[Etherboot-developers] stumbling block for 5.2
Brought to you by:
marty_connor,
stefanhajnoczi
|
From: <ke...@et...> - 2003-07-28 12:54:56
|
Ok, I've fixed up a few more errors in the build rules. The .com rule was missing a start16.bin. Now I can run a .com Etherboot from DOS. The problem I'm facing now is that the Quit function of Etherboot (formerly called Local) doesn't return control to the caller, which is DOS in the case of .com images, and the BIOS in the case of .zrom images. It jumps off into hyperspace and often reboots the machine. .zdsk images don't really have an environment to return to so Quit is arguably undefined. This needs to be fixed before release, otherwise Etherboot doesn't follow the BIOS boot order, which means one cannot quit Etherboot to let the BIOS boot from floppy or disk. Somewhere in start16 or start32, the return path is botched. The C main seems ok, I've printf'ed it to the point of returning from main. Does vmware allow one to trace what's happening in machine code after that? I'm testing with a PCI/PnP BIOS but booting an ISA card if that makes any difference. I'd appreciate some help with this, while I edit the 5.2 docs some more. If somebody wants to get their teeth into this, you'll need the latest from CVS, which has several bug fixes already. |