[10:34] <michael-vb> Hello again.  I am here with a different issue: when I start a virtual machine in VirtualBox with vmgfx emulation and 3D enabled, my whole (Ubuntu 18.10) system hangs.  With tjaalton's updated X server packages, but I doubt they are related.
[10:34] <michael-vb> I see some OOM killer entries in the system log, sometimes for the VM process, sometimes for Xwayland, but neither of those processes looks like it is OOM-ing the system based on the memory usage reported in the log entries.
[10:34] <michael-vb> So wondering if anyone has an idea how I can debug this, or who else I could poke.  The system is a Thinkpad T470 with HD Graphics 620.
[10:46] <tjaalton> michael-vb: is it the host which hangs, or client?
[10:46] <michael-vb> The host.
[10:50] <michael-vb> Oh yes, and it happens with X or Wayland session, no matter.
[10:53] <tjaalton> could be a bug in i915, hard to tell
[10:54] <michael-vb> That was also my first guess.
[10:55] <michael-vb> When the OOM killer hits the VM process the system recovers.  When it hits Xwayland I can at least switch VTs.
[10:56] <tjaalton> ok so it is oom related
[10:56] <tjaalton> limit the display ram in vbox?
[10:58] <michael-vb> The whole VM process is comfortably under host system RAM.
[10:58] <michael-vb> Though it might be interesting to try increasing the VM VRAM.
[10:58] <michael-vb> As I said, the OOM logs suggested that VBox was not using a lot of RAM.
[10:59] <michael-vb> Or Xwayland for that matter.
[10:59] <tjaalton> but sounds like they'd balloon and oom kills them
[11:00] <tjaalton> or something
[11:00] <michael-vb> Right, but it prints the amount of RAM they were using at kill time into the log.
[11:10] <michael-vb> I am guessing that i915 starts swallowing system RAM and that the OOM killer then starts going for small-ish processes.  It has also hit as and cpp.  But that is just a guess so far.
[11:11] <tjaalton> try #intel-gfx
[11:11] <michael-vb> Thanks.