FS-UAE keeps crashing on OSX PPC

Hypex · 12 May 2015, 15:30

Hello. My first post and EAB has already annoyed me by asking me to log in when I wanted to preview it and deleted my entire post! And then did it again after retyping!!! Argh!!!

Anyway similar to what I first put, I find the PPC build of FS-UAE very unstable and it keeps crashing on me. If it doesn't crash then it starts up and starts "thinking" and I have to force quit it. Recently it has locked up OSX straight up with no cursor but using force close hot key eventually brings it back. At times it did work it would eventually loop around thinking again and I have to kill it.

I'm using the tlosm dev29 build here:
https://www.dropbox.com/s/6k29r9zt9b...Dev29.zip?dl=0

My config file:
amiga_model = A4000/OS4
kickstart_file = Kickstart (A4000T).rom
uae_cpuboard_rom_file = cyberstormppc_44.71.rom
hard_drive_0_controller = ide0
hard_drive_0 = OS4.hdf
cdrom_image_0 = ClassicInstallCD.iso
floppy_drive_0 = BootFloppy.adf
uae_a2065 = slirp
automatic_input_grab = 0
fullscreen = 0
slirp_implementations = qemu

Related crash:
Process: fs-uae [16853]
Path: /Users/Damien/Desktop/Amiga/UAE/fs-uaeDev29/FS-UAE.app/Contents/MacOS/fs-uae
Identifier: no.fengestad.fs-uae
Version: 2.5.29 (2.5.29)
Code Type: PPC (Native)
Parent Process: launchd [67]

Interval Since Last Report: 20977 sec
Crashes Since Last Report: 1
Per-App Interval Since Last Report: 2366 sec
Per-App Crashes Since Last Report: 1

Date/Time: 2015-05-12 20:33:37.664 +1000
OS Version: Mac OS X 10.5.8 (9L31a)
Report Version: 6
Anonymous UUID: 4C329DC0-59E3-442D-A263-B83AE0E0D445

Exception Type: EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x00000000c0000003
Crashed Thread: 6

Thread 0:
0 libSystem.B.dylib 0x94b110f8 semaphore_timedwait_signal_trap + 8
1 libSystem.B.dylib 0x94b54214 _pthread_cond_wait + 1320
2 libglib-2.0.0.dylib 0x04138f38 g_cond_wait_until + 312
3 no.fengestad.fs-uae 0x0089bb2c 0x2000 + 9018156
4 no.fengestad.fs-uae 0x00899950 0x2000 + 9009488
5 no.fengestad.fs-uae 0x0089b038 0x2000 + 9015352
6 no.fengestad.fs-uae 0x0088559c 0x2000 + 8926620
7 no.fengestad.fs-uae 0x0000cfd0 0x2000 + 45008
8 no.fengestad.fs-uae 0x000079d0 0x2000 + 22992

Thread 1:
0 libSystem.B.dylib 0x94b17a8c __semwait_signal + 12
1 libSystem.B.dylib 0x94b54318 _pthread_cond_wait + 1580
2 SDL2 0x04291bcc SDL_CondWaitTimeout + 176
3 SDL2 0x042922fc SDL_SemWaitTimeout + 156
4 SDL2 0x042932b4 SDL_RemoveTimer + 1152
5 SDL2 0x042928c4 SDL_TLSGet + 164
6 SDL2 0x0429260c SDL_ThreadID + 644
7 libSystem.B.dylib 0x94b52f70 _pthread_start + 316

Thread 2:
0 libSystem.B.dylib 0x94b17a88 __semwait_signal + 8
1 libSystem.B.dylib 0x94b1789c nanosleep$UNIX2003 + 188
2 no.fengestad.fs-uae 0x00898cc4 0x2000 + 9006276
3 libglib-2.0.0.dylib 0x04119120 g_thread_proxy + 160
4 libSystem.B.dylib 0x94b52f70 _pthread_start + 316

Thread 3:
0 libSystem.B.dylib 0x94b11078 mach_msg_trap + 8
1 libSystem.B.dylib 0x94b17f9c mach_msg + 56
2 com.apple.CoreFoundation 0x924f7394 CFRunLoopRunSpecific + 1812
3 com.apple.audio.CoreAudio 0x925d5fa8 HALRunLoop::OwnThread(void*) + 212
4 com.apple.audio.CoreAudio 0x925d5de4 CAPThread::Entry(CAPThread*) + 104
5 libSystem.B.dylib 0x94b52f70 _pthread_start + 316

Thread 4:
0 libSystem.B.dylib 0x94b110f8 semaphore_timedwait_signal_trap + 8
1 libSystem.B.dylib 0x94b54214 _pthread_cond_wait + 1320
2 com.apple.audio.CoreAudio 0x925e79f0 HP_IOThread::WorkLoop() + 488
3 com.apple.audio.CoreAudio 0x925e77f0 HP_IOThread::ThreadEntry(HP_IOThread*) + 12
4 com.apple.audio.CoreAudio 0x925d5de4 CAPThread::Entry(CAPThread*) + 104
5 libSystem.B.dylib 0x94b52f70 _pthread_start + 316

Thread 5:
0 libSystem.B.dylib 0x94b17a8c __semwait_signal + 12
1 libSystem.B.dylib 0x94b54318 _pthread_cond_wait + 1580
2 libGLProgrammability.dylib 0x92a8ab68 glvmDoWork + 120
3 libSystem.B.dylib 0x94b52f70 _pthread_start + 316

Thread 6 Crashed:
0 OpenAL 0x04203fe4 OALSource::GetQLength() + 48
1 OpenAL 0x041fa588 alSourceUnqueueBuffers + 180
2 no.fengestad.fs-uae 0x000acd18 0x2000 + 699672
3 no.fengestad.fs-uae 0x000e81a8 0x2000 + 942504
4 no.fengestad.fs-uae 0x0016c8e0 0x2000 + 1485024
5 no.fengestad.fs-uae 0x0016d030 0x2000 + 1486896
6 no.fengestad.fs-uae 0x0016c388 0x2000 + 1483656
7 no.fengestad.fs-uae 0x0014c738 0x2000 + 1353528
8 no.fengestad.fs-uae 0x00180ff4 0x2000 + 1568756
9 no.fengestad.fs-uae 0x0000bb38 0x2000 + 39736
10 no.fengestad.fs-uae 0x008852a0 0x2000 + 8925856
11 libglib-2.0.0.dylib 0x04119120 g_thread_proxy + 160
12 libSystem.B.dylib 0x94b52f70 _pthread_start + 316

Thread 7:
0 libSystem.B.dylib 0x94b17a8c __semwait_signal + 12
1 libSystem.B.dylib 0x94b54318 _pthread_cond_wait + 1580
2 SDL2 0x04291bcc SDL_CondWaitTimeout + 176
3 SDL2 0x042922fc SDL_SemWaitTimeout + 156
4 no.fengestad.fs-uae 0x001313dc 0x2000 + 1242076
5 libglib-2.0.0.dylib 0x04119120 g_thread_proxy + 160
6 libSystem.B.dylib 0x94b52f70 _pthread_start + 316

Thread 6 crashed with PPC Thread State 32:
srr0: 0x04203fe4 srr1: 0x0000d030 dar: 0xc0000003 dsisr: 0x40000000
r0: 0x041fa58c r1: 0xf030a5d0 r2: 0xc0000003 r3: 0x08ca56c0
r4: 0x00000000 r5: 0x00000000 r6: 0xffffffff r7: 0x00000001
r8: 0xfa000000 r9: 0x00000008 r10: 0x00000004 r11: 0x08ca1aa0
r12: 0x94b1b734 r13: 0x00969fe0 r14: 0x0097390c r15: 0x00b4390c
r16: 0x0097390c r17: 0x0097390c r18: 0x00b48d94 r19: 0x0097390c
r20: 0x3b9aca00 r21: 0x0097390c r22: 0x00013cf3 r23: 0x01ea38c8
r24: 0x00000000 r25: 0x01f12c6c r26: 0x08ca56c0 r27: 0x00000004
r28: 0xf030a66c r29: 0x00000001 r30: 0x08ca4dd0 r31: 0x041fa4e8
cr: 0x24028224 xer: 0x20000001 lr: 0x041fa598 ctr: 0x94b1b734
vrsave: 0x00000000

tlosm · 12 May 2015, 18:10

Hypex!
it is the scsi endianess i was explain it in the amigaworld thread

:http://amigaworld.net/modules/newbb/...order=0#755982

push fast ESC on the keyboard after set the scsi luns on the cyberstorm bios all set to 0 ... save and reset the emulation. and everything will work.

Hypex · 13 May 2015, 17:27

I'm telling you that's not it! :-)

I set up that ages ago, confirmed it later on and somehow managed to confirm it again tonight. Before it crashed. All my LUNs are off. Makes no difference here. Perhaps they should fix their code so you can recompile a more stable version. ;-)

What I don't understand is how an emulation of a SCSI controller on a big endian CPU can be broke on a real big endian CPU! :-? Either the SCSI controller memory map is native little endian or the code is sloppy! Either way doesn't look good. But it shouldn't matter. :-)

tlosm · 13 May 2015, 20:33

Because the scsi emulation was did by qemu and it is not the best ;-)
Uae is much more better in loading and transfer rate but is not compatible with Os4
Hypex can you sent me your email address thru pm?
need to send you something for check

i think the problem is in the kickstart of amiga 4000t... it try to access by kickstart at the scsi.device and make the endianess exit. the amiga 4000t kickstart have the scsi.device in rom.

Hypex · 14 May 2015, 15:08

What about the IDE emulation? As you could see I changed to cdrom_image_0. As the other cdrom option was for CD32 and CDTV.

But the crash occurs here. So looks like an audio problem.
0 OpenAL 0x04203fe4 OALSource::GetQLength() + 48

tlosm · 14 May 2015, 21:56

let me know after my email pls

Hypex · 15 May 2015, 11:36

So I checked with a few Kickstart ROMs. For some reason it tends to get caught up with some ROMs and goes in an endless loop or crashes. Whatever that is about. But it all ends up the same. The OS4 boot floppy never finds the CDROM. I don't know why! Is there an obvious error in my config?

It would help if the log said file not found but it seems to skip things.

tlosm · 15 May 2015, 14:26

i think everything depend of th fsuae ide cdrom config... the only one who can help about is Mr Frode .
The other solution is use a hdf make by other machine (x86) and see if everything working... i did like this make installation from fsuae on macbook pro and use the hdf on Quad G5

FrodeSolheim · 15 May 2015, 17:08

Regarding config, there's a peculiarity with the ROM config in the latest development versions, so instead of:

Code:

uae_cpuboard_rom_file = cyberstormppc_44.71.rom

... better use this FS-UAE option instead:

Code:

accelerator_rom = cyberstormppc_44.71.rom

Also, do not use the following config:

Code:

hard_drive_0_controller = ide0

... as the hard drive will by default (without this option) be connected to the accelerator SCSI interface.

