Why generic/073 is generic but not btrfs specific?
Subsystems (2)
Hello,
Running generic/073 for the case of HFS+ finishes with volume corruption:
sudo ./check generic/073
FSTYP -- hfsplus
PLATFORM -- Linux/x86_64 hfsplus-testing-0001 6.17.0-rc1+ #4 SMP PREEMPT_DYNAMIC Wed Oct 1 15:02:44 PDT 2025MKFS_OPTIONS -- /dev/loop51MOUNT_OPTIONS -- /dev/loop51 /mnt/scratch
generic/073 _check_generic_filesystem: filesystem on /dev/loop51 is inconsistent (see XFSTESTS-2/xfstests-dev/results//generic/073.full for details)
Ran: generic/073
Failures: generic/073
Failed 1 of 1 tests
sudo fsck.hfsplus -d /dev/loop51
** /dev/loop51
Using cacheBlockSize=32K cacheTotalBlock=1024 cacheSize=32768K. Executing fsck_hfs (version 540.1-Linux).
** Checking non-journaled HFS Plus Volume.
The volume name is untitled
** Checking extents overflow file.
** Checking catalog file.
** Checking multi-linked files.
** Checking catalog hierarchy.
Invalid directory item count
(It should be 1 instead of 0)
** Checking extended attributes file.
** Checking volume bitmap.
** Checking volume information.
Verify Status: VIStat = 0x0000, ABTStat = 0x0000 EBTStat = 0x0000 CBTStat = 0x0000 CatStat = 0x00004000
** Repairing volume.
** Rechecking volume.
** Checking non-journaled HFS Plus Volume.
The volume name is untitled
** Checking extents overflow file.
** Checking catalog file.
** Checking multi-linked files.
** Checking catalog hierarchy.
** Checking extended attributes file.
** Checking volume bitmap.
** Checking volume information.
** The volume untitled was repaired successfully.
Initially, I considered that something is wrong with HFS+ driver logic. But after testing and debugging the issue, I believe that HFS+ logic is correct.
As far as I can see, the generic/073 is checking specific btrfs related case:
# Test file A fsync after moving one other unrelated file B between directories
# and fsyncing B's old parent directory before fsyncing the file A. Check that # after a crash all the file A data we fsynced is available. # # This test is motivated by an issue discovered in btrfs which caused the file # data to be lost (despite fsync returning success to user space). That btrfs # bug was fixed by the following linux kernel patch: # # Btrfs: fix data loss in the fast fsync path
The test is doing these steps on final phase:
mv $SCRATCH_MNT/testdir_1/bar $SCRATCH_MNT/testdir_2/bar $XFS_IO_PROG -c "fsync" $SCRATCH_MNT/testdir_1 $XFS_IO_PROG -c "fsync" $SCRATCH_MNT/foo
So, we move file bar from testdir_1 into testdir_2 folder. It means that HFS+ logic decrements the number of entries in testdir_1 and increments number of entries in testdir_2. Finally, we do fsync only for testdir_1 and foo but not for testdir_2. As a result, this is the reason why fsck.hfsplus detects the volume corruption afterwards. As far as I can see, the HFS+ driver behavior is completely correct and nothing needs to be done for fixing in HFS+ logic here.
But what could be the proper solution? Should generic/073 be excluded from HFS/HFS+ xfstests run? Or, maybe, generic/073 needs to be btrfs specific? Am I missing something here?
Thanks,
Slava.
Quoted 65 lines · Viacheslav Dubeyko
Hello,
Running generic/073 for the case of HFS+ finishes with volume corruption:
sudo ./check generic/073
FSTYP -- hfsplus
PLATFORM -- Linux/x86_64 hfsplus-testing-0001 6.17.0-rc1+ #4 SMPPREEMPT_DYNAMICWed Oct 1 15:02:44 PDT 2025MKFS_OPTIONS--/dev/loop51MOUNT_OPTIONS--/dev/loop51/mnt/scratchgeneric/073
_check_generic_filesystem: filesystem on/dev/loop51is inconsistent (see XFSTESTS-2/xfstests-dev/results//generic/073.full for details)Ran: generic/073
Failures: generic/073
Failed 1 of 1 testssudo fsck.hfsplus -d
/dev/loop51
**/dev/loop51
Using cacheBlockSize=32K cacheTotalBlock=1024 cacheSize=32768K. Executingfsck_hfs(version 540.1-Linux).
