- SarahWalker: What is that, and how do I use it?

Edit: Found what Bitbucket is, but it seems I should have thought of it earlier. I got my copy of the PCem source from here: http://www.retroroms.net/modules/newbb/viewto … id=7657&forum=3 , and not from the repository. I'll see if I can figure out a way to make a patchset from my source folder.

Trio64 32-bit issues should be fixed now. Hopefully.

Submitted my patchset to you. Includes both the S3 changes (forking of Trio64 into three by adding the Phoenix ones, which mainly differ in the VRAM amount, and in case of the Trio32, in how it identifies as), and the keyboard changes.

Also, the 32-bit modes now work fine, though in 1024x768x32-bit on the Trio64, the hardware cursor is rendered incorrectly, look at this screenshot: http://imgur.com/246odQ9 .

Edit: A link to the patchset, in case you did not get it otherwise: http://citadel.ringoflightning.net/92.patch .

Also, again with Little Big Adventure 2, it makes the entire emulator crash with a fatal error saying "bad getppcache B8000" if EMM386 is loaded so that segments B000-B7FF are used for upper memory. This definitely happens neither on real hardware (tested it 12 years ago on a Pentium 100 MHz, as well as on a Celeron 333 MHz), nor in any other emulator or virtualizer that I tested. While without EMM386, it loads but has that render problem I mentioned before (only top quarter of the screen gets drawn).

I compiled PCEm 6579fed92118 on Bitbucket repository, but Windows 3.11 not load anymore.
It hangs on copyright screen.

PCEm 8.1 stable works very good

Configuration:
AMI WinBIOS 486
Number Nine 9FX (S3 Trio64)
Intel i486DX/33 Cache A bit
Fast VLB/PCI
Sound Blaster 16
8 Mega RAM

acd2001 wrote:

I compiled PCEm 6579fed92118 on Bitbucket repository, but Windows 3.11 not load anymore. It hangs on copyright screen. […]
Show full quote

I compiled PCEm 6579fed92118 on Bitbucket repository, but Windows 3.11 not load anymore.
It hangs on copyright screen.

PCEm 8.1 stable works very good

Configuration:
AMI WinBIOS 486
Number Nine 9FX (S3 Trio64)
Intel i486DX/33 Cache A bit
Fast VLB/PCI
Sound Blaster 16
8 Mega RAM

PCEm 6579fed92118

My Microsoft Windows for Worksgroups Version 3.11 is working with the same settings (also with Cache A bit).

acd2001 wrote:

I compiled PCEm 6579fed92118 on Bitbucket repository, but Windows 3.11 not load anymore.
It hangs on copyright screen.

There is an issue currently that loading (some) DOS mouse drivers will cause Windows to hang on load. Is this what you're doing?

SarahWalker wrote:

acd2001 wrote:

I compiled PCEm 6579fed92118 on Bitbucket repository, but Windows 3.11 not load anymore.
It hangs on copyright screen.

There is an issue currently that loading (some) DOS mouse drivers will cause Windows to hang on load. Is this what you're doing?

You are right, if I add the CuteMouse driver v1.9, Windows hangs.

Windows 3.x also hangs if LMOUSE (the Logitech mouse driver) is loaded.

You're right.
I use cutemouse driver.

I try to compile using serial.c and serial.h from PCEm 8.1 and it works again.

for some reason, in my binary of PCem 8.1 (latest revision, 91), cdaudio appears broken, when I play a mixed mode CD in a virtual drive, it just doesn't play any sound and the timer doesn't go forward in any track (except track 1 which is the data one).
EDIT: I think it has something to do with the ioctl implementation at the current time on PCem
EDIT2: Happens on DOS, Win3.x (3.0 and 3.1x), generic CDROM driver (gscdrom, that is oakcdrom compatible and atapi with MSCDEX 2.23) and on 32bit CDROM driver on windows 9x.

Here is a link to my PCem patchset (basically adds Phoenix S3 Trio64 and Phoenix S3 Trio32, makes it so the correct VRAM information is always returned from CRTC 0x36, and rewrites keyboard handling using raw input): http://citadel.ringoflightning.net/92.patch .

apparently the cdaudio counter problem was a subchannel issue and I fixed it.

Patch?

IDE Timer patch for AMI 286/386/486 and PCI related machines + XTIDE.
http://towg.tk/Emulators/ide.diff

CD Audio patch
http://towg.tk/Emulators/cdaudio-ide.diff

hmm Sarah, I also noticed your current IDE implementation is missing the read_capacity command (0x25) which, afaik, is used by NT 3.x and 4.0. It may be a reason why NT doesn't run at the current moment on your emulator. And again, feel free to add it or not 😉, in the meantime, I'm going to add it, borrowing some ideas from MESS.

SA1988 wrote:

hmm Sarah, I also noticed your current IDE implementation is missing the read_capacity command (0x25) which, afaik, is used by NT 3.x and 4.0. It may be a reason why NT doesn't run at the current moment on your emulator. And again, feel free to it or not 😉, in the meantime, I'm going to add it, borrowing some ideas from MESS.

Does anyone have any idea where the read_capacity (0x25) command can be found? :\

Sarah, I've been running PCem on my new computer so far and things are looking great. The PCem emulator has significantly improved since I first used it and I like it! 😀

I have a couple of questions to ask, btw...

In the Award SIS 496/497 and Award 430VX PCI based BIOSes, support for up to four hard drives (secondary master drive E: and secondary slave drive F:) are available and PCem is only capable of emulating up to two hard drives.

1. Would it be possible to add support for up to 4 IDE hard drives (secondary master drive E: and secondary slave drive F:) for the Award SIS 496/497 and Award 430VX PCI-based BIOSes to use?
2. Would it also be possible to implement support for CD-ROM ISOs in the hard disk configuration. That way, I won't have to use the host DVD-ROM drive to insert CDs.

If it is possible, please let me know. I can try to look into the source code of PCem and figure it out.

I think this log provides enough information why NT fails to start the text-based installer (NT 3.1 additionally requires read_capacity in IDE/ATAPI)

1IRET NC DPL  000B   FFFF 0000 9A00 00CF

The log above happens when NT 3.5x is about to start the text-based installer, but throws a black screen. The same log happens when NT 3.1 attempts to get to the GUI installer or to the desktop.

May also be why OS/2 2.0 without Service Paks fails to start too.

EDIT: got a fix, soon posting it 😉

BUG: After opcode 0x3A 0xD8, the instruction pointer is incorrectly increased by only 1 byte, therefore ending up at the 0xD8, rather than correctly by 2 bytes.

In fact, as a general request to everyone, could bug reports please come with as much detail as possible?