Bug#859639: More info
Apparently switching to use the intel driver (instead of modeset) did
not fix anything, just made it worse.
What seems to cause the issue is using Firefox + having a second
instance of Firefox open (using a "Dev" profile, I use this for web
development work) + having a private browsing instance open on the Dev
profile.
With the modesetting driver one of two things would happen. Either one
of the Firefox instances would stop responding, but not due to 100% cpu
lock up, it was as if the firefox instance would just stop responding to
mouse or keyboard inputs for minutes at a time. Normally a 'Ctrl+q'
would fix that (basically quit and restart). I don't actually remember
having this problem up until a couple of weeks ago.
In the other case, I'd minimize firefox to the task bar, do something
else, and then maximize it back. But upon the maximize, the screen on my
external monitor would turn black for a second and then come back on
(laptop is in a docking station with the lid closed and LCD screen off),
if I move my mouse or click, the screen would again turn black for a
second and then come back. Only way to stop this cycle is to either
minimize the firefox instance or 'Ctrl+q' it. Now this particular issue
happens about once a month (sometimes every other month) and had done
that for a few months now (but less than a year). The problem is that
since the last kernel update, it seems to happen everyday. Been that way
for the last 2 weeks or so. But also I updated Firefox from 57 to 57.0.1
about 2 weeks ago as well... so not sure. But either way, a web browser
should not cause a monitor to go blank.
Onto today, about 24 hours after using the intel driver (the current one
in the repos) I got a complete screen freeze. The mouse cursor started
to move very slowly and then stopped. Nothing would respond after that.
There was no hard drive activity either (so nothing ate up all the ram
and started swapping to disk). Keyboard was not responsive either. Not
even pushing capslock would light the capslock led on the keyboard.
Tried ctrl+alt+backspace but got nothing. Tried various sysrq commands
but no response so had to force the whole laptop off by holding down the
power button.
Upon boot up, I removed the intel driver and went back to the
modesetting driver. On my two Firefox instances, I had forced GPU
offloading via the 'layers.acceleration.force-enabled' option. Normally
it is disabled on Linux machines. I went ahead and left it forced on for
my main instance (so no change), but on my Dev instance, I set it back
to default (off). Perhaps running two instances of Firefox with GPU
acceleration is too much, but still should not crash a system...
Not sure what info I should provide, but here's an lspci of my vga:
00:02.0 VGA compatible controller: Intel Corporation 2nd Generation Core
Processor Family Integrated Graphics Controller (rev 09) (prog-if 00
[VGA controller])
Subsystem: Lenovo 2nd Generation Core Processor Family
Integrated Graphics Controller
Flags: bus master, fast devsel, latency 0, IRQ 29
Memory at f0000000 (64-bit, non-prefetchable) [size=4M]
Memory at e0000000 (64-bit, prefetchable) [size=256M]
I/O ports at 4000 [size=64]
[virtual] Expansion ROM at 000c0000 [disabled] [size=128K]
Capabilities: [90] MSI: Enable+ Count=1/1 Maskable- 64bit-
Capabilities: [d0] Power Management version 2
Capabilities: [a4] PCI Advanced Features
Kernel driver in use: i915
Kernel modules: i915
Debating on grabbing the latest kernel from backports.. but really don't
want to have to mess with that. Thinking something is wrong in the i915
driver.
Regards,
Reply to: