Re: [PATCH v2 1/2] md/raid5: freeze reshape when encountering a bad read

[Doug V Johnson]

> Perhaps, check badblocks before starting reshape in the first place can
> solve most problems in this case? We can keep this patch just in case
> new badblocks are set while performing reshape.
>
> Thanks,
> Kuai
This is a really good idea and should mitigate the risk of getting into 
these weird states. I started work on a patch that seems to work, I will 
submit it after I have done some testing.

Thanks for your feedback,
Doug