** Checking non-journaled HFS Plus Volume.
The volume name is untitled
** Checking extents overflow file.
** Checking catalog file.
** Checking multi-linked files.
** Checking catalog hierarchy.
Invalid directory item count
(It should be 1 instead of 0)
** Checking extended attributes file.
** Checking volume bitmap.
** Checking volume information.
Verify Status: VIStat =0x0000, ABTStat =0x0000EBTStat =0x0000CBTStat =0x0000CatStat =0x00004000
** Repairing volume.
** Rechecking volume.
** Checking non-journaled HFS Plus Volume.
The volume name is untitled
** Checking extents overflow file.
** Checking catalog file.
** Checking multi-linked files.
** Checking catalog hierarchy.
** Checking extended attributes file.
** Checking volume bitmap.
** Checking volume information.
** The volume untitled was repaired successfully.Initially, I considered that something is wrong with HFS+ driver logic. But after testing and debugging the issue, I believe that HFS+ logic is correct.
As far as I can see, the generic/073 is checking specific btrfs related case:
# Test file A fsync after moving one other unrelated file B between directories
# and fsyncing B's old parent directory before fsyncing the file A. Check that # after a crash all the file A data we fsynced is available. # # This test is motivated by an issue discovered in btrfs which caused the file # data to be lost (despite fsync returning success to user space). That btrfs # bug was fixed by the following linux kernel patch: # # Btrfs: fix data loss in the fast fsync pathThe test is doing these steps on final phase:
Quoted 8 lines
mv $SCRATCH_MNT/testdir_1/bar $SCRATCH_MNT/testdir_2/bar $XFS_IO_PROG -c "fsync" $SCRATCH_MNT/testdir_1 $XFS_IO_PROG -c "fsync" $SCRATCH_MNT/foo
So, we move file bar from
testdir_1intotestdir_2folder. It means that HFS+ logic decrements the number of entries intestdir_1and increments number of entries intestdir_2. Finally, we do fsync only fortestdir_1and foo but not fortestdir_2.
If the fs is using journal, shouldn't the increments on the testdir_2 already be journaled? Thus after a power loss, the increments on testdir_2/bar should be replayed thus the end user should still see that inode.
To me this looks like a bug in HFS logic where something is not properly journaled (the increment on the testdir_2/bar).
Finally, if you're asking an end user that if it is acceptable that after moving an inode and fsync the old directory, the inode is no longer reachable, I'm pretty sure no end user will think it's acceptable.
Quoted 3 lines
As a result, this is the reason why fsck.hfsplus detects the volume corruption afterwards. As far as I can see, the HFS+ driver behavior is completely correct and nothing needs to be done for fixing in HFS+ logic here.
Then I guess you may also want to ask why EXT4/XFS/Btrfs/F2fs all pass the test case.
Thanks,
Qu
Quoted 7 lines
But what could be the proper solution? Should generic/073 be excluded from HFS/HFS+ xfstests run? Or, maybe, generic/073 needs to be btrfs specific? Am I missing something here?
Thanks,
Slava.
Quoted 76 lines · Qu Wenruo
Quoted 65 lines · Viacheslav Dubeyko
Hello,
Running generic/073 for the case of HFS+ finishes with volume corruption:
sudo ./check generic/073
FSTYP -- hfsplus
PLATFORM -- Linux/x86_64 hfsplus-testing-0001 6.17.0-rc1+ #4 SMPPREEMPT_DYNAMICWed Oct 1 15:02:44 PDT 2025MKFS_OPTIONS--/dev/loop51MOUNT_OPTIONS--/dev/loop51/mnt/scratchgeneric/073
_check_generic_filesystem: filesystem on/dev/loop51is inconsistent (see XFSTESTS-2/xfstests-dev/results//generic/073.full for details)Ran: generic/073
Failures: generic/073
Failed 1 of 1 testssudo fsck.hfsplus -d
/dev/loop51
**/dev/loop51
Using cacheBlockSize=32K cacheTotalBlock=1024 cacheSize=32768K. Executingfsck_hfs(version 540.1-Linux).
