[BUG] Windows 95 desktop corruption with S3 Trio64

Support and general discussion.
Post Reply
Katakis
Posts: 118
Joined: Thu 05 Mar, 2015 6:42 am

[BUG] Windows 95 desktop corruption with S3 Trio64

Post by Katakis »

Sometimes when I boot up Windows 95 with the S3 Trio64 drivers installed, the mouse cursor is replaced by a large black square that can be moved around, the task bar is black except for the clock on the right, and no icons are visible at all. Restarting seems to have the same effect, but exiting PCem and re-running it fixes the issue.

This is a bug, otherwise Win95 complains that the display adapter is not configured properly.

https://youtu.be/gWVvLCAjWO8
SA1988
Posts: 274
Joined: Wed 30 Apr, 2014 9:38 am

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by SA1988 »

What drivers are you using? What machine? I don't encounter this issue, even with built-in drivers.
User avatar
omarsis81
Posts: 945
Joined: Thu 17 Dec, 2015 6:20 pm

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by omarsis81 »

SA1988 wrote:What drivers are you using? What machine? I don't encounter this issue, even with built-in drivers.
It is a Pentium 120 with 16MB RAM on a 430VX, it can be seen in the video...
SA1988
Posts: 274
Joined: Wed 30 Apr, 2014 9:38 am

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by SA1988 »

Ok then, why don't you use the Phoenix S3 Trio64 with 4MB of memory?
User avatar
SarahWalker
Site Admin
Posts: 2054
Joined: Thu 24 Apr, 2014 4:18 pm

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by SarahWalker »

Which drivers?
Katakis
Posts: 118
Joined: Thu 05 Mar, 2015 6:42 am

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by Katakis »

SA1988 wrote:Ok then, why don't you use the Phoenix S3 Trio64 with 4MB of memory?
Because I don't have the Phoenix ones. As can be seen from the video, the config panel shows that I am using the Number 9FX version.
SarahWalker wrote:Which drivers?
http://www.video-drivers.com/drivers/1594/1594343.htm
As per screenshot, the exact display adapter is S3 Inc. Trio 32/64 PCI

Image
Battler
Posts: 793
Joined: Sun 06 Jul, 2014 7:05 pm

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by Battler »

- Katakis: You can find the BIOS for the Phoenix S3 Trio64 easily by Googling. :p

But in any case, the 9FX Vision emulation is not really complete from what I know as it seems to differ in ways from a stock Trio64.
Katakis
Posts: 118
Joined: Thu 05 Mar, 2015 6:42 am

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by Katakis »

Battler wrote:- Katakis: You can find the BIOS for the Phoenix S3 Trio64 easily by Googling. :p
And to use this BIOS, can I use the official version from the downloads page, or do I have to compile the source to get the development build?
Battler
Posts: 793
Joined: Sun 06 Jul, 2014 7:05 pm

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by Battler »

You can use the official version, the Phoenix S3 Trio64 was added, along side the Trio32, per a patch from me, in late 2014. :p
Katakis
Posts: 118
Joined: Thu 05 Mar, 2015 6:42 am

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by Katakis »

Thanks!

Exactly what causes the problems that are shown in the video?
Battler
Posts: 793
Joined: Sun 06 Jul, 2014 7:05 pm

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by Battler »

The 9FX Vision driver might be using blitter features that are not implemented. It's by far not the only thing that doesn't work right - OS/2 has graphics glitches, Windows NT 3.51 caused the video BIOS(!) to be written to the VRAM when I ALT+ENTER the console from window to full screen mode so long as I'm running any DOS program (QBASIC, Norton Commander, etc.), Apple Rhapsody has font glitches, and the Linux Trio64 refuses to accelerate using the card at all.

The generic (non-9FX) drivers that come Windows 9x work, though.

I guess we have to wait for the S3 Trio64 emulation to be complete.
ecksemmess
Posts: 183
Joined: Wed 18 Mar, 2015 5:27 am

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by ecksemmess »

Battler wrote:I guess we have to wait for the S3 Trio64 emulation to be complete.
I'm a bit confused about this comment. I understand that the Number Nine Trio64 emulation is incomplete in various ways, particularly insofar as that board does differ in some ways from a standard Trio64, but what about the Phoenix Trio32/64 emulation that you added? Is that incomplete too? If so, can you summarize what's missing from it?

