common/rc: drop '-f' parameter from fsck.exfat
PATCH 1/1 common/rc: drop '-f' parameter from fsck.exfat
fsck.exfat doesn't support the '-f' flag, so add a special case to _repair_test_fs().
--- common/rc | 8 ++++++++ 1 file changed, 8 insertions(+)
common/rc +8 -0
@@ -1229,6 +1229,14 @@ _repair_test_fs() res=$? fi ;; + exfat) + # exfat doesn't support -f + fsck -t $FSTYP -y $TEST_DEV >$tmp.repair 2>&1 + res=$? + if ((res < 4)); then + res=0 + fi + ;; *) # Let's hope fsck -y suffices... fsck -t $FSTYP -fy $TEST_DEV >$tmp.repair 2>&1
Signature · 1 line
2.35.3
Quoted 2 lines · David Disseldorp
fsck.exfat doesn't support the '-f' flag, so add a special case to
_repair_test_fs().
I'm wondering why _repair_scratch_fs() doesn't have the '-f', but the _repair_test_fs() has it. Looks like the '-f' option was for extN fs originally, it's not a fsck common option, but in fsck.ext4.
So I think the '-f' might not be a necessary option. As _repair_scratch_fs works without it, can we just remove the '-f' from _repair_test_fs()?
As _repair_test_fs was added by Ted, and _repair_scratch_fs was added by Darrick, so CC them to get more review -- do we real need the '-f' or not by default :)
Thanks,
Zorro
Quoted 28 lines
Signed-off-by David Disseldorp <[email protected]>--- common/rc | 8 ++++++++ 1 file changed, 8 insertions(+)common/rc +8
-0@@ -1229,6 +1229,14 @@ _repair_test_fs() res=$? fi ;; + exfat) + # exfat doesn't support -f + fsck -t $FSTYP -y $TEST_DEV >$tmp.repair 2>&1 + res=$? + if ((res < 4)); then + res=0 + fi + ;; *) # Let's hope fsck -y suffices... fsck -t $FSTYP -fy $TEST_DEV >$tmp.repair 2>&1Signature · 1 line
2.35.3
Thanks for the feedback, Zorro...
Quoted 10 lines · Zorro Lang
Quoted 2 lines · David Disseldorp
fsck.exfat doesn't support the '-f' flag, so add a special case to
_repair_test_fs().I'm wondering why
_repair_scratch_fs()doesn't have the '-f', but the_repair_test_fs()has it. Looks like the '-f' option was for extN fs originally, it's not a fsck common option, but in fsck.ext4.So I think the '-f' might not be a necessary option. As
_repair_scratch_fsworks without it, can we just remove the '-f' from_repair_test_fs()?
The fsck.ext4 usage states:
-f Force checking even if filesystem is marked clean
As _repair_test_fs() is only called on _check_test_fs() failure, I suppose '-f' removal should be fine. Still, to avoid any behavioural changes I could also just add an explicit ext case with '-f'.
There's also a question of what btrfs should do here, as calling the "fsck.btrfs" no-op script following btrfs check failure likely doesn't make much sense.
Cheers, David
Quoted 18 lines · David Disseldorp
Thanks for the feedback, Zorro...
Quoted 10 lines · Zorro Lang
Quoted 2 lines · David Disseldorp
fsck.exfat doesn't support the '-f' flag, so add a special case to
_repair_test_fs().I'm wondering why
_repair_scratch_fs()doesn't have the '-f', but the_repair_test_fs()has it. Looks like the '-f' option was for extN fs originally, it's not a fsck common option, but in fsck.ext4.So I think the '-f' might not be a necessary option. As
_repair_scratch_fsworks without it, can we just remove the '-f' from_repair_test_fs()?The fsck.ext4 usage states:
-f Force checking even if filesystem is marked clean
Quoted 7 lines
As
_repair_test_fs()is only called on_check_test_fs()failure, I suppose '-f' removal should be fine. Still, to avoid any behavioural changes I could also just add an explicit ext case with '-f'.There's also a question of what btrfs should do here, as calling the "fsck.btrfs" no-op script following btrfs check failure likely doesn't make much sense.
Hmm, I think you're right. I just checked the commit log, the code (check calls _repair_test_fs if TEST_DEV is corrupted) is added by this commit:
commit c7d81cdecbefd5768163a195e8d5257279216a34
Author: Theodore Ts'o <[email protected]> Date: Fri Apr 21 10:51:31 2023 -0700
check: try to fix the test device if it gets corrupted
That commit cares about xfs and extN more, didn't think about other fs likes btrfs. If btrfs would like to call btrfs-check, not fsck.btrfs, we should do that.
And as the "-f" isn't a common option of fsck, so we'd better to not use it in " *)" branch. If extN hope to use it, we'd better to have a "ext4)" branch.
If btrfs would like to call btrfs-check in _repair_test_fs(), we need to do the same thing in _repair_scratch_fs(). CC btrfs list to get more review/suggestion about that.
Thanks,
Zorro
Quoted 3 lines
Cheers, David
Quoted 8 lines · Zorro Lang
Quoted 2 lines · David Disseldorp
fsck.exfat doesn't support the '-f' flag, so add a special case to
_repair_test_fs().I'm wondering why
_repair_scratch_fs()doesn't have the '-f', but the_repair_test_fs()has it. Looks like the '-f' option was for extN fs originally, it's not a fsck common option, but in fsck.ext4.
Quoted 6 lines
So I think the '-f' might not be a necessary option. As
_repair_scratch_fsworks without it, can we just remove the '-f' from_repair_test_fs()?As
_repair_test_fswas added by Ted, and_repair_scratch_fswas added by Darrick, so CC them to get more review -- do we real need the '-f' or not by default :)
I suspect that was just an ext4ism, since fsck itself doesn't document any such option.
--D
Quoted 32 lines
Thanks,
ZorroQuoted 28 lines
Signed-off-by David Disseldorp <[email protected]>--- common/rc | 8 ++++++++ 1 file changed, 8 insertions(+)common/rc +8
-0@@ -1229,6 +1229,14 @@ _repair_test_fs() res=$? fi ;; + exfat) + # exfat doesn't support -f + fsck -t $FSTYP -y $TEST_DEV >$tmp.repair 2>&1 + res=$? + if ((res < 4)); then + res=0 + fi + ;; *) # Let's hope fsck -y suffices... fsck -t $FSTYP -fy $TEST_DEV >$tmp.repair 2>&1Signature · 1 line
2.35.3