** Checking non-journaled HFS Plus Volume.
The volume name is untitled
** Checking extents overflow file.
** Checking catalog file.
** Checking multi-linked files.
** Checking catalog hierarchy.
Invalid directory item count
(It should be 1 instead of 0)
** Checking extended attributes file.
** Checking volume bitmap.
** Checking volume information.
Verify Status: VIStat =0x0000, ABTStat =0x0000EBTStat =0x0000CBTStat =0x0000CatStat =0x00004000
** Repairing volume.
** Rechecking volume.
** Checking non-journaled HFS Plus Volume.
The volume name is untitled
** Checking extents overflow file.
** Checking catalog file.
** Checking multi-linked files.
** Checking catalog hierarchy.
** Checking extended attributes file.
** Checking volume bitmap.
** Checking volume information.
** The volume untitled was repaired successfully.Initially, I considered that something is wrong with HFS+ driver logic. But after testing and debugging the issue, I believe that HFS+ logic is correct.
As far as I can see, the generic/073 is checking specific btrfs related case:
# Test file A fsync after moving one other unrelated file B between directories
# and fsyncing B's old parent directory before fsyncing the file A. Check that # after a crash all the file A data we fsynced is available. # # This test is motivated by an issue discovered in btrfs which caused the file # data to be lost (despite fsync returning success to user space). That btrfs # bug was fixed by the following linux kernel patch: # # Btrfs: fix data loss in the fast fsync pathThe test is doing these steps on final phase:
Quoted 8 lines
mv $SCRATCH_MNT/testdir_1/bar $SCRATCH_MNT/testdir_2/bar $XFS_IO_PROG -c "fsync" $SCRATCH_MNT/testdir_1 $XFS_IO_PROG -c "fsync" $SCRATCH_MNT/foo
So, we move file bar from
testdir_1intotestdir_2folder. It means that HFS+ logic decrements the number of entries intestdir_1and increments number of entries intestdir_2. Finally, we do fsync only fortestdir_1and foo but not fortestdir_2.
Quoted 5 lines
If the fs is using journal, shouldn't the increments on the
testdir_2already be journaled? Thus after a power loss, the increments on testdir_2/bar should be replayed thus the end user should still see that inode.
Technically speaking, HFS+ is journaling file system in Apple implementation. But we don't have this functionality implemented and fully supported on Linux kernel side. Potentially, it can be done but currently we haven't such functionality yet. So, HFS/HFS+ doesn't use journaling on Linux kernel side and no journal replay could happen. :)
Quoted 4 lines
To me this looks like a bug in HFS logic where something is not properly journaled (the increment on the testdir_2/bar).
This statement could be correct if we will support journaling for HFS+. But HFS never supported the journaling technology.
Quoted 4 lines
Finally, if you're asking an end user that if it is acceptable that after moving an inode and fsync the old directory, the inode is no longer reachable, I'm pretty sure no end user will think it's acceptable.
As far as I can see, the fsck only corrects the number of entries for testdir_2 folder. The rest is pretty OK.
Quoted 7 lines
Quoted 3 lines
As a result, this is the reason why fsck.hfsplus detects the volume corruption afterwards. As far as I can see, the HFS+ driver behavior is completely correct and nothing needs to be done for fixing in HFS+ logic here.
Then I guess you may also want to ask why EXT4/XFS/Btrfs/F2fs all pass the test case.
So, ext4 and xfs support journaling. Maybe, it's interesting to check what is going on with F2FS, FAT, and other file system that hasn't journaling support.
Thanks,
Slava.
