Opened 5 months ago

Last modified 5 months ago

#7 new defect

XTIDECFG hang, possibly regression

Reported by: Plus-Dust Owned by: somebody
Priority: major Milestone:
Component: component1 Version:
Keywords: Cc:

Description

I am having some troubles with my XT-IDE card (a GlitchWorks rev 4b) and wanted to report it here. It's being used in a 286 clone with no other cards (Epson Equity IIe, 8/12mhz 80286 3mb RAM, VGA).

Trouble #1 and the one that I'm furthest along on is that it's difficult to flash the BIOS to the latest r625 because the XTIDECFG.COM immediately hangs when executed before showing the menu or anything, requiring a hard reset to recover.

I've only gotten it to run so far under emulation in DOSBox, on both the XT-IDE machine and another machine I have, a Pentium, it immediately locks up. I've been able to flash the BIOS by using r624, which all earlier versions seem to work fine without the hang, or by running the XTIDECFG from r625 in DOSBox, choosing to save the settings back to the image file, then copying the image onto a floppy and flashing it on the machine using the XTIDECFG from r624, or by just using an EPROM programmer out-of-band.

Trouble #2 is a weird one. In all the firmware versions I've tried so far, there are issues accessing the A: floppy if the system was booted from C:. It accesses the CompactFlash disk C: through the XTIDE just fine, I can boot from C: or A:, if I'm booted from A: everything works fine except of course for DOS periodically having to hit the floppy to reload COMMAND.COM. But if booted from C:, I can "dir A:" just fine, but for some reason any time I try to run or copy a program off the A: drive after having booted from C:, I get a "sector not found" error.

I assume this must have something to do with XUB hooking INT 13h (I assume it has to do that), so I just wrote a simple program in Turbo Pascal that dumps the values of all the interrupt vectors as some kind of place to start, but it always shows INT 13 pointing into low memory with or without the card; I assume that after XUB hooks the BIOS's INT 13 DOS is hooking it as well so it's not as easy to see as just that.

I'm primarily running MS-DOS 5.0 on this system but I've also tried MS-DOS 6.22 with the same effect.

Change History (1)

comment:1 by Plus-Dust, 5 months ago

Should maybe clarify: XTIDECFG hangs as soon as you push enter while still at the prompt, before the menu screen even comes up. That's weird to me because I don't (think?) it's even accessing the card at that point. So iow the hang doesn't occur when trying to flash or anything -- I can't even run the program on any of my real HW. But it works perfectly fine on DOSBox but that's the only place r625 works. Going back to r624 and the config tool works fine everywhere.

For issue 2 I've also tried other disks such as a DOM, in case there's some quirk with the CF card, but no change. And the CF drive works perfectly fine. It's just that having booted from it, you can no longer fully use the floppy anymore. The CF drive was set up by repartioning with fdisk, doing an fdisk /mbr, then a FORMAT C: /S from an MS-DOS 6.22 rescue floppy. That one happens even on the latest r625 *OR* on any earlier version I've tried.

Last edited 5 months ago by Plus-Dust (previous) (diff)
Note: See TracTickets for help on using tickets.