There last natively DOS compatible version of IDA pro I've got. I also have a copy of IDA Free (which doesn't support DOS). I don't know whether I can use it with the DOS plugin for IDA Pro. Might have to look at getting an IDA Pro license.

as i remember jumps should stay intact as they relative to EIP, but call functions may fail, correct me if i understand it wrong.
so changing WORD to DWORD shifts (rise) code size, and every next instruction after modification shifts forward and more forward ?

RaVeN-05 wrote on 2025-02-25, 12:56:

as i remember jumps should stay intact as they relative to EIP, but call functions may fail, correct me if i understand it wrong.
so changing WORD to DWORD shifts (rise) code size, and every next instruction after modification shifts forward and more forward ?

Sorry yes, calls will fail, my bad. My test to see if any part of dos4gw is embedded in icr2 is running. In the meantime I need to go through dos32a code to see if it has a jump table for the subroutines, if there is, I might be able patch dos32a in there and replace or overwrite the jump table.

Relative jumps could also fail if an instruction between the source EIP and dest EIP has become longer.

Die by the Sword seems to be having similar threading problems.

[EDIT] It sets the main thread's (application thread) affinity mask to be 1. For some reason Win11 24H2 doesn't seem to like this being adjusted. There is a warning about SetProcessAffinity for a process other than the DLL owner's, but I think that's just because the parent process might have threads set to run on a core/processor which the DLL might not allow, which ends up freezing the thread.

@Raven-05 where can I find the free dos-plugin for ida pro? I have Ida Free, which is IDA Pro with only x86/x64 decompilation and licensed to individual developers.

[EDIT] The reason I ask is that there're two different ida dos plugins.

[EDITA] Ida free only supports PE (Win32/64 Exes).

[EDIT] Ida Pro Free supports DOS through the last free version 5.0 or 8.0.

sharangad wrote on 2025-02-25, 19:29:

Die by the Sword seems to be having similar threading problems.

[EDIT] It sets the main thread's (application thread) affinity mask to be 1. For some reason Win11 24H2 doesn't seem to like this being adjusted. There is a warning about SetProcessAffinity for a process other than the DLL owner's, but I think that's just because the parent process might have threads set to run on a core/processor which the DLL might not allow, which ends up freezing the thread.

Die by the Sword had another issue where the game engine doesn't refresh unless the CPU keeps working. Does it exist in RReady too?

Actually I tested DBTS again with the Windows Hypervisor stuff uninstalled and the game worked just fine. I think you're talking about the slow mo.frozen menus sometimes, yeah the RRedline version sometime has that. Moving the mouse frantically fixes that.

I *think* Team Apache might work if the Virtualization platform stuff is uninstalled. I can't get rid of it right now because I run IDA in a Win98 VM with VMWare and you need that to run any kind of virtual machines. The last version of IDA Pro compatible with DOS exes (version 5) has rendering issues on newer platforms. If I have time early tomorrow morning I'll give it a go. Die by the Sword doesn't run at all with the Hypervisor stuff installed.

Well as i remember dosboxida plugin works only with IDA Pro (not freeware version)
outdated this i tried to use https://www.romhacking.net/utilities/1129/
latest didn't tried , will do at some time https://github.com/lab313ru/dsbxida

RaVeN-05 wrote on 2025-02-26, 13:25:

Well as i remember dosboxida plugin works only with IDA Pro (not freeware version)
outdated this i tried to use https://www.romhacking.net/utilities/1129/
latest didn't tried , will do at some time https://github.com/lab313ru/dsbxida

Thanks dude!

Time out just to say ...wow! Been testing ICR2 with the latest Alpha build and the Rendition experience/fidelity is extrordinary. Yes, there are currently issues with a lot addon tracks - especially track/carset related conflicts - and it's taking a while to test all the content as there's a ton of it. But for the most part CART seasons 1987-2003 seems within reach +/- in some fashion. Yes ...we finally have THE 90's CART era game in all its flaming Rendition glory!

