Hi,
I have a PCI board that as the MV64360 with a 750GX processor. It doesn't have ethernet interfaces.
u-boot comes up on the board fine, but its behavior is erratic. On the host system, I have a device driver that exposes the SDRAM, flash, SRAM, etc, on the board via some /dev/ devices. When I start u-boot by clearing a bit in the Marvell registers that starts the processor from the host system, u-boot comes up, shows some serial output, and when it comes to autoboot, it just freezes.
My autoboot command is 'bootm 0x40004' and I place the hello_world.bin in the SDRAM at that location, just prior to autoboot via the host machine by stopping u-boot momentarily by putting a getc().
I also tried disabling autoboot and directly jumping to 0x40004. Even this doesn't help. It just shows a bad trap panic.
Once u-boot did manage to go through the 3-second timeout and jump to the binary. However, it reported 'unknown image format' or something like that.
Since then, it hasn't gone to the timeout stage.
I'm wondering if this is related to interrupts or serial output.
I have some debug messages printf()ed in serial_tstc and serial_getchar. When I disable them, it just freezes, so I don't know when and how. When I enable them, I know it did something for some time.
I can't use the BDI reliably too; I start the processor via the /dev interface, get u-boot till the getchar() adn then connect gdb via bdi. However, it'll just keep looping in getchar, and it'll be too long before it gets to the autoboot. I also won't notice if the execution just froze before it timed out.
Any tips where things could be going wrong?
Any debugging tips with BDI?
Thanks for any help, Amit.