weixin_39831242
weixin_39831242
2021-01-03 09:55

Crashing on any game [non-OIT] [ARM]

Updated to the latest lr-reicast 500393a and retroarch master branch. When loading a game, it just crashes. I don't get the mapped controller notification on bottom left and does not get to the Naomi boot.

Previous compiled version 127f845 works fine with the same setup.


Parameters: 
Executing: /opt/retropie/emulators/retroarch/bin/retroarch -L /opt/retropie/libretrocores/lr-reicast/reicast_libretro.so --config /opt/retropie/configs/naomi/retroarch.cfg "/home/pigaming/RetroPie/roms/naomi/Shikigami No Shiro II.bin" --appendconfig /dev/shm/retroarch.cfg
[reicast]: Got size: 640 x 480.
Creating dir: /home/pigaming/RetroPie/BIOS/dc/data
ARM VFP-Run Fast (NFP) enabled !
Linux paging: 00001000 00001000 00000FFF
MAP 00800000 w/ 50331648
MAP 20000000 w/ 50331648
MAP 04000000 w/ 33554432
MAP 06000000 w/ 33554432
MAP 0C000000 w/ 0
A8
vmem reserve: base: 90CB0000, aram: b0cb0000, vram: 94CB0000, ram: 9CCB0000
Resetting mem
Mem alloc successful!Loaded /home/pigaming/RetroPie/BIOS/dc/naomi_boot.bin as bootrom

flash/nvmem is missing, will create new file...Sh4 Reset
recSh4 Init
Freeing fpcb
readm helpers: up to B1782250
nullDC-Naomi rom loader v1.2
File: /home/pigaming/RetroPie/roms/naomi/Shikigami No Shiro II.bin
+1 romfiles, 112.30 MB set size, 112.30 MB set address space
NAOMI GAME ID [SHIKIGAMI NO SHIRO 2]

Mapped ROM Successfully !

EEPROM file : /home/pigaming/RetroPie/roms/naomi/Shikigami No Shiro II.bin.eeprom

     ARM7_TCB addr: 0xb2a82000 | from: 0xb2a81178 | addr here: 0xb2a497c8
Sh4 Reset
reios: IP.bin is ''
reios: Hardware ID is: 
reios: Maker ID is:    
reios: Device info is: 
reios: Area symbols is: 
reios: Peripherals is: 
reios: Product number is: 
reios: Product version is: 
reios: Release date is: 
reios: Boot filename is: 
reios: Software company is: 
reios: Software name is: 
reios: loading ip.bin from FAD: 0
reios: bootfile is ''
context_reset.
/opt/retropie/supplementary/runcommand/runcommand.sh: line 1008:  9917 Segmentation fault      /opt/retropie/emulators/retroarch/bin/retroarch -L /opt/retropie/libretrocores/lr-reicast/reicast_libretro.so --config /opt/retropie/configs/naomi/retroarch.cfg "/home/pigaming/RetroPie/roms/naomi/Shikigami No Shiro II.bin" --appendconfig /dev/shm/retroarch.cfg

该提问来源于开源项目:libretro/flycast

  • 点赞
  • 写回答
  • 关注问题
  • 收藏
  • 复制链接分享
  • 邀请回答

35条回答

  • weixin_39831242 weixin_39831242 4月前

    So do I modify the makefile like?

    CFLAGS += -g -mcpu=cortex-a15 -mtune=cortex-a15.cortex-a7 -mfpu=neon-vfpv4 -mvectorize-with-neon-quad

    From: Galileo Morales Sent: Thursday, November 1, 2018 10:18 AM To: libretro/reicast-emulator; libretro/reicast-emulator Cc: Author Subject: Re: [libretro/reicast-emulator] Crashing on any game [non-OIT] [ARM] (#365)

    Yeah I need help on compiling so I can get you a debug log. Is it a flag I add on the make command?

    From: flyinghead Sent: Thursday, November 1, 2018 10:06 AM To: libretro/reicast-emulator Cc: 6alileo; Author Subject: Re: [libretro/reicast-emulator] Crashing on any game [non-OIT] [ARM] (#365)

    All this points to a problem due to the dc_init() change but we need to understand why it's failing on this platform.

    — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/libretro/reicast-emulator/issues/365#issuecomment-435107982, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AoSfFNT8aOlODs_jAp96fjFZnQX3wWG0ks5uqyovgaJpZM4YBMKd.

    点赞 评论 复制链接分享
  • weixin_39640646 weixin_39640646 4月前

    Build with make DEBUG=1 That should be enough.

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    OK, I've compiled with make DEBUG=1 platform=odroid BOARD="ODROID-XU3" ARCH=arm and ran the gdb command above. See this photo

    点赞 评论 复制链接分享
  • weixin_39640646 weixin_39640646 4月前

    Then type the following commands in gdb:

    
    handle SIGSEGV nostop noprint
    run
    

    Wait for the crash, then:

    
    bt
    

    and report. Thanks

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    I launched a diff game and removed any spaces to avoid using quotes. I was able to run it on terminal via

    /opt/retropie/emulators/retroarch/bin/retroarch -L /opt/retropie/libretrocores/lr-reicast/reicast_libretro.so --config /opt/retropie/configs/atomiswave/retroarch.cfg /home/pigaming/RetroPie/roms/atomiswave/JamboSafari.bin

    Then do gdb --args ... then handle ... then run and it just stops. See below.

    
    pigaming:~$ gdb --args /opt/retropie/emulators/retroarch/bin/retroarch -L /opt/retropie/libretrocores/lr-reicast/reicast_libretro.so --config /opt/retropie/configs/atomiswave/retroarch.cfg /home/pigaming/RetroPie/roms/atomiswave/JamboSafari.bin
    GNU gdb (Ubuntu 8.2-0ubuntu1~16.04.1) 8.2
    Copyright (C) 2018 Free Software Foundation, Inc.
    License GPLv3+: GNU GPL version 3 or later <http:>
    This is free software: you are free to change and redistribute it.
    There is NO WARRANTY, to the extent permitted by law.
    Type "show copying" and "show warranty" for details.
    This GDB was configured as "arm-linux-gnueabihf".
    Type "show configuration" for configuration details.
    For bug reporting instructions, please see:
    <http:></http:>.
    Find the GDB manual and other documentation resources online at:
        <http:></http:>.
    
    For help, type "help".
    Type "apropos word" to search for commands related to "word"...
    Reading symbols from /opt/retropie/emulators/retroarch/bin/retroarch...(no debugging symbols found)...done.
    (gdb) handle SIGSEGV nostop noprint
    Signal        Stop      Print   Pass to program Description
    SIGSEGV       No        No      Yes             Segmentation fault
    (gdb) run
    Starting program: /opt/retropie/emulators/retroarch/bin/retroarch -L /opt/retropie/libretrocores/lr-reicast/reicast_libretro.so --config /opt/retropie/configs/atomiswave/retroarch.cfg /home/pigaming/RetroPie/roms/atomiswave/JamboSafari.bin
    
    [11]+  Stopped                 gdb --args /opt/retropie/emulators/retroarch/bin/retroarch -L /opt/retropie/libretrocores/lr-reicast/reicast_libretro.so --config /opt/retropie/configs/atomiswave/retroarch.cfg /home/pigaming/RetroPie/roms/atomiswave/JamboSafari.bin
    </http:>
    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    Just to add to this - I even compiled retroarch using make DEBUG=1 GL_DEBUG=1 and it outputted (no debugging symbols found)

    点赞 评论 复制链接分享
  • weixin_39640646 weixin_39640646 4月前

    I don't get it. Why would it stop? Are you pressing ctrl+Z? Running in the background with & ??

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    Yeah that's puzzling for me too. I didn't press anything on keyboard or have it running in the background. It also happens instantly as soon as I enter the run command, it stops right away.

    点赞 评论 复制链接分享
  • weixin_39640646 weixin_39640646 4月前

    Are you sure /opt/retropie/emulators/retroarch/bin/retroarch is not a shell script? Now that you mention it I remember there is that runcommand script on Retropie that redirect inputs. It must be getting in the way.

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    Yes I'm sure it's not a shell script. Runcommand is a shell script that gathers different paths and puts into a single long command to launch the game. The configs from emulationstation are not passed along. RA has its own config files.

    The only thing I've noticed is when context_reset. is called earlier rather than later, it works.

    Could you maybe explain why there's only 1 dc_init after dc_prepare_system()? What is working for this platform is the two diff dc_init codes earlier in the code.

    https://github.com/6alileo/reicast-emulator/commit/c317b82197d3ae078643eeb734b016fdf734df6a

    点赞 评论 复制链接分享
  • weixin_39640646 weixin_39640646 4月前

    My guess is retroarch is trying to open the console tty input or something like that and it's stopped because there's already a shell on this tty. The dc_init function was previously called when the emulator was starting (either in retro_run or in the emulator thread depending on whether threaded rendering is on) This causes problems when save state are loaded very early (auto save state) because dc_init does some unconventional things with memory at startup. It is now called in retro_load_game, before the game starts.

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    Hi - would it help if I sent you an XU4 board and power supply? If you could, please DM me on Discord using the same name here ``

    点赞 评论 复制链接分享
  • weixin_39640646 weixin_39640646 4月前

    I can't find you either in the retroarch or reicast Discord servers. Can you me there?

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    Hi there - I have an idea. Would it be possible to create a flag called LAZYLOAD or something to that nature where if it set to LAZYLOAD=1 it would compile libretro.cpp as below?

    https://github.com/6alileo/reicast-emulator/commit/c317b82197d3ae078643eeb734b016fdf734df6a

    I have tested it, even with the mame m2 updates and it works

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    Something in emulationstation caused this crap.

    点赞 评论 复制链接分享
  • weixin_39729784 weixin_39729784 4月前

    I'm hitting the same problem, what was it you had to change in emulationstation?

    点赞 评论 复制链接分享
  • weixin_39640646 weixin_39640646 4月前

    Does it happen also with dreamcast games?

    点赞 评论 复制链接分享
  • weixin_39640646 weixin_39640646 4月前

    Another idea: do you have more than 4 (virtual) controllers connected? Try to use only 4 to see if it makes a difference.

    点赞 评论 复制链接分享
  • weixin_39640646 weixin_39640646 4月前

    I have tested Android, Linux x64 and Linux arm (RPi) without issue. I will need a verbose log or a stack trace to figure it out.

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    It happens on Dreamcast as well. I only have 1 controller connected via Bluetooth during the logs below:

    Naomi: https://pastebin.com/raw/yUqTLt63 Dreamcast: https://pastebin.com/raw/k9GsKcH9

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    Here's another comparison of verbose working vs non-working of the same game (Daytona USA for DC).

    127f845: https://pastebin.com/raw/eaRMW30e 500393a: https://pastebin.com/raw/HfxAMPN5

    Doing a text compare, I notice that the block below gets created before [INFO] Skipping SRAM load.. on the latest version while almost the same block gets created after [INFO] [PulseAudio]: Unpausing. on the previous version.

    
    Creating dir: /home/pigaming/RetroPie/BIOS/dc/data
    ARM VFP-Run Fast (NFP) enabled !
    Linux paging: 00001000 00001000 00000FFF
    MAP 00800000 w/ 25165824
    MAP 20000000 w/ 25165824
    MAP 04000000 w/ 16777216
    MAP 06000000 w/ 16777216
    MAP 0C000000 w/ 0
    A8
    vmem reserve: base: 911D0000, aram: b11d0000, vram: 951D0000, ram: 9D1D0000
    Resetting mem
    Mem alloc successful!Loaded /home/pigaming/RetroPie/BIOS/dc/dc_boot.bin as bootrom
    
    Loaded /home/pigaming/RetroPie/BIOS/dc/dc_nvmem.bin as nvram
    
    [libretro INFO] Using Recompiler
    Sh4 Reset
    recSh4 Init
    Freeing fpcb
    readm helpers: up to B16D0250
    
    Session 1 has 1 track(s)
    Saving  Track:  1  Type: Mode2/2336  Pregap: 150  Size: 57857   LBA: 0       
    Track position: 350400
    
    Session 2 has 1 track(s)
    Saving  Track:  2  Type: Mode2/2336  Pregap: 150  Size: 289537  LBA: 69257   
    Track position: 135854752
    gdrom: Opened image "/home/pigaming/RetroPie/roms/dreamcast/Daytona USA (USA).cdi"
    
         ARM7_TCB addr: 0xb29d0000 | from: 0xb29cf178 | addr here: 0xb29977c8
    Initializing VMU data...
    Initializing VMU data...
    Initializing VMU data...
    Initializing VMU data...
    Sh4 Reset
    reios: IP.bin is 'SEGA SEGAKATANA SEGA ENTERPRISES2BD2 GD-ROM1/1  JUE     2799A10 MK-51037  V1.00420010214        1ST_READ.BIN    themacN         DAYTONAUSA                                                                                                                      /home/pigaming/RetroPie/BIOS/dc/'
    reios: Hardware ID is: SEGA SEGAKATANA 
    reios: Maker ID is:    SEGA ENTERPRISES
    reios: Device info is: 2BD2 GD-ROM1/1  
    reios: Area symbols is: JUE     
    reios: Peripherals is: 2799A10 
    reios: Product number is: MK-51037  
    reios: Product version is: V1.004
    reios: Release date is: 20010214        
    reios: Boot filename is: 1ST_READ.BIN    
    reios: Software company is: themacN         
    reios: Software name is: DAYTONAUSA                                                                                                                      
    reios: loading ip.bin from FAD: 69407
    reios: bootfile is '1ST_READ.BIN    '
    reios: iso9660 PVD found
    reios: iso9660 root_directory, FAD: 69430, len: 28672
    Found 1ST_READ.BIN at 000065
    filename len: 14
    file LBA: 357660
    file LEN: 2010720
    [libretro INFO] [LUT]: Product number: MK-51037  .
    
    点赞 评论 复制链接分享
  • weixin_39640646 weixin_39640646 4月前

    Yes, there has been changes regarding when the core initialization is done. Previously it was done lazily, when the game was starting up. It is now done when the game is being loaded. I'm thinking this issue could be a consequence of these changes but no luck in reproducing it and so far no other user reported it. Any way you could get a stacktrace with gdb? Not sure if -g is enabled by default to get symbols.

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    If I run this manually, it runs on top of EmulationStation very slow with glitches but it does crash when running using runcommand on RetroPie

    /opt/retropie/emulators/retroarch/bin/retroarch -L /opt/retropie/libretrocores/lr-reicast/reicast_libretro.so --config /opt/retropie/configs/naomi/retroarch.cfg "/home/pigaming/RetroPie/roms/naomi/Shikigami No Shiro II.bin" --appendconfig /dev/shm/retroarch.cfg

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    Retropie uses runcommand to execute $COMMAND which is the command above this comment.

    Question - how can I run gdb on this line?

    eval $COMMAND </dev/tty &>>"$LOG"

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    I've noticed something in the logs. On the working one, the order is as follows

    
    [INFO] Written to playlist file: /opt/retropie/configs/dreamcast/content_history.lpl
    Creating dir: /home/pigaming/RetroPie/BIOS/dc/data
    context_reset.
    ARM VFP-Run Fast (NFP) enabled !
    

    The non-working one changes the order of context_reset.

    Can you maybe put it back in between

    Creating dir: /home/pigaming/RetroPie/BIOS/dc/data

    and

    ARM VFP-Run Fast (NFP) enabled !?

    Also just to throw this out there, the working one gets to creating the content_history prior to creating the dir

    [INFO] Written to playlist file: /opt/retropie/configs/dreamcast/content_history.lpl

    Below is the non-working. Notice the order of where context_reset. is

    
    Creating dir: /home/pigaming/RetroPie/BIOS/dc/data
    ARM VFP-Run Fast (NFP) enabled !
    Linux paging: 00001000 00001000 00000FFF
    MAP 00800000 w/ 25165824
    MAP 20000000 w/ 25165824
    MAP 04000000 w/ 16777216
    MAP 06000000 w/ 16777216
    MAP 0C000000 w/ 0
    A8
    vmem reserve: base: 911D0000, aram: b11d0000, vram: 951D0000, ram: 9D1D0000
    Resetting mem
    Mem alloc successful!Loaded /home/pigaming/RetroPie/BIOS/dc/dc_boot.bin as bootrom
    
    Loaded /home/pigaming/RetroPie/BIOS/dc/dc_nvmem.bin as nvram
    
    [libretro INFO] Using Recompiler
    Sh4 Reset
    recSh4 Init
    Freeing fpcb
    readm helpers: up to B16D0250
    
    Session 1 has 1 track(s)
    Saving  Track:  1  Type: Mode2/2336  Pregap: 150  Size: 57857   LBA: 0       
    Track position: 350400
    
    Session 2 has 1 track(s)
    Saving  Track:  2  Type: Mode2/2336  Pregap: 150  Size: 289537  LBA: 69257   
    Track position: 135854752
    gdrom: Opened image "/home/pigaming/RetroPie/roms/dreamcast/Daytona USA (USA).cdi"
    
         ARM7_TCB addr: 0xb29d0000 | from: 0xb29cf178 | addr here: 0xb29977c8
    Initializing VMU data...
    Initializing VMU data...
    Initializing VMU data...
    Initializing VMU data...
    Sh4 Reset
    reios: IP.bin is 'SEGA SEGAKATANA SEGA ENTERPRISES2BD2 GD-ROM1/1  JUE     2799A10 MK-51037  V1.00420010214        1ST_READ.BIN    themacN         DAYTONAUSA                                                                                                                      /home/pigaming/RetroPie/BIOS/dc/'
    reios: Hardware ID is: SEGA SEGAKATANA 
    reios: Maker ID is:    SEGA ENTERPRISES
    reios: Device info is: 2BD2 GD-ROM1/1  
    reios: Area symbols is: JUE     
    reios: Peripherals is: 2799A10 
    reios: Product number is: MK-51037  
    reios: Product version is: V1.004
    reios: Release date is: 20010214        
    reios: Boot filename is: 1ST_READ.BIN    
    reios: Software company is: themacN         
    reios: Software name is: DAYTONAUSA                                                                                                                      
    reios: loading ip.bin from FAD: 69407
    reios: bootfile is '1ST_READ.BIN    '
    reios: iso9660 PVD found
    reios: iso9660 root_directory, FAD: 69430, len: 28672
    Found 1ST_READ.BIN at 000065
    filename len: 14
    file LBA: 357660
    file LEN: 2010720
    [libretro INFO] [LUT]: Product number: MK-51037  .
    [INFO] Skipping SRAM load..
    [INFO] Version of libretro API: 1
    [INFO] Compiled against API: 1
    [INFO] [Audio]: Set audio input rate to: 44055.94 Hz.
    [INFO] [Video]: Video @ 640x480
    [INFO] [Video]: Using HW render, OpenGL driver forced.
    [INFO] [EGL] Falling back to eglGetDisplay
    [INFO] [EGL]: EGL version: 1.4
    [INFO] [GL]: Found GL context: mali-fbdev
    [INFO] [GL]: Detecting screen resolution 0x0.
    [INFO] [EGL]: Created shared context: 0x6422d0.
    [INFO] [EGL]: Current context: 0x5f5b60.
    [INFO] [GL]: Vendor: ARM, Renderer: Mali-T628.
    [INFO] [GL]: Version: OpenGL ES 3.1 v1.r12p0-04rel0.03af15950392f3702b248717f4938b82.
    [INFO] [GL]: Using resolution 1920x1200
    [INFO] [GL]: Default shader backend found: glsl.
    [INFO] [Shader driver]: Using GLSL shader backend.
    [WARN] [GL]: Stock GLSL shaders will be used.
    [INFO] [GLSL]: Found GLSL vertex shader.
    [INFO] [GLSL]: Found GLSL fragment shader.
    [INFO] [GLSL]: Linking GLSL program.
    [INFO] [GLSL]: Found GLSL vertex shader.
    [INFO] [GLSL]: Found GLSL fragment shader.
    [INFO] [GLSL]: Linking GLSL program.
    [INFO] [GLSL]: Found GLSL vertex shader.
    [INFO] [GLSL]: Found GLSL fragment shader.
    [INFO] [GLSL]: Linking GLSL program.
    [INFO] Resetting shader to defaults ... 
    [INFO] [GL]: Using 1 textures.
    [INFO] [GL]: Loaded 1 program(s).
    [INFO] [GL]: Initializing HW render (1024 x 1024).
    [INFO] [GL]: Max texture size: 8192 px, renderbuffer size: 8192 px.
    [INFO] [GL]: Supports FBO (render-to-texture).
    [INFO] [Font]: Using font rendering backend: freetype.
    [INFO] [Video]: Graphics driver did not initialize an input driver. Attempting to pick a suitable driver.
    [WARN] [udev]: Couldn't open any keyboard, mouse or touchpad. Are permissions set correctly for /dev/input/event*?
    [INFO] [udev]: Plugged pad: HORI CO.,LTD. HORIPAD mini4 (3853:238) on port #0.
    [INFO] [udev]: Pad #0 (/dev/input/event1) supports 0 force feedback effects.
    [INFO] [Autoconf]: 2 profiles found.
    [INFO] [autoconf]: selected configuration: /home/pigaming/.config/retroarch/autoconfig/HORI CO.,LTD. HORIPAD mini4.cfg
    [INFO] [Joypad]: Found joypad driver: "udev".
    [WARN] [udev]: Full-screen pointer won't be available.
    [INFO] [Video]: Found display server: null
    [INFO] Found shader "/home/pigaming/.config/retroarch/shaders/bilinear.glslp"
    [INFO] Found shader "/home/pigaming/.config/retroarch/shaders/nearest.glslp"
    [INFO] Found shader "/home/pigaming/.config/retroarch/shaders/slang-conversion-template.glsl"
    [INFO] Found shader "/home/pigaming/.config/retroarch/shaders/stock.glsl"
    context_reset.
    
    点赞 评论 复制链接分享
  • weixin_39640646 weixin_39640646 4月前

    Run it directly. Something like:

    
    gdb --args /opt/retropie/emulators/retroarch/bin/retroarch -L /opt/retropie/libretrocores/lr-reicast/reicast_libretro.so "/home/pigaming/RetroPie/roms/dreamcast/Daytona USA (USA).cdi"
    
    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    Ran it and here's the output

    Reading symbols from /opt/retropie/emulators/retroarch/bin/retroarch...(no debugging symbols found)...done.

    If I ran it via terminal, the game will load on top of emulationstation but very slow, like its fighting for which video is on top.

    If I kill emulationstation, then run it via terminal, the game will load and run normally but I lose the functionality of exiting out and running a diff game or system.

    If I run it via emulationstation, RetroPie uses runcommand.sh on line 1008 as eval $COMMAND </dev/tty &>>"$LOG" where the logs show it segfaults.

    What I'd like to do is modify runcommand.sh with gdb so it outputs it to a log but gdb --args didn't even work via terminal.

    点赞 评论 复制链接分享
  • weixin_39640646 weixin_39640646 4月前

    You would probably have to recompile the core with symbols (-g) to get a meaningful stack trace from gdb.

    The fact that it's working fine if you run it manually probably means that emulationstation is passing arguments / config files that make it crash. I would try to find out what it is. There have been recent changes in the controllers/input code so it could be in that area.

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    If I fork for testing, where would I change the context_reset. so it is run after creating the directory?

    
    Creating dir: /home/pigaming/RetroPie/BIOS/dc/data
    context_reset.
    ARM VFP-Run Fast (NFP) enabled !
    
    点赞 评论 复制链接分享
  • weixin_39640646 weixin_39640646 4月前

    Not sure I understand. context_reset is called to initialize the Open GL context so it doesn't care about the BIOS directory. You shouldn't have to change anything in the code.

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    I compiled a2214fc (save states v3) and this is where the issue began. If I replace the libretro.cpp from 127f845 (linked below), it resolves the issue. There were a ton of changes on this file so not sure what from old file fixed it.

    https://github.com/libretro/reicast-emulator/blob/127f8458bf8cb7817bf1e90349f16664f11240db/core/libretro/libretro.cpp

    点赞 评论 复制链接分享
  • weixin_39640646 weixin_39640646 4月前

    So the crash is due to the change of when dc_init() is called. It would really help to have a gdb stack trace with symbols. Also compiling without #defining RELEASE or NDEBUG could increase the level of logging.

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    I've been compiling to see which will work with the least amount of code changes. Here's what worked for the XU4. I did not test any other games.

    Works fast (fps) for Shikigami No Shiro II https://github.com/6alileo/reicast-emulator/commit/c317b82197d3ae078643eeb734b016fdf734df6a

    Works slow (fps) for Shikigami No Shiro II https://github.com/6alileo/reicast-emulator/commit/26324122ec0edca4382546a42b63673639072cef

    点赞 评论 复制链接分享
  • weixin_39640646 weixin_39640646 4月前

    All this points to a problem due to the dc_init() change but we need to understand why it's failing on this platform.

    点赞 评论 复制链接分享
  • weixin_39831242 weixin_39831242 4月前

    Yeah I need help on compiling so I can get you a debug log. Is it a flag I add on the make command?

    From: flyinghead Sent: Thursday, November 1, 2018 10:06 AM To: libretro/reicast-emulator Cc: 6alileo; Author Subject: Re: [libretro/reicast-emulator] Crashing on any game [non-OIT] [ARM] (#365)

    All this points to a problem due to the dc_init() change but we need to understand why it's failing on this platform.

    — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/libretro/reicast-emulator/issues/365#issuecomment-435107982, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AoSfFNT8aOlODs_jAp96fjFZnQX3wWG0ks5uqyovgaJpZM4YBMKd.

    点赞 评论 复制链接分享