Reborn return to castle Wolfenstein

@RIBDEVIL
Those ones which bring warning about warp3dnova 1.63 on run are prototype, barony and fricking shark. That warning is harmless, everyhting works over 1.58, it's just we put to gl4es for future, and i forgot to change it for game releases. In other words don't worry, with 1.58 all will be fine too.

@Plexus
Thanks but now it freezes at mission briefing.

@Joeled
WHAT IS YOUR GFX CARD?

MAYBE SOMETHING IS WRONG WITH SOUTH ISLAND DRIVER IF YOU HAVE R7, R9...
VRAM MEMORY INCREASE WITH 3MB EVERYTIME IF I QUIT RTCW OR REBORN ENGINE...

IMPORTENT TO SOFTREBOOT EVERTIME YOU MAKE SOME CHANGES IN SETTINGS!

WHEN YOU QUIT RTCW PLEASE DO A SOFTREBOOT OR HARDREBOOT AND TRY START RTCW AGAIN. (VRAM AT NORMAL AGAIN)

(INCREASE OF +3MB VRAM WILL NOT HAPPEN WITH IOQUAKE3 OR EGL SDL1 QUAKE3 ARENA OR ANOTHER GAMES WHEN QUIT.

IDONT KNOW WHY, I AM NOT AT THAT LEVEL OF KNOWLEDGE.

HANS AND HUNO?

@ Huno : can you confirm that we need the data files from the 1.41 version of RTCW ? I dont see any info regarding which version would work or not in the readme or on your site.

@Lio
Yes, use de 1.41 sp.pak files

@HunoPPC

Quote:

What makes you think it's a memory overflow? IIRC, you're still waiting on a new serial cable.

Even if it were a memory overflow, that should *NOT* make your machine spontaneously turn off.

Quote:
That's not a problem because you have v3 of the driver which handles that very well.

Hans

@Plexus

Quote:
If VRAM memory increases with 3MB every time you quit RTCW, then I doubt that it's the drivers. I'd appreciate it if people could stop automatically blaming the drivers. I've ended up debugging other people's code far too often. Just because the code "works on Linux" doesn't mean there are no bugs. I remember one case where there was an uninitialized array bug in a game's code, and it only worked on Linux because Linux zeros the memory.

It *could* be a driver issue, but it could just as easily be something else.

@all
We really need more data to figure this one out, especially actual crash logs. So far we have a few people using an X5000 + RadeonHD.chip v3.6 having it freeze with the HD textures. One person has it working fine with the RadeonRX.chip driver.** Is the HD pack working fine for anyone with a Southern Islands graphics card?

Hans


**NOTE: The RadeonRX and RadeonHD drivers use the same memory management code for using all of VRAM.

@Hans, Huno
I can't say for sure, but as i use gl4es (so ogles2.library + warp3dnova) with SI driver when build some games, and they didn't leak any vram on exit, then it probably or game itself, or some older version of gl4es was used, or if some library like sdl1/sdl2/egl is used something there.

As for freeze, i have x5k with SI driver, so can do tests as well (hope today)

@kas1e

What Q3 sources are used here and have the changes made publicly available?

At least standard ioquake3 doesn't explicitly delete GL context, or close SDL window: https://github.com/ioquake/ioq3/blob/m ... /code/sdl/sdl_glimp.c#L88

@Plexus

I m using AmigaOne X5000 and a RadeonHD 7700.

@Hans
Hi Hans,
Excuse me but at no time I have questioned your magnificent work and I will still allow myself to criticize even less because I do not get to your ankle. When I say that I do not think of an overflow memory is that I already knew this on my first tests with EGL and there I do not have the same symthomes. As for the extinction of my machine it is not all the time that I have this and in general the machine crashes and everything remains frozen. No I still have not received my USB-SERIAL compatible Windows10.
While waiting for my cord I ask myself some simple questions:
The same engine with the same versions of libraries, SFS2, only the Graphic card differs:
On SAM460 functional
On X1000 functional
On X5000 non-functional

A relationship with the system part?

