I could probably have a look at PS/1, doesn't look too difficult.

Sarah, I know the IBM PCjr emulation is preliminary but is it possible to add the 2 cartslots to support the cartridges?
Also some older IBM models have an interface to support the tape like the 5150 or PCjr.

SarahWalker wrote:

I could probably have a look at PS/1, doesn't look too difficult.

In this thread of a spanish site they discuss about PS/1. They also have a dump of PS/1 2011 and 2121 BIOS.

Perhaps the most interesting thing of earlier PS/1 models is that they had a DOS in ROM that booted into a four part screen that let you work with some preinstalled software (BTW only the DOS was stored at ROM, the utilities were stored at HDD).

Also, that computers had their own sound card, as stated on a previous post.

My first PC was an IBM PS/1 with that start screen. Wasn't Prodigy one of the options? I'd love to see that again.

It had a semi-integrated12" VGA monitor with a horrible dot pitch IIRC, something like .39

Damn, any good datasheets or documentations about the NE2000 ISA?

Direct3D in PCemFX doesn't work with MDA, CGA and other pre-VGA adapters, don't know if that's a PCem problem therefore it probably is too immature to include shaders in the main tree.

SA1988 wrote:

Damn, any good datasheets or documentations about the NE2000 ISA?

Not sure entirely which chip you'd need. I haven't seen any photos of an NE2000 good enough to make out all the chip markings, but it looks like the DP8390 is a good place to start, eg http://www.alldatasheet.com/datasheet-pdf/pdf … NSC/DP8390.html.

Well, I got the S3 Trio64 working in all color depths I tested in 800x600 in Windows 3.1. Mainly just some MMIO issues. Actually, just some MMIO issues.

http://pastebin.com/WLzbYScQ

PCemFX is definitely cool, but not ready for prime time -- from what I've seen, there were various bugs and instabilities introduced in the process of patching in the shaders. I want to see shaders in PCem more than anyone, but it's not worth compromising the core emulation quality to do it.

Another vote here for PS/1 support, if it wouldn't be too difficult. That would be really cool to see!

And, with regard to the NE2000, the only documentation anyone should ever need is the readily-available source code to MESS's fantastic implementation of that card. 😀 Not that I'm suggesting actually ripping off that code, but anything that needs to be learned can easily be learned by looking at it.

Alegend45 - great work fixing the S3 Trio MMIO! Hopefully Sarah can patch that into SVN soon?

Btw, Sarah - excellent work fixing the issues with 1.2MB disks and the cascade interrupt! Civ is a must 😀 I've got a good set of patches that will be ready in a couple of weeks.

ecksemmess wrote:

And, with regard to the NE2000, the only documentation anyone should ever need is the readily-available source code to MESS's fantastic implementation of that card. 😀 Not that I'm suggesting actually ripping off that code, but anything that needs to be learned can easily be learned by looking at it.

I'd recommend being extremely cautious here - the MAME license is currently incompatible with the GPL license PCem uses.

Trio64 people : what exact drivers have you been using?

SarahWalker wrote:

I'd recommend being extremely cautious here - the MAME license is currently incompatible with the GPL license PCem uses.

Agreed, of course. Let me more forcefully clarify what I said, to anyone considering working on this: DO NOT ACTUALLY TAKE CODE FROM MESS. However, learning by simply studying existing code is certainly not a license violation, so if there are gaps in the available documentation, they should be easy enough to fill by checking and seeing what MESS is doing.

Alegend45 wrote:

Well, I got the S3 Trio64 working in all color depths I tested in 800x600 in Windows 3.1. Mainly just some MMIO issues. Actually, just some MMIO issues.

http://pastebin.com/WLzbYScQ

is it possible to port the NE2000 ethernet ISA card from DOSBox to PCem? Because Sarah actually ported the OPL3 mode from DOSBox to PCem, so, it should be possible.

SarahWalker wrote:

Trio64 people : what exact drivers have you been using?

I think I used this: http://www.vogonsdrivers.com/getfile.php?fileid=275

ecksemmess wrote:

Agreed, of course. Let me more forcefully clarify what I said, to anyone considering working on this: DO NOT ACTUALLY TAKE CODE FROM MESS. However, learning by simply studying existing code is certainly not a license violation, so if there are gaps in the available documentation, they should be easy enough to fill by checking and seeing what MESS is doing.

Yeah, I had to study the MESS code on the Trio64. I figured they already had 3.1 working on the S3, I just had to take a look at the code to figure out those weird 81xx accesses because VGADOC doesn't seem to document them, at least as far as I could tell.

I also tried a crack at the ViRGE, but no such luck there. It just seems to update the MMIO mapping, but not actually make any accesses, and just futz with the hardware cursor. Bizarre, I tell you.

Alegend45 wrote:

I think I used this: http://www.vogonsdrivers.com/getfile.php?fileid=275 […]
Show full quote

SarahWalker wrote:

Trio64 people : what exact drivers have you been using?

I think I used this: http://www.vogonsdrivers.com/getfile.php?fileid=275

ecksemmess wrote:

Agreed, of course. Let me more forcefully clarify what I said, to anyone considering working on this: DO NOT ACTUALLY TAKE CODE FROM MESS. However, learning by simply studying existing code is certainly not a license violation, so if there are gaps in the available documentation, they should be easy enough to fill by checking and seeing what MESS is doing.

Yeah, I had to study the MESS code on the Trio64. I figured they already had 3.1 working on the S3, I just had to take a look at the code to figure out those weird 81xx accesses because VGADOC doesn't seem to document them, at least as far as I could tell.

I also tried a crack at the ViRGE, but no such luck there. It just seems to update the MMIO mapping, but not actually make any accesses, and just futz with the hardware cursor. Bizarre, I tell you.

about the s3 (either trio64 or vision864), I get red colors with Microsoft Write opened on Windows 3.1 in 256 color mode, and being black (still on Write) in modes higher than 256.

I'll have to look into that. I just tested the basic GUI.

I can't quite figure it out. The only register with a weird value is the foreground color register, which seems to be configured for 16bpp pixels...

EDIT: The offending portion of pclog.txt

1Write S3 accel l 000AA6E8 02580320
2Accel out A6E8 20
3Accel out A6E9 03
4Accel out A6EA 58
5Accel out A6EB 02

SarahWalker wrote:

raidenii wrote:

DOS6.22, 430VX+ET4000 W32P+SB16+16M RAM

Still can't reproduce this I'm afraid. Might be a driver issue - maybe try with Windows configured for VGA and no sound and see if it works then?

If just with VGA then Windows loads without any problems.

Alegend45 wrote:

I think I used this: http://www.vogonsdrivers.com/getfile.php?fileid=275

That's Trio64V+ (86c765). PCem has only ever claimed to emulate the original 86c764 Trio64. Is anyone able to confirm that the V+ drivers work with the original chip? I don't have a Trio64, and I'd prefer to have this confirmed before taking these changes in.

SarahWalker wrote:

Alegend45 wrote:

I think I used this: http://www.vogonsdrivers.com/getfile.php?fileid=275

That's Trio64V+ (86c765). PCem has only ever claimed to emulate the original 86c764 Trio64. Is anyone able to confirm that the V+ drivers work with the original chip? I don't have a Trio64, and I'd prefer to have this confirmed before taking these changes in.

they work fine here, except that foreground issue I have with Microsoft Write in Win3.1 in color modes higher than 16.