Open Medal of Honor (BETA)

Based on the current BETA V0.81.1 release of https://github.com/openmoh/openmohaa.

Here is a BETA-in-itself version of that:

ftp host: tellusregio.duckdns.org
port: 32463
login: anonymous login
file: OMOHAA_v0.81.1_beta_r726.lha

Extract the files in the release folder and put them into your official MOHAA installation directory. For me (and presumably for all), this is the directory that includes the MOHAA.exe file. Finally, run openmohaa.ppc.

Issues:
1. OS Mouse Cursor still visible.
2. Networking; can find servers, but not connect.
3. Working on X1000, not X5000 [requires kernel 54.65] (See Issue Description, below) (Update: should be fixed by latest version)
4. Missing textures on non-X1000 targets

Further issue description (where needed)
3. This is due to the e5500 not being able to handle mis-aligned floats. I now remember why I made a change to ExecSG to emulate - in this case - the lfs instruction - otherwise an alignment exception is raised. I made this change and committed it over a year ago. Either a new updated kernel is released to the public or I need to make manual changes to the MOHAA code base. I can see that with my modified kernel, using lfs emulation, I can play the game on my X5000. Again, please see: Beer.

Beer
Lastly, please buy me beers - in order that I can get more things done - via paypal: 3246251196ryan AT gmail.com.

@rjd324

Congratulations :)

Works on Sam460 (with OS4Beta).. I had to ignore DSI while launching the game, I guess it was released to GetLocalIP..

Now trying the training mission... There are some missing textures when you progress in this mission (missing tank textures and ground/sky textures)

Thanks for the report. I also see this on my X5000. I was worried it was my emulated lfs instructions, but if it is happening on the Sam460, I guess not.

@rjd324

Nice…..!

I will try it out tonight and yes as far as this port has reached a status where we can use it fully I will support their work with donations. Promise 😉

Thanks for working on it again.

@rjd324
Quote:Next to no PowerPC CPUs support it, only exceptions may be the 603/604 CPUs.
AMCC 440/460 (probably all 405 based SoCs) don't support it, at least not if the access crosses a cache line boundary. Such access don't cause crashes as they are handled and emulated by the kernel 4x0 exception handlers, but of course that's very slow.
IIRC G3 and G4 CPUs don't support it either, which was the reason a very large part of the code in my powerpc.library was about emulating such unaligned accesses, not limited to float/double, but emulation of unaligned integer accesses was required as well to be able to run old WarpOS code built for 603/604 CPUs on G3/G4/405/POWER/PowerQUICC/etc. CPUs.

Bottom line: Never ever use any unaligned accesses on PowerPC, POWER and PowerQUICC CPUs, neither for float/double nor integers. Even if it works at all it's very slow.

Thanks again for providing this beta version.

Since I was curious, I tested it of course. At the start there was, as already reported, a DSI that can be ignored, and then the first intro comes, all without sound, but hey, it's only a beta.

The main menu is started and all settings can be made here....resolutions etc. and these are also applied correctly. However, when I start a new game, a DSI appears again, this time it cannot be ignored and the actual game cannot be started.

I tried again with the standard config without changing anything and the same problem again. Even though it may not be useful, I have recorded the GR log.

@Maijestro
Hi. Right. See issue #3.

I have added v0.81.1 to the FTP server. This is the latest BETA version and I have also added a workaround for misaligned floats directly into the code which should prevent the previously mentioned alignment exceptions.

Hello.
Tryd it on my X1000 and radeonrx550 and it works so far.
Is there a way to get Fullscreen?

Send you some Beers ;)


EDIT: To my Problems, it was a faulty installation here. It works now.

@rjd324

Testing on Sam460 (will test on X5000 also)
Level Loading takes much longer because of debug output (I guess) ?

I see following information in shell windows:
origin log:
new log:

Btw previous version runs like 1 fps on A1222. (Menu and videos are fine, but level loading and fps is slow. Maybe it will change with this new version.

@rjd324

How cool is that?

Ok the new version works and I can start a new game. The DSI at the start of the game is still there and there is no sound in the intro.

The colors seem to be correct with my RX580 and the gameplay is smooth. What I have noticed is that the loading times are extremely long, 5-6 minutes.



You are on the right way...thank you very much.

Okay. I will remove the debugging that I forgot to take out. This will be causing the slow down.

@maijestro
1. Is that running on a X5000? You don't see a texture glitch in that area?

2. Can you give me the GR for the initial DSI? I cannot reproduce it.

===

At the moment, I have an X1000 and X5000. A1222 probably needs an SPE build...

One thing at a time, of course. I will at least upload a version without the debugging later today.

@Maijestro

I can confirm that texture problem is also gone on Sam460 + RX-550..I will test on X5000 + RX-580 later today.

About DSI.. Here is my log

tack trace:
module LIBS:bsdsocket.library at 0x7FBC1850 (section 0 @ 0x2582C)
module LIBS:bsdsocket.library at 0x7FBC1690 (section 0 @ 0x2566C)
[/home/rjd/projects/SVN_AMIGA/openmohaa/code/qcommon/net_ip.c:1356] openmohaa.ppc:NET_GetLocalAddress()+0x88 (section 7 @ 0x7B2E8)
[/home/rjd/projects/SVN_AMIGA/openmohaa/code/qcommon/net_ip.c:1439] openmohaa.ppc:NET_OpenIP()+0x28 (section 7 @ 0x7BBAC)
openmohaa.ppc:NET_Config()+0x190 (section 7 @ 0x7BE3C)
[/home/rjd/projects/SVN_AMIGA/openmohaa/code/qcommon/net_ip.c:1645] openmohaa.ppc:NET_Init()+0x14 (section 7 @ 0x7BE78)
[/home/rjd/projects/SVN_AMIGA/openmohaa/code/sys/sys_main.c:866] openmohaa.ppc:main()+0x190 (section 7 @ 0x2F78)
native kernel module newlib.library.kmod+0x00002624
native kernel module newlib.library.kmod+0x00003350
native kernel module newlib.library.kmod+0x00003874
openmohaa.ppc:_start()+0x1e0 (section 7 @ 0x1C3C)
native kernel module kernel+0x00054c94
native kernel module kernel+0x00054d0c

@rjd324

Quote:

I'm not sure, but as far as I can describe it, everything looked normal to me. I played it up to the first tank where the dynamite is supposed to stick and also observed the environment a little. I didn't notice any glitches or missing textures. The first tank in the training mission is in the hangar or is it further out?

If it is normally in the hangar there are at least glitches for me.

I'm no expert, but there seem to be problems with the bsdsocket.library. I'm attaching the log so you can keep an eye on it.

Test system X5000/40 / RX580