We've seen it with several machines. on a SMC93C94 chip, we wrote the driver, it works in the floppy enviornment. It is running in a 128Kbyte flash part. We've moved it in to the system bios flash, since we have 32K available, and it is at E8000, the VGA bios is at E0000 and the system bios is F0000. We've got I believe 3.0 release as well as the newer releases with PnP/PCI support as well. This is a ISA board/chip. The eprom is in the address space. We setup our board with In-Circuit Emulation so we could break-point, we are running a General Software bios, which we have complete source to, we actually see the bios go look at the part, and do the checksum, and on some builds the checksum is correct, sometimes not. I'm wondering if maybe if one of the files that makerom uses is not a multiple then the checksum is calculated incorrectly. We change one minor thing, and the the checksum does not calculate which really mess's up test. Just wondering if anyone has seen this before. We have a secondary problem, is when the code going from real to protected mode, after it adds the RELOC to the stack, then attempts to pop the return address off, we hang. It appears to be talking to non-existent memory. So I suspect the reloc is off. We can only test this with the Checksum passes. I'll have keming, and same forward a few more details. At 05:44 PM 8/22/98 +1000, you wrote: >>We seem to have a problem with makerom, that sometimes it does not generate >>the proper checksum, any ideas. It's driving us crazy. Of course, if >>the checksum is wrong, the BIOS cannot find netboot. > >I assume you mean the makerom from netboot which Gero maintains. Still >I'm sure it would help him if you mentioned which driver, what EPROM, >software version of netboot and any relevant platform details. >
For requests or suggestions regarding this mailing list archive please write to netboot@gkminix.han.de.