[0.14.8] [Oxyd] Linux x64 segfaults

Bugs that are actually features.
golfmiketango
Filter Inserter
Filter Inserter
Posts: 549
Joined: Fri Jan 29, 2016 2:48 am
Contact:

[0.14.8] [Oxyd] Linux x64 segfaults

Post by golfmiketango »

These seem to be triggered semi-randomly in 0.14.8, perhaps once every two hours on average. Not sure if anything I am doing in the game is causing it: in the most recent one, I was just "standing there".

I have a lot of mods... I'll try my vanilla campaign for a while and see if I can get the same results.

Code: Select all

433.025 Error CrashHandler.cpp:230: Received SIGSEGV
Factorio crashed. Generating symbolized stacktrace, please wait ...
 433.983 Warning Logger.cpp:364: Symbols.size() == 15, usedSize == 10
#0  0x82fd5d in Logger::logStacktrace(StackTraceInfo*) at /tmp/factorio-x6usDa/src/Util/Logger.cpp:386
#1  0x9d516b in CrashHandler::writeStackTrace() at /tmp/factorio-x6usDa/src/Util/CrashHandler.cpp:105
#2  0x9d53e4 in CrashHandler::SignalHandler(int) at /tmp/factorio-x6usDa/src/Util/CrashHandler.cpp:235
#3  0x3c44032dd0 in ?? at ??:0
#4  0x8aa6bd in CraftingMachine::draw(DrawQueue&) const at /tmp/factorio-x6usDa/src/Entity/CraftingMachine.cpp:148
#5  0x912182 in prepareRow at /tmp/factorio-x6usDa/src/Graphics/EntityRenderer.cpp:211
#6  0x783044 in threadJob at /tmp/factorio-x6usDa/src/Graphics/EntityRenderer.cpp:161
#7  0x1173040 in swap<std::_Any_data> at /usr/include/c++/5/bits/move.h:186
#8  0x3c44407372 in swap at /usr/include/c++/5/functional:2160
#9  0x3c440e5ced in std::function<void ()>::function(std::function<void ()>&&) at /usr/include/c++/5/functional:2035
 434.750 Warning Logger.cpp:364: Symbols.size() == 14, usedSize == 9
 434.750 Error CrashHandler.cpp:106: Map tick at moment of crash: 16274016
 434.750 Error Util.cpp:77: Unexpected error occurred. If you're running the latest version of the game you can help us solve the problem by posting the contents of the log file on the Factorio forums.
Please also include the save file(s), any mods you may be using, and any steps you know of to reproduce the crash.

Code: Select all

....
7575.085 Error CrashHandler.cpp:230: Received SIGSEGV
Factorio crashed. Generating symbolized stacktrace, please wait ...
7575.802 Warning Logger.cpp:364: Symbols.size() == 17, usedSize == 11
#0  0x82fd5d in Logger::logStacktrace(StackTraceInfo*) at /tmp/factorio-x6usDa/src/Util/Logger.cpp:386
#1  0x9d516b in CrashHandler::writeStackTrace() at /tmp/factorio-x6usDa/src/Util/CrashHandler.cpp:105
#2  0x9d53e4 in CrashHandler::SignalHandler(int) at /tmp/factorio-x6usDa/src/Util/CrashHandler.cpp:235
#3  0x3c44032dd0 in ?? at ??:0
#4  0x8af11b in draw at /tmp/factorio-x6usDa/src/Entity/TransportLine.cpp:302
#5  0x8b0993 in drawItemsOnStraightBelt at /tmp/factorio-x6usDa/src/Entity/TransportBeltConnectable.cpp:609
#6  0x912182 in UndergroundBelt::draw(DrawQueue&) const at /tmp/factorio-x6usDa/src/Entity/UndergroundBelt.cpp:199
#7  0x783044 in prepareRow at /tmp/factorio-x6usDa/src/Graphics/EntityRenderer.cpp:211
#8  0x1173040 in threadJob at /tmp/factorio-x6usDa/src/Graphics/EntityRenderer.cpp:161
#9  0x3c44407372 in swap<std::_Any_data> at /usr/include/c++/5/bits/move.h:186
#10 0x3c440e5ced in swap at /usr/include/c++/5/functional:2160
7576.502 Warning Logger.cpp:364: Symbols.size() == 16, usedSize == 10
7576.502 Error CrashHandler.cpp:106: Map tick at moment of crash: 16697476
7576.502 Error Util.cpp:77: Unexpected error occurred....

Loewchen
Global Moderator
Global Moderator
Posts: 8969
Joined: Wed Jan 07, 2015 5:53 pm
Contact:

Re: 0.14.8 Linux x64 segfaults

Post by Loewchen »

Did they happen all on the same map? If so please upload the save-file.
Please post a complete log.

golfmiketango
Filter Inserter
Filter Inserter
Posts: 549
Joined: Fri Jan 29, 2016 2:48 am
Contact:

Re: 0.14.8 Linux x64 segfaults

Post by golfmiketango »

Yes, same mod-heavy map in both of the above crashes, and the other three or four times I CTD'ed.

