STEMMELIN, FREDERIC (FREDERIC)** CTR ** wrote:
Thank you very much for your help, in the mean time i have very goods news, yes my Tyan s2895 is BOOTING now :)
Nice, but..
kerneloops on screen after kernel was trying to set a IRQ for the intel network card, with error -39 (or -38 i dont remember exactly).
Now after remowing my Intel PRO1000MT server network card from PCI-X slot 6, i can boot without any problems :)
..the PCI-X should also work. I think it may be good to investigate why it gets such a high bus number.
Now here are my results with the patched Memtest86+ version 4.1. Check the attached files.
As you can see it is far way better, even if it is not working, at least i can see that the memory is recognized correctly, wich is pretty good. Memtest86+ is still crashing on test starting (after 3%), but really not important so far, as long as my computer does not crash in normal operation mode.
I think it is important. memtest should not crash, if it does, then I don't think everything is right with the memory setup, and that will lead to your computer crashing, it will just take some time; until you have started filling the RAM with stuff that you don't want to lose.
Another question, what may the consequences be not supporting e820 in bios,
*Please* remember that coreboot is not a BIOS. e820 memory maps are a BIOS-specific interface, and coreboot does not provide this. coreboot still provides a memory map, within the coreboot tables.
SeaBIOS provides BIOS interfaces, among other things it implements "e820" memory maps, based on the coreboot tables memory map.
for other OS like windows for example ? Does another OS use this feature e820 ?
Nearly every OS on the PC so far assumes that the 30-year old BIOS interface will be present.
Because systems do not understand the coreboot memory map format they require legacy BIOS compatibility and can only start using SeaBIOS. In a way this is also true for Linux, but mkelfImage helps bridge the gap so that no BIOS is needed.
Obviously everything that uses libpayload is also free of BIOS requirements.
memtest86 understands both e820 and coreboot tables, but the coreboot support in memtest86 was not updated to interpret the latest version of coreboot tables, or rather the specific high tables feature.
So an unpatched memtest86 will notice that coreboot is available even if you have SeaBIOS, it just does not know how to correctly interpret the coreboot tables.
//Peter