Using the Award 430VX PCI & 440FX PCI bios's, I can't get it to properly detect a cd-rom in the drive when booting from it or swapping it under windows guests. This is using both the ISO method and HOST CD-ROM drive method, and I am not using SCSI. It seems that the only scenario I can get to work with it is when using ISO method for trying to boot from a CD-ROM.

I've tested this against PCem v11 and it all works fine. Therefore I'm assuming this is definitely a bug, correct?

Version c669a05 tested.
Dos navigator detects a sound track, but not playing time remains 0:00
PROBLEM IN PIO , PIO + DMA, SCSI.

The Bochs bios rom seems to have a licensing scheme which may allow bundling of 86Box with the rom. This would also allow the developers here to modify the bios for emulation enhancements. Is this currently being considered? However, I don't know whether the two licenses are compatible with commercial distribution, too, although some versions of the GPL license allows this.

Also, along these lines, is it possible to create a branch of 86Box with enhanced portability to other host operating systems, such as FreeBSD? This would require a focus on a single model type, but would remove the necessity of the dialog boxes and dependency on the external windows. Also, the GUI and sound/video/input interfaces may be made generic for simple replacement to the equivalent functions of another host OS. For the current win32 systems, it should be possible to remove the dinput and d3d dependencies. All the video code could be contained in a single file, for instance.

Have you thought of implementing VHD support?
It's something that was briefly hinted on the PCEM-official forums, but then quickly forgotten.

Being VHDs natively mountable under Windows 7 SP1 and later, this feature would come really in handy.

Just replaced a build from early June which was working fine and seeing all my roms with #891 from CI and it is claiming that it can't find any roms.

The wait state selection does not set its values at all when pressing OK for confirmation, they get resetted to "System default" and the status bar gets messed up (e.g. disappearing icons) when asked for a hard reset upon modifying a hardware component.

The attached floppy works on a physical S3 Trio card. Under 86Box it works with a "Diamond Stealth" cards, but with "Number nine" and "Phoenix" S3 cards it crashes the emulator when accessing the frame buffer.

So there is obviously a bug in 86Box, but also it would be good to know why does the test program crash it. Maybe it works on the physical card just accidentally?

The test has to be executed with Award 440 FX, 32 MiB RAM and a 1.44 Floppy.zip. Select "Diamond Stealth" to see how it's supposed to work, select "Phoenix S3 Trio 64" to see it crash.

I updated to 86Box from an older version of PCEM-X which had no joystick support. The application then crashed on me right after start. I figured out it had something to do with the controllers attached to my machine. I unplugged all controllers and after that 86Box came up fine.
Here is the log generated with the controllers attached:

Raw input registered!
executable_name = C:\Users\DOCa Cola\Desktop\PCem\86Box.exe
path = C:\Users\DOCa Cola\Desktop\PCem
Keyboard initialized!
joystick_enum_callback : found joystick 0 : vJoy Device
joystick_enum_callback : found joystick 1 : vJoy Device
joystick_enum_callback : found joystick 2 : Serial/Keyboard/Mouse/Joystick
joystick_enum_callback : found joystick 3 : MFG Crosswind V2
joystick_init: joysticks_present=4
Joystick 0 :
tszInstanceName = vJoy Device
tszProductName = vJoy Device
Axes = 3
Buttons = 12
POVs = 0
Axis 0 : X-Achse 0 2
Axis 1 : Y-Achse 4 102
Axis 2 : Z-Achse 8 202
Button 0 : Taste 0 c 4
Button 1 : Taste 1 d 104
Button 2 : Taste 2 e 204
Button 3 : Taste 3 f 304
Button 4 : Taste 4 10 404
Button 5 : Taste 5 11 504
Button 6 : Taste 6 12 604
Button 7 : Taste 7 13 704
Button 8 : Taste 8 14 804
Button 9 : Taste 9 15 904
Button 10 : Taste 10 16 a04
Button 11 : Taste 11 17 b04
Joystick 1 :
tszInstanceName = vJoy Device
tszProductName = vJoy Device
Axes = 6
Buttons = 0
POVs = 0
Axis 0 : X-Achse 0 2
Axis 1 : Y-Achse 4 102
Axis 2 : X-Rotation 8 302
Axis 3 : Y-Rotation c 402
Axis 4 : Schieberegler 10 202
Axis 5 : Wählscheibe 14 602
Joystick 2 :
tszInstanceName = Serial/Keyboard/Mouse/Joystick
tszProductName = Serial/Keyboard/Mouse/Joystick
Axes = 6
Buttons = 32
POVs = 1
Button 0 : Taste 0 1c 4
Button 1 : Taste 1 1d 104
Button 2 : Taste 2 1e 204
Button 3 : Taste 3 1f 304
Button 4 : Taste 4 20 404
Button 5 : Taste 5 21 504
Button 6 : Taste 6 22 604
Button 7 : Taste 7 23 704
Button 8 : Taste 8 24 804
Button 9 : Taste 9 25 904
Button 10 : Taste 10 26 a04
Button 11 : Taste 11 27 b04
Button 12 : Taste 12 28 c04
Button 13 : Taste 13 29 d04
Button 14 : Taste 14 2a e04
Button 15 : Taste 15 2b f04
Button 16 : Taste 16 2c 1004
Button 17 : Taste 17 2d 1104
Button 18 : Taste 18 2e 1204
Button 19 : Taste 19 2f 1304
Button 20 : Taste 20 30 1404
Button 21 : Taste 21 31 1504
Button 22 : Taste 22 32 1604
Button 23 : Taste 23 33 1704
Button 24 : Taste 24 34 1804
Button 25 : Taste 25 35 1904
Button 26 : Taste 26 36 1a04
Button 27 : Taste 27 37 1b04
Button 28 : Taste 28 38 1c04
Button 29 : Taste 29 39 1d04
Button 30 : Taste 30 3a 1e04
Button 31 : Taste 31 3b 1f04
POV 0 : Mehrwegeschalter 0 10
Axis 0 : Z-Rotation 4 502
Axis 1 : Z-Achse 8 202
Axis 2 : Y-Achse c 102
Axis 3 : X-Achse 10 2
Axis 4 : Schieberegler 14 602
Axis 5 : Schieberegler 18 702
joystick_init : CreateDevice failed
PIC1 : MASK 00 PEND 00 INS 00 VECTOR 00
PIC2 : MASK 00 PEND 00 INS 00 VECTOR 00
Dumping rram.dmp

