Crisis Recovery - SMART Embedded Computing MVME2500 Installation & Use Manual

Functional Description
At power-up, the selection of the SPI boot device is strictly based upon the Switch Bank
(S2-2) setting. Depending on the S2-2 setting, SPI_SEL0 is routed to one of two SPI
devices. The selected SPI device must contain a boot image. Once the boot image is
copied into memory and executed, the FPGA will wait and once the P20x0 will write on one
bit of the FPGA watchdog register, the FPGA will then pass through the SPI chip select from
the P20x0 to SPI device chip selects. The software can now perform read/write processes
on any SPI device, including copying from one SPI device to another.
With this flexible approach to firmware redundancy, one should always be able to recover
from a corrupt active firmware image, as long as a healthy firmware image is maintained in
single bootable SPI Device.
The MVME2500 supports automatic switch over. If booting one device is not successful,
the watchdog will trigger the board reset and it will automatically boot on the other device.
4.6.4

Crisis Recovery

The MVME2500 provides an independent boot firmware recovery mechanism for the
operating system. The firmware recovery can be performed without leaving the firmware
environment.
During crisis recovery, the healthy boot image contained in SPI Device B is copied to SPI
Device A, replacing the corrupt boot image contained in SPI Device A.
Crisis recovery is performed as follows:
1. Power off the board.
2. Set Switch S2-2 to ON to point to SPI Device B (crisis image).
3. Power on the board.
4. Press <h> key on the keyboard to go to the U-Boot prompt.
5. Type moninit fru to copy the crisis image to SPI Device A.
6. Once the U-Boot prompt is visible, power off the board.
7. Set the S2-2 back to OFF to point to the SPI Device A.
8. Power on the board to boot from the newly recovered image on the SPI Device A.
The board will automatically switch over if one of the devices is corrupted.
MVME2500 Installation and Use (6806800L01S) 75