Skip to contentSkip to conversation

btrfs: use -f by default for mkfs as xfs/extN do

PATCH In review Eric Sandeen 5 participants 8 messages Jun 3 – Jun 5 Btrfs file system

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 btrfs: use -f by default for mkfs as xfs/extN do

Recently saw a case where testing ext4, xfs, then btrfs failed all btrfs tests because mkfs.btrfs refused to overwrite the left-over xfs filesystem on $SCRATCH_DEV.

This seems well-enough fixed by only making the change to _scratch_mkfs, but several other helper functions explicitly use -f / -F for xfs and extN, so I've updated them for btrfs as well.

---

TBH I'm a little baffled, has nobody really run into this before, or am I missing something?

 common/rc | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
common/rc +5 -5
index 9632b211..e77edf97 100644
@@ -871,7 +871,7 @@ _test_mkfs()
         $MKFS_UDF_PROG $MKFS_OPTIONS $* $TEST_DEV > /dev/null
 	;;
     btrfs)
-        $MKFS_BTRFS_PROG $MKFS_OPTIONS $* $TEST_DEV > /dev/null
+        $MKFS_BTRFS_PROG -f $MKFS_OPTIONS $* $TEST_DEV > /dev/null
 	;;
     ext2|ext3|ext4)
 	$MKFS_PROG -t $FSTYP -- -F $MKFS_OPTIONS $* $TEST_DEV
@@ -920,7 +920,7 @@ _try_mkfs_dev()
         $MKFS_UDF_PROG $MKFS_OPTIONS $*
 	;;
     btrfs)
-        $MKFS_BTRFS_PROG $MKFS_OPTIONS $*
+        $MKFS_BTRFS_PROG -f $MKFS_OPTIONS $*
 	;;
     ext2|ext3|ext4)
 	$MKFS_PROG -t $FSTYP -- -F $MKFS_OPTIONS $*
@@ -1010,7 +1010,7 @@ _scratch_mkfs()
 		mkfs_filter="cat"
 		;;
 	btrfs)
-		mkfs_cmd="$MKFS_BTRFS_PROG"
+		mkfs_cmd="$MKFS_BTRFS_PROG -f"
 		mkfs_filter="cat"
 		;;
 	ext3)
@@ -1176,7 +1176,7 @@ _scratch_pool_mkfs()
 {
     case $FSTYP in
     btrfs)
-        $MKFS_BTRFS_PROG $MKFS_OPTIONS $* $SCRATCH_DEV_POOL > /dev/null
+        $MKFS_BTRFS_PROG -f $MKFS_OPTIONS $* $SCRATCH_DEV_POOL > /dev/null
         ;;
     *)
         echo "_scratch_pool_mkfs is not implemented for $FSTYP" 1>&2
@@ -1369,7 +1369,7 @@ _try_scratch_mkfs_sized()
 		# the device is not zoned. Ref: btrfs-progs: btrfs_min_dev_size()
 		(( fssize < $((256 * 1024 * 1024)) )) &&
 			! _scratch_btrfs_is_zoned && mixed_opt='--mixed'
-		$MKFS_BTRFS_PROG $MKFS_OPTIONS $mixed_opt -b $fssize "$@" $SCRATCH_DEV
+		$MKFS_BTRFS_PROG -f $MKFS_OPTIONS $mixed_opt -b $fssize "$@" $SCRATCH_DEV
 		;;
 	jfs)
 		${MKFS_PROG} -t $FSTYP $MKFS_OPTIONS "$@" $SCRATCH_DEV $blocks
Signature · 1 line
2.54.0
Reviewed-by on v1 1/1
#1 lore raw
Quoted 8 lines · Eric Sandeen

Recently saw a case where testing ext4, xfs, then btrfs failed all btrfs tests because mkfs.btrfs refused to overwrite the left-over xfs filesystem on $SCRATCH_DEV.

This seems well-enough fixed by only making the change to _scratch_mkfs, but several other helper functions explicitly use -f / -F for xfs and extN, so I've updated them for btrfs as well.

Quoted 5 lines

---

TBH I'm a little baffled, has nobody really run into this before, or am I missing something?

I found myself wondering btrfs folk test with MKFS_OPTS set, i.e.