I was contributed experimental EMS support on LaserXT series to PCem and merged to mainline. In current branch of PCem can use up to 512kb EMS memory when set memory size to 1152kb but 86Box can still set up only 640kb so can't use EMS memory.

I seem to be having an issue when running any dos program within windows 95 osr2. upon opening it will bring emulation speed to between 0% and 5% until the dos program is closed. I was able to run a native windows game and experienced no issue. This issue presented itself with the new CPU optimization commit Build 779.

I have tested on 2 physical machines and multiple virtual machines.
Current setup

Physical machines tested
i7 6700k 4.5ghz
32gb ddr4 3200
X2 R9 390x
Clean install Windows 10 Enterprise.

i7 4770k 4.3ghz
24gb ddr3 1800 ram
X2 780ti
Clean install Windows 10 Enterprise.

Main Test Virtual Machine config
(Windows 95OSR2)

[Machine]
model = 430vx
mem_size = 65536
cpu_use_dynarec = 1
cpu = 18
cpu_enable_fpu = 1

[Video]
gfxcard = mach64vt2
video_speed = 5
voodoo = 1

[Input devices]
mouse_type = ps2

[Sound]
sndcard = sb16
opl3_type = nukedopl

[Hard disks]
hdd_01_parameters = 63, 16, 2080, 0, ide_pio_and_dma
hdd_01_fn = C:\Users\Brad\Desktop\Win95 OSR2 ATi - Copy\Cdrive.img
hdd_01_ide_channel = 0:0

[Removable devices]
fdd_02_type = 35_2hd
cdrom_01_parameters = 1, atapi_pio_only
cdrom_01_ide_channel = 1:0
cdrom_01_host_drive = 74

[3DFX Voodoo Graphics]
type = 2
framebuffer_memory = 4
texture_memory = 4
render_threads = 2
bilinear = 1
dacfilter = 0
recompiler = 1

I heard better OPL3 than original PCem (maybe PCem doesn't use Nuked OPL3), setting Adlib Gold. Some music player in PCem, loud clipping(popping?) noise from my speaker. but 86Box reduced this. Thanks!!

But 86Box has an incorrect OPL3 playing music with SB series and many OPL3 compatible soundcards like PCem now.
Only Adlib Gold plays correct OPL3.

I want correct OPL3 music and clean PCM sounds with SB series.
Let's fix this problem.

(edit) OPL3 in WSS playing correctly too.
PAS16 can't detect Adlib compatible.

I'm using skylake optimized version of 86box, latest dev build (6cab132). Attached is my config file

• Host CPU i7-6700k
• Tested running game on Windows 98 and MS dos 6.22
• both 3dfx and software rendered versions.