Although I have nowhere near enough hours played to strongly reject the H0 that it's nevertheless prone to the same crash but I simply haven't stumbled across it yet, after some hours spent in my laboratory save (on a separate 0.14.8 installation with roughly the same settings but very few mods and a much smaller map with no production) I have yet to see any signal 11's.

I'm afraid I already blew away those logs above, but I'm quite confident there was nothing interesting in them (I read them fairly carefully), except the standard gobbldeygook, which looks like:

Code: Select all

   0.000 2016-09-21 01:07:25; Factorio 0.14.8 (build 24767, linux64, alpha)
   0.011 Operating system: Linux (Gentoo 2.3)
   0.011 Program arguments: "./bin/x64/factorio" 
   0.011 Read data path: /home/greg/factorio13/data
   0.011 Write data path: /home/greg/factorio13
   0.011 Binaries path: /home/greg/factorio13/bin
   0.019 System info: [CPU: AMD FX(tm)-8350 Eight-Core Processor           , 8 cores, RAM: 31942MB]
   0.020 Display options: [FullScreen: true] [VSync: false] [UIScale: 100%] [MultiSampling: OFF] [Screen: 255]
   0.154 Available display adapters: 3
   0.155  [0]: resolution 1680x1050px at [2560,96]
   0.155  [1]: resolution 2560x1440px at [0,0]
   0.155  [2]: resolution 1680x1344px at [2560,78]
   0.155 Create display on adapter 1. Size 1280x720 at position [630, 342].
   0.480 Initialised OpenGL:[1] Gallium 0.4 on AMD PITCAIRN (DRM 2.45.0 / 4.7.3-aufs-gmt-00, LLVM 3.8.1); driver: 3.0 Mesa 12.0.2
   0.480     Video memory size (dedicated/total available/current available): 2048/4091/2032 MB
   0.533 Graphics options: [Graphics quality: normal] [Video memory usage: high] [Light scale: 25%] [DXT: auto]
   0.673 Loading mod core 0.0.0 (data.lua)
   0.677 Loading mod base 0.14.8 (data.lua)
   0.769 Loading mod blueprint-string 4.0.0 (data.lua)
   0.824 Loading mod StickyNotes 1.0.11 (data.lua)
   0.885 Loading mod TheFatController 1.0.0 (data.lua)
   0.943 Loading mod Advanced-Drills 0.1.8 (data.lua)
   1.015 Loading mod Advanced-Electric 0.1.3 (data.lua)
   1.079 Loading mod advanced-logistics-system-update-by-andrew 0.14.1 (data.lua)
   1.156 Loading mod Advanced-Tanks 0.1.2 (data.lua)
   1.227 Loading mod Advanced_Personal_Defense 0.1.3 (data.lua)
   1.302 Loading mod AfraidOfTheDark 1.0.15 (data.lua)
   1.384 Loading mod autofill 1.4.5 (data.lua)
   1.453 Loading mod blueprint-book-2 0.14.0 (data.lua)
   1.519 Loading mod Blueprint Items Request 0.1.3 (data.lua)
   1.603 Loading mod Bottleneck 0.3.2 (data.lua)
   1.666 Loading mod CircuitAlerter 0.0.4 (data.lua)
   1.738 Loading mod color-coding 1.1.1 (data.lua)
   1.827 Loading mod compoundsplitters 0.2.8 (data.lua)
   1.900 Loading mod concrete-edges 1.2.1 (data.lua)
   1.986 Loading mod creative-mode 0.1.5 (data.lua)
   2.058 Loading mod Enhanced_Map_Colors 1.3.8 (data.lua)
   2.131 Loading mod EvoGUI 0.4.108 (data.lua)
   2.209 Loading mod Explosive Excavation 1.1.0 (data.lua)
   2.319 Loading mod extinguisher 0.14.0 (data.lua)
   2.400 Loading mod Factorissimo 0.2.0 (data.lua)
   2.485 Loading mod FARL 0.6.0 (data.lua)
   2.588 Loading mod filtered-deconstruction-planner 0.4.8 (data.lua)
   2.670 Loading mod FLAN 1.0.3 (data.lua)
   2.751 Loading mod Flare Stack 2.0.1 (data.lua)
   2.856 Loading mod Flow Control 2.1.1 (data.lua)
   2.937 Loading mod Foreman 1.0.1 (data.lua)
   3.023 Loading mod FusionRobots 0.0.3 (data.lua)
   3.117 Loading mod GroundPrinter 1.0.5 (data.lua)
   3.201 Loading mod HandyHands 1.4.0 (data.lua)
   3.313 Loading mod helmod 0.2.13 (data.lua)
   3.405 Loading mod HotCraft 1.0.2 (data.lua)
   3.490 Loading mod KBlueprints 1.14.1 (data.lua)
   3.601 Loading mod Lejv-armor 0.1.1 (data.lua)
   3.688 Loading mod manual-inventory-sort 1.5.0 (data.lua)
   3.792 Loading mod original-music-hd 1.0.2 (data.lua)
   3.878 Loading mod Orphan Finder 1.0.3 (data.lua)
   3.986 Loading mod PersonalTeleporter 0.1.8 (data.lua)
   4.073 Loading mod PipeCleaner 1.0.2 (data.lua)
   4.159 Loading mod pushbutton 0.0.2 (data.lua)
   4.271 Loading mod RadarGridGuide 0.1.1 (data.lua)
   4.365 Loading mod RailTanker 1.4.0 (data.lua)
   4.470 Loading mod Reinforced-Walls 0.14.10 (data.lua)
   4.559 Loading mod ScoreExtended 1.0.19 (data.lua)
   4.672 Loading mod ShuttleTrain 1.1.4 (data.lua)
   4.764 Loading mod Side Inserters 2.2.2 (data.lua)
   4.855 Loading mod SmartDisplay 1.0.17 (data.lua)
   4.987 Loading mod Tape Measure 1.0.3 (data.lua)
   5.089 Loading mod Turtle_Speed 0.1.2 (data.lua)
   5.196 Loading mod upgrade-planner 1.2.7 (data.lua)
   5.340 Loading mod Vehicle_Radar 0.2.0 (data.lua)
   5.341 Script data.lua:6: [RT] energy_usage: 1000 - refresh_time: 1 - scanned_area: 10 - precognotion: 40 - energy_per_sector: 1000
   5.444 Loading mod WaterWell 1.0.15 (data.lua)
