Hello
I have biostar M6TLD, it survives already raminit, patch is in attachment, also log from boot and config. Payload is uncompressed since i had problems with error in decompression from lzma, and current payload is coreinfo.
Machine reboots at point indicated in log but only difference between next loops is like that one below, diff between two restarts from same log.
Jumping to image. coreboot-2.3 Thu Dec 3 00:30:36 CET 2009 booting... Calibrating delay loop... -end 1ee10434f, start 1c570413d +end 2c5cfc8a0, start 29d2fc6b1 32-bit delta 650 calibrate_tsc 32-bit result is 650 clocks_per_usec: 650
Anyone has idea how to debug why it restarts? I am not placing signed off yet as it don't work unless producing endless stream of coreboot attempts to start is working.
best regards Maciej
Maciej Pijanka wrote:
I have biostar M6TLD, it survives already raminit, patch is in attachment, also log from boot and config.
Nice work so far!
Payload is uncompressed since i had problems with error in decompression from lzma,
What errors did you get?
and current payload is coreinfo.
Machine reboots at point indicated in log but only difference between next loops is like that one below, diff between two restarts from same log.
Jumping to image. coreboot-2.3 Thu Dec 3 00:30:36 CET 2009 booting... Calibrating delay loop... -end 1ee10434f, start 1c570413d +end 2c5cfc8a0, start 29d2fc6b1 32-bit delta 650 calibrate_tsc 32-bit result is 650 clocks_per_usec: 650
This difference makes sense if it is using something like RDTSC to calibrate. It runs from power on, so will be higher after a reset.
Anyone has idea how to debug why it restarts?
I think the problem may be that RAM is not completely functional. Decompression needs RAM, and the payload also needs RAM.
Did you try calling the very simple ram_check() function to test your RAM configuration? And how does the RAM configuration done by coreboot compare with that done by the factory BIOS?
//Peter
Anyone has idea how to debug why it restarts?
I think the problem may be that RAM is not completely functional. Decompression needs RAM, and the payload also needs RAM.
I agree. It might be interesting to disable caching and see if you still get so far. I've been fooled before because the cache was big enough to fit the working set.
Thanks, Myles
On Thu, 03 Dec 2009, Peter Stuge wrote:
Maciej Pijanka wrote:
I have biostar M6TLD, it survives already raminit, patch is in attachment, also log from boot and config.
Nice work so far!
Payload is uncompressed since i had problems with error in decompression from lzma,
What errors did you get?
just decompression error and then hang or reboot, it was some days ago, i can recheck in next hours i think
Machine reboots at point indicated in log but only difference between next loops is like that one below, diff between two restarts from same log.
Jumping to image. coreboot-2.3 Thu Dec 3 00:30:36 CET 2009 booting... Calibrating delay loop... -end 1ee10434f, start 1c570413d +end 2c5cfc8a0, start 29d2fc6b1 32-bit delta 650 calibrate_tsc 32-bit result is 650 clocks_per_usec: 650
This difference makes sense if it is using something like RDTSC to calibrate. It runs from power on, so will be higher after a reset.
I am aware that this change is harmless, but its only change diff found.
Anyone has idea how to debug why it restarts?
I think the problem may be that RAM is not completely functional. Decompression needs RAM, and the payload also needs RAM.
Did you try calling the very simple ram_check() function to test your RAM configuration? And how does the RAM configuration done by coreboot compare with that done by the factory BIOS?
yep, both low 1M, and 1M->memsize, even few times in rows (few times added in C code)
Mostly configuration (register contents) is similar or very, only thing i can't change is APBASE register thus i disabled AGP interface.
best regards Maciej
-
Maciej Pijanka -
Myles Watson -
Peter Stuge