VOGONS

Search

Common searches

Search results

Display options

Display options

Re: x86 Carry and Auxiliary/Adjust flags?

Posted on 2017-08-12, 15:06

I've tried making a simple truth table to use for calculating carry/adjust flags, based on the 7/15/31th bit of v1(the value being added/substracted to/from), v2(the value being added/substracted) and the result bit: ADD: CARRY(7/15/31)/AUX(3): V1 V2 RES FLAG 0 0 0 0 0 0 1 0 0 1 0 1 0 1 1 0 1 0 0 1 …

Re: x86 Carry and Auxiliary/Adjust flags?

Posted on 2017-08-11, 19:52

So the carry bit is set on toggling bit 8, while the aux bit does the same for bit 4? So it's quite litterally a half-carry(4-bits carry into bit 4 or borrow from bit 4)? Edit: I've changed the bits to check from bits 3(auxiliary) and 7/15/31(carry) to bits 4(auxiliary) and 8/16/32. The test suite's …

Re: Finding bugs in 8086-80186 emulation core?

Posted on 2017-08-11, 18:27

I think the problem is with the CPU itself, seeing as strange (un)real mode problems occur on 80286+, as well as problems with the Compaq BIOS locking up on CGA. Also, I know that there's problems with add/sub instructions using carry/overflow/adjust flags(I'm currently trying to adopt those based …

Re: Finding bugs in 8086-80186 emulation core?

Posted on 2017-08-11, 09:34

I've tested the following cases until now: - IBM XT 8086/88 with CGA installs and runs Windows 3.0 without problems. - IBM XT NEC V30(or 386, but don't remember for sure) with VGA running CGA windows that's already installed without problems. - IBM AT with 80286 with VGA crashes Windows 3.0 in real …

x86 Carry and Auxiliary/Adjust flags?

Posted on 2017-08-10, 16:40

I'm trying to get the flags properly implemented in UniPCemu, but the 80186(or 80286, since it expects the high 4 bits to be zeroed?) testsuite says the carry and adjust flags are incorrect? //Addition Carry, Overflow, Adjust logic //Tables based on http://teaching.idallen.com/dat2343/10f/notes/040_ …

Re: Finding bugs in 8086-80186 emulation core?

Posted on 2017-08-09, 21:16

Somehow, it now gets up to the phase of copying the second font(helvetica), but then entering the path, the keyboard input is 100% unresponsive(should still generate packets though). Same for the MS-DOS 5.0 prompt, but not for the mouse input(which responds fine(serial mouse on COM1). I'm testing …

Re: PS/2 mouse on Compaq Deskpro 386?

Posted on 2017-08-08, 11:56

@vladstamate: Yes. I've implemented a simply non-cycle accurate PS/2 mouse(the PS/2 keyboard uses timings, the PS/2 mouse doesn't yet). The PS/2 mouse still doesn't report any other thing with timings, like the PS/2 keyboard does, but does currently(not in released versions, just the development …

Re: Finding bugs in 8086-80186 emulation core?

Posted on 2017-08-05, 20:32

Just tried the 80186 testsuite again. The control.bin ROM fails. So there's definitely a problem with some control transfers. Edit: And by pure accident I find that I still haven't implemented the Resume Flag into the debugger exceptions. Implemented it right away. I'm wondering about something: …

Re: Finding bugs in 8086-80186 emulation core?

Posted on 2017-08-05, 17:36

Just tried Megarace again on the Compaq Deskpro 386 with the latest fixes(the 32-bit instructions being reimplemented and PS/2 mouse changed to Serial mouse). It now has a responsive menu with mouse reponse(no keyboard reponse until inputting the keys for mapping to fire etc.). When trying to map …

Re: PS/2 mouse on Compaq Deskpro 386?

Posted on 2017-08-05, 14:38

OK. Changed the PS/2 mouse to be disabled and used a serial mouse on all supported architectures(XT, AT, PS/2 and Compaq Deskpro 386). What was the first 386-based Compaq Deskpro 386-compatible 'modern-style' PC with PS/2 mouse support(on the Second PS/2 port)?

PS/2 mouse on Compaq Deskpro 386?

Posted on 2017-08-05, 13:23

Does the Compaq Deskpro 386 have a PS/2 mouse port? It probably has a PS/2 keyboard, but is a PS/2 mouse port installed as well? When I try to use UniPCemu, which uses a PS/2 mouse instead of a serial mouse on Compaq Deskpro 386 and PS/2 architectures, Cutemouse won't detect the PS/2 mouse(on the …

Re: Finding bugs in 8086-80186 emulation core?

Posted on 2017-08-04, 22:32

Gotten the 8086, 80186 and 80286 cores working now. I've adjusted the 80386 and 80486 timing tables to fix some errors and reconverted the cycle-accurate 8086 code to the 32-bit 80386 instructions(80186+ instructions still need to be converted). So far, the converted 8086 32-bit instructions seem to …

Re: Finding bugs in 8086-80186 emulation core?

Posted on 2017-08-03, 12:39

After using VBinDiff to compare the two logs, I've finally found a difference: F000:E180 (26C70608005FF8)MOVW AX,F85F Which is supposed to be: F000:E180 (26C70608005FF8)MOVW [ES:0008],F85F So one down, maybe some to go? Edit: Just a disassembler error, but it's a start. Next... Edit: Next up: the …

Re: Finding bugs in 8086-80186 emulation core?

Posted on 2017-08-03, 09:20

I'm trying to get the sed command to remove the logged timestamp(format: "%02u:%02u:%02u:%02u.%05u: ", without "-marks, as inputted into sprintf). I've currently end up with: $ sed -e 's/\(\[0-9\]\{2,3\}\):\(\[0-9\]\{2\}\):\(\[0-9\]\{2\}\):\(\[0-9\]\{2\}\).\(\[0-9\]\{5\}\)://g' debugger.log But it …

Re: Finding bugs in 8086-80186 emulation core?

Posted on 2017-08-03, 05:05

With the current commit, that's impossible, as it requires a running BIOS that can already boot from a harddisk to run the executables(which need to be stored inside a hard disk image for the emulator to run the .exe files). It's 80486 NOFPU version has always crashed on UniPCemu, not generating any …

Page 143 of 229

Go to top of page Go to top of page
Close window
Close window