Re: [PATCH v4 2/2] md: allow configuring logical_block_size

[Li Nan <linan666@huaweicloud.com>]

在 2025/9/15 16:50, Xiao Ni 写道:
> On Mon, Sep 15, 2025 at 10:15 AM Li Nan <linan666@huaweicloud.com> wrote:
>>
>>
>>
>> 在 2025/9/15 8:33, Xiao Ni 写道:
>>> Hi Nan
>>>
>>> On Thu, Sep 11, 2025 at 3:41 PM <linan666@huaweicloud.com> wrote:
>>>>
>>>> From: Li Nan <linan122@huawei.com>
>>>>
>>>> Previously, raid array used the maximum logical_block_size (LBS) of
>>>> all member disks. Adding a larger LBS during disk at runtime could
>>>> unexpectedly increase RAID's LBS, risking corruption of existing
>>>> partitions.
>>>
>>> Could you describe more about the problem? It's better to give some
>>> test steps that can be used to reproduce this problem.
>>
>> Thanks for your review. I will add reproducer in the next version.
> 
> Thanks.
>>
>>>>
>>>> Simply restricting larger-LBS disks is inflexible. In some scenarios,
>>>> only disks with 512 LBS are available currently, but later, disks with
>>>> 4k LBS may be added to the array.
>>>>
>>>> Making LBS configurable is the best way to solve this scenario.
>>>> After this patch, the raid will:
>>>>     - stores LBS in disk metadata.
>>>>     - add a read-write sysfs 'mdX/logical_block_size'.
>>>>
>>>> Future mdadm should support setting LBS via metadata field during RAID
>>>> creation and the new sysfs. Though the kernel allows runtime LBS changes,
>>>> users should avoid modifying it after creating partitions or filesystems
>>>> to prevent compatibility issues.
>>>
>>> Because it only allows setting when creating an array. Can this be
>>> done automatically in kernel space?
>>>
>>> Best Regards
>>> Xiao
>>
>> The kernel defaults LBS to the max among all rdevs. When creating RAID
>> with mdadm, if mdadm doesn't set LBS explicitly, how does the kernel
>> learn the intended value?
>>
>> Gunaghao previously submitted a patch related to mdadm:
>> https://lore.kernel.org/all/3a9fa346-1041-400d-b954-2119c1ea001c@huawei.com/
> 
> Thanks for reminding me about this patch. First I still need to
> understand the problem. It may be a difficult thing for a user to


Thank you for your response.

Reproducer:
https://lore.kernel.org/all/3e26e8a3-e0c1-cd40-af79-06424fb2d54b@huaweicloud.com/

I will add it to messge log in the next version.

> choose the logcial block size. They don't know why they need to
> consider this value, right? If we only need a default value, the
> kernel space should be the right place?
> 

The scenario of our product is that they have a mix of 4k LBS and 512 LBS
disks. For most users, they do not need to modify the default values. This
is difficult to solve through default values.

> Regards
> Xiao
>>
>> --
>> Thanks,
>> Nan
>>
> 
> 
> .

-- 
Thanks,
Nan