Quoted 83 lines · Viacheslav Dubeyko
Quoted 76 lines · Qu Wenruo
Quoted 65 lines · Viacheslav Dubeyko
Hello,
Running generic/073 for the case of HFS+ finishes with volume corruption:
sudo ./check generic/073
FSTYP -- hfsplus
PLATFORM -- Linux/x86_64 hfsplus-testing-0001 6.17.0-rc1+ #4 SMPPREEMPT_DYNAMICWed Oct 1 15:02:44 PDT 2025MKFS_OPTIONS--/dev/loop51MOUNT_OPTIONS--/dev/loop51/mnt/scratchgeneric/073
_check_generic_filesystem: filesystem on/dev/loop51is inconsistent (see XFSTESTS-2/xfstests-dev/results//generic/073.full for details)Ran: generic/073
Failures: generic/073
Failed 1 of 1 testssudo fsck.hfsplus -d
/dev/loop51
**/dev/loop51
Using cacheBlockSize=32K cacheTotalBlock=1024 cacheSize=32768K. Executingfsck_hfs(version 540.1-Linux).
** Checking non-journaled HFS Plus Volume.
The volume name is untitled
** Checking extents overflow file.
** Checking catalog file.
** Checking multi-linked files.
** Checking catalog hierarchy.
Invalid directory item count
(It should be 1 instead of 0)
** Checking extended attributes file.
** Checking volume bitmap.
** Checking volume information.
Verify Status: VIStat =0x0000, ABTStat =0x0000EBTStat =0x0000CBTStat =0x0000CatStat =0x00004000
** Repairing volume.
** Rechecking volume.
** Checking non-journaled HFS Plus Volume.
The volume name is untitled
** Checking extents overflow file.
** Checking catalog file.
** Checking multi-linked files.
** Checking catalog hierarchy.
** Checking extended attributes file.
** Checking volume bitmap.
** Checking volume information.
** The volume untitled was repaired successfully.Initially, I considered that something is wrong with HFS+ driver logic. But after testing and debugging the issue, I believe that HFS+ logic is correct.
As far as I can see, the generic/073 is checking specific btrfs related case:
# Test file A fsync after moving one other unrelated file B between directories
# and fsyncing B's old parent directory before fsyncing the file A. Check that # after a crash all the file A data we fsynced is available. # # This test is motivated by an issue discovered in btrfs which caused the file # data to be lost (despite fsync returning success to user space). That btrfs # bug was fixed by the following linux kernel patch: # # Btrfs: fix data loss in the fast fsync pathThe test is doing these steps on final phase:
Quoted 8 lines
mv $SCRATCH_MNT/testdir_1/bar $SCRATCH_MNT/testdir_2/bar $XFS_IO_PROG -c "fsync" $SCRATCH_MNT/testdir_1 $XFS_IO_PROG -c "fsync" $SCRATCH_MNT/foo
So, we move file bar from
testdir_1intotestdir_2folder. It means that HFS+ logic decrements the number of entries intestdir_1and increments number of entries intestdir_2. Finally, we do fsync only fortestdir_1and foo but not fortestdir_2.Quoted 5 lines
If the fs is using journal, shouldn't the increments on the
testdir_2already be journaled? Thus after a power loss, the increments on testdir_2/bar should be replayed thus the end user should still see that inode.
Quoted 5 lines
Technically speaking, HFS+ is journaling file system in Apple implementation. But we don't have this functionality implemented and fully supported on Linux kernel side. Potentially, it can be done but currently we haven't such functionality yet. So, HFS/HFS+ doesn't use journaling on Linux kernel side and no journal replay could happen. :)
That's fine, btrfs doesn't support (traditional) journal either, since its metadata is already fully COW protected.
The journal-like part is called log-tree, which is only utilized for speeding up fsync() so no full fs sync is needed (which is super expensive for btrfs)
That's the reason why btrfs' fsync() is super complex and Filipe spent tons of his time on this field, and it is btrfs that exposed this problem.
But if HFS+ on linux doesn't support metadata journal, I'm afraid you will need to go the hard path just like btrfs, to manually check if an inode needs its parent inodes updated during fsync.
Quoted 9 lines
Quoted 5 lines
To me this looks like a bug in HFS logic where something is not properly journaled (the increment on the testdir_2/bar).
This statement could be correct if we will support journaling for HFS+. But HFS never supported the journaling technology.
Quoted 7 lines
Quoted 4 lines
Finally, if you're asking an end user that if it is acceptable that after moving an inode and fsync the old directory, the inode is no longer reachable, I'm pretty sure no end user will think it's acceptable.