Code:

slirp_implementations = qemu

slirp_implementations is a mis-spelling of slirp_implementation. You also do not need to specify this option anyway.

tlosm · 15 May 2015, 18:21

Frode if not on Ide0 setting the fsuae crash because the indianess of SCSI on PPC port .

Hypex · 17 May 2015, 06:43

Thanks for the tips Frode. I changed it to suit. Unfortunately it just made it crash! So I removed the 2065 options altogether and it was better again. But then later it crashed again. Seriously, this audio bug needs to be fixed. Does it only affect OSX PPC? Because with it hanging around I cannot trust FS-UAE, it can and does crash at any moment.

I removed the ide0 setting but had to put it back as it defaulted to the CPU board. Also I have some good news. I got the CDROM working! I had to use an option I made up that doesn't exist on the FS-UAE page! But it works. This. :-)

cdrom_drive_0_controller = ide1

So OS4 started to boot off CD. And then it crashed on a red and blue screen. Sigh...

But, it all ended in tears. I made the mistake of telling it to go full screen and grab input. Well it took over the screen and then went into an endless loop. I couldn't force it out. I had to force shutdown the computer. Well turn it off really.

Typical, Everything works on Intel. But on OSX PPC everything breaks. I'm so over it!

FrodeSolheim · 17 May 2015, 13:20

The reason the it works when you use cdrom_drive_0_controller = ide1 is that with this option, AmigaOS will see the CD device. But, since the problem was probably, and still is, that the CyberStorm device isn't fully initialized, it fails to switch to PPC mode. If the CyberStorm device had been initialized, it would have booted properly in the first place.

As for PPC hosts, I do not officially support this. This means that I do not don't test that stuff works on PPC hosts, but I do not try to prevent it from working there either. I'll fix PPC host issues if they are trivial, and other people can easily test it, but I won't do any real/hard debugging to solve PPC issues. This will have to be done by other skilled volunteers, if any!

tlosm · 17 May 2015, 20:34

Frode on me everything working right , like to some others guys who send my build of FS-Uae ... i think there is something missed on the Hypex system or a miss configuration issue. The G5 is really hard to optimize because full big endian .. i think on G4 dual (because little and big endian) or on CoriQ / Power.x serie (because only little endian) problably everything can work right

Hypex · 18 May 2015, 14:47

In any case, miss configured or not, a crash should not happen when interpreting a text file. If a file is missing it should be reported. If it is a binary it should be checked, most likely by the OS, to see if it loads. Using different Kickstart ROM files also causes it to be unstable. An incorrect ROM crashing in the emulation/Amiga side is a possible and acceptable outcome, but crashing the emulator itself is not an acceptable outcome pointing to a bug. Something is obviously tripping it up. And as demonstrated from my first post, that something is calling the OpenAL audio API functions and it's always thread 6. However, it really needs FS-UAE debugging symbols, otherwise it's too cryptic. .

Perhaps the QEMU SCSI code should be patched. PPC is bi-endian and it's ridiculous that little-endian specific code doesn't work. I know hacking little-endian specific C code to be endian agnostic is a bitch. In my experience a program can appear to work okay bur won't crash. Here it looks looks like it has more dependance on things being backwards. :-)

Hypex · 26 May 2015, 11:48

Well I've tried with ROM tlsom has. But it never even boots here. FS-UAE just hangs on a black screen and I have to force close it. So went back to the ROM I had. I also have an A4000T ROM that acred the same.

Something is wrong. It's almost acting as if the PPC CPU is missing or broken. So I wonder if qemu is setting up the PPC emulation at all. I tried PPC test disk which supports both PUP and WOS. Trying either one freezes the emulation. Or locks up the virtual Amiga anyway.

Sp here is my current config with other settings I've tried. Anything stand out. Is a simple setting crashing it? Well to the point it can run without a crash since I'm still testing it with an unstable build. BTW OS4.hdf is blank formatted HD image from HiToro.

amiga_model = A4000/OS4
kickstart_file = amiga-os-310-a4000.rom
#kickstart_file = Kickstart v3.1 rev 40.70 (1993)(Commodore)(A4000).rom
accelerator_rom = cyberstormppc_44.71.rom
hard_drive_0_controller = ide0
#hard_drive_1_controller = ide1
hard_drive_0 = OS4.hdf
cdrom_drive_0_controller = ide1
cdrom_image_0 = ClassicInstallCD-53.69.iso
#floppy_drive_0 = BootFloppy.adf
floppy_drive_0 = ppctest.adf
#uae_a2065 = slirp
automatic_input_grab = 0
fullscreen = 0
#slirp_implementations = qemu

Hypex · 29 August 2015, 16:10

So i cleaned up my config file and almost got things working. But it eventually ends up in a crash. It looks like the PPC is missing or gone dead in the emulation. I'm sure I have all the files. So here is another crash log. It crashes in a different place.

Code:

Process:         fs-uae [344]
Path:            /Users/Damien/Desktop/Amiga/UAE/fs-uaeDev29/FS-UAE.app/Contents/MacOS/fs-uae
Identifier:      no.fengestad.fs-uae
Version:         2.5.29 (2.5.29)
Code Type:       PPC (Native)
Parent Process:  launchd [85]

Interval Since Last Report:          425445 sec
Crashes Since Last Report:           88
Per-App Interval Since Last Report:  497 sec
Per-App Crashes Since Last Report:   1

Date/Time:       1970-01-01 11:02:16.573 +1000
OS Version:      Mac OS X 10.5.8 (9L31a)
Report Version:  6
Anonymous UUID:  4C329DC0-59E3-442D-A263-B83AE0E0D445

Exception Type:  EXC_BAD_ACCESS (SIGBUS)
Exception Codes: KERN_PROTECTION_FAILURE at 0x0000000000000000
Crashed Thread:  6

Thread 0:
0   libgcc_s.1.dylib              	0x9016c298 __floatundidf + 4
1   libglib-2.0.0.dylib           	0x040efbfc g_get_monotonic_time + 176
2   libglib-2.0.0.dylib           	0x04138e28 g_cond_wait_until + 40
3   no.fengestad.fs-uae           	0x0089bb2c 0x2000 + 9018156
4   no.fengestad.fs-uae           	0x00899950 0x2000 + 9009488
5   no.fengestad.fs-uae           	0x0089b038 0x2000 + 9015352
6   no.fengestad.fs-uae           	0x0088559c 0x2000 + 8926620
7   no.fengestad.fs-uae           	0x0000cfd0 0x2000 + 45008
8   no.fengestad.fs-uae           	0x000079d0 0x2000 + 22992

Thread 1:
0   libSystem.B.dylib             	0x94b17a8c __semwait_signal + 12
1   libSystem.B.dylib             	0x94b54318 _pthread_cond_wait + 1580
2   SDL2                          	0x04291bcc SDL_CondWaitTimeout + 176
3   SDL2                          	0x042922fc SDL_SemWaitTimeout + 156
4   SDL2                          	0x042932b4 SDL_RemoveTimer + 1152
5   SDL2                          	0x042928c4 SDL_TLSGet + 164
6   SDL2                          	0x0429260c SDL_ThreadID + 644
7   libSystem.B.dylib             	0x94b52f70 _pthread_start + 316

Thread 2:
0   libSystem.B.dylib             	0x94b17a88 __semwait_signal + 8
1   libSystem.B.dylib             	0x94b1789c nanosleep$UNIX2003 + 188
2   no.fengestad.fs-uae           	0x00898cc4 0x2000 + 9006276
3   libglib-2.0.0.dylib           	0x04119120 g_thread_proxy + 160
4   libSystem.B.dylib             	0x94b52f70 _pthread_start + 316

Thread 3:
0   libSystem.B.dylib             	0x94b11078 mach_msg_trap + 8
1   libSystem.B.dylib             	0x94b17f9c mach_msg + 56
2   com.apple.CoreFoundation      	0x924f7394 CFRunLoopRunSpecific + 1812
3   com.apple.audio.CoreAudio     	0x925d5fa8 HALRunLoop::OwnThread(void*) + 212
4   com.apple.audio.CoreAudio     	0x925d5de4 CAPThread::Entry(CAPThread*) + 104
5   libSystem.B.dylib             	0x94b52f70 _pthread_start + 316

Thread 4:
0   libSystem.B.dylib             	0x94b110f8 semaphore_timedwait_signal_trap + 8
1   libSystem.B.dylib             	0x94b54214 _pthread_cond_wait + 1320
2   com.apple.audio.CoreAudio     	0x925e79f0 HP_IOThread::WorkLoop() + 488
3   com.apple.audio.CoreAudio     	0x925e77f0 HP_IOThread::ThreadEntry(HP_IOThread*) + 12
4   com.apple.audio.CoreAudio     	0x925d5de4 CAPThread::Entry(CAPThread*) + 104
5   libSystem.B.dylib             	0x94b52f70 _pthread_start + 316

Thread 5:
0   libSystem.B.dylib             	0x94b17a8c __semwait_signal + 12
1   libSystem.B.dylib             	0x94b54318 _pthread_cond_wait + 1580
2   libGLProgrammability.dylib    	0x92a8ab68 glvmDoWork + 120
3   libSystem.B.dylib             	0x94b52f70 _pthread_start + 316

Thread 6 Crashed:
0   ???                           	0000000000 0 + 0
1   no.fengestad.fs-uae           	0x00199b10 0x2000 + 1669904
2   no.fengestad.fs-uae           	0x0014f984 0x2000 + 1366404
3   no.fengestad.fs-uae           	0x00150660 0x2000 + 1369696
4   no.fengestad.fs-uae           	0x0007a15c 0x2000 + 491868
5   no.fengestad.fs-uae           	0x0007aac8 0x2000 + 494280
6   no.fengestad.fs-uae           	0x00247d38 0x2000 + 2383160
7   no.fengestad.fs-uae           	0x0016cfd0 0x2000 + 1486800
8   no.fengestad.fs-uae           	0x0016c38c 0x2000 + 1483660
9   no.fengestad.fs-uae           	0x0014c73c 0x2000 + 1353532
10  no.fengestad.fs-uae           	0x00180ff8 0x2000 + 1568760
11  no.fengestad.fs-uae           	0x0000bb3c 0x2000 + 39740
12  no.fengestad.fs-uae           	0x008852a4 0x2000 + 8925860
13  libglib-2.0.0.dylib           	0x04119124 g_thread_proxy + 164
14  libSystem.B.dylib             	0x94b52f74 _pthread_start + 320
15  ???                           	0000000000 0 + 0

Thread 7:
0   libSystem.B.dylib             	0x94b17a8c __semwait_signal + 12
1   libSystem.B.dylib             	0x94b54318 _pthread_cond_wait + 1580
2   SDL2                          	0x04291bcc SDL_CondWaitTimeout + 176
3   SDL2                          	0x042922fc SDL_SemWaitTimeout + 156
4   no.fengestad.fs-uae           	0x001313dc 0x2000 + 1242076
5   libglib-2.0.0.dylib           	0x04119120 g_thread_proxy + 160
6   libSystem.B.dylib             	0x94b52f70 _pthread_start + 316

