linux-kernel - Re: [PATCH] md: prevent incoreect update of resync/recovery offset

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: <40fa8f33-22d8-b0f1-7d78-8961bee5b1f3@huaweicloud.com>
Date: Mon, 1 Sep 2025 09:26:43 +0800
From: Li Nan <linan666@...weicloud.com>
To: Yu Kuai <hailan@...uai.org.cn>, Paul Menzel <pmenzel@...gen.mpg.de>,
 Li Nan <linan666@...weicloud.com>
Cc: song@...nel.org, yukuai3@...wei.com, linux-raid@...r.kernel.org,
 linux-kernel@...r.kernel.org, yangerkun@...wei.com, yi.zhang@...wei.com
Subject: Re: [PATCH] md: prevent incoreect update of resync/recovery offset



在 2025/8/30 19:04, Yu Kuai 写道:
> Hi,
> 
> 在 2025/8/30 17:41, Paul Menzel 写道:
>> Dear Nan,
>>
>>
>> Thank you for your patch. I have some formal comments. In the 
>> summary/title: incor*r*ect.
>>
>> Am 30.08.25 um 11:02 schrieb linan666@...weicloud.com:
>>> From: Li Nan <linan122@...wei.com>
>>>
>>> In md_do_sync(), when md_sync_action returns ACTION_FROZEN, subsequent
>>> call to md_sync_position() will return MaxSector. This causes
>>> 'curr_resync' (and later 'recovery_offset') to be set to MaxSector too,
>>> which incorrectly signals that recovery/resync has completed even though
>>> disk data has not actually been updated.
>>>
>>> To fix this issue, skip updating any offset values when the sync acion
>>
>> ac*t*ion
>>
>>> is either FROZEN or IDLE.
>>
>> Maybe state that the same holds true for IDLE?
>>
>> Should these two cases be handled differently in `md_sync_position()`. 
>> Does it semantically make sense, that the default of MaxSector is returned?
>>
> Max sectors will return, however, following procedures will update
> resync_offset to MaxSector, while recovery can be interuppted by the
> concurrent frozen, can this will cause data lost.
> 
>>> Fixes: 7d9f107a4e94 ("md: use new helpers in md_do_sync()")
>>> Signed-off-by: Li Nan <linan122@...wei.com>
>>> ---
>>>   drivers/md/md.c | 3 +++
>>>   1 file changed, 3 insertions(+)
>>>
>>> diff --git a/drivers/md/md.c b/drivers/md/md.c
>>> index e78f80d39271..6828a569e819 100644
>>> --- a/drivers/md/md.c
>>> +++ b/drivers/md/md.c
>>> @@ -9397,6 +9397,9 @@ void md_do_sync(struct md_thread *thread)
>>>       }
>>>         action = md_sync_action(mddev);
>>> +    if (action == ACTION_FROZEN || action == ACTION_IDLE)
>>> +        goto skip;
>>> +
>>>       desc = md_sync_action_name(action);
>>>       mddev->last_sync_action = action;
>>
> Please send a new verison with the typo fixed.
> 
> Thanks,
> Kuai
> 

Sorry for my carelessness. I will send v2 later.

-- 
Thanks,
Nan

>>
>> Kind regards,
>>
>> Paul
>>
> 
> .