All of lore.kernel.org
 help / color / mirror / Atom feed
From: Eryu Guan <guaneryu@gmail.com>
To: Lukas Czerner <lczerner@redhat.com>
Cc: fstests@vger.kernel.org
Subject: Re: [PATCH] fstests: add support for fuse.ext4
Date: Mon, 22 Oct 2018 00:40:47 +0800	[thread overview]
Message-ID: <20181021164047.GA3876@desktop> (raw)
In-Reply-To: <20181011134153.722-1-lczerner@redhat.com>

On Thu, Oct 11, 2018 at 03:41:53PM +0200, Lukas Czerner wrote:
> Add support for ext4 file system in userspace. The new file system type
> (FSTYP) is fuse.ext4 and it allows the device to be either a regular
> block device, or a regular file.
> 
> Signed-off-by: Lukas Czerner <lczerner@redhat.com>
> ---
>  common/config |  6 +++++
>  common/rc     | 65 ++++++++++++++++++++++++++++++++++++++++++++-------
>  2 files changed, 62 insertions(+), 9 deletions(-)
> 
> diff --git a/common/config b/common/config
> index 1ba8d96c..37b66d31 100644
> --- a/common/config
> +++ b/common/config
> @@ -232,6 +232,7 @@ case "$HOSTOS" in
>  	export E2FSCK_PROG=$(type -P e2fsck)
>  	export TUNE2FS_PROG=$(type -P tune2fs)
>  	export FSCK_OVERLAY_PROG=$(type -P fsck.overlay)
> +	export FUSE2FS=$(type -P fuse2fs)
>          ;;
>  esac
>  
> @@ -470,6 +471,11 @@ _check_device()
>  			_fatal "common/config: $name ($dev) is not a character device"
>  		fi
>  		;;
> +	fuse.ext4)
> +		if [ ! -b "$dev" ] && [ ! -f "$dev" ]; then
> +			__fatal "common/config: $name ($dev) is not a block device nor a regular file"

Typo here, should be "_fatal".

> +		fi
> +		;;
>  	*)
>  		_fatal "common/config: $name ($dev) is not a block device or a network filesystem"
>  	esac
> diff --git a/common/rc b/common/rc
> index 90dc3002..2f5a3463 100644
> --- a/common/rc
> +++ b/common/rc
> @@ -130,6 +130,10 @@ case "$FSTYP" in
>  
>  	 . ./common/btrfs
>  	 ;;
> +    fuse.ext4)
> +	 [ "$FUSE2FS" = "" ] && _fatal "fuse2fs not found"
> +	 [ "$MKFS_EXT4_PROG" = "" ] && _fatal "mkfs.ext4 not found"
> +	 ;;
>      ext4)
>  	 [ "$MKFS_EXT4_PROG" = "" ] && _fatal "mkfs.ext4 not found"
>  	 ;;
> @@ -170,6 +174,22 @@ _mount()
>      $MOUNT_PROG `_mount_ops_filter $*`
>  }
>  
> +_mount_fstyp()
> +{
> +	ret=0
> +	case "$FSTYP" in
> +		fuse.ext4)
> +			$FUSE2FS `_mount_ops_filter $*` 2>&1> /dev/null

Hmm, fuse2fs expects the first argument to be the device/image file, if
we pass mount options first, /proc/mounts shows weird "device", e.g.

mount -o ro /dev/sda5 /mnt/test, then /proc/mounts shows

-o on /mnt/test type fuse.ext4 (rw,nosuid,nodev,relatime,user_id=0,group_id=0)

I think it might be better to update _mount() instead of introducing a
new _mount_fstyp() function, so that we keep the _mount() interface
unchanged and don't need to update all callers of _mount() (either in
common/rc or in tests).

> +			ret=$?
> +			;;
> +		*)
> +			_mount -t $FSTYP $*
> +			ret=$?
> +			;;
> +	esac
> +	return $ret
> +}
> +
>  # Call _mount to do mount operation but also save mountpoint to
>  # MOUNTED_POINT_STACK. Note that the mount point must be the last parameter
>  _get_mount()
> @@ -305,7 +325,7 @@ _supports_filetype()
>  	xfs)
>  		$XFS_INFO_PROG $dir | grep -q "ftype=1"
>  		;;
> -	ext2|ext3|ext4)
> +	ext2|ext3|ext4|fuse.ext4)
>  		local dev=`$DF_PROG $dir | tail -1 | $AWK_PROG '{print $1}'`
>  		tune2fs -l $dev | grep -q filetype
>  		;;
> @@ -328,7 +348,7 @@ _try_scratch_mount()
>  		_overlay_scratch_mount $*
>  		return $?
>  	fi
> -	_mount -t $FSTYP `_scratch_mount_options $*`
> +	_mount_fstyp `_scratch_mount_options $*`
>  }
>  
>  # mount scratch device with given options and _fail if mount fails
> @@ -400,7 +420,7 @@ _test_mount()
>          return $?
>      fi
>      _test_options mount
> -    _mount -t $FSTYP $TEST_OPTIONS $TEST_FS_MOUNT_OPTS $SELINUX_MOUNT_OPTIONS $* $TEST_DEV $TEST_DIR
> +    _mount_fstyp $TEST_OPTIONS $TEST_FS_MOUNT_OPTS $SELINUX_MOUNT_OPTIONS $* $TEST_DEV $TEST_DIR
>  }
>  
>  _test_unmount()
> @@ -610,6 +630,9 @@ _test_mkfs()
>      btrfs)
>          $MKFS_BTRFS_PROG $MKFS_OPTIONS $* $TEST_DEV > /dev/null
>  	;;
> +    fuse.ext4)
> +	$MKFS_PROG -t ext4 -- -F $MKFS_OPTIONS $* $TEST_DEV
> +	;;
>      ext2|ext3|ext4)
>  	$MKFS_PROG -t $FSTYP -- -F $MKFS_OPTIONS $* $TEST_DEV
>  	;;
> @@ -641,7 +664,7 @@ _mkfs_dev()
>      btrfs)
>          $MKFS_BTRFS_PROG $MKFS_OPTIONS $* 2>$tmp.mkfserr 1>$tmp.mkfsstd
>  	;;
> -    ext2|ext3|ext4)
> +    ext2|ext3|ext4|fuse.ext4)
>  	$MKFS_PROG -t $FSTYP -- -F $MKFS_OPTIONS $* \

Introduce a new "fuse.ext4" case as in _test_mkfs(), otherwise we'll do
"mkfs -t fuse.ext4 ..." here.

>  		2>$tmp.mkfserr 1>$tmp.mkfsstd
>  	;;
> @@ -709,7 +732,7 @@ _scratch_mkfs()
>  		$UBIUPDATEVOL_PROG ${SCRATCH_DEV} -t
>  		return 0
>  		;;
> -	ext4)
> +	ext4|fuse.ext4)
>  		_scratch_mkfs_ext4 $*
>  		return $?
>  		;;
> @@ -906,7 +929,7 @@ _scratch_mkfs_sized()
>      xfs)
>  	def_blksz=`echo $MKFS_OPTIONS|sed -rn 's/.*-b ?size= ?+([0-9]+).*/\1/p'`
>  	;;
> -    ext2|ext3|ext4|ext4dev|udf|btrfs|reiser4|ocfs2|reiserfs)
> +    ext2|ext3|ext4|ext4dev|udf|btrfs|reiser4|ocfs2|reiserfs|fuse.ext4)
>  	def_blksz=`echo $MKFS_OPTIONS| sed -rn 's/.*-b ?+([0-9]+).*/\1/p'`
>  	;;
>      jfs)
> @@ -943,6 +966,9 @@ _scratch_mkfs_sized()
>  		_scratch_mkfs_xfs -d size=$fssize -b size=$blocksize
>  	fi
>  	;;
> +    fuse.ext4)
> +	${MKFS_PROG}.ext4 -F $MKFS_OPTIONS -b $blocksize $SCRATCH_DEV $blocks
> +	;;
>      ext2|ext3|ext4|ext4dev)
>  	${MKFS_PROG}.$FSTYP -F $MKFS_OPTIONS -b $blocksize $SCRATCH_DEV $blocks
>  	;;
> @@ -1020,7 +1046,7 @@ _scratch_mkfs_geom()
>      xfs)
>  	MKFS_OPTIONS+=" -b size=$blocksize, -d su=$sunit_bytes,sw=$swidth_mult"
>  	;;
> -    ext4|ext4dev)
> +    ext4|ext4dev|fuse.ext4)
>  	MKFS_OPTIONS+=" -b $blocksize -E stride=$sunit_blocks,stripe_width=$swidth_blocks"
>  	;;
>      *)
> @@ -1045,6 +1071,9 @@ _scratch_mkfs_blocksized()
>      xfs)
>  	_scratch_mkfs_xfs $MKFS_OPTIONS -b size=$blocksize
>  	;;
> +    fuse.ext4)
> +	${MKFS_PROG}.ext4 -F $MKFS_OPTIONS -b $blocksize $SCRATCH_DEV
> +	;;
>      ext2|ext3|ext4)
>  	${MKFS_PROG}.$FSTYP -F $MKFS_OPTIONS -b $blocksize $SCRATCH_DEV
>  	;;
> @@ -1514,6 +1543,15 @@ _require_scratch_nocheck()
>  			_notrun "this test requires a valid \$SCRATCH_MNT"
>  		fi
>  		;;
> +	fuse.ext4)
> +		if [ ! -b "$SCRATCH_DEV" ] && [ ! -f "$SCRATCH_DEV" ]; then
> +		     _notrun "this test requires a valid \$SCRATCH_DEV"
> +		fi
> +		if [ ! -d "$SCRATCH_MNT" ]
> +		then