   5.548 Loading mod Zoom 1.0.5 (data.lua)
   5.658 Loading mod SmartTrains 1.1.2 (data.lua)
   5.817 Loading mod rso-mod 2.2.3 (data-updates.lua)
   5.939 Loading mod AfraidOfTheDark 1.0.15 (data-updates.lua)
   6.054 Loading mod Enhanced_Map_Colors 1.3.8 (data-updates.lua)
   6.199 Loading mod FARL 0.6.0 (data-updates.lua)
   6.317 Loading mod manual-inventory-sort 1.5.0 (data-updates.lua)
   6.454 Loading mod Battle-Train-MOD 0.1.2 (data-final-fixes.lua)
   6.578 Loading mod creative-mode 0.1.5 (data-final-fixes.lua)
   6.689 Loading mod DoubleReach 0.14.1 (data-final-fixes.lua)
   6.829 Loading mod FARL 0.6.0 (data-final-fixes.lua)
   6.937 Loading mod Flare Stack 2.0.1 (data-final-fixes.lua)
   7.053 Loading mod RadarGridGuide 0.1.1 (data-final-fixes.lua)
   7.202 Loading mod RailTanker 1.4.0 (data-final-fixes.lua)
   7.314 Checksum for core: 1620335853
   7.314 Checksum for mod base: 1378211439
   7.314 Checksum for mod blueprint-string: 62577300
   7.314 Checksum for mod rso-mod: 1874398137
   7.314 Checksum for mod StickyNotes: 395363666
   7.314 Checksum for mod TheFatController: 3783390293
   7.314 Checksum for mod Advanced-Drills: 232610411
   7.314 Checksum for mod Advanced-Electric: 1348850227
   7.314 Checksum for mod advanced-logistics-system-update-by-andrew: 4004148590
   7.314 Checksum for mod Advanced-Tanks: 720432334
   7.314 Checksum for mod Advanced_Personal_Defense: 723261564
   7.314 Checksum for mod AfraidOfTheDark: 3041156566
   7.315 Checksum for mod AutoDeconstruct: 0
   7.315 Checksum for mod autofill: 61278364
   7.315 Checksum for mod Battle-Train-MOD: 756234979
   7.315 Checksum for mod blueprint-book-2: 157980193
   7.315 Checksum for mod Blueprint Items Request: 2634176260
   7.315 Checksum for mod Bottleneck: 1776934153
   7.315 Checksum for mod CircuitAlerter: 1878379684
   7.315 Checksum for mod color-coding: 1364593114
   7.315 Checksum for mod compoundsplitters: 2681990732
   7.315 Checksum for mod concrete-edges: 2741158827
   7.315 Checksum for mod creative-mode: 2911277995
   7.315 Checksum for mod DoubleReach: 3180674491
   7.315 Checksum for mod Enhanced_Map_Colors: 112677526
   7.315 Checksum for mod EvoGUI: 1040096666
   7.315 Checksum for mod Explosive Excavation: 2050893711
   7.315 Checksum for mod extinguisher: 2987997061
   7.315 Checksum for mod Factorissimo: 708485668
   7.315 Checksum for mod FARL: 1093886923
   7.315 Checksum for mod filtered-deconstruction-planner: 750010405
   7.315 Checksum for mod FLAN: 574514439
   7.315 Checksum for mod Flare Stack: 2975512954
   7.315 Checksum for mod Flow Control: 129829411
   7.315 Checksum for mod Foreman: 2641235779
   7.315 Checksum for mod FusionRobots: 987373360
   7.315 Checksum for mod GroundPrinter: 194626391
   7.315 Checksum for mod HandyHands: 1828761107
   7.315 Checksum for mod helmod: 2029106788
   7.315 Checksum for mod HotCraft: 4188998374
   7.315 Checksum for mod KBlueprints: 2250577784
   7.315 Checksum for mod Lejv-armor: 2007586398
   7.315 Checksum for mod manual-inventory-sort: 1547411448
   7.315 Checksum for mod original-music-hd: 2740735163
   7.315 Checksum for mod Orphan Finder: 169567895
   7.315 Checksum for mod PersonalTeleporter: 751202290
   7.315 Checksum for mod PipeCleaner: 2420122462
   7.315 Checksum for mod ProgressiveRunning: 0
   7.315 Checksum for mod pushbutton: 1043420401
   7.315 Checksum for mod RadarGridGuide: 1498028621
   7.315 Checksum for mod RailTanker: 1260574978
   7.315 Checksum for mod Reinforced-Walls: 633795303
   7.315 Checksum for mod ScoreExtended: 1611246975
   7.315 Checksum for mod ShuttleTrain: 2475647514
   7.315 Checksum for mod Side Inserters: 3560953205
   7.315 Checksum for mod SmartDisplay: 2685622438
   7.315 Checksum for mod Tape Measure: 284499724
   7.315 Checksum for mod Turtle_Speed: 507660349
   7.315 Checksum for mod upgrade-planner: 696037660
   7.315 Checksum for mod Vehicle_Radar: 391950733
   7.315 Checksum for mod WaterWell: 3229134519
   7.315 Checksum for mod Zoom: 2782134040
   7.315 Checksum for mod SmartTrains: 3977623989
   8.106 Info PlayerData.cpp:45: Local player-data.json available, timestamp 1474444905
   8.106 Info PlayerData.cpp:52: Cloud player-data.json unavailable
   8.327 Initial atlas bitmap size is 16384
   8.389 Created atlas bitmap 16384x4449
   8.404 Created atlas bitmap 4096x4080
   8.419 Created atlas bitmap 4096x4092
   8.420 Created atlas bitmap 4096x3476
   8.424 Created atlas bitmap 4096x1224
   8.424 Created atlas bitmap 4096x3660
  22.734 Sprites loaded
  22.734 Convert atlas 4096x4080 to: compressed 
  24.023 Convert atlas 4096x4092 to: compressed 
  25.263 Convert atlas 4096x3476 to: compressed 
  26.253 Convert atlas 4096x1224 to: trilinear-filtering 
  26.294 Convert atlas 4096x3660 to: mipmap 
  26.399 Loading sounds...
  27.796 Custom inputs active: 28