86box.zip

In the last couple of builds I wasn't able to load the Rhapsody Developer Release isos to it it would just show that There'S nothing in the emulated cd drive

This video card this with an error in to double the screen. (tested in Win98se)

Compiling 64-bit version does not work in full screen mode directdraw. 32bits version is ok.

I tested about standalone MPU-401, and SB16's MPU-401, with rev#793.
and I found some issues.

1. Standalone MPU-401 checkbox save problem
   Standalone MPU-401 turned to uncheck after quit and restart 86Box.

2. Standalone MPU-401 can execute with another soundcard
   I tested after change soundcard to Windows Sound System, MPU-401 worked.
   Is this bug? What's model name of this MPU-401 interface card? I don't know. but...

I saw an ability of really standalone MPU-401.
I think it will can work with any soundcard, maybe.
(with SB16/AWE32 too, if standalone MPU-401's base address not same with SB16's MPU-401.)
and, this can separate MIDI out device between standalone MPU-401 and SB... (or use same device)

for example...
SB16's MIDI base address : 300 (UART)
SB16's MIDI out device : MIDI Port A
standalone MPU-401 base address : 330
standalone MPU-401 IRQ : 9 (Intelligent)
standalone MPU-401 MIDI out device : MIDI Port B

3. Problem with SB16/AWE32's MPU-401 UART mode
   SB16's MPU-401 Intelligent mode, and SBMIDI works well.
   but UART mode didn't work well.

This started to port by TC1995, I know porting from DOSBox. Thanks to him.
but PCem's SB16 UART mode was right.

I tested two DOS MIDI player, MEGAMID and DOSMID.
and add argument /mpu when run these with midi file.

MEGAMID can't detect MPU-401 with DOSBox and current 86Box, but PCem did.
DOSMID can detect and play MPU-401 with all.

I asked who had real machines and SB16/AWE32.
They told me MIDI plays OK both MEGAMID and DOSMID.
Can revert back for MPU-401 UART mode's?

Thanks!

I was trying to load CD image with Chinese characters in path, but it fails to load which acts like nothing happened. It can be fixed by renaming folders to not containing unicode characters, which leads me to guess that this project currently still lacks proper unicode support. So my question is, will we get unicode support in path eventually?

If you set up 86Box to use a Tyan Titan-Pro AT/Tyan Titan-Pro ATX and proceed to install Windows 95/98/ME, then the keyboard will start to work incorrectly when setup enters the GUI part, since the arrow keys no longer work and Enter starts to act like the Tab key, switching between two (only two) different GUI elements.

When I use 286/386SX PC emulation like IBM AT, 86Box crashes when memory size change. After crashes, memory size shows minimum size in 86Box Setting dialog but always maximum size applied.

And mem_remap_top_384k() function should not be called AT NEAT, AT SCAT, AT Headland chipset initialization in model.c because that chipsets support shadow ram and can alter memory remapping feature via chipset-spectific register.

I'm trying to put this diplomatically but your automake/autoconf system appears to be fundamentally broken. You're calling aclocal/autoconf/automake from the Makefile, when they should be run by autoconf when it generates the configure file, and should not be run at all during the make process, not least because someone might (and indeed, for many purposes, should) not have automake, autoconf, etc installed. Even if you want to go down this route, the fact you're hard-coding in automake-1.11 breaks systems such as mine which use automake-1.15, requiring manual adjusting of the Makefiles. I also don't quite understand what you're doing with depcomp, as retargeting it at automake-1.15's depcomp simply causes it to error out with "No command. Try depcomp --help for more information"

Can we please decide on one code style and all use it? I personally use tabs rather than spaces, braces on separate lines, and braces for nested single line statements too, but I see that they're being actively replaced with a different coding style.
I believe a discussion is needed so that we have one style we all agree on, which would ensure consistency.

The sound is silent could it amplify it or the possibility of manual increase?

Thank you for answer

When installing a cdrom game and moving in the menu, the keyboard stops responding as darts and more. The same thing in the menus in the game until only the mouse is used can then get to play the race. But if you want to climb out of the race, "escape" works but enter no.
It's not the last version I did in all I've ever tried.

