WinUAE 3.4.0 beta series

[arti]

Try to uncapture mouse before showing crash report dialog.

I can now click on crash report dialog but nothing happens if I press Close button.

[Dr.Venom]

Originally Posted by Toni Wilen

I don't see any difference between 5 and 6. Both have same problem, it is just
random if it starts quickly or later.

I've gone back and retested 3.0.0, 3.1.0, 3.2.0, 3.3.0, 3.4.0beta1 to beta 5 and they are all OK. The problem is introduced with beta 6. Sorry can't make anything else from it.

I did make a new config/testcase. This makes use of Lionheart intro/options screen, as it is static interlaced, and will not switch back to progressive, such that there is enough time for proper testing. It's also done with the Lionheart HDF for some quicker loading (upped by ransom in "ftp://grandis.nu//~Uploads/ransom").

Please see also attached logs from both beta 5 and 6. I tried various times going in and out the GUI, but you can see from the logs that beta5 has no issue getting the right field synchronization, while beta 6 goes into the infinite field type (mis)matching loop every time.

You should be able to use the attached config with minimal adjustment (a500i.uae), only change the display to your interlaced screen, and get the lionheart HDF from above link. When Lionheart loads up to the intro screen (scrolling background) press fire and you'll be in the interlaced options screen.

Hope this helps uncovering the interlace frame type mismatch behaviour in >=beta6 releases.

Edit: All tests done were done with the 64-bit versions, in case that could make any difference.

[dirkies]

interesting bug here:
just run x64 version of beta12, and browse around a bit in the settings (rescan roms, save config, select adf, etc)... winaue just crashes for no reason after a minute or 2 (?). you do NOT have to actually start emulation, just the settings screens of WinUAE suffice to break it.

Or just open the adf select window and leave it stay open for a while, same crash.
Seems it will simply crash whatever you do

Using Win2012R2.

Always EXACT same crash message, no dumpfile created:

Problem signature:
  Problem Event Name:	APPCRASH
  Application Name:	winuae64.exe
  Application Version:	3.4.0.0
  Application Timestamp:	58026562
  Fault Module Name:	StackHash_c6ea
  Fault Module Version:	6.3.9600.18438
  Fault Module Timestamp:	57ae642e
  Exception Code:	c0000374
  Exception Offset:	PCH_E5_FROM_ntdll+0x0000000000090C6A
  OS Version:	6.3.9600.2.0.0.272.7
  Locale ID:	1033
  Additional Information 1:	c6ea
  Additional Information 2:	c6ea0d23c6d913ee3697ef85ce3f7f60
  Additional Information 3:	a230
  Additional Information 4:	a230d60bbaa73edaa3a619043caae73b

Does NOT happen with x86 version, only the x64 version seem to have an issue with what I think is the drive tree in the OS (because it contains a lot of mapped network drives?)

At least it is better than previous beta's, there WinUAE would crash asa the file selection window was opened.

[Toni Wilen]

100% not 3.4 beta related. You have buggy shell extension(s). http://eab.abime.net/showthread.php?t=80216

[dirkies]

ok, but have a lot of extensions of well known software here (7zip, Google, Dropbox, etc)
but also some less used ones.

If I need to remove one that I often use, it will be x64 WinUAE that will not be used
anymore instead...

If I find it I will let you know here which one it was.

[mark_k]

Originally Posted by emufan

tried to provoke a crash, no luck so far

Copy the file from the attached archive to plugins/filtershaders/direct3d/ in your WinUAE program directory and rename it to _winuae.fx

WinUAE should crash on starting emulation then.

[emufan]

good work mark - it does crash
mouse pointer is untraped, but i cannot close crash window, nor the winuae-window.
you cannot activate the little crash window - no effect on pushing [ ok ] button in the crash window.

[Toni Wilen]

I'll add TerminateProcess() call after the crash dump dialog. It will guarantee nothing will remain
(But if the dialog won't close, there is not much to do, and I want to keep the dialog because no one will notice the crash dump files without it)

[emufan]

hmm, did you fix mark's crash test dummy? it does not crash anymore.
i wait for another big bang and report back.
the diallog itself is not a problem at all. better than a silent bye-bye.

#1) now it does crash again (mark's dummy) - rather strange.
crash window now "active", but "close" is still not working (win7 x64).

==> the untraped mouse is ok for me, so one may open the taskmanager more easily.

[Toni Wilen]

http://www.winuae.net/files/b/winuae_3400b13.7z
http://www.winuae.net/files/b/winuae64_3400b13.7z

Beta 13:

- Fixed double free in RTL8019 9346 EEPROM emulation. Caused random crashes.
- Implemented DP8390 NIC support (Same register set as NE2000 but without on board RAM "fake DMA" support)
- Added DP8390 based NICs, ASDG LAN Rover/EB920 and Hydra AmigaNET.
- Moved slirp/winpcap option to Expansions -> Network Adapters. Note: Still can't have more than 1 network device active.
- Removed A2065, NE2000 PCMCIA and Realtek 8029 PCI GUI checkboxes. Everything is now under Network Adapters.
- Expansion device select menu sorting fixed.
- winpcap mode now by default uses generated local MAC address (aa-82-8a-xx-xx-xx xx=from physical NIC MAC), or if A2065, it uses real A2065 00-80-10-xx-xx-xx MAC. This fixes conflicts with Windows' use of same physical NIC. Custom winpcap MAC address can be forced by editing config file (*_rom_options=mac=xx.xx.xx.xx.xx.xx). GUI is not yet supported. Physical NIC is also switched to promiscous mode automatically if NIC MAC does not match Amiga side MAC.
- Ignore all received winpcap packets when emulation is paused.
- FFS partition HDF (with WB 1.3 disk L:FastFileSystem) now fully boots under KS 1.0.

WinPCap note: Use http://www.win10pcap.org/ if you have Windows 10. NOTE2: It has a bug that causes "FATAL Bad Memory Block." errors to appear in winuae's log. It is fortunately harmless. https://nmap.org/npcap/ may also work. (npcap DLL path locations are now supported)

NOTE: Only 3.3 (and older) config files with old style A2065 or RTL8029 PCI config entries are loaded. 3.4 beta created config files need manual A2065/RTL8029/NE2000 PCMCIA reconfiguration. Old config entries are still created for backwards compatibility but they are ignored when loading.

[honx]

Originally Posted by Toni Wilen

Expect lots of breakage.

will this also apply to 3.4.0 final version? should i start to worry?

[Toni Wilen]

Originally Posted by honx

will this also apply to 3.4.0 final version? should i start to worry?

"It depends"

Seriously, whole point of betas is to test and find the problems before final version.

[Toni Wilen]

http://www.winuae.net/files/b/winuae_3400b14.7z
http://www.winuae.net/files/b/winuae64_3400b14.7z

Beta 14:

- Ariadne network adapter emulation added. (Am79c960 based, improved Am79c90)
- uaesnd update, sample format byte changed, added support for little-endian and signed samples.
- Old NE2000 config entries overrode new style entries until config file was rewritten at least once.
- Sound output frequency max limit increased from 96KHz to 192KHz. (Useless but if you really want to do it..)
- Added missing drive type/feature level options to add harddrive dialog.
- Opalvision updates, OpalPaint palette is now fully visible, interlace mode missing lines fixed. King of Karate confirmed working. PlayAnim24 (which uses sort of polarity inverted control line) works.

[Pheonix]

The latest betas have been corrupting ADF files when they are written to. Included below are a log & a corrupted ADF file. I was working with Beta 13 when I discovered it, and have since upgraded to Beta 14 to confirm that it is still occurring.

It could very easily have something to do with the ADF create tool, as it seems to stop happening if I do a format on the created ADF prior to copying files to it. Those could have been extremely lucky coincidences, though, as I only tried it 3 times.

If I copy enough files to fill most of the disk (~75%) it has always happened so far, with a freshly created ADF file.

[Toni Wilen]

Disk creation cleanup few betas ago broke disk bitmap creation, it thought adf was hd image which caused root block not being marked as allocated..

[amilo3438]

Originally Posted by Toni Wilen

- Compiled again with pre-release MSVC compiler. Does speed now match 3.3.0? (Few reported that 3.4 betas have slower performance)

Unfortunately no, the speed problem is still there in beta6 !

I have also compared WinUAE 3300 with WinUAE 3400b14 on Jim Power title screen.
Find used config + logs + screen shoots attached below.

[Toni Wilen]

Same CPU% values?

(Unfortunately this is a bit late now and unless someone else reports similar slowdown, nothing will happen. It is too difficult to narrow it down to something without reports from different PC configs. It even may be as simple as some loop missing the cache by few bytes when previously it fit perfectly in CPU's cache)

[amilo3438]

Originally Posted by Toni Wilen

Same CPU% values?

(Unfortunately this is a bit late now and unless someone else reports similar slowdown, nothing will happen. It is too difficult to narrow it down to something without reports from different PC configs. It even may be as simple as some loop missing the cache by few bytes when previously it fit perfectly in CPU's cache)

CPU% values was shown on attached pictures, so..

WinUAE3300 -> CPU usage = 96% on Jim Power title screen (Edit: CPU usage = 90% when Sound = Disabled, but emulated)
WinUAE3400b14 -> CPU usage = 120% on Jim Power title screen (Edit: CPU usage = 110% when Sound = Disabled, but emulated)

My PC CPU is Pentium M 725 (1.6GHz FSB 400MHz o/c to 2.13 GHz FSB 533MHz) codenamed "Dothan" with cache 2MB L2 !!! (so I guess the cashe should not be a problem)

p.s.
I am already fine with WinUAE3300, but wonder why new versions is over 20% slower in performance.
Also, I still have an older test version WinUAE 3.3.1 (Public Beta 0, 2016.06.19) that works fine !?

[Toni Wilen]

http://www.winuae.net/files/b/winuae2.7z has 3 executables. Does any of them run faster or slower?

[amilo3438]

Originally Posted by Toni Wilen

http://www.winuae.net/files/b/winuae2.7z has 3 executables. Does any of them run faster or slower?

With the same config and Jim Power title screen CPU usage:

winuae.exe = 96%
winuae_14_pre.exe = 120%
winuae_14.exe = 122%
winuae_15.exe = 122%

(note that the % is taken around most stable value)

p.s.
Btw. regarding previous WinUAE3300 and WinUAE3400b14 test:
By default the nVidia PowerMizer Control was on "Balance", but when is set on "Maximum performance"
then on WinUAE3300 CPU usage = 93% and on WinUAE3400b14 CPU usage =115%
(CPU Speed settings was on "Max. Performance")

Edit p.s.:
When change Graphics API from D3D to DirectDraw got on WinUAE3300 CPU usage = 90% and on WinUAE3400b14 CPU usage =113% !
(no matter of nVidia PowerMizer Control if "Balance" or "Maximum performance").