WinUAE 3.6.1 beta series - Page 4

Retro-Nerd · 11 February 2018, 20:33

Alt+Tab is broken for me since 3.6.1 beta 1. Alt+Tab works first, but when you try to return to WinUAE it crashes the emulator. No crash dump though inside my folder (portable ticked). Are logs enough?

edit: Argh, it's a shader issue again. Without a guest CRT shader WinUAE doesn't crash.

Toni Wilen · 11 February 2018, 20:44

Quote:

Originally Posted by Retro-Nerd

Alt+Tab is broken for me since 3.6.1 beta 1. Alt+Tab works first, but when you try to return to WinUAE it crashes the emulator. No crash dump though inside my folder (portable ticked). Are logs enough?

edit: Argh, it's a shader issue again. Without a guest CRT shader WinUAE doesn't crash.

Include your config file + all shaders/extra files used. It shouldn't cause crash.

Also log with -log -logflush command line parameters (without it last few lines will be lost)

Retro-Nerd · 11 February 2018, 21:00

Done.

Seems also D3D11 only.

Toni Wilen · 11 February 2018, 21:18

As expected, it works fine here but it probably is driver/gpu manufacturer specific as usually.

Probably not much help but I noticed one D3D11 debug mode warning which is now fixed in winuae.7z.

Does windowed mode resize also cause same crash?

Retro-Nerd · 11 February 2018, 21:29

The debug mode fix doesn't help with the crash. Just tried resize in windowed mode. This works fine though. No crashes.

AMIGASYSTEM · 13 February 2018, 15:10

Toni I found on the beta 3.6.1 a problem in the assembly of the SD, with 3.6.0 assembly is immediate, 3.6.1 and very slow, to understand watch the attached videos.

Toni Wilen · 13 February 2018, 19:31

Quote:

Originally Posted by AMIGASYSTEM

Toni I found on the beta 3.6.1 a problem in the assembly of the SD, with 3.6.0 assembly is immediate, 3.6.1 and very slow, to understand watch the attached videos.

Does winuae.7z fix it? (and config and log is required if not!)

White · 13 February 2018, 20:28

Hi Toni, I'm sorry
I have seen that sometimes the mickjT-mplayer nonealtivec when it switches to full screen
It remains all black.
With the RTG scale if smaller than display size setting the problem is solved.
But sometimes this also does not seem to work, but it happens very rarely.
Here is the dump and my configuration.
I can not tell if this is due to the player or to winuae but this also happens with emotion player.
This problem happens with WinUAE 3.6.0 final and with WinUAE 3.6.1 dated 11/02/2018
Thank you and I'm sorry if the report is not part of the development of WinUAE
Thanks again.

Toni Wilen · 13 February 2018, 20:42

Real question: does it happen with pre-3.6.0? (I am quite sure it does!).

White · 13 February 2018, 20:59

Toni,
version 3.5 beta 14 seems immune, I have tried 20 times continuously and the screen has never appeared black.
With the RTG option scale if smaller activated.
I can do other tests if it can serve.
However, this rarely occurs with other versions 3.6

AMIGASYSTEM · 13 February 2018, 21:07

Quote:

Originally Posted by Toni Wilen

Does winuae.7z fix it? (and config and log is required if not!)

The problem is also present in the winuae.7z fix

I attach config and log 3.6.0 (OK) and 3.6.1 (Problem)

Toni Wilen · 13 February 2018, 21:10

Quote:

Originally Posted by White

Toni,
version 3.5 beta 14 seems immune, I have tried 20 times continuously and the screen has never appeared black.
With the RTG option scale if smaller activated.
I can do other tests if it can serve.
However, this rarely occurs with other versions 3.6

Attached dump is NOT generated by official beta. winuae.7z is not supported.

White · 13 February 2018, 21:23

Toni, you're right, I'm sorry, I must have used the previous 64bit version,
sorry again.
With the last 32bit beta it works perfectly.
I apologize, thank you

Toni Wilen · 14 February 2018, 16:12

Quote:

Originally Posted by AMIGASYSTEM

Toni I found on the beta 3.6.1 a problem in the assembly of the SD, with 3.6.0 assembly is immediate, 3.6.1 and very slow, to understand watch the attached videos.

Your USB adapter has issues with SCSI direct which is now used to attempt to read IDE identity data, instead of immediately rejecting it, it times out.

I reduced the delay (and do not attempt again if it fails) but there is no way to get rid of it completely, especially because sane hardware just works.

Which adapter and what kind of drive is connected to it?

AMIGASYSTEM · 14 February 2018, 17:47

Quote:

Originally Posted by Toni Wilen

I reduced the delay (and do not attempt again if it fails) but there is no way to get rid of it completely, especially because sane hardware just works.

Ok Toni with the last fix has improved now is better, although in all previous versions this problem did not exist

Quote:

Which adapter and what kind of drive is connected to it?

Is a Pendrive adapter (USB / SD)

Toni Wilen · 14 February 2018, 18:07

Of course it didn't exists previously because it didn't need that info!

I don't generally support crap hardware (that does not work like spec says) but click identify-button in add harddrive panel and include VID and PID values so that I can blacklist it.

AMIGASYSTEM · 14 February 2018, 18:40

Quote:

Originally Posted by Toni Wilen

but click identify-button in add harddrive panel and include VID and PID values so that I can blacklist it.

BusType: 0x07
Vendor: 'Mass'
Product: 'Storage Device'
Revision: ''
Serial: ''
Size: 1981808640
VID: 14cd PID: 125c

INQUIRY:
008000011F0000004D6173732020202053746F72616765204465766963652020 ........Mass Storage Device
01000001 ....

READ CAPACITY:
003B0FFF00000200 .;......

INQUIRY 1F:
008000011F0000004D6173732020202053746F72616765204465766963652020 ........Mass Storage Device
01000001 ....

MODE SENSE:
00000000 ....

I hope I understand your request

if you need more information from the device, I attach the results of Chip Genius

Toni Wilen · 14 February 2018, 19:34

http://www.winuae.net/files/b/winuae_3610b4.7z
http://www.winuae.net/files/b/winuae64_3610b4.7z

Beta 4:

- Cirrus Logic emulation blit out of bounds checks added, previously it usually caused a crash, for example OS4 + Picasso IV sometimes does small blits that cross end of VRAM. (Original QEMU blit checks were very broken)
- Yet another interrupt handling rewrite to properly support cross thread interrupt activation (PPC emulation).
- Added more validation to clipboard sharing IFF parsing to prevent crashes if data is truncated.
- RTG multithread + hardware mouse cursor + no OSD: mouse cursor movement didn't re-draw the screen.
- If SCSI ATA passtrough returns weird timeout error (instead of immediate failure like it is supposed to do), don't attempt it again to reduce delay when selecting real harddrive.
- Fixed D3D 11 mode debug mode warning. Most likely it was harmless.

AMIGASYSTEM · 14 February 2018, 20:25

Perfect Toni, with beta4 solved the problem, thanks

Toni Wilen · 15 February 2018, 11:32

Quote:

Originally Posted by White

Toni, you're right, I'm sorry, I must have used the previous 64bit version,
sorry again.
With the last 32bit beta it works perfectly.
I apologize, thank you

Do you mean you still get problems (crash?) but only when using 64-bit version? (using b4)

13 February 2018, 15:10	#66
AMIGASYSTEM Registered User Join Date: Aug 2014 Location: Brindisi (Italy) Age: 70 Posts: 8,248	Toni I found on the beta 3.6.1 a problem in the assembly of the SD, with 3.6.0 assembly is immediate, 3.6.1 and very slow, to understand watch the attached videos. Last edited by AMIGASYSTEM; 16 February 2018 at 00:26.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
WinUAE 2.8.0 beta series (Was 2.7.1)	Toni Wilen	support.WinUAE	214	01 May 2014 11:28
WinUAE 2.4.0 beta series	Toni Wilen	support.WinUAE	342	29 March 2012 09:02
WinUAE 2.3.0 beta series (was 2.2.1)	Toni Wilen	support.WinUAE	229	22 September 2010 19:20
WinUAE 2.0.1 beta series	Toni Wilen	support.WinUAE	14	22 December 2009 18:46
WinUAE 1.5.1 beta series	Toni Wilen	support.WinUAE	242	12 August 2008 12:42

11 February 2018, 21:18	#64
Toni Wilen WinUAE developer Join Date: Aug 2001 Location: Hämeenlinna/Finland Age: 49 Posts: 26,502	As expected, it works fine here but it probably is driver/gpu manufacturer specific as usually. Probably not much help but I noticed one D3D11 debug mode warning which is now fixed in winuae.7z. Does windowed mode resize also cause same crash?

11 February 2018, 21:29	#65
Retro-Nerd Missile Command Champion Join Date: Aug 2005 Location: Germany Age: 52 Posts: 12,435	The debug mode fix doesn't help with the crash. Just tried resize in windowed mode. This works fine though. No crashes.

13 February 2018, 20:42	#69
Toni Wilen WinUAE developer Join Date: Aug 2001 Location: Hämeenlinna/Finland Age: 49 Posts: 26,502	Real question: does it happen with pre-3.6.0? (I am quite sure it does!).

13 February 2018, 20:59	#70
White Registered User Join Date: Sep 2016 Location: italy Posts: 1,480	Toni, version 3.5 beta 14 seems immune, I have tried 20 times continuously and the screen has never appeared black. With the RTG option scale if smaller activated. I can do other tests if it can serve. However, this rarely occurs with other versions 3.6

13 February 2018, 21:23	#73
White Registered User Join Date: Sep 2016 Location: italy Posts: 1,480	Toni, you're right, I'm sorry, I must have used the previous 64bit version, sorry again. With the last 32bit beta it works perfectly. I apologize, thank you

14 February 2018, 18:07	#76
Toni Wilen WinUAE developer Join Date: Aug 2001 Location: Hämeenlinna/Finland Age: 49 Posts: 26,502	Of course it didn't exists previously because it didn't need that info! I don't generally support crap hardware (that does not work like spec says) but click identify-button in add harddrive panel and include VID and PID values so that I can blacklist it.

14 February 2018, 19:34	#78
Toni Wilen WinUAE developer Join Date: Aug 2001 Location: Hämeenlinna/Finland Age: 49 Posts: 26,502	http://www.winuae.net/files/b/winuae_3610b4.7z http://www.winuae.net/files/b/winuae64_3610b4.7z Beta 4: - Cirrus Logic emulation blit out of bounds checks added, previously it usually caused a crash, for example OS4 + Picasso IV sometimes does small blits that cross end of VRAM. (Original QEMU blit checks were very broken) - Yet another interrupt handling rewrite to properly support cross thread interrupt activation (PPC emulation). - Added more validation to clipboard sharing IFF parsing to prevent crashes if data is truncated. - RTG multithread + hardware mouse cursor + no OSD: mouse cursor movement didn't re-draw the screen. - If SCSI ATA passtrough returns weird timeout error (instead of immediate failure like it is supposed to do), don't attempt it again to reduce delay when selecting real harddrive. - Fixed D3D 11 mode debug mode warning. Most likely it was harmless.

14 February 2018, 20:25	#79
AMIGASYSTEM Registered User Join Date: Aug 2014 Location: Brindisi (Italy) Age: 70 Posts: 8,248	Perfect Toni, with beta4 solved the problem, thanks

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)