No the code is not public for RTCW-HD in any case all that is part of the audio rendering and Input

HunoPPC

@Hans

My X5000 has a RadeonHD 7700.
My X1000 has a Radeon 280 TOXIC

HunoPPC

@HunoPPC

Hi Huno,
Quote:
Relax. I don't think you're being critical. I was just asking why you thought it was a memory overflow. Memory overflow is quite specific, so I wanted to know what you were basing that on.


Quote:
Okay, I think it's best we wait until you (or someone else) gets a crash log. We're just guessing right now.

Yes, since it's only happening with the X5000 it's likely to be a system module, although that isn't guaranteed. I've seen bugs that only get exposed on certain systems because the memory layout is different. Bear in mind that there's no X5000-specific code in the RadeonHD.chip driver, or Warp3D Nova.

Hope you get that USB to serial device soon...

Hans

@HunnoPPC & Hans

Have not had any feed back for Sam460ex + Radeon SI configurations .

Only from me with Sam460ex + Radeon RX.

@hans

You can find some crashlogs at the links below, that I had on my system when I tried to run Reborn.

My system is an AmigaOne X5000/40 with a Radeon HD 7700 with 1GB VRam

http://os4coding.net/node/2312
http://os4coding.net/node/2313

I have the following versions installed:
ATIRadeon.chip 53.30
RadeonHD.chip 3.6
ogles2.library 1.22
Warp3D.library 53.27
Warp3DNova.library 1.61
Warp3DPPC.library 4.3

@walkero

Thanks.

I just looked at your crash logs. That's definitely *not* what I was expecting. None of the crashes directly involve RCTW or any part of the graphics system.

It's crashing with a null pointer in the p50x0sata.device multiple times, followed by another null pointer crash in the USB stack. Both at exactly the same point in the kernel, although the p50x0sata.device also crashes elsewhere too (yet another null pointer crash).

It's possible that there's a bug in the p50x0sata.device, but it could also be another module trashing memory, in which case it could be caused by anything on the system, including RCTW itself.** If it's the p50x0sata.device, then loading the larger textures and data files may be hitting some special case.


I see you used the debug kernel with those logs. Could you try again using the release kernel?

It's also a good idea to ask Steven Solie if he could check out the debug logs. He wrote the p50x0sata driver, so may be able to figure out what it's doing at that point. Please tell him which version of the driver you have.

Hans


** With memory trashing like this, it may only take down the system on the X5000 + RadeonHD SI card due to bad luck with where things are located in memory.

Sam460 2 gig ram, Radeon HD 7770 1 gig ram

ATIRadeon.chip 53.29
RadeonHD.chip 3.6
ogles2.library 1.22
Warp3D.library 53.27
Warp3DNova.library 1.58
Warp3DPPC.library 4.3


Errors after the mission statement first day I tried.
Errors could be ignored and game then played.
On the first day I was switching between computers using a USB KVM switch.
I also tried cancelling the intro.

Today I didn't switch machines and let the intro run.
No errors and game played.

Possibly just needed a cold reboot.
Also possible dropping and adding USB devices during game loading was an issue.

Cheers

@Hans

Hello ,

Now my debug


[SDL_SYS_JoystickInit] ENUM RETURNED: 1
[SDL_SYS_JoystickInit] Found 1 joysticks
[SDL_SYS_JoystickOpen] Opening joystick #0 (AI ID=4096)
[SDL_SYS_JoystickOpen] Successful
[HAL_DfltTrapHandler] *** Warning: Fatal exception in task 0x60692930 (Background CLI, etask = 0xEFB2FC10) at ip 0x018030AC
[HAL_Alert] ALERT! 0x80000003
[HAL_SuspendTaskWithUpdate] Suspending 0x60692930 with state=2, DAR=0x00001824, at IP=0x018030AC, LR=0x01802EC4
Dump of context at 0xEFAF9BA0
Trap type: DSI exception
Exception Syndrome Register (ESR): 0x00000000
Machine State (raw): 0x00023032
Machine State (verbose): [Critical Ints on] [Super] [FPU on] [IAT on] [DAT on] [Recoverable]
DSISR: 00000000 DAR: 00001824
No matching page found
Temporary stack trace:
#0: in module kernel.debug+0x000030AC (0x018030AC)

Crashed process: reborn-sp (0x60692930)
DSI verbose error description: Page not found in hash table (page fault)
Access was a load operation
0: 00000000 02182210 00000002 02182220 43933D48 00001824 00000001 00004004
8: 08000000 08000000 43933E30 7ECE67D8 35953333 3D373CAC 00002901 00000000
16: 00000000 3FFC3A28 00000000 00000000 3FEC3168 55766[HAL_DfltTrapHandler] *** Warning: Fatal exception in task 0x60692930 (Background CLI, etask = 0xEFB2FC10) at ip 0x018030AC
[HAL_Alert] ALERT! 0x80000003
[HAL_SuspendTaskWithUpdate] Suspending 0x60692930 with state=7, DAR=0x00001018, at IP=0x018030AC, LR=0x01802EC4
[HAL_SuspendTaskWithUpdate] *** Warning: Suspending an already suspended task
[HAL_SuspendTaskWithUpdate] Aborting to prevent infinite recursion
[HAL_DfltTrapHandler] *** Warning: Fatal exception in task 0x5F313BB0 (CPUInfo.CPUTask, etask = 0xEFB2F250) at ip 0x018030AC
[HAL_Alert] ALERT! 0x80000003
[HAL_SuspendTaskWithUpdate] Suspending 0x5F313BB0 with state=2, DAR=0x00001018, at IP=0x018030AC, LR=0x01802EC4
Dump of context at 0xEFB11000
Trap type: DSI exception
Exception Syndrome Register (ESR): 0x00000000
Machine State (raw): 0x00003032
Machine State (verbose): [Super] [FPU on] [IAT on] [DAT on] [Recoverable]
DSISR: 00000000 DAR: 00001018
No matching page found
Temporary stack trace:
#0: in module kernel.debug+0x000030AC (0x018030AC)

Crashed task: CPUInfo.CPUTask (0x5F313BB0)
DSI verbose error description: Page not found in hash table (page fault)
Access was a load operation
0: 00000000 0217C210 00000002 0217C220 00000000 00001018 00000000 00000000
8: 08000000 08000000 5C100000 5C100000 000002E0 0000000D 0000000E 0000000F
16: 00000010 00000011 00000012 4E007164 00000014 00005F15 00005F16 00005F17
24: 00000018 00000019 0000001A 0000001B 0000001C 0000001D 0000001E 0000001F
CR: 5BADCAF3 XER: A000007E CTR: ABADCAFE LR: 01802EC4
DSISR: 00000000 DAR: 00001018

FP0 : FFF80000A2004100 40C3008000000000 405F000000000000 0000000000000000
FP4 : 3FF0000000000000 4330000080000000 401E000000000000 4330000080000000
FP8 : 4330000080000000 4051800000000000 4070000000000000 4070000000000000
FP12: 3FF0000000000000 4137290040362900 160298CAC1597666 A6C8A8006FF46726
FP16: F6E1D67606C8E316 00E6A352E3D84A18 0B1D5A034E861070 F3884402717AF5FD
FP20: 0248E7CC24C06173 2659B949C4CB2D27 2243F0C966EAC112 012A059036E82213
FP24: 0660170444A97195 4E688914644EE947 0A71A492F4EA6152 424834C872F8E173
FP28: 126A340F25EA2174 A662214C356B6B16 629064C073EC7367 40C3008000000000
FPSCR: A2004100

