WinUAE 4.9.2+ beta series

[Toni Wilen]

Originally Posted by thomas

CyberVision64 Z3 shows as defective on the Expansion Board Diagnostic screen, both CV64 and CV64/3D.

Fixed. Got re-broken when another fix was added.

[White]

Toni,
OS4

Currently all Video Player applications (all also locally)
Games etc.
As soon as the UAEGFX "graphics card" or PicassoIV "winuae" is invoked it completely freezes.

I am using "windowed" so I can kill winua and else I should always reboot the PC completely.

ok add all applications as amigaamp
tunet etc.
:-)

[White]

Toni,
I'm sorry but I do the tests quickly
now everything works.
Basically the sound was disabled in the "winuae" options
And this bore all the problems listed in the fix you released

here the configuration that now works:


strangely usually if you turn off the audio winuae shouldn't crash but you just can't hear the audio.

besides the fact that if you use the full screen it leads to the complete block of windows 10 at least with OS4

[AMIGASYSTEM]

Toni, AROS 68k still does not work, only black screen.

SDL on OS3 also does not work, executables go in a loop

[Viceroy]

Originally Posted by Toni Wilen

More details needed. Does it only happen when using your CWB config? If it does, enable logging and attach/email it. (winuaelog.txt)

I'm away from computer for a couple of days I'll get on it as soon as I get back. Cheers Toni

[White]

Toni,
here a setup for the 68k (usually works with everything as a base)
with audio disabled on purpose.

In this case "winuae" does not block and windows 10 is not blocked as it happens instead for OS4 by completely turning off the sound.

[Toni Wilen]

Originally Posted by White

Toni,
here a setup for the 68k (usually works with everything as a base)
with audio disabled on purpose.

In this case "winuae" does not block and windows 10 is not blocked as it happens instead for OS4 by completely turning off the sound.

Is this really 4.9.2 beta bug? If audio emulation is disabled, there is no audio interrupts and program will hang if it expects audio interrupts.

[White]

Toni,

I am currently unable to reproduce exactly what is going on.
But I can cause the crash on purpose.
This also applies to the latest Beta with the "resync" fix

But the steps are these that lead to the crash:

1) launch winuae purposely disable audio in winuae
2) I completely quit winuae
3) I fully relaunch winuae and load the configuration where the audio is selected correctly
4) winuae does not load the audio correctly as it is in the configuration ( should it also load the audio as per configuration ? )
5) winuae crashes in "windowed" mode
6) if you are going to use the full-screen mode windows 10 would crash completely.

The steps are these
If you think this is useful I can do exactly the steps that cause it to crash more precisely.
If, on the other hand, this is foreseen by the "winuae" features then it is not a bug.

[White]

Toni,
here a video just simply download it
I did it as fast as possible
the crash is seen at the end.
ignore the board error because the crash occurs even without it.
If you need something more specific I can do it

(( clearly this happens without using video capture ))

https://we.tl/t-ixW05RiX8T

[Toni Wilen]

It still sounds like nothing to do with this beta. Standard requirement is to ALWAYS confirm with previous official version first. And if it happens with old version: this is wrong thread.

And as usual: videos are mostly useless. Logs and config and the usual needed.

[White]

Toni,
I'll try to get you a .log about this
And surely you can get a more complete picture
But as in his time I had said the only version "immune" to this crash is the winuae 3.6.0 version.

Now I will test the latest official version released. (4.9.1)

But I repeat only winuae version 3.6.0 is immune to this crash
obviously it is clear that there are more advantages in the versions you developed later.

And even without debugging which I obviously can't do
you can perceive with the naked eye the speed of execution of OS4 with winuae 3.6.0
in comparison to the versions released later.

To give an example, opening the OS4 windows while the audio is running is practically instantaneous.
While with the versions released later of winuae a slowdown is perceived to the naked eye.

Clearly winuae is developed for 68k I will do some tests also on 68k maybe the same happens.
But the 68k is so fast that it is probably not perceived.
But I will try to crash with 68k as well.

A little note on windows 10
I use the 1809 ltsc version because I like it

I used the win 10 PRO latest build
And also the win 10 ltsc 2021 version with 21H2 (becomes a brick)

because I have the licenses of these systems
but I went back to 1809 for me the best.

The winuae crash problem still occurred with these versions as well.
I say this because at the time I was told that the 1809 was buggy as a version :-)
Not from you but from other people, now I don't remember who :-)

the log you need is winuaelog.txt?
Thank you.

