Re: Memory leak when raid10 takeover raid0

[Guoqing Jiang <guoqing.jiang@xxxxxxxxx>]

On 10/10/22 8:40 AM, Guoqing Jiang wrote:
> diff --git a/drivers/md/md.c b/drivers/md/md.c
> index afaf36b2f6ab..5a7134e3c182 100644
> --- a/drivers/md/md.c
> +++ b/drivers/md/md.c
> @@ -3979,6 +3979,9 @@ level_store(struct mddev *mddev, const char 
> *buf, size_t len)
>                 goto out_unlock;
>         }
>
> +       /* Only accept IO after takeover is done */
> +       mddev->pers = NULL;
> +
>         /* Looks like we have a winner */
>         mddev_suspend(mddev); 

mddev->pers->quiesce is called in suspend/resume, so we need relevant 
checking in them
accordingly, maybe other places as well.

Thanks,
Guoqing