Hi all,
At last... Finally I have had time to look at LinuxBIOS... I have been playing around with Etherboot and Filo for booting a development target, making good progress.
Up to now I have managed to add Etherboot to the original BIOS and have it working fine.
But now I want to LinuxBIOS my system ;-)
I have a Lippert Cool Roadrunner II with 64MB memory and a 128MB CompactFlash for disk, and I have tried to use LinuxBIOS v1 with its support for this board, but no success.
The serial port is very quite, and unfortunately I have no POST card. I have verified the serial connection: Linux on target "$ echo Hello
/dev/ttyS0" gives a nice "Hello" on the receiving end (no handshakes),
so something seems odd with the LinuxBIOS image...
Host is i686-Linux using gcc-4.0.3
--- config ------------ target /home/pdj/tuxway/builds/roadrunner2 mainboard lippert/roadrunner2
# Enable Serial Console for debugging option SERIAL_CONSOLE=1 option TTYS0_BAUD=57600
option DEFAULT_CONSOLE_LOGLEVEL=9 option DEBUG=1
option RAMTEST=1 option USE_GENERIC_ROM=1 option USE_ELF_BOOT=1 option ROM_SIZE=262144 option STD_FLASH=1
payload ~/tuxway/etherboot-5.0.6/src/bin32/eepro100.elf
option PAYLOAD_SIZE=196608 ---------------
I've hexdump'd the resulting romimage, and it looks fine as far as I can tell: (jmp far 0xf000:0x0004 at address 0x0003fff0, and payload image starting at 0x00000000).
I've used the flashrom from LinuxBIOSv2-2191 for burning: "# flashrom -w romimage"
Anyone able to give me a clue of what I've missed?
Anyone had success with LinuxBIOSv2 on a Roadrunner2 ?
~Per -----------------
PS: --- My ultimate goal is fast boot to either a remote image (through Etherboot) or a local linux on CompactFlash (filo and ext2).
My plan was to use LinuxBIOS with a combined filo and etherboot image as payload. Like this: 0x000000 FILO 0x010000 ETHERBOOT 0x030000 LINUXBIOS 0x040000 --- I have already modified filo so an external switch can select the boot-method.
Resounding silence...
Well I guess I will have to acquire a POST card and do some digging then.
~Per
Per Dalgas Jakobsen wrote:
Hi all,
At last... Finally I have had time to look at LinuxBIOS... I have been playing around with Etherboot and Filo for booting a development target, making good progress.
Up to now I have managed to add Etherboot to the original BIOS and have it working fine.
But now I want to LinuxBIOS my system ;-)
I have a Lippert Cool Roadrunner II with 64MB memory and a 128MB CompactFlash for disk, and I have tried to use LinuxBIOS v1 with its support for this board, but no success.
The serial port is very quite, and unfortunately I have no POST card. I have verified the serial connection: Linux on target "$ echo Hello
/dev/ttyS0" gives a nice "Hello" on the receiving end (no handshakes),
so something seems odd with the LinuxBIOS image...
Host is i686-Linux using gcc-4.0.3
--- config ------------ target /home/pdj/tuxway/builds/roadrunner2 mainboard lippert/roadrunner2
# Enable Serial Console for debugging option SERIAL_CONSOLE=1 option TTYS0_BAUD=57600
option DEFAULT_CONSOLE_LOGLEVEL=9 option DEBUG=1
option RAMTEST=1 option USE_GENERIC_ROM=1 option USE_ELF_BOOT=1 option ROM_SIZE=262144 option STD_FLASH=1
payload ~/tuxway/etherboot-5.0.6/src/bin32/eepro100.elf
option PAYLOAD_SIZE=196608
I've hexdump'd the resulting romimage, and it looks fine as far as I can tell: (jmp far 0xf000:0x0004 at address 0x0003fff0, and payload image starting at 0x00000000).
I've used the flashrom from LinuxBIOSv2-2191 for burning: "# flashrom -w romimage"
Anyone able to give me a clue of what I've missed?
Anyone had success with LinuxBIOSv2 on a Roadrunner2 ?
~Per
PS:
My ultimate goal is fast boot to either a remote image (through Etherboot) or a local linux on CompactFlash (filo and ext2).
My plan was to use LinuxBIOS with a combined filo and etherboot image as payload. Like this: 0x000000 FILO 0x010000 ETHERBOOT 0x030000 LINUXBIOS 0x040000
I have already modified filo so an external switch can select the boot-method.
On 3/14/06, Per Dalgas Jakobsen pdj@knaldgas.dk wrote:
Resounding silence...
Sorry, your post probally just got lost in all the stuff surrounding the GX2 and OLPC.
Well I guess I will have to acquire a POST card and do some digging then.
Until you get a post card my advice would be to edit the start up assembly and add a series of 5 outs to port 80 in a row and then hang in a spin loop. Then you can look at the ISA IO lines with a scope and see if you are getting the 5 outs.
No outs then your image needs a closer look.
If you are then walk the outs and hang pattern up the boot chain until you figure out why your serial port isn't up.
You can also fast track this by doing the same pattern in the early serial init code and see if you are getting to that code. if you are then your super IO is just not getting set up correctly. If not then you have to start earlier.
If you don't have access to a scope, well then you may just have to wait until you get a post card.
-- Richard A. Smith
On 3/14/06, Richard Smith smithbone@gmail.com wrote:
Well I guess I will have to acquire a POST card and do some digging then.
The thought just occured to me. How are the legacy ranges handled on that platform? Do you perhaps have to have some VSA magic to get the legacy serial range up?
-- Richard A. Smith
Resounding silence...
Sorry, your post probally just got lost in all the stuff surrounding the GX2 and OLPC.
Thanks for the feedback - I felt all alone ;-)
Until you get a post card my advice would be to edit the start up assembly and add a series of 5 outs to port 80 in a row and then hang in a spin loop. Then you can look at the ISA IO lines with a scope and see if you are getting the 5 outs.
I have a scope at hand, but I just found a place where I can borrow a PC104-development kit with a POST card :-D So I think I will fetch that kit and get going (I don't wan't to do too much hot-flash-switching if I can avoid it).
To be continued ;-)
~Per
you can look at the web article I did on debugging when things are going wrong.
Basically, infinite loops are your friend. :-)
ron