EDIT: Perhaps I've misunderstood--did you in fact add your own separate emulation for the Phoenix Trio32/64, or did you just patch in the ability to use those BIOSes with Sarah's original core Trio64 emulation? If the latter, are you saying that the whole long list of Trio64 bugs/glitches you just mentioned affects BOTH the Number Nine and Phoenix Trio64 emulation, because those problems stem from inadequacies in the "shared" Trio64 emulation that powers both?
Battler
Posts: 793
Joined: Sun 06 Jul, 2014 7:05 pm

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by Battler »

The whole long list of Trio64 bugs/glitches I just mentioned, affects not only the Number Nine and Phoenix Trio64 emulation, but also the Phoenix Trio32 emulation and the Paradise Bahamas 64 (S3 Vision864) emulation, which is no surprise since all of them use the same emulation core, which is Sarah's.
User avatar
SarahWalker
Site Admin
Posts: 2054
Joined: Thu 24 Apr, 2014 4:18 pm

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by SarahWalker »

Were you planning on mentioning any of these bugs to me at any point?
User avatar
SarahWalker
Site Admin
Posts: 2054
Joined: Thu 24 Apr, 2014 4:18 pm

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by SarahWalker »

In any case, I can't reproduce the reported bug. It looks to me like the Trio64 memory mapping has totally broken for some reason. Could you provide the config file you're using and the CMOS RAM file (430vx.nvr)? Also, could you boot Windows into safe mode, look at the Trio64 in Device Manager and post what resources it claims to be using?
Battler
Posts: 793
Joined: Sun 06 Jul, 2014 7:05 pm

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by Battler »

- SarahWalker: The OS/2 bug was reported here on this forum two years ago, you said it affects all accelerated cards (which is wrong - we tested the Virge and its drivers do not glitch OS/2) and that was it, and I suspect the Apple Rhapsody font glitches are related to this bug.

The NT 3.51 bug I was going to post about it here, but I was waiting for you to fix the Windows 95 OSR 2 DOS prompt in a final way first.

As for Linux, it seems to refuse to accelerate using any of the emulated accelerated cards, or so I was told, except for surprisingly Alegend45's unfinished Riva.
Katakis
Posts: 118
Joined: Thu 05 Mar, 2015 6:42 am

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by Katakis »

SarahWalker wrote:Were you planning on mentioning any of these bugs to me at any point?
Well, if by "me" you are suggesting that I should report any future bugs by sending you a private message, then I am happy to do that.
SarahWalker wrote:In any case, I can't reproduce the reported bug. It looks to me like the Trio64 memory mapping has totally broken for some reason. Could you provide the config file

Code: Select all

gameblaster = 0
gus = 0
ssi2001 = 0
voodoo = 0
model = 39
cpu_manufacturer = 0
cpu = 4
cpu_use_dynarec = 1
gfxcard = 8
video_speed = 5
sndcard = 8
cpu_speed = 14
has_fpu = 1
slow_video = 1
cache = 3
cga_composite = 0
disc_a = 
disc_b = 
mem_size = 16384
cdrom_drive = 200
cdrom_enabled = 1
cdrom_channel = 2
cdrom_path = C:\Program Files\PCem v12\PCem Files\CD-ROMS\Shivers.iso
vid_resize = 1
vid_api = 1
video_fullscreen_scale = 0
video_fullscreen_first = 1
hdc_sectors = 63
hdc_heads = 16
hdc_cylinders = 2082
hdc_fn = C:\Program Files\PCem v12\PCem Files\1GB - Windows 95 + Drivers (S3Trio + CD-ROM + Mouse + SBAWE32).img
hdd_sectors = 32821164
hdd_heads = 1349784
hdd_cylinders = 32820200
hdd_fn = 
hde_sectors = 1127964904
hde_heads = 28
hde_cylinders = 16
hde_fn = 
hdf_sectors = 65566
hdf_heads = 92
hdf_cylinders = 32820072
hdf_fn = 
drive_a_type = 5
drive_b_type = 0
cpu_waitstates = 0
window_w = 1093
window_h = 802
window_x = 296
window_y = 28
window_remember = 1
joystick_type = 0
mouse_type = 0
enable_sync = 1
midi = 0

[Joysticks]
joystick_0_nr = 0
joystick_1_nr = 0

[Sound Blaster AWE32]
onboard_ram = 2048
SarahWalker wrote:the CMOS RAM file (430vx.nvr)?
done
SarahWalker wrote:Also, could you boot Windows into safe mode, look at the Trio64 in Device Manager and post what resources it claims to be using?
When I click the Resources tab in safe mode, it says "This device isn't using any resources because it is not currently enabled or has a problem."

However, if I click Set Configuration Manually…, it gives me the following resource types:

Memory Range 000C0000 - 000C7FFF
Memory Range 000A0000 - 000AFFFF
Memory Range 000B0000 - 000BFFFF
Input/Output Range 03B0 - 03BB
Input/Output Range 03C0 - 03DF
Attachments
430vx.nvr
(128 Bytes) Downloaded 300 times
Battler
Posts: 793
Joined: Sun 06 Jul, 2014 7:05 pm

Re: [BUG] Windows 95 desktop corruption with S3 Trio64

Post by Battler »

- Katakis: She was asking that question to me, asking me why I haven't reported any of those bugs on this forum. :p
Post Reply