MKFS_OPTS="-f" ./check -g quick

common/rc _test_mkfs()

Quoted 9 lines
 common/rc | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
common/rc +0 -0
index 9632b211..e77edf97 100644
@@ -871,7 +871,7 @@ _test_mkfs()

Not part of your change, but it feels like _test_mkfs() could simplify to

_test_mkfs()
{
try_mkfs_dev $TEST_DEV > /dev/null
}

But simple is good, so

-- Don

Quoted 44 lines
common/rc +5 -5
         $MKFS_UDF_PROG $MKFS_OPTIONS $* $TEST_DEV > /dev/null
 	;;
     btrfs)
-        $MKFS_BTRFS_PROG $MKFS_OPTIONS $* $TEST_DEV > /dev/null
+        $MKFS_BTRFS_PROG -f $MKFS_OPTIONS $* $TEST_DEV > /dev/null
 	;;
     ext2|ext3|ext4)
 	$MKFS_PROG -t $FSTYP -- -F $MKFS_OPTIONS $* $TEST_DEV
@@ -920,7 +920,7 @@ _try_mkfs_dev()
         $MKFS_UDF_PROG $MKFS_OPTIONS $*
 	;;
     btrfs)
-        $MKFS_BTRFS_PROG $MKFS_OPTIONS $*
+        $MKFS_BTRFS_PROG -f $MKFS_OPTIONS $*
 	;;
     ext2|ext3|ext4)
 	$MKFS_PROG -t $FSTYP -- -F $MKFS_OPTIONS $*
@@ -1010,7 +1010,7 @@ _scratch_mkfs()
 		mkfs_filter="cat"
 		;;
 	btrfs)
-		mkfs_cmd="$MKFS_BTRFS_PROG"
+		mkfs_cmd="$MKFS_BTRFS_PROG -f"
 		mkfs_filter="cat"
 		;;
 	ext3)
@@ -1176,7 +1176,7 @@ _scratch_pool_mkfs()
 {
     case $FSTYP in
     btrfs)
-        $MKFS_BTRFS_PROG $MKFS_OPTIONS $* $SCRATCH_DEV_POOL > /dev/null
+        $MKFS_BTRFS_PROG -f $MKFS_OPTIONS $* $SCRATCH_DEV_POOL > /dev/null
         ;;
     *)
         echo "_scratch_pool_mkfs is not implemented for $FSTYP" 1>&2
@@ -1369,7 +1369,7 @@ _try_scratch_mkfs_sized()
 		# the device is not zoned. Ref: btrfs-progs: btrfs_min_dev_size()
 		(( fssize < $((256 * 1024 * 1024)) )) &&
 			! _scratch_btrfs_is_zoned && mixed_opt='--mixed'
-		$MKFS_BTRFS_PROG $MKFS_OPTIONS $mixed_opt -b $fssize "$@" $SCRATCH_DEV
+		$MKFS_BTRFS_PROG -f $MKFS_OPTIONS $mixed_opt -b $fssize "$@" $SCRATCH_DEV
 		;;
 	jfs)
 		${MKFS_PROG} -t $FSTYP $MKFS_OPTIONS "$@" $SCRATCH_DEV $blocks
#2 lore raw
In reply to Eric Sandeen
Quoted 7 lines · Eric Sandeen

Recently saw a case where testing ext4, xfs, then btrfs failed all btrfs tests because mkfs.btrfs refused to overwrite the left-over xfs filesystem on $SCRATCH_DEV.

This seems well-enough fixed by only making the change to _scratch_mkfs, but several other helper functions explicitly use -f / -F for xfs and extN, so I've updated them for btrfs as well.

I haven't run into this yet, but it seem like btrfs needs this. But maybe you should also Cc the btrfs list?

Quoted 63 lines

---

TBH I'm a little baffled, has nobody really run into this before, or am I missing something?

