu, Dec 05, 2013 at 07:58:35AM -0800, Christoph Hellwig wrote:
> Make sure that xfs_bmapi_read has the ilock held in some way, and that
> xfs_bmapi_write, xfs_bmapi_delay and xfs_iread_extents are called with
> the ilock held exclusively.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> 
> Index: xfs/fs/xfs/xfs_bmap.c
> ===================================================================
> --- xfs.orig/fs/xfs/xfs_bmap.c	2013-11-29 14:25:12.172459195 +0100
> +++ xfs/fs/xfs/xfs_bmap.c	2013-12-05 10:03:28.243801633 +0100
> @@ -3997,6 +3997,7 @@ xfs_bmapi_read(
>  	ASSERT(*nmap >= 1);
>  	ASSERT(!(flags & ~(XFS_BMAPI_ATTRFORK|XFS_BMAPI_ENTIRE|
>  			   XFS_BMAPI_IGSTATE)));
> +	ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED|XFS_ILOCK_EXCL));
>  
>  	if (unlikely(XFS_TEST_ERROR(
>  	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
> @@ -4191,6 +4192,7 @@ xfs_bmapi_delay(
>  	ASSERT(*nmap >= 1);
>  	ASSERT(*nmap <= XFS_BMAP_MAX_NMAP);
>  	ASSERT(!(flags & ~XFS_BMAPI_ENTIRE));
> +	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
>  
>  	if (unlikely(XFS_TEST_ERROR(
>  	    (XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) != XFS_DINODE_FMT_EXTENTS &&
> @@ -4484,6 +4486,7 @@ xfs_bmapi_write(
>  	ASSERT(tp != NULL);
>  	ASSERT(len > 0);
>  	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL);
> +	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
>  
>  	if (unlikely(XFS_TEST_ERROR(
>  	    (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
> Index: xfs/fs/xfs/xfs_inode_fork.c
> ===================================================================
> --- xfs.orig/fs/xfs/xfs_inode_fork.c	2013-12-05 09:57:24.347809100 +0100
> +++ xfs/fs/xfs/xfs_inode_fork.c	2013-12-05 09:59:04.767807039 +0100
> @@ -431,6 +431,8 @@ xfs_iread_extents(
>  	xfs_ifork_t	*ifp;
>  	xfs_extnum_t	nextents;
>  
> +	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
> +
>  	if (unlikely(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE)) {
>  		XFS_ERROR_REPORT("xfs_iread_extents", XFS_ERRLEVEL_LOW,
>  				 ip->i_mount);

Looks good, but can we add an assert to xfs_bunmapi() at the same
time just to cover all the public bmapi interfaces with locking
requirements?

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs