Giter VIP home page Giter VIP logo

Comments (14)

Awakened0 avatar Awakened0 commented on July 24, 2024

Two users have said on Discord they can reproduce this issue with glcore on Windows.

from melonds.

bslenul avatar bslenul commented on July 24, 2024

Yeah, happens to me with glcore too, I have absolutely no clue what's causing this, I was super confused because I had the crash everytime 2 days ago but now it works fine for some reason, so I made a test yesterday on Windows: https://youtu.be/uc-qFQkaGMc

1st attempt I have RetroArch in G:\RA_TEST\RetroArch-Win64\ and no crash.
2nd attempt in G:\RA_TEST\test1\RetroArch-Win64\ = crash
G:\RA_TEST\test2\RetroArch-Win64\ = crash
G:\RA_TEST\bleh\RetroArch-Win64\ = no crash
G:\RA_TEST\waat\RetroArch-Win64\ = crash

And I'm just moving the whole folder, so it's the exact same settings/core options.
And then if somehow I can get it to work inside a specific folder, then it will always work after that from this folder (even after a reboot of the PC).

I tried with retroarch_debug.exe but it didn't create the crash log on crash :/

edit: Also, on my Linux VM it doesn't crash, but if I build RetroArch with ASAN it crashes on close content too. Maybe unrelated since it doesn't crash if I build it normally but still, here's what it says:

AddressSanitizer:DEADLYSIGNAL
=================================================================
==8749==ERROR: AddressSanitizer: SEGV on unknown address 0x7f992f000345 (pc 0x7f99de3e8989 bp 0x7ffeaf1c4ad0 sp 0x7ffeaf1c4ab8 T0)
==8749==The signal is caused by a READ memory access.
    #0 0x7f99de3e8988 in __pthread_detach /build/glibc-ZN95T4/glibc-2.31/nptl/pthread_detach.c:30
    #1 0x7f99be6cae09  (/home/bobby/.config/retroarch/cores/melonds_libretro.so+0x43e09)
    #2 0x7f99be72d926  (/home/bobby/.config/retroarch/cores/melonds_libretro.so+0xa6926)
    #3 0x7f99be6cb92a  (/home/bobby/.config/retroarch/cores/melonds_libretro.so+0x4492a)
    #4 0x557878e3cd8d in core_unload_game /home/bobby/Documents/RetroArch/retroarch.c:38307
    #5 0x557878d1e16a in command_event_deinit_core /home/bobby/Documents/RetroArch/retroarch.c:11858
    #6 0x557878d2d361 in command_event /home/bobby/Documents/RetroArch/retroarch.c:13725
    #7 0x557878e1be4c in rarch_ctl /home/bobby/Documents/RetroArch/retroarch.c:35396
    #8 0x557878e5eedc in content_load tasks/task_content.c:602
    #9 0x557878e66e1c in task_push_start_dummy_core tasks/task_content.c:1556
    #10 0x557878d2a5e9 in command_event /home/bobby/Documents/RetroArch/retroarch.c:13375
    #11 0x557879993b29 in generic_action_ok_command menu/cbs/menu_cbs_ok.c:1628
    #12 0x5578799a93a3 in action_ok_close_content menu/cbs/menu_cbs_ok.c:4860
    #13 0x557878cbdda7 in generic_menu_entry_action /home/bobby/Documents/RetroArch/retroarch.c:2103
    #14 0x55787981bed3 in ozone_menu_entry_action menu/drivers/ozone/ozone.c:644
    #15 0x557878cc1fd2 in menu_entry_action /home/bobby/Documents/RetroArch/retroarch.c:2386
    #16 0x557878cbcc29 in generic_menu_iterate /home/bobby/Documents/RetroArch/retroarch.c:2028
    #17 0x557878cd14eb in menu_driver_iterate /home/bobby/Documents/RetroArch/retroarch.c:3995
    #18 0x557878e2bcdd in runloop_check_state /home/bobby/Documents/RetroArch/retroarch.c:36911
    #19 0x557878e34822 in runloop_iterate /home/bobby/Documents/RetroArch/retroarch.c:37588
    #20 0x557878d372f3 in rarch_main /home/bobby/Documents/RetroArch/retroarch.c:14990
    #21 0x557879279652 in main ui/drivers/qt/ui_qt_application.cpp:152
    #22 0x7f99db82d0b2 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x270b2)
    #23 0x557878c994ed in _start (/home/bobby/Documents/RetroArch/retroarch+0x4bfe4ed)

AddressSanitizer can not provide additional info.
SUMMARY: AddressSanitizer: SEGV /build/glibc-ZN95T4/glibc-2.31/nptl/pthread_detach.c:30 in __pthread_detach
==8749==ABORTING

from melonds.

Myaats avatar Myaats commented on July 24, 2024

@bslenul

What version of the RetroArch frontend are you using, I tried with master and it seemed to be fairly unstable often crashing with null pointer dereferences, stable 1.8.0 seems to run fine for me.

from melonds.

bslenul avatar bslenul commented on July 24, 2024

That was with the latest RA nightly from yesterday (1.9.0 libretro/RetroArch@9a44722), but just tried with 1.8.0 and same result:

2021-01-20.22-25-15.mp4

That's with everything to default for both RA and the core (expect RA UI language). Tried the _debug.exe as well, still no crash log :/

from melonds.

Immersion95 avatar Immersion95 commented on July 24, 2024

I can also add these issues :
-Opt Config file isn't created/updated
-Core doesn't change from Vulkan to GL/GLCore

from melonds.

Ryunam avatar Ryunam commented on July 24, 2024

I finally tested the latest code and can confirm what others have said. Core tends to be unstable right now, often crashing upon hitting the Close Content option on RA.

I've also noticed the following issue:

  • With the OpenGL renderer, using all default settings other than the rendering resolution I'm getting 1 more frame of input latency compared to how it was before the last rebase. I have an earlier version of the core backed up, dating back to December 3 2020, and with that version the GL renderer with the same settings reacts 1 frame earlier than the latest version.
  • This is easily verifiable with the "pause and advance" (P+K) method on RetroArch. I have used the Japanese version of New Super Mario Bros., played with the OpenGL renderer at 2x the resolution and all the other core options set to their default values. When jumping, Mario reacts on the third frame with the December backup and on the fourth frame with the latest core from master.
  • I've also had the core crash RA once while performing the P+K method to measure the latency frames.

I'm on Windows 10 x64. using the glcore video driver in RA.

If anyone else wants to quickly compare the internal latency when using the OpenGL renderer, here is the core that I had backed up in December: melonds_libretro.zip

from melonds.

Myaats avatar Myaats commented on July 24, 2024

Is anyone interested in compiling the core themselves with the newest mingw release? I just noticed someone replaced the entire Windows crossbuilding toolchain around new year. I remember the core used to have similar issues before due to outdated GCC versions emitting broken code when compiling using -O3.

from melonds.

Ryunam avatar Ryunam commented on July 24, 2024

I just performed a full update of MSYS2 and mingw, then compiled the core locally. Same issues as before: getting consistent crashes whenever hitting Close Content, as well as 1 frame of added input latency when using the GL renderer.

from melonds.

bslenul avatar bslenul commented on July 24, 2024

Updated my MSYS2 setup as well (it says gcc version 10.2.0 (Rev6, Built by MSYS2 project)), I re-built the core and still the same crashes on close content for me too.

from melonds.

trulystupidnewb avatar trulystupidnewb commented on July 24, 2024

I have this issue as well, regardless of using gl or vulkan. I'm using a Samsung Galaxy A51, and I noticed that melonds does not save core configurations. For example, I changed the screen layout to left/right instead of top/bottom, and it doesn't save on reboot. I also noticed the app crashes on closing content.

The app crashing is actually my lesser problematic issue. The main issue I have is that I cannot save core settings.

from melonds.

Awakened0 avatar Awakened0 commented on July 24, 2024

The latest Win64 nightly build isn't crashing anymore (closes immediately, changed core options save correctly). f6b775c is probably what fixed it. Anyone who can test Linux and Android should see if it's fixed there too.

from melonds.

Ryunam avatar Ryunam commented on July 24, 2024

Crashes are gone here too! I'll soon open a separate issue about the extra flame of latency with the OpenGL renderer.

from melonds.

bslenul avatar bslenul commented on July 24, 2024

Yeah no crash anymore on Windows 10, thank you m4tsa! 👍

No crash on Android 32bit for me (but can't remember if it crashed before...) and no crash on my Linux VM with a "normal" build of RetroArch but it didn't crash previously so no changes.

I still get the crash on my Linux VM if I use RA built with make DEBUG=1 SANITIZER=address,undefined, but like I said in a previous reply this is maybe completely unrelated to this issue.

from melonds.

Awakened0 avatar Awakened0 commented on July 24, 2024

Alright, will close this then.

from melonds.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.