[AMIGASYSTEM]

Thanks Toni, with your latest fix today, AROS One 68k works perfectly, I would say much faster and smoother graphics, also SDL works well, see screnshot (NetSurf SDL)

[ceztko]

Originally Posted by Toni Wilen

Time to finally start next round of betas..
Later ones might get version bump.

In the previous thread:

Originally Posted by coldacid

May I suggest version 4.10.0?

If you are going to stay sub 5.0, I also second this request. 4.9.2 is confusing to me.

[Viceroy]

Originally Posted by Toni Wilen

More details needed. Does it only happen when using your CWB config? If it does, enable logging and attach/email it. (winuaelog.txt)

I've just tried this again after 5 days away from PC, it now works fine? Sorry for wasting your time Toni ;-)

[Toni Wilen]

Originally Posted by ceztko

In the previous thread:



If you are going to stay sub 5.0, I also second this request. 4.9.2 is confusing to me.

Final version won't be 4.9.2. I still haven't decided what it is but probably 4.9.9 or 4.9.10.

Originally Posted by Viceroy

I've just tried this again after 5 days away from PC, it now works fine? Sorry for wasting your time Toni ;-)

Keep logging active if it happens again

Originally Posted by AMIGASYSTEM

Thanks Toni, with your latest fix today, AROS One 68k works perfectly, I would say much faster and smoother graphics, also SDL works well, see screnshot (NetSurf SDL)

It was serial port logging that hung because serial port status bits didn't work correctly in all CPU speed modes.

[Ponki1986]

Will the new version work properly on Windows 7 SP1 x86? Do I need to install new Visual C libraries?

[OlafSch]

Originally Posted by Toni Wilen

Final version won't be 4.9.2. I still haven't decided what it is but probably 4.9.9 or 4.9.10.



Keep logging active if it happens again



It was serial port logging that hung because serial port status bits didn't work correctly in all CPU speed modes.

Where do I find current version?

[Toni Wilen]

Originally Posted by Ponki1986

Will the new version work properly on Windows 7 SP1 x86? Do I need to install new Visual C libraries?

Yes. B1 has unnecessary direct3d requirement (GDI mode works without it) which will be fixed in next beta.

[Toni Wilen]

https://download.abime.net/winuae/fi...nuae_4920b2.7z
https://download.abime.net/winuae/fi...ae64_4920b2.7z

Beta 2:

- Display resync blanking was stuck enabled (~1s blank screen when mode switches)
- Internal emulated serial devices (for example laserdisc player) data transfer was unstable.
- Z3 hardware emulated RTG board autoconfig size data was fixed and then was broken again.
- Some NCR710+ SCSI controllers hung.
- Serial port status bits were not correct in all CPU modes.
- Internal CPU instruction emulation table refactoring, use separate table for instructions that return total cycles and instruction that count cycles internally (cycle-exact modes).
- Quickstart panel floppy type dialog was partially clipped if GUI was using smaller default size.
- D3D9 full install was accidentally required (d3dcompiler_47.dll error message).
- Automatic fall back to GDI mode (if both D3D11 and D3D9 is not available) retried forever.

Pre-4.9.2 fixes:

- Box art path detection didn't parse cdimage0 paths correctly if path had "," character. ("," needs special case because path can be optionally followed by ",delay")
- If only one autoconfig device was enabled and it was removed when emulator was running and then system was hard reset: old autoconfig data was not cleared causing disabled device to get misdetected by AmigaOS which usually caused hang.

[Rock'n Roll]

Hi Toni,

i checked a little bit the winuae debugger funcionalaty.
Some point are new and good. Thanks.

I found also some bugs:
1.) Wf <addr> <endaddr> 'string" don't works anymore
2.) the dma debugger output is wrong formated or little bit mess
there is also often a '|' symbol or free lines
3.) Debugger sp command parsed sprites incorrectly
that is incorrect for normal sprites and attached sprites
(Action replay is correct) eg. spr 0 or spr 0 1 (for attached)
WinUAE debugger
sp addr(sprite0) or sp addr(sprite0) addr(sprite1)

for normal sprite: eg mouspointer
43 000C84 1111110000000000 ; 0 --> 0 correct
44 000C88 1333331000000000 ; 1 --> has to be 2
45 000C8C 1222231000000000 ; 2 --> has to be 1
46 000C90 1222310000000000 ; 3 --> 3 correct
47 000C94 1222231000000000

1A0 COLOR16 ; 0 ; binary 00=COLOR 0
1A2 COLOR17 ; 1 ; binary 10=COLOR 1
1A4 COLOR18 ; 2 ; binary 01=COLOR 2
1A6 COLOR19 ; 3 ; binary 11=COLOR 3
...
1BE COLOR31 ; F