[PATCH] xfs/604: Make test as _notrun for higher blocksizes filesystem

[PATCH] xfs/604: Make test as _notrun for higher blocksizes filesystem

[Ritesh Harjani (IBM)]

If we have filesystem with blocksize = 64k, then the falloc value will
be huge which makes fallocate fail hence causing the test to fail.
Instead make the testcase "_notrun" if the fallocate itself fails.

Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
---
 tests/xfs/604 | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/tests/xfs/604 b/tests/xfs/604
index bb6db797..40596a28 100755
--- a/tests/xfs/604
+++ b/tests/xfs/604
@@ -35,7 +35,8 @@ allocbt_node_maxrecs=$(((dbsize - alloc_block_len) / 12))
 # Create a big file with a size such that the punches below create the exact
 # free extents we want.
 num_holes=$((allocbt_leaf_maxrecs * allocbt_node_maxrecs - 1))
-$XFS_IO_PROG -c "falloc 0 $((9 * dbsize + num_holes * dbsize * 2))" -f "$SCRATCH_MNT/big"
+$XFS_IO_PROG -c "falloc 0 $((9 * dbsize + num_holes * dbsize * 2))" -f "$SCRATCH_MNT/big" ||
+				_notrun "Not enough space on device for bs=$dbsize"
 
 # Fill in any small free extents in AG 0. After this, there should be only one,
 # large free extent.
-- 
2.43.0

Re: [PATCH] xfs/604: Make test as _notrun for higher blocksizes filesystem

[Darrick J. Wong]

On Fri, Jan 19, 2024 at 02:57:45PM +0530, Ritesh Harjani (IBM) wrote:
> If we have filesystem with blocksize = 64k, then the falloc value will
> be huge which makes fallocate fail hence causing the test to fail.
> Instead make the testcase "_notrun" if the fallocate itself fails.

How much space is it asking for?

--D

> Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
> ---
>  tests/xfs/604 | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/tests/xfs/604 b/tests/xfs/604
> index bb6db797..40596a28 100755
> --- a/tests/xfs/604
> +++ b/tests/xfs/604
> @@ -35,7 +35,8 @@ allocbt_node_maxrecs=$(((dbsize - alloc_block_len) / 12))
>  # Create a big file with a size such that the punches below create the exact
>  # free extents we want.
>  num_holes=$((allocbt_leaf_maxrecs * allocbt_node_maxrecs - 1))
> -$XFS_IO_PROG -c "falloc 0 $((9 * dbsize + num_holes * dbsize * 2))" -f "$SCRATCH_MNT/big"
> +$XFS_IO_PROG -c "falloc 0 $((9 * dbsize + num_holes * dbsize * 2))" -f "$SCRATCH_MNT/big" ||
> +				_notrun "Not enough space on device for bs=$dbsize"
>  
>  # Fill in any small free extents in AG 0. After this, there should be only one,
>  # large free extent.
> -- 
> 2.43.0
> 
> 

Re: [PATCH] xfs/604: Make test as _notrun for higher blocksizes filesystem

[Ritesh Harjani]

"Darrick J. Wong" <djwong@kernel.org> writes:

> On Fri, Jan 19, 2024 at 02:57:45PM +0530, Ritesh Harjani (IBM) wrote:
>> If we have filesystem with blocksize = 64k, then the falloc value will
>> be huge which makes fallocate fail hence causing the test to fail.
>> Instead make the testcase "_notrun" if the fallocate itself fails.
>
> How much space is it asking for?

~5451.33 GB for 64k bs v/s ~1.29GB for 4k.

Let me modify the patch to also print the requested falloc size in GB.
That will be much better.

-ritesh