  27.796 Info Updater.cpp:749: Downloading https://updater.factorio.com/get-available-versions?username=golfmiketango&token=<private>&apiVersion=2
  28.433 0 packages available to download (experimental updates enabled).
  28.474 Factorio initialised
  40.699 Loading map /home/greg/factorio13/saves/_autosave3.zip
  40.717 Info Scenario.cpp:148: Map version 0.14.8-1
  50.961 Checksum for script /home/greg/factorio13/temp/currently-playing/control.lua: 900459546
  50.968 Checksum for script __blueprint-string__/control.lua: 1646716633
  50.976 Checksum for script __rso-mod__/control.lua: 3542210094
  50.977 Checksum for script __StickyNotes__/control.lua: 3099992545
  50.982 Checksum for script __TheFatController__/control.lua: 577423359
  50.989 Checksum for script __advanced-logistics-system-update-by-andrew__/control.lua: 2064035505
  50.989 Checksum for script __Advanced-Tanks__/control.lua: 1898714372
  50.990 Checksum for script __Advanced_Personal_Defense__/control.lua: 1319372541
  50.991 Checksum for script __AfraidOfTheDark__/control.lua: 4151387163
  50.993 Checksum for script __AutoDeconstruct__/control.lua: 611373444
  50.996 Checksum for script __autofill__/control.lua: 3388995140
  50.997 Checksum for script __Blueprint Items Request__/control.lua: 4052098442
  50.999 Checksum for script __Bottleneck__/control.lua: 1675732387
  51.006 Checksum for script __CircuitAlerter__/control.lua: 3406867794
  51.007 Checksum for script __compoundsplitters__/control.lua: 982241740
  51.013 Checksum for script __creative-mode__/control.lua: 1991832476
  51.015 Checksum for script __Enhanced_Map_Colors__/control.lua: 3155565040
  51.018 Checksum for script __EvoGUI__/control.lua: 3075659584
  51.019 Checksum for script __Explosive Excavation__/control.lua: 2926481450
  51.020 Checksum for script __extinguisher__/control.lua: 4129171009
  51.022 Checksum for script __Factorissimo__/control.lua: 841786998
  51.029 Checksum for script __FARL__/control.lua: 2419079015
  51.031 Checksum for script __filtered-deconstruction-planner__/control.lua: 1149518439
  51.032 Checksum for script __FLAN__/control.lua: 1226841149
  51.033 Checksum for script __Flow Control__/control.lua: 629381566
  51.042 Checksum for script __Foreman__/control.lua: 3611386632
  51.043 Checksum for script __FusionRobots__/control.lua: 3262095305
  51.046 Checksum for script __GroundPrinter__/control.lua: 4045728384
  51.047 Checksum for script __HandyHands__/control.lua: 922340473
  51.057 Checksum for script __helmod__/control.lua: 2916360666
  51.058 Checksum for script __HotCraft__/control.lua: 1466120655
  51.066 Checksum for script __KBlueprints__/control.lua: 1498718714
  51.067 Checksum for script __Lejv-armor__/control.lua: 1517568801
  51.069 Checksum for script __manual-inventory-sort__/control.lua: 787075625
  51.070 Checksum for script __Orphan Finder__/control.lua: 155176166
  51.071 Checksum for script __PersonalTeleporter__/control.lua: 252882551
  51.072 Checksum for script __PipeCleaner__/control.lua: 2704372211
  51.074 Checksum for script __ProgressiveRunning__/control.lua: 3383724527
  51.074 Checksum for script __pushbutton__/control.lua: 1466110861
  51.075 Checksum for script __RadarGridGuide__/control.lua: 1599608961
  51.077 Checksum for script __RailTanker__/control.lua: 2344758156
  51.079 Checksum for script __ScoreExtended__/control.lua: 3139141152
  51.080 Checksum for script __ShuttleTrain__/control.lua: 581151770
  51.081 Checksum for script __Side Inserters__/control.lua: 260054353
  51.083 Checksum for script __SmartDisplay__/control.lua: 969901242
  51.084 Checksum for script __Tape Measure__/control.lua: 3300237787
  51.085 Checksum for script __Turtle_Speed__/control.lua: 3015026008
  51.087 Checksum for script __upgrade-planner__/control.lua: 3319451699
  51.088 Checksum for script __Vehicle_Radar__/control.lua: 2396222382
  51.089 Checksum for script __WaterWell__/control.lua: 2317626080
  51.090 Checksum for script __Zoom__/control.lua: 540596016
  51.098 Checksum for script __SmartTrains__/control.lua: 2258680208
  52.051 Script control.lua:359: AutoFill: Defaults Updated
  52.058 Script control.lua:26: [RT] Found 60 trains for the player-force
  53.574 Script control.lua:45: [RT] Found 60 trains on surface
  53.577 Script control.lua:26: [RT] Found 0 trains for the enemy-force
  55.093 Script control.lua:45: [RT] Found 0 trains on surface
  55.097 Script control.lua:26: [RT] Found 0 trains for the neutral-force
  56.609 Script control.lua:45: [RT] Found 0 trains on surface
  96.795 Script control.lua:1217: Checking line, needs update: false
  96.796 Script control.lua:1288:  
The last two lines then repeat ad infinitum (those, I assume, are from Smart Trains) and that's it right up to the text in my previous post.

