Re: [PATCH] md: do not require mddev_lock() for all options

[Yu Kuai <yukuai1@huaweicloud.com>]

在 2023/09/23 5:04, Song Liu 写道:
> Hi Mariusz,
> 
> Sorry for the late reply.
> 
> On Wed, Sep 13, 2023 at 1:55 AM Mariusz Tkaczyk
> <mariusz.tkaczyk@linux.intel.com> wrote:
>>
>> We don't need to lock device to reject not supported request
>> in array_state_store().
>> Main motivation is to make a room for action does not require lock yet,
>> like prepare to stop (see md_ioctl()).
> 
> I made some changes to the commit log:
> 
>      md: do not require mddev_lock() for all options
> 
>      We don't need to lock device to reject not supported request
>      in array_state_store().
>      Main motivation is to make a room for action does not require lock yet,
>      like prepare to stop (see md_ioctl()).
> 
> But I am not sure what you meant by "make a room for action does not
> require lock yet". Could you please explain?

Yes, this sounds confusing, if 'action does not require lock', then it
shound not be blocked by array_state_store() with or without this patch.

> 
> Otherwise, the code looks reasonable to me.

Changes look good to me, after clarify commit message, feel free to add

Reviewed-by: Yu Kuai <yukuai3@huawei.com>
> 
> Thanks,
> Song
> .
>