linux-kernel - Re: [PATCH v2 2/2] mtd: spinand: micron: fixing the offset for OOB

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: <90a53eda-bbeb-dffa-0f90-6fd229f555e6@sberdevices.ru>
Date:   Mon, 4 Sep 2023 17:20:59 +0300
From:   Martin Kurbanov <mmkurbanov@...rdevices.ru>
To:     Miquel Raynal <miquel.raynal@...tlin.com>
CC:     Richard Weinberger <richard@....at>,
        Vignesh Raghavendra <vigneshr@...com>,
        <linux-kernel@...r.kernel.org>, <linux-mtd@...ts.infradead.org>,
        <kernel@...rdevices.ru>
Subject: Re: [PATCH v2 2/2] mtd: spinand: micron: fixing the offset for OOB

Hi Miquel,

On 24.08.2023 12:35, Martin Kurbanov wrote:
> 
> 
> On 23.08.2023 14:39, Miquel Raynal wrote:
>> Hi Martin,
>>
>> mmkurbanov@...rdevices.ru wrote on Wed, 23 Aug 2023 14:33:57 +0300:
>>
>>> Hi Miquel,
>>>
>>> On 23.08.2023 11:41, Miquel Raynal wrote:
>>>> Hi Martin,
>>>>
>>>> I don't think the four bytes have any "bad block specific" meaning. In
>>>> practice, the datasheet states:
>>>>
>>>> 	Value programmed for bad block at the first byte of spare
>>>> 	area: 00h
>>>>
>>>> So only the first byte is used to mark the block bad, the rest is
>>>> probably marked "reserved" for simplicity. I believe we should keep the
>>>> current layout because it would otherwise break users for no real
>>>> reason.  
>>>
>>> I agree with you that this can break the work of users who use OOB.
>>> However, I believe it would be more appropriate to use an offset of 4,
>>> as the micron chip can use all 4 bytes for additional data about the
>>> bad block. So, there is a non-zero probability of losing OOB data in
>>> the reserved area (2 bytes) when the hardware chip attempts to mark
>>> the block as bad.
>>
>> Is this really a process the chip can do? Aren't bad blocks factory
>> marked only?
> 
> Actually, there is my understanding, I’m not sure exactly.

I tested with an offset of 2, no read/write errors were detected
(including read/write to OOB). But I don't have a flash chip with
factory bad blocks yet, when I find such a flash, I will report the
results.
Do I need to send the v3 of the patch with only first commit ("correct
bitmask for ecc status")?

-- 
Best Regards,
Martin Kurbanov