Here's a save -- unfortunately I seem to have lost a bunch of progress, because I had fixed all those trains not getting automated fuel deliveries. Kind of weird that I don't have any autosaves from that, not sure what's going on there, maybe I ran it again and left it sitting there...

All of the mods are stock zips straight from the mod portal. I'd be happy to upload a my entire factorio installation and/or (within reason) any bits of my /usr/lib64 or /etc you'd like to see. Just pm me your e-mail and a laundry list of whatever you need and I'll shoot it all to you on google drive.
Attachments
fourth_try_030.zip
Save from map where I experienced problem (divergent timeline)
(21.44 MiB) Downloaded 65 times

Oxyd
Former Staff
Former Staff
Posts: 1428
Joined: Thu May 07, 2015 8:42 am
Contact:

Re: 0.14.8 Linux x64 segfaults

Post by Oxyd »

You seem to be using Gallium and the two log fragments you posted show a crash in a draw call. Could you try with AMD's driver, just to rule out a bug in Gallium?

golfmiketango
Filter Inserter
Filter Inserter
Posts: 549
Joined: Fri Jan 29, 2016 2:48 am
Contact:

Re: 0.14.8 Linux x64 segfaults

Post by golfmiketango »

Oxyd wrote:You seem to be using Gallium and the two log fragments you posted show a crash in a draw call. Could you try with AMD's driver, just to rule out a bug in Gallium?
Actually I took a look at this already. Turns out, I would have to downgrade to xorg 1.17 which would break all kinds of stuff in my system, including the xrandr configuration which I rely on extensively, now that it finally works, (I have my third monitor using xrandr transforms to display the same content as my second; it's in a different room of my house). So I decided the burden was just too great given that even if it did work it would not be feasible to maintain it going forward. But I do like your thinking.

