 No.23379
cuavas Administrator
14 days ago
|
I can't reproduce this. I test that the screen chains setting in the CFG file is applied on every release candidate (it's one of my standard tests). Are you sure it isn't a GPU driver issue? |
|---|---|
 No.23380
Robbbert Moderator
14 days ago
|
Unable to replicate this crash in Windows. Maybe someone with unix can try. |
 No.23381
Kale Developer
13 days ago
|
Can't repro either. |
 No.23383
hap Developer
13 days ago
edited on: 13 days ago
|
It's the crash I mentioned here: https://mametesters.org/view.php?id=9186 How I can repro it: open mame.ini and change video to bgfx, and bgfx_screen_chains to literally nothing To be clear, surrounding contents look like this. bgfx_path bgfx bgfx_backend auto bgfx_debug 0 bgfx_screen_chains bgfx_shadow_mask slot-mask.png bgfx_lut lut-default.png bgfx_avi_name auto delete cfg/pacman.cfg mame pacman, exit mame the newly saved pacman.cfg should include this now: <bgfx> <window index="0"> <screen index="0" chain="default" /> </window> </bgfx> change mame.ini bgfx_screen_chains to none mame pacman mame crashes if I do this after: mame pacman -bgfx_screen_chains none it does not crash FWIW I'm also getting the crash on 0.275, contradicting what Augusto claims. I tried as far back as 0.256 (oldest version I have here), and it also crashes with this method. |
 No.23384
Augusto Tester
13 days ago
edited on: 13 days ago
|
@cuavas "Are you sure it isn't a GPU driver issue?" GT 640 using driver 470.256. Not any driver issue. @hap You have figured the issue and details about the crash. "FWIW I'm also getting the crash on 0.275, contradicting what Augusto claims. I tried as far back as 0.256 (oldest version I have here), and it also crashes with this method." I remember here that had happened only recently so for me was 0.277 being the version updated then 0.275. However is better follow the informations detailed by hap that has replied more details than my report. |
|
No.23386
Robbbert Moderator
12 days ago
|
The procedure from hap is reproducible, and we finally have the elusive dump. Not entirely sure if this is what Augusto is referring to though. ----------------------------------------------------- Exception at EIP=00007ff7c5175d57 (bgfx_chain_entry::setup_view(texture_manager&, int, unsigned short, unsigned short, int, unsigned short&, unsigned short&) const+0x05a7): ACCESS VIOLATION While attempting to read memory at 0000000000000000 ----------------------------------------------------- RAX=0000000000000000 RBX=0000028eb1702ae0 RCX=0000000000000000 RDX=000001c078d43060 RSI=0000000000000000 RDI=000000ce39af8f90 RBP=000000ce39af8f80 RSP=000000ce39af8f10 R8=0000000000000006 R9=0000000000000000 R10=000000ce39af8f98 R11=0000000000000008 R12=000000ce39af8fc0 R13=000000ce39af8fb0 R14=0000000000000000 R15=0000028ea13f2630 ----------------------------------------------------- Stack crawl: 000000ce39af90e0: 00007ff7c5175d57 (luaopen_lfs+0x7e4c77) 000000ce39af9200: 00007ff7c517638e (luaopen_lfs+0x7e52ae) 000000ce39af9310: 00007ff7c4ff209a (luaopen_lfs+0x660fba) 000000ce39af9390: 00007ff7c4b5e7d2 (luaopen_lfs+0x1cd6f2) 000000ce39af9430: 00007ff7c4b5e955 (luaopen_lfs+0x1cd875) 000000ce39af9590: 00007ff7c39b087b (renderer_bgfx::draw(int)+0x044b) 000000ce39af95f0: 00007ff7bfb7e29b (win_window_info::draw_video_contents(HDC__*, bool)+0x005b) 000000ce39af96f0: 00007ff7bfb82806 (win_window_info::video_window_proc(HWND__*, unsigned int, unsigned long long, long long)+0x0266) 000000ce39af9880: 00007ffeccc2ef5c (CallWindowProcW+0x060c) 000000ce39af9920: 00007ffeccc2dfbb (SendMessageW+0x092b) 000000ce39af9980: 00007ffeccc2d814 (SendMessageW+0x0184) 000000ce39af99c0: 00007ff7bfb8d3ea (windows_osd_interface::update(bool)+0x003a) 000000ce39af9a50: 00007ff7bfcd8c59 (video_manager::frame_update(bool)+0x00a9) 000000ce39af9ab0: 00007ff7bc81f63a (screen_device::vblank_begin(int)+0x022a) 000000ce39af9b30: 00007ff7bc80e6fb (device_scheduler::timeslice()+0x013b) 000000ce39af9cb0: 00007ff7bc80b925 (running_machine::run(bool)+0x01c5) 000000ce39aff370: 00007ff7bfb623cc (mame_machine_manager::execute()+0x024c) 000000ce39aff760: 00007ff7c38ffc5a (cli_frontend::start_execution(mame_machine_manager*, std::vector<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::allocator<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > > > const&)+0x03ea) 000000ce39affa70: 00007ff7c390028a (cli_frontend::execute(std::vector<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::allocator<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > > >&)+0x007a) 000000ce39affad0: 00007ff7bfb5d167 (emulator_info::start_frontend(emu_options&, osd_interface&, std::vector<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::allocator<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > > >&)+0x0027) 000000ce39affea0: 00007ff7c56c4801 (luaopen_lfs+0xd33721) 000000ce39affef0: 00007ff7b8a312ee (__tmainCRTStartup+0x016e) 000000ce39afff20: 00007ff7b8a31406 (mainCRTStartup+0x0016) 000000ce39afff50: 00007ffecc587374 (BaseThreadInitThunk+0x0014) 000000ce39afffd0: 00007ffecd11cc91 (RtlUserThreadStart+0x0021) |
|
No.23387
cuavas Administrator
12 days ago
|
It still doesn't reproduce for me on Windows with Intel or NVIDIA or Linux with llvmpipe. |
|
No.23388
Robbbert Moderator
11 days ago
|
Error occurs on line 309 of src/osd/modules/render/bgfx/chainentry.cpp Referring to any of the textures.provider(name) elements crashes. Perhaps the structure is null. name = "screen0" |
|
No.23389
Augusto Tester
11 days ago
|
@Robbbert "The procedure from hap is reproducible, and we finally have the elusive dump. Not entirely sure if this is what Augusto is referring to though." hap are correct with my bug report. IMO is better hap continue keeping control over that topic issue than me. |