allan@dhw68k.cts.com (Alan Perry) (02/05/89)
Well, despite my continued (though slow) efforts, Minix still does not run on my Unisys PW2 850 (Mitsubishi MP386). I am running with 1.1 with two patches: a) someone suggested adding a routine to print an asterisk and calling it into the kernel initialization between where the interrupt vectors for 0 - 15 are changed and where the ones for 16 - 255 are changed; I also scattered calls to this print_star routine throughout main() and set up things so INT 10H was not changed (so I could continue to print asterisks up to the call to restart()). b) the recent posted floppy.c (I believe from 1.4?) was added (modified to remove the reference to ps). When the machine is reset, it starts fsck just fine. When I put in the root file system disk and press "=", the calls to print an asterisk that I scattered through main() display the asterisks up to the one just before restart(), the floppy indicator light comes on, the screen blanks and the world stops. Does anyone have a clue where I should go from here? Two things complicate my work: a) the technical reference that I ordered is on back order (last count there were 137 on back order) b) the machine that Minix runs on (where I do my compiles and builds) is 15 miles from the machine that I am trying to get Minix to run on. ANY clues will be appreciated. -- alan perry UUCP: {trwrb,hplabs}!felix!dhw68k!allan Internet: allan@dhw68k.cts.com
evans@ditsyda.oz (Bruce Evans) (02/07/89)
in article <19562@dhw68k.cts.com>, allan@dhw68k.cts.com (Alan Perry) says: >Well, despite my continued (though slow) efforts, Minix still does not run >on my Unisys PW2 850 (Mitsubishi MP386). I am running with 1.1 with two Whatever happened to the kernel debugger I posted last December? After a year's practice with it :-), I can usually locate kernel bugs within a few minutes. I rarely need or use debugging printfs. The debugger has a good chance of working whether Minix does or not, since it is essentially independent. >['=' key works but Minix crashes soon after the floppy light turns on] The '=' key "always" works unless you're enhancing and putting bugs into the bootstrap programs. Getting the floppy light shows that all the hardware tasks have at least started. If the keyboard is still alive an F1 dump will probably show FS stuck waiting for WINCHE or FLOPPY. Otherwise the interrupt handling is probably inconsistent as discussed for PS/50's. I suspect the floppy. What about converting bios_wini.c to bios_floppy.c? Bruce Evans evans@ditsyda.oz.au D