linux-kernel - Re: [PATCH 11/13] ext4: correct gdblock calculation in add_new_gdb_meta

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <29c9e94f-63b3-e757-9d6d-c9beaa0e0c19@huaweicloud.com>
Date:   Thu, 17 Aug 2023 11:38:34 +0800
From:   Kemeng Shi <shikemeng@...weicloud.com>
To:     Theodore Ts'o <tytso@....edu>
Cc:     adilger.kernel@...ger.ca, linux-ext4@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 11/13] ext4: correct gdblock calculation in
 add_new_gdb_meta_bg to support non first group



on 8/16/2023 11:45 AM, Theodore Ts'o wrote:
> On Thu, Jun 29, 2023 at 08:00:42PM +0800, Kemeng Shi wrote:
>> In add_new_gdb_meta_bg, we assume that group could be non first
>> group in meta block group as we call ext4_meta_bg_first_block_no
>> to get first block of meta block group rather than call
>> ext4_group_first_block_no for passed group directly. Then ext4_bg_has_super
>> should be called with first group in meta group rather than new added
>> group. Or we can call ext4_group_first_block_no instead of
>> ext4_meta_bg_first_block_no to assume only first group of
>> meta group will be passed.
>> Either way, ext4_meta_bg_first_block_no will be useless and
>> could be removed.
> 
> Unfortunately, I spent more time trying to understand the commit
> description than the C code.  Perhaps this might be a better way of
> describing the situation?
> 
Sorry for my poor language again, :( I will try to improve this.
> The ext4_new descs() function calls ext4_meta_bg_first_block_no() with
> the group paramter when the group is the first group of a meta_bg
> (e.g., when (group % EXT4_DESC_PER_BLOCK) is zero.  So we can simplify
> things a bit by removing ext4_meta_bg_first_block_no() and an open
> coding its logic.
> 
> Does this make more sense to tou?
> 
This patch tries to correct gdbblock calculation in add_new_gdb_meta_bg
in case group from caller is not the first group of meta_bg which is
supposed to be handled by add_new_gdb_meta_bg.
We should call ext4_bg_has_super with first group in meta_bg instead
of group which could be non first group in meta_bg to calculate gdb
of meta_bg.
Fortunately, the only caller ext4_add_new_descs always call
add_new_gdb_meta_bg with first group of meta_bg and no real issue
will happen.


> 					- Ted
>