As far as I can see, the fsck only corrects the number of entries for
testdir_2folder. The rest is pretty OK.
So it means the nlink number on the child inode is correct. I guess since testdir_2 is not the directory where fsync is called, thus its number of references is not updated.
With journal it will be much easier, but without it, you have to iterate through the changed inode backrefs, and update the involved directories one by one during fsync(), just like btrfs...
Quoted 11 lines
Quoted 8 lines
Quoted 4 lines
As a result, this is the reason why fsck.hfsplus detects the volume corruption afterwards. As far as I can see, the HFS+ driver behavior is completely correct and nothing needs to be done for fixing in HFS+ logic here.
Then I guess you may also want to ask why EXT4/XFS/Btrfs/F2fs all pass the test case.
So, ext4 and xfs support journaling. Maybe, it's interesting to check what is going on with F2FS, FAT, and other file system that hasn't journaling support.
IIRC F2FS also support journals.
For VFAT the test case is rejected due to the lack of journal support.
On the other hand, if HFS+ doesn't support journal at all, you may want to add hfs/hfs+ to _has_metadata_journaling(), so that the test case will be skipped.
But still, since this test case already leads to fsck error reports, it is worthy some improvement to avoid such problem.
Thanks,
Qu
Quoted 3 lines
Thanks,
Slava.
Quoted 98 lines · Qu Wenruo
Quoted 83 lines · Viacheslav Dubeyko
Quoted 76 lines · Qu Wenruo
Quoted 65 lines · Viacheslav Dubeyko
Hello,
Running generic/073 for the case of HFS+ finishes with volume corruption:
sudo ./check generic/073
FSTYP -- hfsplus
PLATFORM -- Linux/x86_64 hfsplus-testing-0001 6.17.0-rc1+ #4 SMPPREEMPT_DYNAMICWed Oct 1 15:02:44 PDT 2025MKFS_OPTIONS--/dev/loop51MOUNT_OPTIONS--/dev/loop51/mnt/scratchgeneric/073
_check_generic_filesystem: filesystem on/dev/loop51is inconsistent (see XFSTESTS-2/xfstests-dev/results//generic/073.full for details)Ran: generic/073
Failures: generic/073
Failed 1 of 1 testssudo fsck.hfsplus -d
/dev/loop51
**/dev/loop51
Using cacheBlockSize=32K cacheTotalBlock=1024 cacheSize=32768K. Executingfsck_hfs(version 540.1-Linux).
** Checking non-journaled HFS Plus Volume.
The volume name is untitled
** Checking extents overflow file.
** Checking catalog file.
** Checking multi-linked files.
** Checking catalog hierarchy.
Invalid directory item count
(It should be 1 instead of 0)
** Checking extended attributes file.
** Checking volume bitmap.
** Checking volume information.
Verify Status: VIStat =0x0000, ABTStat =0x0000EBTStat =0x0000CBTStat =0x0000CatStat =0x00004000
** Repairing volume.
** Rechecking volume.
** Checking non-journaled HFS Plus Volume.
The volume name is untitled
** Checking extents overflow file.
** Checking catalog file.
** Checking multi-linked files.
** Checking catalog hierarchy.
** Checking extended attributes file.
** Checking volume bitmap.
** Checking volume information.
** The volume untitled was repaired successfully.Initially, I considered that something is wrong with HFS+ driver logic. But after testing and debugging the issue, I believe that HFS+ logic is correct.