 common/rc | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
common/rc +5 -5
index 9632b211..e77edf97 100644
@@ -871,7 +871,7 @@ _test_mkfs()
         $MKFS_UDF_PROG $MKFS_OPTIONS $* $TEST_DEV > /dev/null
 	;;
     btrfs)
-        $MKFS_BTRFS_PROG $MKFS_OPTIONS $* $TEST_DEV > /dev/null
+        $MKFS_BTRFS_PROG -f $MKFS_OPTIONS $* $TEST_DEV > /dev/null
 	;;
     ext2|ext3|ext4)
 	$MKFS_PROG -t $FSTYP -- -F $MKFS_OPTIONS $* $TEST_DEV
@@ -920,7 +920,7 @@ _try_mkfs_dev()
         $MKFS_UDF_PROG $MKFS_OPTIONS $*
 	;;
     btrfs)
-        $MKFS_BTRFS_PROG $MKFS_OPTIONS $*
+        $MKFS_BTRFS_PROG -f $MKFS_OPTIONS $*
 	;;
     ext2|ext3|ext4)
 	$MKFS_PROG -t $FSTYP -- -F $MKFS_OPTIONS $*
@@ -1010,7 +1010,7 @@ _scratch_mkfs()
 		mkfs_filter="cat"
 		;;
 	btrfs)
-		mkfs_cmd="$MKFS_BTRFS_PROG"
+		mkfs_cmd="$MKFS_BTRFS_PROG -f"
 		mkfs_filter="cat"
 		;;
 	ext3)
@@ -1176,7 +1176,7 @@ _scratch_pool_mkfs()
 {
     case $FSTYP in
     btrfs)
-        $MKFS_BTRFS_PROG $MKFS_OPTIONS $* $SCRATCH_DEV_POOL > /dev/null
+        $MKFS_BTRFS_PROG -f $MKFS_OPTIONS $* $SCRATCH_DEV_POOL > /dev/null
         ;;
     *)
         echo "_scratch_pool_mkfs is not implemented for $FSTYP" 1>&2
@@ -1369,7 +1369,7 @@ _try_scratch_mkfs_sized()
 		# the device is not zoned. Ref: btrfs-progs: btrfs_min_dev_size()
 		(( fssize < $((256 * 1024 * 1024)) )) &&
 			! _scratch_btrfs_is_zoned && mixed_opt='--mixed'
-		$MKFS_BTRFS_PROG $MKFS_OPTIONS $mixed_opt -b $fssize "$@" $SCRATCH_DEV
+		$MKFS_BTRFS_PROG -f $MKFS_OPTIONS $mixed_opt -b $fssize "$@" $SCRATCH_DEV
 		;;
 	jfs)
 		${MKFS_PROG} -t $FSTYP $MKFS_OPTIONS "$@" $SCRATCH_DEV $blocks
Signature · 1 line
2.54.0

---end quoted text---

#3 lore raw
Quoted 11 lines · Christoph Hellwig
Quoted 7 lines · Eric Sandeen

Recently saw a case where testing ext4, xfs, then btrfs failed all btrfs tests because mkfs.btrfs refused to overwrite the left-over xfs filesystem on $SCRATCH_DEV.

This seems well-enough fixed by only making the change to _scratch_mkfs, but several other helper functions explicitly use -f / -F for xfs and extN, so I've updated them for btrfs as well.

I haven't run into this yet, but it seem like btrfs needs this. But maybe you should also Cc the btrfs list?

This is the wrong fix.

It's caused by a mkfs.btrfs help string output change, and it's properly fixed by this patch which is already merged for a while:

https://lore.kernel.org/linux-btrfs/[email protected]/

Thanks,
Qu

Quoted 66 lines
Quoted 64 lines

---

TBH I'm a little baffled, has nobody really run into this before, or am I missing something?

  common/rc | 10 +++++-----
  1 file changed, 5 insertions(+), 5 deletions(-)
common/rc +5 -5
index 9632b211..e77edf97 100644
@@ -871,7 +871,7 @@ _test_mkfs()
          $MKFS_UDF_PROG $MKFS_OPTIONS $* $TEST_DEV > /dev/null
  	;;
      btrfs)
-        $MKFS_BTRFS_PROG $MKFS_OPTIONS $* $TEST_DEV > /dev/null
+        $MKFS_BTRFS_PROG -f $MKFS_OPTIONS $* $TEST_DEV > /dev/null
  	;;
      ext2|ext3|ext4)
  	$MKFS_PROG -t $FSTYP -- -F $MKFS_OPTIONS $* $TEST_DEV
@@ -920,7 +920,7 @@ _try_mkfs_dev()
          $MKFS_UDF_PROG $MKFS_OPTIONS $*
  	;;
      btrfs)
-        $MKFS_BTRFS_PROG $MKFS_OPTIONS $*
+        $MKFS_BTRFS_PROG -f $MKFS_OPTIONS $*
  	;;
      ext2|ext3|ext4)
  	$MKFS_PROG -t $FSTYP -- -F $MKFS_OPTIONS $*
@@ -1010,7 +1010,7 @@ _scratch_mkfs()
  		mkfs_filter="cat"
  		;;
  	btrfs)
-		mkfs_cmd="$MKFS_BTRFS_PROG"
+		mkfs_cmd="$MKFS_BTRFS_PROG -f"
  		mkfs_filter="cat"
  		;;
  	ext3)
@@ -1176,7 +1176,7 @@ _scratch_pool_mkfs()
  {
      case $FSTYP in
      btrfs)
-        $MKFS_BTRFS_PROG $MKFS_OPTIONS $* $SCRATCH_DEV_POOL > /dev/null
+        $MKFS_BTRFS_PROG -f $MKFS_OPTIONS $* $SCRATCH_DEV_POOL > /dev/null
          ;;
      *)
          echo "_scratch_pool_mkfs is not implemented for $FSTYP" 1>&2
@@ -1369,7 +1369,7 @@ _try_scratch_mkfs_sized()
  		# the device is not zoned. Ref: btrfs-progs: btrfs_min_dev_size()
  		(( fssize < $((256 * 1024 * 1024)) )) &&
  			! _scratch_btrfs_is_zoned && mixed_opt='--mixed'
-		$MKFS_BTRFS_PROG $MKFS_OPTIONS $mixed_opt -b $fssize "$@" $SCRATCH_DEV
+		$MKFS_BTRFS_PROG -f $MKFS_OPTIONS $mixed_opt -b $fssize "$@" $SCRATCH_DEV
  		;;
  	jfs)
  		${MKFS_PROG} -t $FSTYP $MKFS_OPTIONS "$@" $SCRATCH_DEV $blocks
Signature · 1 line
2.54.0

