View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0004694 | Spring engine | Gfx driver issues | public | 2015-03-08 15:22 | 2016-01-22 03:21 |
| Reporter | lhog | Assigned To | abma | ||
| Priority | normal | Severity | crash | Reproducibility | unable to reproduce |
| Status | resolved | Resolution | unable to reproduce | ||
| Product Version | 98.0.1+git | ||||
| Summary | 0004694: ATI stack gets consistently crashed when something gets into camera viewport (UnitDrawerState.cpp:105) | ||||
| Description | Game consistently crahes ingame (see replay http://zero-k.info/Battles/Detail/336996) after player ivand lost his comm and camera is moved somewhere to center of the map (I'll try to narrow it down myself). | ||||
| Steps To Reproduce | Not clear. Happened first time. Ptreviously had similar crashes if Zero-K solar was reclaimed (on 98.0.1+). No longer the case, but cases might be related. | ||||
| Additional Information | Latest ATI drivers, W7x64. | ||||
| Tags | No tags attached. | ||||
| Attached Files | |||||
| Checked infolog.txt for Errors | |||||
|
|
Crashes is guaranteed after 9:00 ingame time, even on pause. Happens if camera flies around center bottom part of map (around "dang it" sign and screamer under construction - both should be in the top part of the camera viewport), on the middle level zoom. Crash happens if zoom-in/zoom-out around this area or if one just moves camera from side to the target area. Crash does NOT happen if same area is viewed on pretty close zoom-in. |
|
|
Attaching exact position of camera, zooming out from which (two or three mouse wheel steps to zoom out) causes crash. http://snag.gy/WbD7s.jpg |
|
|
add DebugGL = 1 to your config and see if it gives warnings |
|
|
"translated": C:\Windows\system32\atioglxx.dll 0x041AE3EF ??:0 C:\Windows\system32\atioglxx.dll 0x0417AFE1 ??:0 C:\Windows\system32\atioglxx.dll 0x03C0AEA6 ??:0 E:\Games\Spring\engine\98.0.1-398-g80e6cfa\spring.exe 0x006CDF8F /opt/mxe/usr/lib/gcc/i686-w64-mingw32/4.8.1/include/c++/bits/stl_vector.h:646 E:\Games\Spring\engine\98.0.1-398-g80e6cfa\spring.exe 0x006D2C5E rts/Rendering/UnitDrawerState.cpp:105 |
|
|
Added infolog_glDebug.txt. Crash has happened on pause on 0016368 frame. |
|
|
k I see a correlation between warnings & crash now run with `DebugGLStacktraces = 1` |
|
|
spring with 'DebugGLStacktraces = 1' has never started for me (hangs in the last percentages of "loading...") If "DebugGLStacktraces = true", file is not much different... See infolog_DGLST.txt <Added Later> Seems like this crashes game: [f=0016368] Error: OpenGL: source<API> type<error> id<2269> severity<high>: glUniformMatrix4fv has failed to set the uniform at location '25': the uniform's type, component size, or matrix dimensions are incompatible with the function variant (GL_INVALID_OPERATION) Also bad event happens like between 8.55 and 9.00 game time. Once I stayed zoomed it till around 11.00, then zoomed out and caught crash, so after it appears it stays for a while. |
|
|
i need an infolog with DebugGLStacktraces = 1 and no, it doesn't hang with it, it's just DAMN slow. |
|
|
Pretty sure it hangs as I waited more than 25 minutes (usually it takes <1min). See infolog_hangs.txt |
|
|
I implemented a workaround gameside https://github.com/ZeroK-RTS/Zero-K/commit/76f2237be685eb9367f0c365f7f2a2b88ae850b0#diff-ab54cea9b3a3e4edeadcc210c8f0355eL376 When the crash occurred for most ATI hardware, everyone else saw invisible units. It seemed to only occur for obj models. The bug was caused by doing Spring.UnitRendering.DeactivateMaterial(unitID,3) to a nanoframe. Crashing/causing invisible units still seems like a bug. |
|
|
Disregard last 4 files: posted into wrong topic. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2015-03-08 15:22 | lhog | New Issue | |
| 2015-03-08 15:22 | lhog | File Added: infolog.txt | |
| 2015-03-08 15:58 | lhog | Note Added: 0014117 | |
| 2015-03-08 16:03 | lhog | Note Edited: 0014117 | |
| 2015-03-08 17:49 | lhog | Note Added: 0014118 | |
| 2015-03-08 17:54 | lhog | File Added: stack_trace.png | |
| 2015-03-08 18:18 | jK | Note Added: 0014119 | |
| 2015-03-08 19:05 | abma | Note Added: 0014120 | |
| 2015-03-08 20:18 | lhog | File Added: infolog_glDebug.txt | |
| 2015-03-08 20:19 | lhog | Note Added: 0014121 | |
| 2015-03-08 21:45 | jK | Note Added: 0014122 | |
| 2015-03-08 22:31 | lhog | Note Added: 0014123 | |
| 2015-03-08 22:31 | lhog | File Added: infolog_DGLST.txt | |
| 2015-03-08 22:38 | lhog | Note Edited: 0014123 | |
| 2015-03-08 22:39 | lhog | Note Edited: 0014123 | |
| 2015-03-08 22:40 | lhog | Note Edited: 0014123 | |
| 2015-03-09 10:29 | jK | Note Added: 0014125 | |
| 2015-03-09 14:13 | lhog | File Added: infolog_hangs.txt | |
| 2015-03-09 14:14 | lhog | Note Added: 0014127 | |
| 2015-03-09 23:48 | abma | Summary | ATI stack gets consistently crashed when something gets into camera viewport => ATI stack gets consistently crashed when something gets into camera viewport (UnitDrawerState.cpp:105) |
| 2015-03-10 11:26 | Google_Frog | Note Added: 0014143 | |
| 2015-03-30 21:05 | lhog | File Added: infolog_adv_shading_shadows-2.txt | |
| 2015-03-30 21:05 | lhog | File Added: infolog_adv_shading_shadows-0.txt | |
| 2015-03-30 21:06 | lhog | File Added: screen00024_shadows-0.png | |
| 2015-03-30 21:06 | lhog | File Added: screen00025_shadows-2.png | |
| 2015-03-30 21:10 | lhog | Note Added: 0014257 | |
| 2016-01-22 03:21 | abma | Status | new => resolved |
| 2016-01-22 03:21 | abma | Resolution | open => unable to reproduce |
| 2016-01-22 03:21 | abma | Assigned To | => abma |
