FWIW, our hardware solution was to replace motherboard what seem to fix the problem. Maybe there was some bug in the chipset and the PC bios had some patch in it to fix it, as the old MBO would run fine with PCBIOS classic. The old MBO was winfast6300.
We have fairly advanced prototype system running house and we exepect release fairly soon.
seems like we are just down to one (stowstopper) bug.
It seems that we must disable MTRR, or else the stuff just don't work right (ex: it will fail to boot kernel on switch to protected mode). On the other hand if we do disable MTRRS stuff work pretty briskily with lilo/grub/linux/openbsd/etc working.
The system is winfast 6300 with sis630 chipset and pentium III and 128mb SDRAM.
does any one has any idea what sort of interaction could cause such problems?
MTRRs are reported by linux bios are:
Setting fixed MTRRs(0-88) type: UC Setting fixed MTRRs(0-16) type: WB DONE fixed MTRRs Setting variable MTRR 0, base: 0MB, range: 64MB, type WB Setting variable MTRR 1, base: 64MB, range: 32MB, type WB Setting variable MTRR 2, base: 96MB, range: 16MB, type WB Setting variable MTRR 3, base: 112MB, range: 8MB, type WB DONE variable MTRRs Clear out the extra MTRR's call intel_enable_fixed_mtrr() call intel_enable_var_mtrr() Leave setup_mtrrs done.