jjw@igloo.Scum.COM (John Welch) (04/16/89)
Some time ago, Igloo ran on an old 286 with Phoenix BIOS. Microport had trouble accessing the floppy drive on this system, and DOS likewise had problems. When we substituted a different rev of Phoenix BIOS, the problems went away for both DOS and Microport. Now igloo is on another machine, this time with AMI BIOS. Same trouble with the floppies, under both DOS and UNIX. We are hoping to try a BIOS swap to get rid of this trouble. My question is - I thought that BIOS was real-mode code, and not re-entrant, and all of those bad things that scre UNIX up. If that's the case, what explains the problems going away for both DOS *and* UNIX if the BIOS ONLY was changed??? Seems to me that Microport, in some way or another, is running BIOS to do disk i/o. They could be running it under protected mode, or they could be dropping back into real mode to do this. Either way, it seems kind of funny, and if they are dropping into real mode to do this that could explain character loss and other strange disk behavior. Any ideas, world? -- ========================================================================== John Welch <backbone>!killer!jjw@igloo "Oh, reality - it's not for me, and it makes me laugh; but fantasy world, and those Disney girls... I'm coming back!"
plocher%sally@Sun.COM (John Plocher) (04/16/89)
From article <1346@igloo.Scum.COM>, by jjw@igloo.Scum.COM (John Welch): > Some time ago, Igloo ran on an old 286 with Phoenix BIOS. Microport > Now igloo is on another machine, this time with AMI BIOS. Same trouble > My question is - I thought that BIOS was real-mode code, and not > the BIOS ONLY was changed??? Seems to me that Microport, in some way or > another, is running BIOS to do disk i/o. They could be running it under > protected mode, or they could be dropping back into real mode to do this. BullShit! The ONLY way that the BIOS interfaces to Microport Unix is thru whatever the BIOS does at POST time. After the big switch into protected mode (about 50 lines of code after you see the boot(system5): prompt) the BIOS is never used by unix. I know, I worked on the driver. DosMerge WHEN RUNNING DOS allows the DOS task full access to the BIOS (after all, the machine switches back to REAL mode to run the DOS task) This is a DOS only thing, though. Whenever an interrupt comes in, the machine switches into protected mode, handles the interrupt, and then switches back to real mode. -John Plocher
tneff@well.UUCP (Tom Neff) (04/17/89)
In article <1346@igloo.Scum.COM> jjw@igloo.UUCP (John Welch) writes: > [286 machine has floppy problems under DOS and uPort V/AT with some > BIOS's, not with others] > My question is - I thought that BIOS was real-mode code, and not >re-entrant, and all of those bad things that scre UNIX up. If that's the >case, what explains the problems going away for both DOS *and* UNIX if >the BIOS ONLY was changed??? I would be surprised if V/AT were mode switching for floppy I/O, you would think that would run like a pig. Maybe for the floppy they don't care? Still doubtful. Another thought is that they might steal the disk drive parameters, including spin-up times and such, from the BIOS at the beginning of the world, and use them to drive their own protected mode code. Then you might have floppy problems with some BIOS's even thought the driver code itself never changed. -- Tom Neff tneff@well.UUCP or tneff@dasys1.UUCP