Notably, however, if the crash is literally in gallium user code, everything on my system has full -ggdb split debug symbols, so we might expect your unwinder to find them, depending on how it works.

Anyhow I was thinking along these same lines. I removed the aufs patchset from my kernel, migrated from kernel 4.7.3 to kernel 4.7.4 (which includes some radeon fixes), and from xorg 1.18.3 to 1.18.4, and mesa 12.0.2 to 12.0.3. Since then, I've been unable to repro the problem despite quite a bit of factorio gaming on the same map. Given the frequency I was experiencing the crashes the other day, I'd guesstimate I can reject the null hypothesis that the bug still exists but has not yet occurred with about 90% confidence. I'd say I probably need another day or two of gaming without a repro to reach a two-standard-deviation 96% confidence level, so my plan is to keep playing factorio :)

Regardless, I strongly suspect one of the above changes did indeed address some kind of problem.

Sometimes I used to run factorio and my whole system would simply freeze. I had to use magic sysrq to reboot. The freeze always happened during "cropping bitmaps." When my system came back I would discover NMI soft lockup backtraces like these made it into the journal:

Code: Select all

Sep 14 11:17:45 moneypit kernel:  [<ffffffff81134c17>] ? get_page_from_freelist+0x977/0xac0
Sep 14 11:17:45 moneypit kernel:  [<ffffffff81134c17>] ? get_page_from_freelist+0x977/0xac0
Sep 14 11:17:45 moneypit kernel:  [<ffffffff81135422>] __alloc_pages_nodemask+0x142/0xe90
Sep 14 11:17:45 moneypit kernel:  [<ffffffff81135422>] ? __alloc_pages_nodemask+0x142/0xe90
Sep 14 11:17:45 moneypit kernel:  [<ffffffff81134c17>] ? get_page_from_freelist+0x977/0xac0
Sep 14 11:17:45 moneypit kernel:  [<ffffffff81136187>] __get_free_pages+0x17/0x48
Sep 14 11:17:45 moneypit kernel:  [<ffffffff815a5a2d>] ? alloc_new_range+0x2ed/0x3d8
Sep 14 11:17:45 moneypit kernel:  [<ffffffff815a60ea>] ? __map_single+0x51a/0x610
Sep 14 11:17:45 moneypit kernel:  [<ffffffff8159f499>] ? iommu_get_domain_for_dev+0x31/0x48
Sep 14 11:17:45 moneypit kernel:  [<ffffffff815a639d>] ? map_page+0x75/0x98
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa0acfe14>] ? radeon_ttm_tt_populate+0xdc/0x2b8 [radeon]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa02f1c9f>] ? ttm_tt_bind+0x37/0x70 [ttm]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa02f3c8f>] ? ttm_bo_handle_move_mem+0x39f/0x420 [ttm]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa02f4316>] ? ttm_bo_mem_space+0x136/0x3a8 [ttm]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa02f496a>] ? ttm_bo_validate+0x13a/0x150 [ttm]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa04c0043>] ? drm_encoder_init+0xdb/0x138 [drm]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa02f4c9c>] ? ttm_bo_init+0x31c/0x3e0 [ttm]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa0ad13fb>] ? radeon_bo_create+0x19b/0x260 [radeon]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa0ad0f30>] ? radeon_update_memory_usage.isra.0+0x50/0x50 [radeon]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa0ae3e7b>] ? radeon_gem_object_create+0xc3/0x190 [radeon]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa0ae42af>] ? radeon_gem_create_ioctl+0x5f/0xf0 [radeon]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa04b47fa>] ? drm_ioctl+0x152/0x538 [drm]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa0ae4250>] ? radeon_gem_pwrite_ioctl+0x28/0x28 [radeon]
Sep 14 11:17:45 moneypit kernel:  [<ffffffff8113d7a6>] ? lru_cache_add_active_or_unevictable+0x36/0xa8
Sep 14 11:17:45 moneypit kernel:  [<ffffffff8115f930>] ? handle_mm_fault+0xc50/0x1cd0
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa0ab404c>] ? radeon_drm_ioctl+0x4c/0x80 [radeon]
Sep 14 11:17:45 moneypit kernel:  [<ffffffff811ae3a2>] ? do_vfs_ioctl+0x92/0x578
Sep 14 11:17:45 moneypit kernel:  [<ffffffff811b8cf7>] ? __fget+0x77/0xa8
Sep 14 11:17:45 moneypit kernel:  [<ffffffff811ae8c9>] ? SyS_ioctl+0x41/0x78
Sep 14 11:17:45 moneypit kernel:  [<ffffffff81848adf>] ? entry_SYSCALL_64_fastpath+0x17/0x93
Haven't seen those either, since the above changes were made. I used to repeatably cause this 100% of the time if I tried the high resolution textures mod -- so maybe I should give that a try; if it doesn't crash that would clearly indicate something fundamental has improved.

