[PATCH] xfstests: enable test 032 (mkfs overwrite) for btrfs

[PATCH] xfstests: enable test 032 (mkfs overwrite) for btrfs

[Eric Sandeen]

Now that btrfs has an "-f" arg, we can test that it doesn't
improperly overwrite other filesystems in 032 like we do 
for xfs.

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

diff --git a/032 b/032
index bf88492..89fb225 100755
--- a/032
+++ b/032
@@ -38,12 +38,18 @@ rm -f $seq.full
 . ./common.filter
 
 # real QA test starts here
-_supported_fs xfs
+_supported_fs xfs btrfs
 _supported_os Linux
 
 _require_nobigloopfs
 _require_scratch
 
+# mkfs.btrfs did not have overwrite detection at first
+if [ "$FSTYP" == "btrfs" ]; then
+	grep -q 'force overwrite' `echo $MKFS_BTRFS_PROG | awk '{print $1}'` || \
+		_notrun "Installed mkfs.btrfs does not support -f option"
+fi
+
 echo "Silence is golden."
 for fs in `echo ${MKFS_PROG}.* | sed -e 's/.sbin.mkfs.//g'`
 do
@@ -61,7 +67,7 @@ do
 	[ $fs = gfs2 ] && preop="echo y |" && preargs="-p lock_nolock -j 1"
 	[ $fs = reiserfs ] && preop="echo y |" && preargs="-f"
 	# cramfs mkfs requires a directory argument
-	[ $fs = cramfs ] && preargs=/proc/fs/xfs
+	[ $fs = cramfs ] && preargs=/proc/fs
 	[ $fs = ext2 ] && preargs="-F"
 	[ $fs = ext3 ] && preargs="-F"
 	[ $fs = ext4 ] && preargs="-F"
@@ -77,8 +83,8 @@ do
 
 	if [ $? -eq 0 ] ; then
 		# next, ensure we don't overwrite it
-		echo "=== Attempting XFS overwrite of $fs..." >>$seq.full
-		${MKFS_PROG}.xfs $SCRATCH_DEV >>$seq.full 2>&1
+		echo "=== Attempting $FSTYP overwrite of $fs..." >>$seq.full
+		${MKFS_PROG}.$FSTYP $SCRATCH_DEV >>$seq.full 2>&1
 
 		[ $? -eq 0 ] && echo "Failed - overwrote fs type ${fs}!"
 	else

Re: [PATCH] xfstests: enable test 032 (mkfs overwrite) for btrfs

[Rich Johnston]

Eric,

Thanks once more for the patch.  Once I finally got the correct version 
of mkfs.btrfs this looks good.

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

Interesting that the default make install of btrfs-progs did not install 
over my btrfsprogs rpm.

Was running SLES11 SP2 with btrfsprogs-0.19-5.31.1.x86_64.

It did not put mkfs.btrfs in /sbin.  This did however, validate all the 
error paths ;).

--Rich




On 03/13/2013 11:01 AM, Eric Sandeen wrote:
> Now that btrfs has an "-f" arg, we can test that it doesn't
> improperly overwrite other filesystems in 032 like we do
> for xfs.
>
> Signed-off-by: Eric Sandeen <sandeen@redhat.com>
> ---
>
> diff --git a/032 b/032
> index bf88492..89fb225 100755
> --- a/032
> +++ b/032
> @@ -38,12 +38,18 @@ rm -f $seq.full
>   . ./common.filter
>
>   # real QA test starts here
> -_supported_fs xfs
> +_supported_fs xfs btrfs
>   _supported_os Linux
>
>   _require_nobigloopfs
>   _require_scratch
>
> +# mkfs.btrfs did not have overwrite detection at first
> +if [ "$FSTYP" == "btrfs" ]; then
> +	grep -q 'force overwrite' `echo $MKFS_BTRFS_PROG | awk '{print $1}'` || \
> +		_notrun "Installed mkfs.btrfs does not support -f option"
> +fi
> +
>   echo "Silence is golden."
>   for fs in `echo ${MKFS_PROG}.* | sed -e 's/.sbin.mkfs.//g'`
>   do
> @@ -61,7 +67,7 @@ do
>   	[ $fs = gfs2 ] && preop="echo y |" && preargs="-p lock_nolock -j 1"
>   	[ $fs = reiserfs ] && preop="echo y |" && preargs="-f"
>   	# cramfs mkfs requires a directory argument
> -	[ $fs = cramfs ] && preargs=/proc/fs/xfs
> +	[ $fs = cramfs ] && preargs=/proc/fs
>   	[ $fs = ext2 ] && preargs="-F"
>   	[ $fs = ext3 ] && preargs="-F"
>   	[ $fs = ext4 ] && preargs="-F"
> @@ -77,8 +83,8 @@ do
>
>   	if [ $? -eq 0 ] ; then
>   		# next, ensure we don't overwrite it
> -		echo "=== Attempting XFS overwrite of $fs..." >>$seq.full
> -		${MKFS_PROG}.xfs $SCRATCH_DEV >>$seq.full 2>&1
> +		echo "=== Attempting $FSTYP overwrite of $fs..." >>$seq.full
> +		${MKFS_PROG}.$FSTYP $SCRATCH_DEV >>$seq.full 2>&1
>
>   		[ $? -eq 0 ] && echo "Failed - overwrote fs type ${fs}!"
>   	else
>
>
> _______________________________________________
> xfs mailing list
> xfs@oss.sgi.com
> http://oss.sgi.com/mailman/listinfo/xfs
>

Re: [PATCH] xfstests: enable test 032 (mkfs overwrite) for btrfs

[Rich Johnston]

This patch has been commited.

Thanks
--Rich

commit 99eb53da9c08148132cdf3cd401b248e6a0d96c5
Author: Eric Sandeen <sandeen@redhat.com>
Date:   Wed Mar 13 16:01:00 2013 +0000

     xfstests: enable test 032 (mkfs overwrite) for btrfs

     Now that btrfs has an "-f" arg, we can test that it doesn't
     improperly overwrite other filesystems in 032 like we do
     for xfs.

     Signed-off-by: Eric Sandeen <sandeen@redhat.com>
     Reviewed-by: Rich Johnston <rjohnston@sgi.com>
     Signed-off-by: Rich Johnston <rjohnston@sgi.com>

Re: [PATCH] xfstests: enable test 032 (mkfs overwrite) for btrfs

[Eric Sandeen]

On 3/15/13 12:07 PM, Rich Johnston wrote:
> Eric,
> 
> Thanks once more for the patch.  Once I finally got the correct version of mkfs.btrfs this looks good.
> 
> Reviewed-by: Rich Johnston <rjohnston@sgi.com>
> 
> Interesting that the default make install of btrfs-progs did not install over my btrfsprogs rpm.
> 
> Was running SLES11 SP2 with btrfsprogs-0.19-5.31.1.x86_64.
> 
> It did not put mkfs.btrfs in /sbin.  This did however, validate all the error paths ;).

Heh ;)

Well, mixing & matching upstream installs from source w/ rpm-packaged binaries is usually asking for trouble.

I don't think it's a bug, just bad administrative practice ;)

-Eric

> --Rich
> 
> 
> 
> 
> On 03/13/2013 11:01 AM, Eric Sandeen wrote:
>> Now that btrfs has an "-f" arg, we can test that it doesn't
>> improperly overwrite other filesystems in 032 like we do
>> for xfs.
>>
>> Signed-off-by: Eric Sandeen <sandeen@redhat.com>
>> ---
>>
>> diff --git a/032 b/032
>> index bf88492..89fb225 100755
>> --- a/032
>> +++ b/032
>> @@ -38,12 +38,18 @@ rm -f $seq.full
>>   . ./common.filter
>>
>>   # real QA test starts here
>> -_supported_fs xfs
>> +_supported_fs xfs btrfs
>>   _supported_os Linux
>>
>>   _require_nobigloopfs
>>   _require_scratch
>>
>> +# mkfs.btrfs did not have overwrite detection at first
>> +if [ "$FSTYP" == "btrfs" ]; then
>> +    grep -q 'force overwrite' `echo $MKFS_BTRFS_PROG | awk '{print $1}'` || \
>> +        _notrun "Installed mkfs.btrfs does not support -f option"
>> +fi
>> +
>>   echo "Silence is golden."
>>   for fs in `echo ${MKFS_PROG}.* | sed -e 's/.sbin.mkfs.//g'`
>>   do
>> @@ -61,7 +67,7 @@ do
>>       [ $fs = gfs2 ] && preop="echo y |" && preargs="-p lock_nolock -j 1"
>>       [ $fs = reiserfs ] && preop="echo y |" && preargs="-f"
>>       # cramfs mkfs requires a directory argument
>> -    [ $fs = cramfs ] && preargs=/proc/fs/xfs
>> +    [ $fs = cramfs ] && preargs=/proc/fs
>>       [ $fs = ext2 ] && preargs="-F"
>>       [ $fs = ext3 ] && preargs="-F"
>>       [ $fs = ext4 ] && preargs="-F"
>> @@ -77,8 +83,8 @@ do
>>
>>       if [ $? -eq 0 ] ; then
>>           # next, ensure we don't overwrite it
>> -        echo "=== Attempting XFS overwrite of $fs..." >>$seq.full
>> -        ${MKFS_PROG}.xfs $SCRATCH_DEV >>$seq.full 2>&1
>> +        echo "=== Attempting $FSTYP overwrite of $fs..." >>$seq.full
>> +        ${MKFS_PROG}.$FSTYP $SCRATCH_DEV >>$seq.full 2>&1
>>
>>           [ $? -eq 0 ] && echo "Failed - overwrote fs type ${fs}!"
>>       else
>>
>>
>> _______________________________________________
>> xfs mailing list
>> xfs@oss.sgi.com
>> http://oss.sgi.com/mailman/listinfo/xfs
>>
> 
> 

Re: [PATCH] xfstests: enable test 032 (mkfs overwrite) for btrfs

[Linda Walsh]

Eric Sandeen wrote:

>>
>> It did not put mkfs.btrfs in /sbin.  This did however, validate all the error paths ;).
> Heh ;)
> Well, mixing & matching upstream installs from source w/ rpm-packaged binaries is usually asking for trouble. 
> I don't think it's a bug, just bad administrative practice ;)
----
	SuSE is moving most if not all of their boot-related binaries off of "/" and
putting
them in /usr, requiring that you either have 1 partition for / and /usr or you run
their initrd that will pre-mount /usr on "/".  Supposedly this is a requirement
of moving the the MS-compat boot architecture, "systemd" and is being done by all
the distros... (supposedly)...  At least RedHat and SuSE are going that way...

	It's great to boot up in single-user from your hard disk and be told mount
needs libs on a yet-to-be mounted partition...

;-/