From: Jan Kara <jack@suse.cz>
To: Zhang Yi <yi.zhang@huawei.com>
Cc: Jan Kara <jack@suse.cz>, "Theodore Y. Ts'o" <tytso@mit.edu>,
Ext4 Developers List <linux-ext4@vger.kernel.org>,
yangerkun <yangerkun@huawei.com>
Subject: Re: [BUG && Question] question of SB_ACTIVE flag in ext4_orphan_cleanup()
Date: Mon, 22 Mar 2021 18:25:51 +0100 [thread overview]
Message-ID: <20210322172551.GJ31783@quack2.suse.cz> (raw)
In-Reply-To: <8a6864dd-7e6c-5268-2b5b-1010f99d2a1b@huawei.com>
Hi!
On Mon 22-03-21 23:24:23, Zhang Yi wrote:
> We find a use after free problem when CONFIG_QUOTA is enabled, the detail of
> this problem is below.
>
> mount_bdev()
> ext4_fill_super()
> sb->s_root = d_make_root(root);
> ext4_orphan_cleanup()
> sb->s_flags |= SB_ACTIVE; <--- 1. mark sb active
> ext4_orphan_get()
> ext4_truncate()
> ext4_block_truncate_page()
> mark_buffer_dirty <--- 2. dirty inode
> iput()
> iput_final <--- 3. put into lru list
> ext4_mark_recovery_complete <--- 4. failed and return error
> sb->s_root = NULL;
> deactivate_locked_super()
> kill_block_super()
> generic_shutdown_super()
> <--- 5. did not evict_inodes
> put_super()
> __put_super()
> <--- 6. put super block
>
> Because of the truncated inodes was dirty and will write them back later, it
> will trigger use after free problem. Now the question is why we need to set
> SB_ACTIVE bit when enable CONFIG_QUOTA below?
>
> #ifdef CONFIG_QUOTA
> /* Needed for iput() to work correctly and not trash data */
> sb->s_flags |= SB_ACTIVE;
>
> This code was merged long long ago in v2.6.6, IIUC, it may not affect
> the quota statistics it we evict inode directly in the last iput.
> In order to slove this UAF problem, I'm not sure is there any side effect
> if we just remove this code, or remove SB_ACTIVE and call evict_inodes()
> in the error path of ext4_fill_super().
>
> Could you give some suggestions?
That's a very good question. I do remember that I've added this code back
then because otherwise orphan cleanup was loosing updates to quota files.
But you're right that now I don't see how that could be happening and it
would be nice if we could get rid of this hack (and even better if it also
fixes the problem you've found). I guess I'll just try and test this change
with various quota configurations to see whether something still breaks or
not. Thanks report!
Honza
--
Jan Kara <jack@suse.com>
SUSE Labs, CR
next prev parent reply other threads:[~2021-03-22 17:26 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-03-22 15:24 [BUG && Question] question of SB_ACTIVE flag in ext4_orphan_cleanup() Zhang Yi
2021-03-22 17:25 ` Jan Kara [this message]
2021-03-29 9:20 ` Zhang Yi
2021-03-30 15:02 ` Jan Kara
2021-03-31 3:11 ` Zhang Yi
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20210322172551.GJ31783@quack2.suse.cz \
--to=jack@suse.cz \
--cc=linux-ext4@vger.kernel.org \
--cc=tytso@mit.edu \
--cc=yangerkun@huawei.com \
--cc=yi.zhang@huawei.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).