golfmiketango
Filter Inserter
Filter Inserter
Posts: 549
Joined: Fri Jan 29, 2016 2:48 am
Contact:

Re: 0.14.8 Linux x64 segfaults

Post by golfmiketango »

golfmiketango wrote: Sometimes I used to run factorio and my whole system would simply freeze. I had to use magic sysrq to reboot. The freeze always happened during "cropping bitmaps." When my system came back I would discover NMI soft lockup backtraces like these made it into the journal:

Code: Select all

Sep 14 11:17:45 moneypit kernel:  [<ffffffff81134c17>] ? get_page_from_freelist+0x977/0xac0
Sep 14 11:17:45 moneypit kernel:  [<ffffffff81134c17>] ? get_page_from_freelist+0x977/0xac0
Sep 14 11:17:45 moneypit kernel:  [<ffffffff81135422>] __alloc_pages_nodemask+0x142/0xe90
Sep 14 11:17:45 moneypit kernel:  [<ffffffff81135422>] ? __alloc_pages_nodemask+0x142/0xe90
Sep 14 11:17:45 moneypit kernel:  [<ffffffff81134c17>] ? get_page_from_freelist+0x977/0xac0
Sep 14 11:17:45 moneypit kernel:  [<ffffffff81136187>] __get_free_pages+0x17/0x48
Sep 14 11:17:45 moneypit kernel:  [<ffffffff815a5a2d>] ? alloc_new_range+0x2ed/0x3d8
Sep 14 11:17:45 moneypit kernel:  [<ffffffff815a60ea>] ? __map_single+0x51a/0x610
Sep 14 11:17:45 moneypit kernel:  [<ffffffff8159f499>] ? iommu_get_domain_for_dev+0x31/0x48
Sep 14 11:17:45 moneypit kernel:  [<ffffffff815a639d>] ? map_page+0x75/0x98
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa0acfe14>] ? radeon_ttm_tt_populate+0xdc/0x2b8 [radeon]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa02f1c9f>] ? ttm_tt_bind+0x37/0x70 [ttm]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa02f3c8f>] ? ttm_bo_handle_move_mem+0x39f/0x420 [ttm]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa02f4316>] ? ttm_bo_mem_space+0x136/0x3a8 [ttm]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa02f496a>] ? ttm_bo_validate+0x13a/0x150 [ttm]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa04c0043>] ? drm_encoder_init+0xdb/0x138 [drm]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa02f4c9c>] ? ttm_bo_init+0x31c/0x3e0 [ttm]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa0ad13fb>] ? radeon_bo_create+0x19b/0x260 [radeon]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa0ad0f30>] ? radeon_update_memory_usage.isra.0+0x50/0x50 [radeon]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa0ae3e7b>] ? radeon_gem_object_create+0xc3/0x190 [radeon]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa0ae42af>] ? radeon_gem_create_ioctl+0x5f/0xf0 [radeon]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa04b47fa>] ? drm_ioctl+0x152/0x538 [drm]
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa0ae4250>] ? radeon_gem_pwrite_ioctl+0x28/0x28 [radeon]
Sep 14 11:17:45 moneypit kernel:  [<ffffffff8113d7a6>] ? lru_cache_add_active_or_unevictable+0x36/0xa8
Sep 14 11:17:45 moneypit kernel:  [<ffffffff8115f930>] ? handle_mm_fault+0xc50/0x1cd0
Sep 14 11:17:45 moneypit kernel:  [<ffffffffa0ab404c>] ? radeon_drm_ioctl+0x4c/0x80 [radeon]
Sep 14 11:17:45 moneypit kernel:  [<ffffffff811ae3a2>] ? do_vfs_ioctl+0x92/0x578
Sep 14 11:17:45 moneypit kernel:  [<ffffffff811b8cf7>] ? __fget+0x77/0xa8
Sep 14 11:17:45 moneypit kernel:  [<ffffffff811ae8c9>] ? SyS_ioctl+0x41/0x78
Sep 14 11:17:45 moneypit kernel:  [<ffffffff81848adf>] ? entry_SYSCALL_64_fastpath+0x17/0x93
Haven't seen those either, since the above changes were made. I used to repeatably cause this 100% of the time if I tried the high resolution textures mod -- so maybe I should give that a try; if it doesn't crash that would clearly indicate something fundamental has improved.
Well I tried that -- still crashed. However, I had a brainstorm -- since this seemed to be some kind of interaction between ttm and the iommu, I tried turning off the iommu in BIOS -- I don't use it these days anyhow. Sure enough, that fixed the crashing. Yay! rip, gpu passthrough fantasy, for now at least.

I suspect the two problems are unrelated, however; like I said, I'll keep playing. If I can't repro, I guess we can presume it was some kind of non-factorio problem.