This problem is in MSDOS 6.22 testing

hi

when i try to compile 86box i get that error

cdrom_image.cc: In function 'int image_open(uint8_t, wchar_t*)':
cdrom_image.cc:988:31: error: 'wcstombs' was not declared in this scope
wcstombs(afn, fn, sizeof(afn));
^
Makefile.mingw:274: recipe for target 'cdrom_image.o' failed

i have try with gcc 6.3 and 7.1 (that gcc compiled good mame or other project)

thanks in advance for you help

I'm trying to run 86box on a very old machine (a 2007 MacBook to be precise) running Windows 7 x86. While I was able to get mainline PCem running, the minute I try to start 86box, I get a window for about 3 seconds and then before I can interact with anything it crashes with an access violation. I'm looking at other issue posts where people talk about getting a log, but I'm not sure how to get my copy of 86box to generate a log. Do I have to compile it myself in debug mode?

I'm using the 86box build #304 from polar.rol.im

86Box seems to crash a lot when doing network operations, such as downloading a file or accessing smb shares.

Host OS: Windows 10 x64
86Box Version: 2.00 (from your builds)
Guest OS: Windows 98 SE
Host Network card: Atheros Ar81xx
Guest Network card: Novell NE1000 (default settings), Network type SLiRP
Machine: Intel Advanced/ATX, Pentium MMX 233, 256MB memory, S3 ViRGE/DX Fast VLB/PCI, Voodoo 2, Sound Blaster 16 with Nuked OPL, hard disk and cdrom on IDE (PIO+DMA)

According to Windows, the crash is a memory violation, and it's always at the same instruction

Faulting application name: 86Box.exe, version: 2.0.0.0, time stamp: 0x00000000
Faulting module name: 86Box.exe, version: 2.0.0.0, time stamp: 0x00000000
Exception code: 0xc0000005
Fault offset: 0x001b6ac4
Faulting process id: 0x1980
Faulting application start time: 0x01d2ef487911582d
Faulting application path: C:\Users\Federico\Desktop\86box\86Box.exe
Faulting module path: C:\Users\Federico\Desktop\86box\86Box.exe
Report Id: 5e560710-03a0-443a-ac23-c25ae6a825d1
Faulting package full name: 
Faulting package-relative application ID: 

Using WinPcap on both AMD/Intel platforms & both Win7/Win10, I cannot get LAN to work under Windows 95,98 & ME. I've even tried disabling/enabling Network bios. I'm using a wired Ethernet cable that attaches to my host PC with WinPcap installed (I'm NOT trying to get this to work over a wi-fi connection.) Can anyone else confirm this or is anyone able to get this to work with the latest 86Box build? If so, then how?

Hello,
I am using 86Box (64-bit version) on a Windows 10 64-bit laptop, and 486 emulation appears to be working just fine, but for whatever reason, no matter what Pentium motherboard I use, all I get is a screen that looks like this:

Pentium II emulation works, why not Pentium 1? Does anyone know what could be causing this and maybe a possible solution?

Tested with r805 : black screen at start even with VGA only adapter.
Pclog stops here : ps2_read: port=0096 temp=7a
ps2_write: port=0096 val=00 f000:03d8
Tested fine with r749, failed since r750

Paradise Bahamas (s3 Vision864) in 86box (build 270 and later) has problems in windows 98SE.
After windows loading i get this screen......

This is from build 282
No arrow on screen and all icons looks weird,
Maybe this problem in in other cards too.
Just have a check.
Thank you.

3dmark99 no run benchmark with voodoo.

Hi,
Disk does not load. data is OK, but audio not function.

Running 86box commit 3ea92db, host Win7 Pro, guest Windows for Workgroups 3.11 on DOS 6.22
Locking the host using Win+L will make guest OS freeze, can only do hard reset. Happened on both WfW 3.11 and DOS 6.22.

Config file:

[Machine]
model = p55r2p4
mem_size = 32768
nvr_path = C:\86box\nvr\
cpu_use_dynarec = 1
cpu = 6
cpu_enable_fpu = 1

[Video]
gfxcard = px_trio64

[Input devices]
mouse_type = msserial

[Novell NE2000]
mac = c7:44:e2

[Sound]
sndcard = sb16
midi_device = mt32

[Network]
net_type = slirp
net_card = ne2k

[Hard disks]
hdd_01_parameters = 63, 15, 2492, 0, ide_pio_and_dma
hdd_01_fn = C:\86box\images\wfw311.img
hdd_01_ide_channel = 0:0

[Removable devices]
fdd_01_type = 35_2hd
fdd_02_type = none
cdrom_01_parameters = 1, atapi_pio_and_dma
cdrom_01_ide_channel = 0:1

[Phoenix S3 Trio64]
memory = 4

Tried multiple motherboards/cpu combinations, behaves the same.

The project is missing some common style guidelines that all developers abide by.

There isn't a nvr for that machine so it is created automatically once it runs. But when enter the BIOS setup, I'm asked a passwork while I've no idea what it is.
Is there a initial password or a cracked nvr file available?

