Re: UBIFS errors since Linux 5.15

[Eddie James <eajames@linux.ibm.com>]

On 11/7/25 2:03 PM, Richard Weinberger wrote:
> Eddie,
>
> ----- Ursprüngliche Mail -----
>> Von: "Eddie James" <eajames@linux.ibm.com>
>> An: "linux-mtd" <linux-mtd@lists.infradead.org>, "richard" <richard@nod.at>, "chengzhihao1" <chengzhihao1@huawei.com>,
>> "OpenBMC Maillist" <openbmc@lists.ozlabs.org>, "linux-kernel" <linux-kernel@vger.kernel.org>, sethjenkins@google.com,
>> "adobriyan" <adobriyan@gmail.com>
>> Gesendet: Freitag, 7. November 2025 18:38:54
>> Betreff: UBIFS errors since Linux 5.15
>> Hello,
>>
>> We are updating some BMC (Aspeed AST2600, ARM32) systems that were on
>> Linux 5.4 but observe consistent UBIFS errors with Linux 5.15 and later.
>> The errors seem to vary; there is a sampling below.
>>
>>
>> We performed a git bisect and came up with the following commit, which
>> of course is completely unrelated to UBIFS. Reverting it does prevent
>> the issue... Does anyone have any suggestions?
>>
>> commit 33fc9e26b7cb39f0d4219c875a2451802249c225
>> Author: Seth Jenkins <sethjenkins@google.com>
>> Date:   Thu Oct 27 11:36:52 2022 -0400
> That's indeed strange.
> Are you absolutely sure that this commit triggers the issue?
> E.g. if you revert it on top of 5.15 the issue vanishes and when you
> re-apply it, the issue is back?


Thanks for your quick reply. Yes, that is the case; it's not very 
logical. We are currently running more extended tests with 5.15 with 
33fc9e26b reverted. I assume the potential for corruption is there 
without this commit too, but somehow it doesn't happen.


Eddie


>
> Thanks,
> //richard
>