Oxyd
Former Staff
Former Staff
Posts: 1428
Joined: Thu May 07, 2015 8:42 am
Contact:

Re: 0.14.8 Linux x64 segfaults

Post by Oxyd »

Glad to hear those changes seem to have an effect. :)

Since we haven't had any other bug reports about that crash – or the lockup during bitmap cropping – I'd say that's additional support for the hypothesis that it's something in your system. Definitely keep us updated on how the testing goes.

golfmiketango
Filter Inserter
Filter Inserter
Posts: 549
Joined: Fri Jan 29, 2016 2:48 am
Contact:

Re: 0.14.8 Linux x64 segfaults

Post by golfmiketango »

Oxyd wrote:Glad to hear those changes seem to have an effect. :)

Since we haven't had any other bug reports about that crash – or the lockup during bitmap cropping – I'd say that's additional support for the hypothesis that it's something in your system. Definitely keep us updated on how the testing goes.
I think it's a combination of factors.

One is almost surely this:

Code: Select all

greg@moneypit /goodz/dl/android/honor5x/opengapps64 $ xrandr | grep -v '^\s'
Screen 0: minimum 320 x 200, current 4240 x 1440, maximum 16384 x 16384
DisplayPort-0 connected primary 1680x1050+2560+96 (normal left inverted right x axis y axis) 433mm x 270mm
HDMI-0 disconnected (normal left inverted right x axis y axis)
DVI-0 connected 2560x1440+0+0 (normal left inverted right x axis y axis) 597mm x 336mm
DVI-1 connected 1680x1344+2560+78 (normal left inverted right x axis y axis) 338mm x 270mm

greg@moneypit /goodz/dl/android/honor5x/opengapps64 $ xrandr --verbose
Screen 0: minimum 320 x 200, current 4240 x 1440, maximum 16384 x 16384
       .----.
      / .--. \
     | (    ) |              ___...---.
      \ '--'  '----._..---"""___...---'`
 ------'-.--:------|_()---"""__ -------------------
      / .--.  .----'`''---...__"""---..
     | (    ) |                """---''`
      \ '--' /
       '----'

DVI-1 connected 1680x1344+2560+78 (0x5c) normal (normal left inverted right x axis y axis) 338mm x 270mm
        Identifier: 0x58
        Timestamp:  41842
        Subpixel:   horizontal rgb
        Gamma:      1.0:1.0:1.0
        Brightness: 1.0
        Clones:    
        CRTC:       2
        CRTCs:      0 1 2 3 4 5
        Transform:  1.312500 0.000000 0.000000
                    0.000000 1.312500 0.000000
                    0.000000 0.000000 1.000000
 
There is some weird pair of scissors in my xrandr output! Clearly something wrong there.

Just kidding. The real issue I'm pointing out is, my aging 2GB HD7870 probably struggles to keep my 1440p monitor going and keep the xrandr projection of my second monitor onto my third (note the completely different sizes and proportions of the displays. Until 1.18 xorg, it didn't work, period, not even in stretch-o-vision; now it works, almost, but definitely takes a toll if I watch video on monitors 2 & 3 while playing factorio. I'd turn them off for factorio, but my window manager gets all confused if I make geometry changes and crashes xorg, so I just leave it.

When I play factorio after disabling my MMU, after many hours of gameplay, instead of crashing, I eventually get 2FPS for a while and then see something like:

Code: Select all

Sep 27 08:04:16 moneypit kernel: [TTM] Failed to find memory space for buffer 0xffff8805b745c868 eviction
Sep 27 08:04:16 moneypit kernel: [TTM] No space for ffff8805b745c868 (215040 pages, 860160K, 840M)
Sep 27 08:04:16 moneypit kernel: [TTM]   placement[0]=0x00060002 (1)
Sep 27 08:04:16 moneypit kernel: [TTM]     has_type: 1
Sep 27 08:04:16 moneypit kernel: [TTM]     use_type: 1
Sep 27 08:04:16 moneypit kernel: [TTM]     flags: 0x0000000A
Sep 27 08:04:16 moneypit kernel: [TTM]     gpu_offset: 0x80000000
Sep 27 08:04:16 moneypit kernel: [TTM]     size: 524288
Sep 27 08:04:16 moneypit kernel: [TTM]     available_caching: 0x00070000
Sep 27 08:04:16 moneypit kernel: [TTM]     default_caching: 0x00010000
Which looks to me (I'm no expert) like some kind of kernel leak and not a factorio problem.

Meanwhile, although by now I have changed multiple variables which makes it not terribly helpful information, I am no long able to repro the original crash.

Oxyd
Former Staff
Former Staff
Posts: 1428
Joined: Thu May 07, 2015 8:42 am
Contact:

Re: 0.14.8 Linux x64 segfaults

Post by Oxyd »

golfmiketango wrote:Meanwhile, although by now I have changed multiple variables which makes it not terribly helpful information, I am no long able to repro the original crash.
I guess we'll call that good enough then, and move this to Not a Bug under the assumption that it was your system and not Factorio.

Post Reply

Return to “Not a bug”