[Bug 200635] Oops error in refresh_sit_entry() while unmounting a crafted F2FS image

[Bug 200635] Oops error in refresh_sit_entry() while unmounting a crafted F2FS image

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=200635

Chao Yu (chao@kernel.org) changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |chao@kernel.org
           Assignee|fs_other@kernel-bugs.osdl.o |filesystem_f2fs@kernel-bugs
                   |rg                          |.kernel.org

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

[Bug 200635] Oops error in refresh_sit_entry() while unmounting a crafted F2FS image

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=200635

--- Comment #1 from Chao Yu (chao@kernel.org) ---
It has been fixed by below commit, can you please check it?

f2fs: fix to do sanity check with cp_pack_start_sum

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=e494c2f995d6181d6e29c4927d68e0f295ecf75b

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

[Bug 200635] Oops error in refresh_sit_entry() while unmounting a crafted F2FS image

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=200635

Stathis Maneas (smaneas@cs.toronto.edu) changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |CODE_FIX

--- Comment #2 from Stathis Maneas (smaneas@cs.toronto.edu) ---
I can confirm that the specified commit actually resolves the error. Therefore,
I am marking the bug as resolved.

Thanks a lot,
Stathis Maneas

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

[Bug 200635] Oops error in refresh_sit_entry() while unmounting a crafted F2FS image

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=200635

Stathis Maneas (smaneas@cs.toronto.edu) changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|CODE_FIX                    |---

--- Comment #3 from Stathis Maneas (smaneas@cs.toronto.edu) ---
I have reopened the bug, because indeed, the fix included in the aforementioned
commit prevents the file system from being mounted after reading an invalid
value from the superblock.

Nonetheless, after invoking fsck to fix the partition, the file system still
cannot be mounted. Would the patch that introduces the checksum mechanism for
the superblock completely resolve the error?

Thanks,
Stathis Maneas

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

[Bug 200635] Oops error in refresh_sit_entry() while unmounting a crafted F2FS image

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=200635

--- Comment #4 from Chao Yu (chao@kernel.org) ---
Thanks for reminding that, IMO, yes, after we enable superblock checksum
feature, we can know which sb is corrupted, and try to recovery corrupted sb
with another good one.

Although, for your case, I guess the only corrupted field is
f2fs_super_block.cp_payload, which should be 0 instead of 1, as it can be
correctly hinted from checkpoint package fields that are under checksum
protection already.

So, w/o sb checksum feature, do you wish that we can have a try with recovering
sb fields from checkpoint pack?

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

[Bug 200635] Oops error in refresh_sit_entry() while unmounting a crafted F2FS image

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=200635

--- Comment #5 from Stathis Maneas (smaneas@cs.toronto.edu) ---
For the case where the checksum mechanism is not enabled for the superblock, it
would be nice to fix the error and make the file system able to be mounted once
again, provided that there is a safe way to restore the values stored in the
superblock using any of the underlying checkpoints.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

[Bug 200635] Oops error in refresh_sit_entry() while unmounting a crafted F2FS image

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=200635

--- Comment #6 from Chao Yu (chao@kernel.org) ---
SB checksum feature is not upstreamed yet, and there is one bug in the patch,
could you fix this manually and apply that patchset?

https://sourceforge.net/p/linux-f2fs/mailman/message/36402182/

I wrote a patch, which is based on sb checksum feature, could you have a try?

https://sourceforge.net/p/linux-f2fs/mailman/message/36402198/

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

[Bug 200635] Oops error in refresh_sit_entry() while unmounting a crafted F2FS image

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=200635

Stathis Maneas (smaneas@cs.toronto.edu) changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|---                         |CODE_FIX

--- Comment #7 from Stathis Maneas (smaneas@cs.toronto.edu) ---
I can verify that the latter patch is able to resolved the error and thus, I am
marking the issue as resolved.

Thanks a lot!

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

[Bug 200635] Oops error in refresh_sit_entry() while unmounting a crafted F2FS image

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=200635

--- Comment #8 from Chao Yu (chao@kernel.org) ---
(In reply to Stathis Maneas from comment #7)
> I can verify that the latter patch is able to resolved the error and thus, I
> am marking the issue as resolved.
> 
> Thanks a lot!

You're welcome, it't very good that we can dig one more fsck requirement from
this issue, thanks for the report and test. :)

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot