Re: [PATCH RESEND 2/2] generic/470: Replace thin volume with blkdiscard -z

["Darrick J. Wong" <djwong@kernel.org>]

On Sun, Oct 23, 2022 at 06:48:13AM +0000, yangx.jy@fujitsu.com wrote:
> generic/470 was original designed to verify mmap(MAP_SYNC) which
> is only vaild to the DAX capable device(e.g. PMEM). Thin volume[1] was
> introduced to fix the inconsistent filesystem issue[2] but it make
> the test become not run because it doesn't support DAX. As Darrick
> mentioned[3], discarding the entire mapped range of scartch device
> can fix the issue as well, so I try to use blkdiscard -z instead.
> 
> [1]: https://git.kernel.org/pub/scm/fs/xfs/xfstests-dev.git/commit/?id=fc5870da485aec0f9196a0f2bed32f73f6b2c664
> [2]: https://lore.kernel.org/fstests/20190227061529.GF16436@dastard/
> [3]: https://lore.kernel.org/linux-xfs/Y1NRNtToQTjs0Dbd@magnolia/T/#me0e77cb0ecd80bf4b5109e4433cb4863ae6e6727
> 
> Signed-off-by: Xiao Yang <yangx.jy@fujitsu.com>
> ---
>  tests/generic/470 | 31 ++++++++++++-------------------
>  1 file changed, 12 insertions(+), 19 deletions(-)
> 
> diff --git a/tests/generic/470 b/tests/generic/470
> index f3407511..d2846a38 100755
> --- a/tests/generic/470
> +++ b/tests/generic/470
> @@ -15,38 +15,29 @@ _cleanup()
>  {
>  	cd /
>  	_log_writes_cleanup
> -	_dmthin_cleanup
>  	rm -f $tmp.*
>  }
>  
>  # Import common functions.
>  . ./common/filter
> -. ./common/dmthin
>  . ./common/dmlogwrites
>  
>  # real QA test starts here
>  _supported_fs generic
> -_require_scratch_nocheck
> +_require_scratch
>  _require_no_logdev
>  _require_log_writes_dax_mountopt "dax"
> -_require_dm_target thin-pool
>  _require_xfs_io_command "mmap" "-S"
>  _require_xfs_io_command "log_writes"
> +_require_command "$BLKDISCARD_PROG" blkdiscard
>  
> -devsize=$((1024*1024*200 / 512))        # 200m phys/virt size
> -csize=$((1024*64 / 512))                # 64k cluster size
> -lowspace=$((1024*1024 / 512))           # 1m low space threshold
> -
> -# Use a thin device to provide deterministic discard behavior. Discards are used
> -# by the log replay tool for fast zeroing to prevent out-of-order replay issues.
> -_dmthin_init $devsize $devsize $csize $lowspace
> +RANGE=$((512 * 1024 * 1024)) # 512 MiB
> +LEN=$((1024 * 1024)) # 1 MiB
>  
> -_log_writes_init $DMTHIN_VOL_DEV
> +_log_writes_init $SCRATCH_DEV $RANGE
>  _log_writes_mkfs >> $seqres.full 2>&1
>  _log_writes_mount -o dax
>  
> -LEN=$((1024 * 1024)) # 1 MiB
> -
>  $XFS_IO_PROG -t -c "truncate $LEN" -c "mmap -S 0 $LEN" -c "mwrite 0 $LEN" \
>  	-c "log_writes -d $LOGWRITES_NAME -m preunmap" \
>  	-f $SCRATCH_MNT/test
> @@ -54,14 +45,16 @@ $XFS_IO_PROG -t -c "truncate $LEN" -c "mmap -S 0 $LEN" -c "mwrite 0 $LEN" \
>  # Unmount the scratch dir and tear down the log writes target
>  _log_writes_unmount
>  _log_writes_remove
> -_dmthin_check_fs
> +_check_scratch_fs
>  
> -# destroy previous filesystem so we can be sure our rebuild works
> -_mkfs_dev $DMTHIN_VOL_DEV >> $seqres.full 2>&1
> +# Discard the mapped range of scratch device and destroy previous filesystem
> +# so we can be sure our rebuild works
> +$BLKDISCARD_PROG -fzl $RANGE $SCRATCH_DEV >> $seqres.full 2>&1

You used -z, which means the comment should say "Forcibly zero the
mapped range of the scratch device..." and note that blkdiscard -z will
fall back to writing buffers of zeroes if the device doesn't support
WRITE SAME with zeroes.

I think the logic in here looks good though.

--D

> +_scratch_mkfs >> $seqres.full 2>&1
>  
>  # check pre-unmap state
> -_log_writes_replay_log preunmap $DMTHIN_VOL_DEV
> -_dmthin_mount
> +_log_writes_replay_log preunmap $SCRATCH_DEV
> +_scratch_mount
>  
>  # We should see $SCRATCH_MNT/test as having 1 MiB in block allocations
>  du -sh $SCRATCH_MNT/test | _filter_scratch | _filter_spaces
> -- 
> 2.34.1