Make "if" and "then" in the same line as in above if check.

> +		     _notrun "this test requires a valid \$SCRATCH_MNT"
> +		fi
> +		 ;;
>  	*)
>  		 if [ -z "$SCRATCH_DEV" -o "`_is_block_dev "$SCRATCH_DEV"`" = "" ]
>  		 then
> @@ -1635,6 +1673,15 @@ _require_test()
>  			_notrun "this test requires a valid \$TEST_DIR"
>  		fi
>  		;;
> +	fuse.ext4)
> +		if [ ! -b "$TEST_DEV" ] && [ ! -f "$TEST_DEV" ]; then
> +		     _notrun "this test requires a valid \$TEST_DEV"
> +		fi
> +		if [ ! -d "$TEST_DIR" ]
> +		then

Same here.

> +		     _notrun "this test requires a valid \$TEST_DIR"
> +		fi
> +		 ;;
>  	*)
>  		 if [ -z "$TEST_DEV" ] || [ "`_is_block_dev "$TEST_DEV"`" = "" ]
>  		 then
> @@ -1961,7 +2008,7 @@ _filesystem_timestamp_range()
>  {
>  	local device=${1:-$TEST_DEV}
>  	case $FSTYP in
> -	ext4)
> +	ext4|fuse.ext4)
>  		if [ $(dumpe2fs -h $device 2>/dev/null | grep "Inode size:" | cut -d: -f2) -gt 128 ]; then
>  			echo "-2147483648 15032385535"
>  		else
> @@ -2458,7 +2505,7 @@ _mount_or_remount_rw()
>  
>  	if [ $USE_REMOUNT -eq 0 ]; then
>  		if [ "$FSTYP" != "overlay" ]; then
> -			_mount -t $FSTYP $mount_opts $device $mountpoint
> +			_mount_fstyp $mount_opts $device $mountpoint

This (fuse2fs) won't remount the filesystem, mount command is still
needed here.

Thanks,
Eryu

>  		else
>  			_overlay_mount $device $mountpoint
>  		fi
> -- 
> 2.17.1
> 

  reply	other threads:[~2018-10-22  0:55 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-10-11 13:41 [PATCH] fstests: add support for fuse.ext4 Lukas Czerner
2018-10-21 16:40 ` Eryu Guan [this message]
2018-10-24 11:12   ` Lukas Czerner

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20181021164047.GA3876@desktop \
    --to=guaneryu@gmail.com \
    --cc=fstests@vger.kernel.org \
    --cc=lczerner@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.