Disassembly of crash site:
[HAL_DfltTrapHandler] *** Warning: Fatal exception in task 0x5F313BB0 (CPUInfo.CPUTask, etask = 0xEFB2F250) at ip 0x018030AC
[HAL_Alert] ALERT! 0x80000003
[HAL_SuspendTaskWithUpdate] Suspending 0x5F313BB0 with state=7, DAR=0x00000934, at IP=0x018030AC, LR=0x01802EC4
[HAL_SuspendTaskWithUpdate] *** Warning: Suspending an already suspended task
[HAL_SuspendTaskWithUpdate] Aborting to prevent infinite recursion
[HAL_DfltTrapHandler] *** Warning: Fatal exception in task 0x6FFAA360 (idle.task, etask = 0xEFFF40C0) at ip 0x018030AC
[HAL_Alert] ALERT! 0x80000003
[HAL_SuspendTaskWithUpdate] Suspending 0x6FFAA360 with state=2, DAR=0x00001018, at IP=0x018030AC, LR=0x01802EC4
Dump of context at 0xEFFF33E0
Trap type: DSI exception
Exception Syndrome Register (ESR): 0x00000000
Machine State (raw): 0x00003032
Machine State (verbose): [Super] [FPU on] [IAT on] [DAT on] [Recoverable]
DSISR: 00000000 DAR: 00001018
No matching page found
Temporary stack trace:
#0: in module kernel.debug+0x000030AC (0x018030AC)

Crashed task: idle.task (0x6FFAA360)
DSI verbose error description: Page not found in hash table (page fault)
Access was a load operation
0: 00000000 02176210 00000002 02176220 00004000 00001018 00000006 00000007
8: 08000000 08000000 0000000A 0000000B 00000000 0000000D 0000000E 0000000F
16: 00000010 00000011 00000012 4E007164 00000014 00005F15 00005F16 00005F17
24: 00000018 00000019 0000001A 0000001B 0000001C 0000001D 0000001E 0000001F
CR: 5EADBEE5 XER: C000006F CTR: 018187AC LR: 01802EC4
DSISR: 00000000 DAR: 00001018

FP0 : FFF80000A2004100 40C3008000000000 405F000000000000 0000000000000000
FP4 : 3FF0000000000000 4330000080000000 401E000000000000 4330000080000000
FP8 : 4330000080000000 4051800000000000 4070000000000000 4070000000000000
FP12: 3FF0000000000000 4137290040362900 160298CAC1597666 A6C8A8006FF46726
FP16: F6E1D67606C8E316 00E6A352E3D84A18 0B1D5A034E861070 F3884402717AF5FD
FP20: 0248E7CC24C06173 2659B949C4CB2D27 2243F0C966EAC112 012A059036E82213
FP24: 0660170444A97195 4E688914644EE947 0A71A492F4EA6152 424834C872F8E173
FP28: 126A340F25EA2174 A662214C356B6B16 629064C073EC7367 40C3008000000000
FPSCR: A2004100

Disassembly of crash site:
[HAL_DfltTrapHandler] *** Warning: Fatal exception in task 0x6FFAA360 (idle.task, etask = 0xEFFF40C0) at ip 0x018023FC
[HAL_Alert] ALERT! 0x80000003
[HAL_SuspendTaskWithUpdate] Suspending 0x6FFAA360 with state=7, DAR=0x0180022C, at IP=0x018023FC, LR=0x018023C0
[HAL_SuspendTaskWithUpdate] *** Warning: Suspending an already suspended task
[HAL_SuspendTaskWithUpdate] Aborting to prevent infinite recursion


HunoPPC

@HunoPPC

Well, at least this time one of the crashes is of reborn-sp itself. However, you've got a stream of crashes happening so fast that the crash log for each gets cut off, missing the most important parts: the disassembly and actual stack trace. The temporary stack trace is clearly wrong.

Walkero's crash log also features the same temporary stack trace, so it's the same issue in both cases.

Not sure where to go from here...

Hans

@hans and @hunoppc

Today I tested again the game on my X5000. Before I executed reborn, I closed every program from Exchange, like AmiDock, Screenblanker etc. I even disabled MassStorage. I did that because I believe when the crash happens, it makes all these programs crash and the serial port gets full of trash. Also deleted the config file from the main folder.

All the crashes happen after the first intro, and the output console's last step was "Loading weapons", while the first instructions were shown at the screen.

The following crashlog came up when I executed the game from an NGFS partition
http://os4coding.net/node/2314

The following crashlog came up when I executed the game from an SFS 2 partition
http://os4coding.net/node/2315

If you check both crashlogs, they are identical. The only differences you will find are on memory addresses. That makes me think that the filesystem doesn't have to do anything with the crash. What do you believe?

Hope these can help.