I've been following the modder chatter and hopefully your RR updates will cure the remaining issues - rather than having to edit all the addon tracks/carsets themselves.

This is the real deal. I'm getting my retail copy soon.

sharangad wrote on 2025-02-26, 09:20:

Actually I tested DBTS again with the Windows Hypervisor stuff uninstalled and the game worked just fine. I think you're talking about the slow mo.frozen menus sometimes, yeah the RRedline version sometime has that. Moving the mouse frantically fixes that.

I *think* Team Apache might work if the Virtualization platform stuff is uninstalled. I can't get rid of it right now because I run IDA in a Win98 VM with VMWare and you need that to run any kind of virtual machines. The last version of IDA Pro compatible with DOS exes (version 5) has rendering issues on newer platforms. If I have time early tomorrow morning I'll give it a go. Die by the Sword doesn't run at all with the Hypervisor stuff installed.

I don't have any kind of hypervirtualisation installer right now, but Team Apache still behaves the same. Although I have AMD V enabled but is that an issue?

Regarding Interstate 76, it broke again in DxWnd, just at the time OpenGL hook is getting added. It works most of the time. I noticed a common pattern: in I76 initialisation there is a RReady screen, then it moves to the DirectDraw menus, and back to Redline for the mission. Whenever the game gets the Software error, it already missed the initial Redline initialisation screen.

The other issue is, non 640x480 resolutions cause a black screen now rather than an enlarged screen, but we are working on it through a lot of OGL games (Quake, Daikatana, Wizball, CS and CZERO...)

tincup wrote on 2025-02-26, 17:05:

Time out just to say ...wow! Been testing ICR2 with the latest Alpha build and the Rendition experience/fidelity is extrordinary. Yes, there are currently issues with a lot addon tracks - especially track/carset related conflicts - and it's taking a while to test all the content as there's a ton of it. But for the most part CART seasons 1987-2003 seems within reach +/- in some fashion. Yes ...we finally have THE 90's CART era game in all its flaming Rendition glory!

I've been following the modder chatter and hopefully your RR updates will cure the remaining issues - rather than having to edit all the addon tracks/carsets themselves.

This is the real deal. I'm getting my retail copy soon.

Thank you! Hopefully I can overcome the texture memory limit soon.

BEEN_Nath_58 wrote on 2025-02-26, 17:59:

I don't have any kind of hypervirtualisation installer right now, but Team Apache still behaves the same. Although I have AMD V […]
Show full quote

sharangad wrote on 2025-02-26, 09:20:

Actually I tested DBTS again with the Windows Hypervisor stuff uninstalled and the game worked just fine. I think you're talking about the slow mo.frozen menus sometimes, yeah the RRedline version sometime has that. Moving the mouse frantically fixes that.

I *think* Team Apache might work if the Virtualization platform stuff is uninstalled. I can't get rid of it right now because I run IDA in a Win98 VM with VMWare and you need that to run any kind of virtual machines. The last version of IDA Pro compatible with DOS exes (version 5) has rendering issues on newer platforms. If I have time early tomorrow morning I'll give it a go. Die by the Sword doesn't run at all with the Hypervisor stuff installed.

I don't have any kind of hypervirtualisation installer right now, but Team Apache still behaves the same. Although I have AMD V enabled but is that an issue?

Regarding Interstate 76, it broke again in DxWnd, just at the time OpenGL hook is getting added. It works most of the time. I noticed a common pattern: in I76 initialisation there is a RReady screen, then it moves to the DirectDraw menus, and back to Redline for the mission. Whenever the game gets the Software error, it already missed the initial Redline initialisation screen.

The other issue is, non 640x480 resolutions cause a black screen now rather than an enlarged screen, but we are working on it through a lot of OGL games (Quake, Daikatana, Wizball, CS and CZERO...)

AMD-V should be fine. The thing is i76 works on some machines and not on others, sometimes with the same hardware. I've been testing it with every single release of RReady on my dev machine and it's pretty much worked fine.

Well, what do you know, Descent 2 runs in DMA mode with Dosbox-staging-Rendition once Dos32A is permanently patched into d2verite with and run with this batch file (this is the important bit: set SPEEDY3D_DMA=1). It was only runnable in FIFO mode before, even on real hardware, unless run from Win98:

REM @echo off
CD D2
set SPEEDY3D_DMA=1
set SPEEDY3D=V:\D2
rem Batch file to run Descent 2
if exist *.swp del *.swp
eregcard
Dos32A d2verite %1 %2 %3 %4 %5 %6 %7 %8 %9
if exist *.swp del *.swp

Dosbox-Rendition uses a different mem write model for Descent 2. I'm going to try disabling it to see if it runs.

[EDIT] It does run with Dosbox-Rendition, but tends to freeze every once in a while.

The Dos32A patched version of Descent II is here, in case anyone is interested:
https://nirvtek.com/downloads/D2VERITE-DOS32A.7z
MD5: 1f81b930438ba88391ea1bd58a4ea311

I think I've figured out why this game didn't run in DMA mode and why ICR2 won't see more than 64 MB of RAM. They're using 16-bit DPMI extenders, well at least Descent 2 is. The reason Descent 2 aborted with an error in DMA mode was that it called a DPMI function with the upper half of one of 32 bit registers set to non-zero . 16-bit DPMI extenders don't like having the upper half of parameter registers set. Apps generally have to use BX:CX and with 32 bit DPMI extenders you could use BX:ECX. 16 bit extenders also ran in 32 bit mode, so I'm not entirely sure why this distinction exists.

I still think my best bet is patching ICR2 with Dos32-A somehow.

Scorched Planet videos work 100%:

https://youtu.be/6uSY5rp_SGQ

I actually fixed this last week but was too busy with the ICR2 track support to actually upload this and the corresponding alpha release. The ICR2 work is still ongoing.

This is an untested alpha for supporting Scorched Planet in its current state. The Game must be run in FIFO mode (Set SPEEDY3D_DMA=0 before running the game):
https://nirvtek.com/downloads/RReady.Alpha.20 … PlanetVideos.7z
MD5: 4b942ba81ac74a658993975c309a1373

In one of the last builds the "Software mode" bug reappeared in I76. Now its gone again.

However the black screen is black. What I think is gho is struggling to put I76 to fake fullscreen mode. Since it presents better control over the game through DxWnd, instead of regular fullscreen mode.

I don't know if you can help him on it.

Basically the idea is to make the game work in the same way as the "Run in Window" setting, except bypassing the resolution changes to Windows

BEEN_Nath_58 wrote on 2025-03-03, 09:00:

In one of the last builds the "Software mode" bug reappeared in I76. Now its gone again. […]
Show full quote

In one of the last builds the "Software mode" bug reappeared in I76. Now its gone again.

However the black screen is black. What I think is gho is struggling to put I76 to fake fullscreen mode. Since it presents better control over the game through DxWnd, instead of regular fullscreen mode.

I don't know if you can help him on it.

Basically the idea is to make the game work in the same way as the "Run in Window" setting, except bypassing the resolution changes to Windows

Just tick the "windowed" checkbox under Rendition settings and the game will run in a window at a resolution set in the second resolution selector from the top.

Somehow I never managed to notice this.

The "Failed to run 3D hardware acceleration" is coming from RReady itself. I let RReady use the Launch function to play, in which case the game *crashed* most times (ran 3 out of 20 Launches).

Edit: Not to mention the game will crash anyways on Windows 11 even if RReady init is successful, since EmulateHeap is pulling it from behind... the motive here is to let the init happen correctly and to let the game crash automatically, not at the Redline error

NB: Game crashes more often if you setup a higher resolution in RReady, such as 1360x768