---end quoted text---

#4 lore raw
Quoted 15 lines · Qu Wenruo
Quoted 11 lines · Christoph Hellwig
Quoted 7 lines · Eric Sandeen

Recently saw a case where testing ext4, xfs, then btrfs failed all btrfs tests because mkfs.btrfs refused to overwrite the left-over xfs filesystem on $SCRATCH_DEV.

This seems well-enough fixed by only making the change to _scratch_mkfs, but several other helper functions explicitly use -f / -F for xfs and extN, so I've updated them for btrfs as well.

I haven't run into this yet, but it seem like btrfs needs this. But maybe you should also Cc the btrfs list?

Quoted 6 lines

This is the wrong fix.

It's caused by a mkfs.btrfs help string output change, and it's properly fixed by this patch which is already merged for a while:

https://lore.kernel.org/linux-btrfs/[email protected]/

In other words: we already get -f, but it got broken recently. That explains why things always worked for me.

I took me a bit to reverse-engineers this from your answer :)

#5 lore raw
Quoted 18 lines · Christoph Hellwig
Quoted 17 lines · Qu Wenruo
Quoted 11 lines · Christoph Hellwig
Quoted 7 lines · Eric Sandeen

Recently saw a case where testing ext4, xfs, then btrfs failed all btrfs tests because mkfs.btrfs refused to overwrite the left-over xfs filesystem on $SCRATCH_DEV.

This seems well-enough fixed by only making the change to _scratch_mkfs, but several other helper functions explicitly use -f / -F for xfs and extN, so I've updated them for btrfs as well.

I haven't run into this yet, but it seem like btrfs needs this. But maybe you should also Cc the btrfs list?

This is the wrong fix.

Quoted 7 lines
Quoted 4 lines

It's caused by a mkfs.btrfs help string output change, and it's properly fixed by this patch which is already merged for a while:

https://lore.kernel.org/linux-btrfs/[email protected]/

In other words: we already get -f, but it got broken recently. That explains why things always worked for me.

Hey, this shows you're running on a proper server distro, not some rolling bleeding edge one :)

And thanks for CCing the btrfs list.

Quoted 4 lines

I took me a bit to reverse-engineers this from your answer :)

#6 lore raw
Quoted 5 lines · Qu Wenruo
Quoted 2 lines

In other words: we already get -f, but it got broken recently. That explains why things always worked for me.

Hey, this shows you're running on a proper server distro, not some rolling bleeding edge one :)

Actually I run self-built btrfsprogs in my test VMs. I just didn't have a need to update it recently.

#7 lore raw
In reply to Qu Wenruo
Quoted 8 lines · Qu Wenruo
Quoted 5 lines · Christoph Hellwig
Quoted 4 lines · Eric Sandeen

Recently saw a case where testing ext4, xfs, then btrfs failed all btrfs tests because mkfs.btrfs refused to overwrite the left-over xfs filesystem on $SCRATCH_DEV.

Quoted 6 lines
Quoted 6 lines
Quoted 3 lines

This seems well-enough fixed by only making the change to _scratch_mkfs, but several other helper functions explicitly use -f / -F for xfs and extN, so I've updated them for btrfs as well.

I haven't run into this yet, but it seem like btrfs needs this. But maybe you should also Cc the btrfs list?

Yeah, sorry.

Quoted 5 lines

This is the wrong fix.

It's caused by a mkfs.btrfs help string output change, and it's properly fixed by this patch which is already merged for a while:

https://lore.kernel.org/linux-btrfs/[email protected]/

Oh, ugh. I did not know (ok: remember, because I wrote it in 2013) that we had a special helper function used only for a subset of filesystems, while hardcoding -f / -F on other filesystems. :(

Anyway, this fix is still on for-next and I missed it, sorry about that.

Thanks for the heads up.

-Eric

Quoted 2 lines

Thanks,
Qu