Thread 8:
0   libSystem.B.dylib             	0x94b17a8c __semwait_signal + 12
1   libSystem.B.dylib             	0x94b54318 _pthread_cond_wait + 1580
2   SDL2                          	0x04291bcc SDL_CondWaitTimeout + 176
3   SDL2                          	0x042922fc SDL_SemWaitTimeout + 156
4   no.fengestad.fs-uae           	0x00051d40 0x2000 + 326976
5   libglib-2.0.0.dylib           	0x04119120 g_thread_proxy + 160
6   libSystem.B.dylib             	0x94b52f70 _pthread_start + 316

Thread 6 crashed with PPC Thread State 32:
  srr0: 0x00000000  srr1: 0x4200d030   dar: 0xf0697050 dsisr: 0x42000000
    r0: 0x00000000    r1: 0xf030a910    r2: 0xffffffff    r3: 0x00000000
    r4: 0x00200000    r5: 0x00000000    r6: 0x00000000    r7: 0x00000001
    r8: 0x00b2eae8    r9: 0x00000000   r10: 0x03dfa0f4   r11: 0x000000c0
   r12: 0x00000000   r13: 0x00b4bb88   r14: 0x000109c7   r15: 0x02d1abcc
   r16: 0x01cdbb88   r17: 0x02cfabcc   r18: 0x000109aa   r19: 0x03ebadf4
   r20: 0x000109be   r21: 0x01cdbb88   r22: 0x00b4bb88   r23: 0x00b3e2a8
   r24: 0x02d0abcc   r25: 0x02cfabcc   r26: 0x00000000   r27: 0x00bfa108
   r28: 0x00000000   r29: 0x00000001   r30: 0x01d46ea4   r31: 0x00199ab0
    cr: 0x28022244   xer: 0x20000007    lr: 0x00199b10   ctr: 0x00000000
vrsave: 0x00000000

FrodeSolheim · 29 August 2015, 17:40

By the way, the initial crash report from your first post indicates a crash/bug in the FS-UAE OpenAL audio output code, and not in PPC emulation.

Does FS-UAE only crash when you try to run PPC emulation, or does it always crash (including when running stock A500 model)?

Hypex · 25 September 2015, 15:33

Hi Frode. Yes I did see a crash in some audio code and as pointed out in post #5 it was coming from here:
0 OpenAL 0x04203fe4 OALSource::GetQLength() + 48

I looked for options to disable audio but could see none.

I setup a basic A500 emulation as below that booted with AROS Kickstart. At first it seemed to be okay after some thinking but then started thinking endlessly and locked the cursor. I was able to break it. It didn't crash but it did hang.

amiga_model = A500
floppy_drive_0 = Swooper.adf

tlosm · 27 September 2015, 10:51

Quote:

Originally Posted by Hypex

Hi Frode. Yes I did see a crash in some audio code and as pointed out in post #5 it was coming from here:
0 OpenAL 0x04203fe4 OALSource::GetQLength() + 48

I looked for options to disable audio but could see none.

I setup a basic A500 emulation as below that booted with AROS Kickstart. At first it seemed to be okay after some thinking but then started thinking endlessly and locked the cursor. I was able to break it. It didn't crash but it did hang.

amiga_model = A500
floppy_drive_0 = Swooper.adf

Try to check if the openal is installed in your frameworks and sdl2 too ;-)

Hypex · 27 September 2015, 15:47

Quote:

Originally Posted by tlosm

Try to check if the openal is installed in your frameworks and sdl2 too ;-)

I don't have OpenAL in my Frameworks. Only SDL, I presume to be SDL 1.

But it is opening OpenAL from somewhere since it did crash in an OpenAL method. ;-)

The hangs appear to be different.