You can try to make a sound panel for audio; it comes to me that audio for cdrom is louder than normal sound.
I tested windows sound system

When Tyan Titan-Pro AT or ATX model chosen on the config window, the whole emulator crashes immediately. Also, the emulator encounters an AppCrash error as soon as it launches when model in 86box.cfg is changed to 440fx/tpatx.
I use the two roms: NTMAW501.BIN and S1668P.ROM and the build is downloaded from http://dome.rol.im/job/86Box-Optimized/

Hi, new version 2.0 044e3ba some problems.

1.I tested game Spellcross
The 86box will hang up after a while. Statistics are changing. But the mouse is stuck picture is standing.
2. CDROM mount iso not function play audio tracks.
Check or uncheck "mute" - not function

I'd like to help. How to use debug mod and where to see the listing?

I compiled last version

DEBUG = y
OPTIM = n
COPTIM = -O1

On the last couple versions of 86Box, if you choose System MIDI as your MIDI Out Device, it still reports it as Roland CM-32L Emulation the next time you look at it.

I would like to see the Si54P-AIO motherboard from ECS on 86Box,
but seems like something is missing...

It's not clear what version to use. The configure script looks for allegro-config, which Allegro 5 (here on Github) doesn't seem to provide. Allegro 5 installs pkg-config files so you can use pkg-config instead.

I am unable to compile and develop this project until I can get this project's major dependency (Allegro) to compile and install for this project.

I know bus mouse uses IRQ2. However, 86box set to IRQ5 default. It seems to be the reason why the DOS mouse drivers does not recognize the bus mouse.

Mach64 only work in 256 color mode. (tested in Win98se)

Hi,

First sorry for my poor english. Hope my question will be understanding.
I just install 86box on my computer and download image form MSDOS 6.22.
Someone can tell me how to mount it for 86box boot on it at starting ?
do I need a specific directory to place image files ?

Thx

Cédric

When the user saves a config or a hard disk image and enter a file name without an extension, 86box should append a default extension that corresponds to the chosen file type. To create a file without an extension, the user would have to choose All files (*.*) in the file type listbox. This is the consistent and expected behavior in any Windows program, therefore 86box should follow this guideline as well.

I seem to be encountering an error when switching to full screen. There is a black bar on the left hand side of the screen and everything is shifted off to the right leaving the far right hand side of the screen not visible. This happens both in full screen stretch and 4-3 aspect modes.
The problem started with build #647 and seems to be there no matter the configuration I use.
I have tested this on 2 separate physical machines and multiple virtual machines as well, and the problem persists. Thanks for any help in advance.

Everytime I try to run the Windows XP Setup from CD boot I get this no matter If I use the 440FX or 430VX

and that is the BSOD I get when I try to run the Setup

Recent builds of 86box doesn't see ROMs for Cirrus Logic and nVidia cards (build from March 2017 doesn't have that problem).

at first, Thank you for your help
VERY!! VERY!!! :)

recently, I launched 86box, and checked some modified function with SPC-4200P Bios

Samsung Aladdin 286 bios SPC-4200P
turn on 86box using SPC-4200P bios.
"keyboard stuck..." message occured with BEEP

when I pressed keyboard alt+tab to change windows, in 86box many tab key event is occured.

on Hercules with GREEN/AMBOR/WHITE
I changed color GREEN in Hercules Graphic Mode, I launched the Game as example : Lost Ducth Man :)
game screen was chaning original color WHITE....
but exit game....screen color was GREEN...

after I launched SIMCGA.COM in Hercules Mode
some program was freezing...example...Norton Commander...

and When I used Turbo C++ for DOS, it shows "dpmi server initialization error - machine not in database"

Plz...I want to your help.
and thank you for your help...everytime

at first, Thank you for your help
VERY!! VERY!!! :)

recently, I launched 86box, and checked some modified function with SPC-4200P Bios

1. Samsung Aladdin 286 bios SPC-4200P
   turn on 86box using SPC-4200P bios.
   "keyboard stuck..." message occured with BEEP

2. when I pressed keyboard alt+tab to change windows, in 86box many tab key event is occured.

3. on Hercules with GREEN/AMBOR/WHITE
   I changed color GREEN in Hercules Graphic Mode, I launched the Game as example : Lost Ducth Man :)
   game screen was chaning original color WHITE....
   but exit game....screen color was GREEN...

4. after I launched SIMCGA.COM in Hercules Mode
   some program was freezing...example...Norton Commander...

Plz...I want to your help.
and thank you for your help...everytime