bkukanov
commented
7 years ago Points needed to be clarified:
- What have to see the Linux driver: new device with total space or identify each device and see the HW Raid device which will control the other eMMC cards
- If linux driver identify just HW raid 0 device, where should be stored super block informations
- How the driver should identify in which card has error or how it should react for errors
- How the data should be stored and addresses generated
We have used SW Raid 0 with 2 eMMC cards and speed has been grown just for 50%, now we are going to implement the HW raid 0. First off all we need to analyse the mdadm working principles.