As far as I can see, the generic/073 is checking specific btrfs related case:
# Test file A fsync after moving one other unrelated file B between directories
# and fsyncing B's old parent directory before fsyncing the file A. Check that # after a crash all the file A data we fsynced is available. # # This test is motivated by an issue discovered in btrfs which caused the file # data to be lost (despite fsync returning success to user space). That btrfs # bug was fixed by the following linux kernel patch: # # Btrfs: fix data loss in the fast fsync pathThe test is doing these steps on final phase:
Quoted 8 lines
mv $SCRATCH_MNT/testdir_1/bar $SCRATCH_MNT/testdir_2/bar $XFS_IO_PROG -c "fsync" $SCRATCH_MNT/testdir_1 $XFS_IO_PROG -c "fsync" $SCRATCH_MNT/foo
So, we move file bar from
testdir_1intotestdir_2folder. It means that HFS+ logic decrements the number of entries intestdir_1and increments number of entries intestdir_2. Finally, we do fsync only fortestdir_1and foo but not fortestdir_2.Quoted 5 lines
If the fs is using journal, shouldn't the increments on the
testdir_2already be journaled? Thus after a power loss, the increments on testdir_2/bar should be replayed thus the end user should still see that inode.Quoted 5 lines
Technically speaking, HFS+ is journaling file system in Apple implementation. But we don't have this functionality implemented and fully supported on Linux kernel side. Potentially, it can be done but currently we haven't such functionality yet. So, HFS/HFS+ doesn't use journaling on Linux kernel side and no journal replay could happen. :)
That's fine, btrfs doesn't support (traditional) journal either, since its metadata is already fully COW protected.
The journal-like part is called log-tree, which is only utilized for speeding up
fsync()so no full fs sync is needed (which is super expensive for btrfs)
Quoted 7 lines
That's the reason why btrfs'
fsync()is super complex and Filipe spent tons of his time on this field, and it is btrfs that exposed this problem.But if HFS+ on linux doesn't support metadata journal, I'm afraid you will need to go the hard path just like btrfs, to manually check if an inode needs its parent inodes updated during fsync.
Yeah, you are right, this is one of the possible way. Let me check the complexity of this change.
Quoted 39 lines
Quoted 9 lines
Quoted 5 lines
To me this looks like a bug in HFS logic where something is not properly journaled (the increment on the testdir_2/bar).
This statement could be correct if we will support journaling for HFS+. But HFS never supported the journaling technology.
Quoted 7 lines
Quoted 4 lines
Finally, if you're asking an end user that if it is acceptable that after moving an inode and fsync the old directory, the inode is no longer reachable, I'm pretty sure no end user will think it's acceptable.
As far as I can see, the fsck only corrects the number of entries for
testdir_2folder. The rest is pretty OK.So it means the nlink number on the child inode is correct. I guess since
testdir_2is not the directory where fsync is called, thus its number of references is not updated.With journal it will be much easier, but without it, you have to iterate through the changed inode backrefs, and update the involved directories one by one during
fsync(), just like btrfs...Quoted 11 lines
Quoted 8 lines
Quoted 4 lines
As a result, this is the reason why fsck.hfsplus detects the volume corruption afterwards. As far as I can see, the HFS+ driver behavior is completely correct and nothing needs to be done for fixing in HFS+ logic here.
Then I guess you may also want to ask why EXT4/XFS/Btrfs/F2fs all pass the test case.
So, ext4 and xfs support journaling. Maybe, it's interesting to check what is going on with F2FS, FAT, and other file system that hasn't journaling support.
IIRC F2FS also support journals.
Quoted 9 lines
For VFAT the test case is rejected due to the lack of journal support.
On the other hand, if HFS+ doesn't support journal at all, you may want to add hfs/hfs+ to
_has_metadata_journaling(), so that the test case will be skipped.But still, since this test case already leads to fsck error reports, it is worthy some improvement to avoid such problem.
Make sense. Let me check the complexity of HFS+ modification to support correctly such situation. But HFS should be added to _has_metadata_journaling() if it is not already there.
Thanks,
Slava.
Quoted 5 lines · Viacheslav Dubeyko
Quoted 5 lines
Quoted 5 lines
Technically speaking, HFS+ is journaling file system in Apple implementation. But we don't have this functionality implemented and fully supported on Linux kernel side. Potentially, it can be done but currently we haven't such functionality yet. So, HFS/HFS+ doesn't use journaling on Linux kernel side and no journal replay could happen. :)
If the implementation of HJFJS+ in Linux doesn't support metadata consistency after a crash, I'd suggest adding HFS+ to _has_metadat_journalling(). This will suppress a number of test failures so you can focus on other issues which arguably is probably higher priority for you to fix.
After you get HFS+ to run clean with the journalling tesets skipped, then you can focus on adding that guarantee at that point, perhaps?
