On Sat, Sep 5, 2020 at 12:25 AM Anand Jain <anand.jain@xxxxxxxxxx> wrote:
>
> Make this test case inline with the kernel patch [1] changes
> [1] btrfs: fix replace of seed device
Same comment as in the previous patch. Now that this is in Linus'
tree, it would be good to mention the commit id too.
>
> So use the sprout device as the replace target instead of the seed device.
> This change is compatible with the older kernels.
>
> While at this, this patch also fixes a typo fix as well.
>
> Signed-off-by: Anand Jain <anand.jain@xxxxxxxxxx>
> ---
> tests/btrfs/163 | 21 ++++++++++++++++-----
> tests/btrfs/163.out | 5 ++++-
> 2 files changed, 20 insertions(+), 6 deletions(-)
>
> diff --git a/tests/btrfs/163 b/tests/btrfs/163
> index 24c725afb6b9..354d88502d47 100755
> --- a/tests/btrfs/163
> +++ b/tests/btrfs/163
> @@ -4,11 +4,15 @@
> #
> # FS QA Test 163
> #
> -# Test case to verify that a seed device can be replaced
> +# Test case to verify that a sprouted device can be replaced
> # Create a seed device
> # Create a sprout device
> # Remount RW
> -# Run device replace on the seed device
> +# Run device replace on the sprout device
> +#
> +# Depends on the kernel patch
> +# btrfs: fail replace of seed device
> +
> seq=`basename $0`
> seqres=$RESULT_DIR/$seq
> echo "QA output created by $seq"
> @@ -39,6 +43,7 @@ _supported_fs btrfs
> _supported_os Linux
> _require_command "$BTRFS_TUNE_PROG" btrfstune
> _require_scratch_dev_pool 3
> +_require_btrfs_forget_or_module_loadable
>
> _scratch_dev_pool_get 3
>
> @@ -52,7 +57,7 @@ create_seed()
> run_check _mount $dev_seed $SCRATCH_MNT
> $XFS_IO_PROG -f -d -c "pwrite -S 0xab 0 4M" $SCRATCH_MNT/foobar >\
> /dev/null
> - echo -- gloden --
> + echo -- golden --
> od -x $SCRATCH_MNT/foobar
> _run_btrfs_util_prog filesystem show -m $SCRATCH_MNT
> _scratch_unmount
> @@ -64,22 +69,28 @@ add_sprout()
> {
> _run_btrfs_util_prog device add -f $dev_sprout $SCRATCH_MNT
> _run_btrfs_util_prog filesystem show -m $SCRATCH_MNT
> + _mount -o remount,rw $dev_sprout $SCRATCH_MNT
> + $XFS_IO_PROG -f -d -c "pwrite -S 0xcd 0 4M" $SCRATCH_MNT/foobar2 >\
> + /dev/null
Same comment as for the other patch.
Why the direct IO? The test passes with buffered IO as well.
If there's a reason for direct IO, it should be mentioned in a comment
and _require_odirect added above.
> }
>
> replace_seed()
> {
> - _run_btrfs_util_prog replace start -fB $dev_seed $dev_replace_tgt $SCRATCH_MNT
> + _run_btrfs_util_prog replace start -fB $dev_sprout $dev_replace_tgt $SCRATCH_MNT
So now the function should be renamed from replace_seed() to
replace_sprout() as well. Shouldn't it?
Other than that, it looks good and it works as expected.
Thanks.
> _run_btrfs_util_prog filesystem show -m $SCRATCH_MNT
> _scratch_unmount
> - run_check _mount $dev_replace_tgt $SCRATCH_MNT
> + _btrfs_forget_or_module_reload
> + run_check _mount -o device=$dev_seed $dev_replace_tgt $SCRATCH_MNT
> echo -- sprout --
> od -x $SCRATCH_MNT/foobar
> + od -x $SCRATCH_MNT/foobar2
> _scratch_unmount
>
> }
>
> seed_is_mountable()
> {
> + _btrfs_forget_or_module_reload
> run_check _mount $dev_seed $SCRATCH_MNT
> _run_btrfs_util_prog filesystem show -m $SCRATCH_MNT
> _scratch_unmount
> diff --git a/tests/btrfs/163.out b/tests/btrfs/163.out
> index 91f6f5b6f48a..351ef7b040b2 100644
> --- a/tests/btrfs/163.out
> +++ b/tests/btrfs/163.out
> @@ -1,5 +1,5 @@
> QA output created by 163
> --- gloden --
> +-- golden --
> 0000000 abab abab abab abab abab abab abab abab
> *
> 20000000
> @@ -7,3 +7,6 @@ QA output created by 163
> 0000000 abab abab abab abab abab abab abab abab
> *
> 20000000
> +0000000 cdcd cdcd cdcd cdcd cdcd cdcd cdcd cdcd
> +*
> +20000000
> --
> 2.25.1
>
--
Filipe David Manana,
“Whether you think you can, or you think you can't — you're right.”
