Xilinx ML300 u-boot question
Does anybody know what parameters I should use to
load u-boot, using u-boot. That's right, I want to
build a generic u-boot that goes on the Microdrive of a Xilinx ML300 development board. I have several of them to manage. I want the generic u-boot to go and download a machine-specific u-boot that I can change on my workstation without having to write it to the microdrive.
Or can I just have u-boot start the kernel with no parameters and have all the initial args for the kernel built into the kernel?
-mike wellington
[email protected]
** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Mike,
for TFTP boot normally the HW MAC address of the target device is used for identification. On the ML300 the HW MAC address is stored inside the IIC EEPROM. Since accessing this EEPROM is not yet supported in the pusblished U-Boot port for ML300 the MAC address is hardcoded in board/xilinx/xilinx_enet/emac_adapter.c.
So, to get a machine specific u-boot for every board change the MAC address and recompile. I'm not sure that you would need the secondary u-boot then as you could then directly load the Linux kernel in the second step.
- Peter
Quoted 9 lines · Mike Wellington
Does anybody know what parameters I should use to
load u-boot, using u-boot. That's right, I want to
build a generic u-boot that goes on the Microdrive of a Xilinx ML300 development board. I have several of them to manage. I want the generic u-boot to go and download a machine-specific u-boot that I can change on my workstation without having to write it to the microdrive.
Quoted 10 lines
Or can I just have u-boot start the kernel with no parameters and have all the initial args for the kernel built into the kernel?
-mike wellington
[email protected]
** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Peter-
That solves part of the problem. I'd still have to either
download a new u-boot via XPS or copy it to the MicroDrive every time I changed u-boot. I'll do it this way for now, but it would be better if I built two different u-boots that are made to execute from different RAM locations. Then I can download the (changeable) second-stage u-boot via tftp.
-mike wellington
[email protected]
[email protected]
Quoted 14 lines · Peter Ryser
Mike,
for TFTP boot normally the HW MAC address of the target device is used for identification. On the ML300 the HW MAC address is stored inside the IIC EEPROM. Since accessing this EEPROM is not yet supported in the pusblished U-Boot port for ML300 the MAC address is hardcoded in
board/xilinx/xilinx_enet/emac_adapter.c.So, to get a machine specific u-boot for every board change the MAC address and recompile. I'm not sure that you would need the secondary u-boot then as you could then directly load the Linux kernel in the second step.
Quoted 26 lines
- Peter
Quoted 19 lines · Mike Wellington
Does anybody know what parameters I should use to
load u-boot, using u-boot. That's right, I want to
build a generic u-boot that goes on the Microdrive of a Xilinx ML300 development board. I have several of them to manage. I want the generic u-boot to go and download a machine-specific u-boot that I can change on my workstation without having to write it to the microdrive.Or can I just have u-boot start the kernel with no parameters and have all the initial args for the kernel built into the kernel?
-mike wellington
[email protected]
** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/