- Ted
Quoted 7 lines · Theodore Ts'o
Quoted 5 lines · Viacheslav Dubeyko
Quoted 5 lines
Quoted 5 lines
Technically speaking, HFS+ is journaling file system in Apple implementation. But we don't have this functionality implemented and fully supported on Linux kernel side. Potentially, it can be done but currently we haven't such functionality yet. So, HFS/HFS+ doesn't use journaling on Linux kernel side and no journal replay could happen. :)
Quoted 10 lines
If the implementation of HJFJS+ in Linux doesn't support metadata consistency after a crash, I'd suggest adding HFS+ to
_has_metadat_journalling(). This will suppress a number of test failures so you can focus on other issues which arguably is probably higher priority for you to fix.After you get HFS+ to run clean with the journalling tesets skipped, then you can focus on adding that guarantee at that point, perhaps?
Yes, it makes sense. It's really good strategy. But I've decided to spend couple of days on the fix of this issue. If I am not lucky to find the quick fix, then I'll follow this strategy. :)
Thanks,
Slava.
Quoted 14 lines · Viacheslav Dubeyko
Quoted 13 lines · Theodore Ts'o
Quoted 5 lines · Viacheslav Dubeyko
Quoted 5 lines
Quoted 5 lines
Technically speaking, HFS+ is journaling file system in Apple implementation. But we don't have this functionality implemented and fully supported on Linux kernel side. Potentially, it can be done but currently we haven't such functionality yet. So, HFS/HFS+ doesn't use journaling on Linux kernel side and no journal replay could happen. :)
If the implementation of HJFJS+ in Linux doesn't support metadata consistency after a crash, I'd suggest adding HFS+ to
_has_metadat_journalling(). This will suppress a number of test failures so you can focus on other issues which arguably is probably higher priority for you to fix.
Quoted 8 lines
Quoted 4 lines
After you get HFS+ to run clean with the journalling tesets skipped, then you can focus on adding that guarantee at that point, perhaps?
Yes, it makes sense. It's really good strategy. But I've decided to spend couple of days on the fix of this issue. If I am not lucky to find the quick fix, then I'll follow this strategy. :)
Hi Slava,
fstests doesn't have an offical HFS+ supporting report (refer to README), so if you find some helpers/cases can't work on hfs/hfsplus well, please feel free to modify them to support hfs/hfsplus, then add hfs/hfsplus to the supporting list (in README) after we make sure it works :)
Thanks,
Zorro
Quoted 3 lines
Thanks,
Slava.
Quoted 24 lines · Zorro Lang
Quoted 14 lines · Viacheslav Dubeyko
Quoted 13 lines · Theodore Ts'o
Quoted 5 lines · Viacheslav Dubeyko
Quoted 5 lines
Quoted 5 lines
Technically speaking, HFS+ is journaling file system in Apple implementation. But we don't have this functionality implemented and fully supported on Linux kernel side. Potentially, it can be done but currently we haven't such functionality yet. So, HFS/HFS+ doesn't use journaling on Linux kernel side and no journal replay could happen. :)
If the implementation of HJFJS+ in Linux doesn't support metadata consistency after a crash, I'd suggest adding HFS+ to
_has_metadat_journalling(). This will suppress a number of test failures so you can focus on other issues which arguably is probably higher priority for you to fix.Quoted 8 lines
Quoted 4 lines
After you get HFS+ to run clean with the journalling tesets skipped, then you can focus on adding that guarantee at that point, perhaps?
Yes, it makes sense. It's really good strategy. But I've decided to spend couple of days on the fix of this issue. If I am not lucky to find the quick fix, then I'll follow this strategy. :)
Quoted 7 lines
Hi Slava,
fstests doesn't have an offical HFS+ supporting report (refer to README), so if you find some helpers/cases can't work on hfs/hfsplus well, please feel free to modify them to support hfs/hfsplus, then add hfs/hfsplus to the supporting list (in README) after we make sure it works :)
Hi Zorro,
Sounds good! I am working on it. Currently, I am trying to identify such test- cases. I have several in mind but, currently, I am focused on test-cases that require the fixes on HFS/HFS+ side.
I'll share the patches when I'll completely sure that HFS/HFS+ cannot manage some xfstests cases.
Thanks,
Slava.