[PATCH] xfstests: don't assume that falloc_punch implies falloc in test 255

[PATCH] xfstests: don't assume that falloc_punch implies falloc in test 255

[Theodore Ts'o]

As of Linux 3.9-rc1, ext4 will support the punch operation on file
systems using indirect blocks, but it can not support the fallocate
operation (since there is no way to mark a block as uninitialized
using indirect block scheme).  This caused test 255 to fail, since it
only used _require_xfS_io_falloc_punch assuming that all file systems
which supported punch can also support fallocate.  Fix this.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
---
 255 | 1 +
 1 file changed, 1 insertion(+)

diff --git a/255 b/255
index 0083963..ae1d8e0 100755
--- a/255
+++ b/255
@@ -48,6 +48,7 @@ _supported_fs generic
 _supported_os Linux
 
 _require_xfs_io_falloc_punch
+_require_xfs_io_falloc
 _require_xfs_io_fiemap
 
 testfile=$TEST_DIR/255.$$
-- 
1.7.12.rc0.22.gcdd159b

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

Re: [PATCH] xfstests: don't assume that falloc_punch implies falloc in test 255

[Rich Johnston]

On 03/05/2013 11:59 AM, Theodore Ts'o wrote:
> As of Linux 3.9-rc1, ext4 will support the punch operation on file
> systems using indirect blocks, but it can not support the fallocate
> operation (since there is no way to mark a block as uninitialized
> using indirect block scheme).  This caused test 255 to fail, since it
> only used _require_xfS_io_falloc_punch assuming that all file systems
> which supported punch can also support fallocate.  Fix this.
>
> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
> ---
>   255 | 1 +
>   1 file changed, 1 insertion(+)
>
> diff --git a/255 b/255
> index 0083963..ae1d8e0 100755
> --- a/255
> +++ b/255
> @@ -48,6 +48,7 @@ _supported_fs generic
>   _supported_os Linux
>
>   _require_xfs_io_falloc_punch
> +_require_xfs_io_falloc
#rcj looks reasonable to me to add this requirement
>   _require_xfs_io_fiemap
>
>   testfile=$TEST_DIR/255.$$
>

Reviewed-by: Rich Johnston <rjohnston@sgi.com>

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

Re: [PATCH] xfstests: don't assume that falloc_punch implies falloc in test 255

[Rich Johnston]

This patch has been committed.

Thanks
--Rich

commit 864688d368d6781c3f6d60bc55b5e3591953e462
Author: Theodore Ts'o <tytso@mit.edu>
Date:   Tue Mar 5 17:59:42 2013 +0000

     xfstests: don't assume that falloc_punch implies falloc in test 255

     As of Linux 3.9-rc1, ext4 will support the punch operation on file
     systems using indirect blocks, but it can not support the fallocate
     operation (since there is no way to mark a block as uninitialized
     using indirect block scheme).  This caused test 255 to fail, since it
     only used _require_xfS_io_falloc_punch assuming that all file systems
     which supported punch can also support fallocate.  Fix this.

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

Re: [PATCH] xfstests: don't assume that falloc_punch implies falloc in test 255

[Eric Sandeen]

On 3/5/13 11:59 AM, Theodore Ts'o wrote:
> As of Linux 3.9-rc1, ext4 will support the punch operation on file
> systems using indirect blocks, but it can not support the fallocate
> operation (since there is no way to mark a block as uninitialized
> using indirect block scheme).  This caused test 255 to fail, since it
> only used _require_xfS_io_falloc_punch assuming that all file systems
> which supported punch can also support fallocate.  Fix this.

Seems fine to avoid the incorrect failure, so as far as that goes:

Reviewed-by: Eric Sandeen <sandeen@redhat.com>

But we probably can & should still test punch in this situation,
so we need a new test to exercise that I guess.

-Eric

> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
> ---
>  255 | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/255 b/255
> index 0083963..ae1d8e0 100755
> --- a/255
> +++ b/255
> @@ -48,6 +48,7 @@ _supported_fs generic
>  _supported_os Linux
>  
>  _require_xfs_io_falloc_punch
> +_require_xfs_io_falloc
>  _require_xfs_io_fiemap
>  
>  testfile=$TEST_DIR/255.$$
> 

Re: [PATCH] xfstests: don't assume that falloc_punch implies falloc in test 255

[Zheng Liu]

On Wed, Mar 06, 2013 at 10:10:09AM -0600, Eric Sandeen wrote:
> On 3/5/13 11:59 AM, Theodore Ts'o wrote:
> > As of Linux 3.9-rc1, ext4 will support the punch operation on file
> > systems using indirect blocks, but it can not support the fallocate
> > operation (since there is no way to mark a block as uninitialized
> > using indirect block scheme).  This caused test 255 to fail, since it
> > only used _require_xfS_io_falloc_punch assuming that all file systems
> > which supported punch can also support fallocate.  Fix this.
> 
> Seems fine to avoid the incorrect failure, so as far as that goes:
> 
> Reviewed-by: Eric Sandeen <sandeen@redhat.com>
> 
> But we probably can & should still test punch in this situation,
> so we need a new test to exercise that I guess.

Hi Eric,

I have sent a patch set to add a test case for punching hole.  You can
find it in this link [1].  Sorry I don't finish the second version
according to Mark's comment.

1. http://www.spinics.net/lists/xfs/msg16234.html

Regards,
                                                - Zheng

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