12 May 2015, 15:30	#1
Hypex Registered User Join Date: May 2015 Location: Australia Posts: 131	FS-UAE keeps crashing on OSX PPC Hello. My first post and EAB has already annoyed me by asking me to log in when I wanted to preview it and deleted my entire post! And then did it again after retyping!!! Argh!!! Anyway similar to what I first put, I find the PPC build of FS-UAE very unstable and it keeps crashing on me. If it doesn't crash then it starts up and starts "thinking" and I have to force quit it. Recently it has locked up OSX straight up with no cursor but using force close hot key eventually brings it back. At times it did work it would eventually loop around thinking again and I have to kill it. I'm using the tlosm dev29 build here: https://www.dropbox.com/s/6k29r9zt9b...Dev29.zip?dl=0 My config file: amiga_model = A4000/OS4 kickstart_file = Kickstart (A4000T).rom uae_cpuboard_rom_file = cyberstormppc_44.71.rom hard_drive_0_controller = ide0 hard_drive_0 = OS4.hdf cdrom_image_0 = ClassicInstallCD.iso floppy_drive_0 = BootFloppy.adf uae_a2065 = slirp automatic_input_grab = 0 fullscreen = 0 slirp_implementations = qemu Related crash: Process: fs-uae [16853] Path: /Users/Damien/Desktop/Amiga/UAE/fs-uaeDev29/FS-UAE.app/Contents/MacOS/fs-uae Identifier: no.fengestad.fs-uae Version: 2.5.29 (2.5.29) Code Type: PPC (Native) Parent Process: launchd [67] Interval Since Last Report: 20977 sec Crashes Since Last Report: 1 Per-App Interval Since Last Report: 2366 sec Per-App Crashes Since Last Report: 1 Date/Time: 2015-05-12 20:33:37.664 +1000 OS Version: Mac OS X 10.5.8 (9L31a) Report Version: 6 Anonymous UUID: 4C329DC0-59E3-442D-A263-B83AE0E0D445 Exception Type: EXC_BAD_ACCESS (SIGSEGV) Exception Codes: KERN_INVALID_ADDRESS at 0x00000000c0000003 Crashed Thread: 6 Thread 0: 0 libSystem.B.dylib 0x94b110f8 semaphore_timedwait_signal_trap + 8 1 libSystem.B.dylib 0x94b54214 _pthread_cond_wait + 1320 2 libglib-2.0.0.dylib 0x04138f38 g_cond_wait_until + 312 3 no.fengestad.fs-uae 0x0089bb2c 0x2000 + 9018156 4 no.fengestad.fs-uae 0x00899950 0x2000 + 9009488 5 no.fengestad.fs-uae 0x0089b038 0x2000 + 9015352 6 no.fengestad.fs-uae 0x0088559c 0x2000 + 8926620 7 no.fengestad.fs-uae 0x0000cfd0 0x2000 + 45008 8 no.fengestad.fs-uae 0x000079d0 0x2000 + 22992 Thread 1: 0 libSystem.B.dylib 0x94b17a8c __semwait_signal + 12 1 libSystem.B.dylib 0x94b54318 _pthread_cond_wait + 1580 2 SDL2 0x04291bcc SDL_CondWaitTimeout + 176 3 SDL2 0x042922fc SDL_SemWaitTimeout + 156 4 SDL2 0x042932b4 SDL_RemoveTimer + 1152 5 SDL2 0x042928c4 SDL_TLSGet + 164 6 SDL2 0x0429260c SDL_ThreadID + 644 7 libSystem.B.dylib 0x94b52f70 _pthread_start + 316 Thread 2: 0 libSystem.B.dylib 0x94b17a88 __semwait_signal + 8 1 libSystem.B.dylib 0x94b1789c nanosleep$UNIX2003 + 188 2 no.fengestad.fs-uae 0x00898cc4 0x2000 + 9006276 3 libglib-2.0.0.dylib 0x04119120 g_thread_proxy + 160 4 libSystem.B.dylib 0x94b52f70 _pthread_start + 316 Thread 3: 0 libSystem.B.dylib 0x94b11078 mach_msg_trap + 8 1 libSystem.B.dylib 0x94b17f9c mach_msg + 56 2 com.apple.CoreFoundation 0x924f7394 CFRunLoopRunSpecific + 1812 3 com.apple.audio.CoreAudio 0x925d5fa8 HALRunLoop::OwnThread(void) + 212 4 com.apple.audio.CoreAudio 0x925d5de4 CAPThread::Entry(CAPThread) + 104 5 libSystem.B.dylib 0x94b52f70 _pthread_start + 316 Thread 4: 0 libSystem.B.dylib 0x94b110f8 semaphore_timedwait_signal_trap + 8 1 libSystem.B.dylib 0x94b54214 _pthread_cond_wait + 1320 2 com.apple.audio.CoreAudio 0x925e79f0 HP_IOThread::WorkLoop() + 488 3 com.apple.audio.CoreAudio 0x925e77f0 HP_IOThread::ThreadEntry(HP_IOThread) + 12 4 com.apple.audio.CoreAudio 0x925d5de4 CAPThread::Entry(CAPThread) + 104 5 libSystem.B.dylib 0x94b52f70 _pthread_start + 316 Thread 5: 0 libSystem.B.dylib 0x94b17a8c __semwait_signal + 12 1 libSystem.B.dylib 0x94b54318 _pthread_cond_wait + 1580 2 libGLProgrammability.dylib 0x92a8ab68 glvmDoWork + 120 3 libSystem.B.dylib 0x94b52f70 _pthread_start + 316 Thread 6 Crashed: 0 OpenAL 0x04203fe4 OALSource::GetQLength() + 48 1 OpenAL 0x041fa588 alSourceUnqueueBuffers + 180 2 no.fengestad.fs-uae 0x000acd18 0x2000 + 699672 3 no.fengestad.fs-uae 0x000e81a8 0x2000 + 942504 4 no.fengestad.fs-uae 0x0016c8e0 0x2000 + 1485024 5 no.fengestad.fs-uae 0x0016d030 0x2000 + 1486896 6 no.fengestad.fs-uae 0x0016c388 0x2000 + 1483656 7 no.fengestad.fs-uae 0x0014c738 0x2000 + 1353528 8 no.fengestad.fs-uae 0x00180ff4 0x2000 + 1568756 9 no.fengestad.fs-uae 0x0000bb38 0x2000 + 39736 10 no.fengestad.fs-uae 0x008852a0 0x2000 + 8925856 11 libglib-2.0.0.dylib 0x04119120 g_thread_proxy + 160 12 libSystem.B.dylib 0x94b52f70 _pthread_start + 316 Thread 7: 0 libSystem.B.dylib 0x94b17a8c __semwait_signal + 12 1 libSystem.B.dylib 0x94b54318 _pthread_cond_wait + 1580 2 SDL2 0x04291bcc SDL_CondWaitTimeout + 176 3 SDL2 0x042922fc SDL_SemWaitTimeout + 156 4 no.fengestad.fs-uae 0x001313dc 0x2000 + 1242076 5 libglib-2.0.0.dylib 0x04119120 g_thread_proxy + 160 6 libSystem.B.dylib 0x94b52f70 _pthread_start + 316 Thread 6 crashed with PPC Thread State 32: srr0: 0x04203fe4 srr1: 0x0000d030 dar: 0xc0000003 dsisr: 0x40000000 r0: 0x041fa58c r1: 0xf030a5d0 r2: 0xc0000003 r3: 0x08ca56c0 r4: 0x00000000 r5: 0x00000000 r6: 0xffffffff r7: 0x00000001 r8: 0xfa000000 r9: 0x00000008 r10: 0x00000004 r11: 0x08ca1aa0 r12: 0x94b1b734 r13: 0x00969fe0 r14: 0x0097390c r15: 0x00b4390c r16: 0x0097390c r17: 0x0097390c r18: 0x00b48d94 r19: 0x0097390c r20: 0x3b9aca00 r21: 0x0097390c r22: 0x00013cf3 r23: 0x01ea38c8 r24: 0x00000000 r25: 0x01f12c6c r26: 0x08ca56c0 r27: 0x00000004 r28: 0xf030a66c r29: 0x00000001 r30: 0x08ca4dd0 r31: 0x041fa4e8 cr: 0x24028224 xer: 0x20000001 lr: 0x041fa598 ctr: 0x94b1b734 vrsave: 0x00000000

13 May 2015, 17:27	#3
Hypex Registered User Join Date: May 2015 Location: Australia Posts: 131	I'm telling you that's not it! :-) I set up that ages ago, confirmed it later on and somehow managed to confirm it again tonight. Before it crashed. All my LUNs are off. Makes no difference here. Perhaps they should fix their code so you can recompile a more stable version. ;-) What I don't understand is how an emulation of a SCSI controller on a big endian CPU can be broke on a real big endian CPU! :-? Either the SCSI controller memory map is native little endian or the code is sloppy! Either way doesn't look good. But it shouldn't matter. :-) Last edited by Hypex; 14 May 2015 at 08:34.

13 May 2015, 20:33	#4
tlosm Registered User Join Date: Sep 2014 Location: italy Posts: 189	Because the scsi emulation was did by qemu and it is not the best ;-) Uae is much more better in loading and transfer rate but is not compatible with Os4 Hypex can you sent me your email address thru pm? need to send you something for check i think the problem is in the kickstart of amiga 4000t... it try to access by kickstart at the scsi.device and make the endianess exit. the amiga 4000t kickstart have the scsi.device in rom. Last edited by tlosm; 14 May 2015 at 07:12.

15 May 2015, 17:08	#9
FrodeSolheim FS-UAE Developer Join Date: Dec 2011 Location: Førde, Norway Age: 43 Posts: 4,043	Regarding config, there's a peculiarity with the ROM config in the latest development versions, so instead of: Code: uae_cpuboard_rom_file = cyberstormppc_44.71.rom ... better use this FS-UAE option instead: Code: accelerator_rom = cyberstormppc_44.71.rom Also, do not use the following config: Code: hard_drive_0_controller = ide0 ... as the hard drive will by default (without this option) be connected to the accelerator SCSI interface. Code: slirp_implementations = qemu slirp_implementations is a mis-spelling of slirp_implementation. You also do not need to specify this option anyway.

17 May 2015, 06:43	#11
Hypex Registered User Join Date: May 2015 Location: Australia Posts: 131	Thanks for the tips Frode. I changed it to suit. Unfortunately it just made it crash! So I removed the 2065 options altogether and it was better again. But then later it crashed again. Seriously, this audio bug needs to be fixed. Does it only affect OSX PPC? Because with it hanging around I cannot trust FS-UAE, it can and does crash at any moment. I removed the ide0 setting but had to put it back as it defaulted to the CPU board. Also I have some good news. I got the CDROM working! I had to use an option I made up that doesn't exist on the FS-UAE page! But it works. This. :-) cdrom_drive_0_controller = ide1 So OS4 started to boot off CD. And then it crashed on a red and blue screen. Sigh... But, it all ended in tears. I made the mistake of telling it to go full screen and grab input. Well it took over the screen and then went into an endless loop. I couldn't force it out. I had to force shutdown the computer. Well turn it off really. Typical, Everything works on Intel. But on OSX PPC everything breaks. I'm so over it! Last edited by Hypex; 18 May 2015 at 14:21.

12 May 2015, 18:10	#2
tlosm Registered User Join Date: Sep 2014 Location: italy Posts: 189	Hypex! it is the scsi endianess i was explain it in the amigaworld thread :http://amigaworld.net/modules/newbb/...order=0#755982 push fast ESC on the keyboard after set the scsi luns on the cyberstorm bios all set to 0 ... save and reset the emulation. and everything will work.

14 May 2015, 15:08	#5
Hypex Registered User Join Date: May 2015 Location: Australia Posts: 131	What about the IDE emulation? As you could see I changed to cdrom_image_0. As the other cdrom option was for CD32 and CDTV. But the crash occurs here. So looks like an audio problem. 0 OpenAL 0x04203fe4 OALSource::GetQLength() + 48

14 May 2015, 21:56	#6
tlosm Registered User Join Date: Sep 2014 Location: italy Posts: 189	let me know after my email pls

15 May 2015, 11:36	#7
Hypex Registered User Join Date: May 2015 Location: Australia Posts: 131	So I checked with a few Kickstart ROMs. For some reason it tends to get caught up with some ROMs and goes in an endless loop or crashes. Whatever that is about. But it all ends up the same. The OS4 boot floppy never finds the CDROM. I don't know why! Is there an obvious error in my config? It would help if the log said file not found but it seems to skip things.

15 May 2015, 14:26	#8
tlosm Registered User Join Date: Sep 2014 Location: italy Posts: 189	i think everything depend of th fsuae ide cdrom config... the only one who can help about is Mr Frode . The other solution is use a hdf make by other machine (x86) and see if everything working... i did like this make installation from fsuae on macbook pro and use the hdf on Quad G5

15 May 2015, 18:21	#10
tlosm Registered User Join Date: Sep 2014 Location: italy Posts: 189	Frode if not on Ide0 setting the fsuae crash because the indianess of SCSI on PPC port .

17 May 2015, 13:20	#12
FrodeSolheim FS-UAE Developer Join Date: Dec 2011 Location: Førde, Norway Age: 43 Posts: 4,043	The reason the it works when you use cdrom_drive_0_controller = ide1 is that with this option, AmigaOS will see the CD device. But, since the problem was probably, and still is, that the CyberStorm device isn't fully initialized, it fails to switch to PPC mode. If the CyberStorm device had been initialized, it would have booted properly in the first place. As for PPC hosts, I do not officially support this. This means that I do not don't test that stuff works on PPC hosts, but I do not try to prevent it from working there either. I'll fix PPC host issues if they are trivial, and other people can easily test it, but I won't do any real/hard debugging to solve PPC issues. This will have to be done by other skilled volunteers, if any!

17 May 2015, 20:34	#13
tlosm Registered User Join Date: Sep 2014 Location: italy Posts: 189	Frode on me everything working right , like to some others guys who send my build of FS-Uae ... i think there is something missed on the Hypex system or a miss configuration issue. The G5 is really hard to optimize because full big endian .. i think on G4 dual (because little and big endian) or on CoriQ / Power.x serie (because only little endian) problably everything can work right

18 May 2015, 14:47	#14
Hypex Registered User Join Date: May 2015 Location: Australia Posts: 131	In any case, miss configured or not, a crash should not happen when interpreting a text file. If a file is missing it should be reported. If it is a binary it should be checked, most likely by the OS, to see if it loads. Using different Kickstart ROM files also causes it to be unstable. An incorrect ROM crashing in the emulation/Amiga side is a possible and acceptable outcome, but crashing the emulator itself is not an acceptable outcome pointing to a bug. Something is obviously tripping it up. And as demonstrated from my first post, that something is calling the OpenAL audio API functions and it's always thread 6. However, it really needs FS-UAE debugging symbols, otherwise it's too cryptic. . Perhaps the QEMU SCSI code should be patched. PPC is bi-endian and it's ridiculous that little-endian specific code doesn't work. I know hacking little-endian specific C code to be endian agnostic is a bitch. In my experience a program can appear to work okay bur won't crash. Here it looks looks like it has more dependance on things being backwards. :-)

26 May 2015, 11:48	#15
Hypex Registered User Join Date: May 2015 Location: Australia Posts: 131	Well I've tried with ROM tlsom has. But it never even boots here. FS-UAE just hangs on a black screen and I have to force close it. So went back to the ROM I had. I also have an A4000T ROM that acred the same. Something is wrong. It's almost acting as if the PPC CPU is missing or broken. So I wonder if qemu is setting up the PPC emulation at all. I tried PPC test disk which supports both PUP and WOS. Trying either one freezes the emulation. Or locks up the virtual Amiga anyway. Sp here is my current config with other settings I've tried. Anything stand out. Is a simple setting crashing it? Well to the point it can run without a crash since I'm still testing it with an unstable build. BTW OS4.hdf is blank formatted HD image from HiToro. amiga_model = A4000/OS4 kickstart_file = amiga-os-310-a4000.rom #kickstart_file = Kickstart v3.1 rev 40.70 (1993)(Commodore)(A4000).rom accelerator_rom = cyberstormppc_44.71.rom hard_drive_0_controller = ide0 #hard_drive_1_controller = ide1 hard_drive_0 = OS4.hdf cdrom_drive_0_controller = ide1 cdrom_image_0 = ClassicInstallCD-53.69.iso #floppy_drive_0 = BootFloppy.adf floppy_drive_0 = ppctest.adf #uae_a2065 = slirp automatic_input_grab = 0 fullscreen = 0 #slirp_implementations = qemu

29 August 2015, 17:40	#17
FrodeSolheim FS-UAE Developer Join Date: Dec 2011 Location: Førde, Norway Age: 43 Posts: 4,043	By the way, the initial crash report from your first post indicates a crash/bug in the FS-UAE OpenAL audio output code, and not in PPC emulation. Does FS-UAE only crash when you try to run PPC emulation, or does it always crash (including when running stock A500 model)?

25 September 2015, 15:33	#18
Hypex Registered User Join Date: May 2015 Location: Australia Posts: 131	Hi Frode. Yes I did see a crash in some audio code and as pointed out in post #5 it was coming from here: 0 OpenAL 0x04203fe4 OALSource::GetQLength() + 48 I looked for options to disable audio but could see none. I setup a basic A500 emulation as below that booted with AROS Kickstart. At first it seemed to be okay after some thinking but then started thinking endlessly and locked the cursor. I was able to break it. It didn't crash but it did hang. amiga_model = A500 floppy_drive_0 = Swooper.adf

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
FS-UAE for G4 ppc OSX 10.5.8	elowan	support.FS-UAE	35	10 November 2014 19:22
OSX (PPC) build and CD32 audio tracks	snakecoils	support.FS-UAE	9	28 March 2014 00:04
[Solved] FS-UAE 2.1.10 build fails on OSX PPC	mr_a500	support.FS-UAE	6	12 April 2013 18:40
FS-UAE (OSX-PPC) - No OSD menu?	snakecoils	support.FS-UAE	10	22 March 2013 22:51
FS-UAE 2.0: How make an OSX PPC build? [Solved]	snakecoils	support.FS-UAE	8	04 November 2012 17:06