Bug 92523

Does this happen without suspend/resume? If not, how is this different than bug #92522 which you filed earlier today?

(In reply to Ilia Mirkin from comment #1)
> Does this happen without suspend/resume? If not, how is this different than
> bug #92522 which you filed earlier today?

Yes this happens without suspend/resume. While the other bug was a desktop freeze after resume from suspension, this happens using the desktop normally, it gets slow all of a sudden, freezes, I open terminal (CTRL+ALT+F1) and with the top command I can see compiz process  using 100% of CPU.

After like 3-4 minutes this process uses less CPU and desktop becomes usable again.

Do the set_domain messages in dmesg coincide with the freezes?

(In reply to Ilia Mirkin from comment #3)
> Do the set_domain messages in dmesg coincide with the freezes?

I will have to check this out, never noticed this before.
I will check this the next time my desktop gets freezed and let you know.

Please give the patch at http://lists.freedesktop.org/archives/nouveau/2015-October/022732.html a shot -- it will hopefully help with the suspend/resume issues, and who knows -- perhaps even fix these -- you only have 256MB of vram, so similar situations will occur as with suspend/resume, with buffers getting swapped out.

(In reply to Ilia Mirkin from comment #5)
> Please give the patch at
> http://lists.freedesktop.org/archives/nouveau/2015-October/022732.html a
> shot -- it will hopefully help with the suspend/resume issues, and who knows
> -- perhaps even fix these -- you only have 256MB of vram, so similar
> situations will occur as with suspend/resume, with buffers getting swapped
> out.



Ok I will give it a shot and let you know.
Meanwhile it has just happened again, I will attach two screenshots with top command and one with dmesg, just for you info.

Created attachment 119169 [details]
Top screen shot n. 1

Top command while desktop is freezed with compiz process using 100% (more or less) CPU

Created attachment 119170 [details]
Top screen shot n. 2

Created attachment 119171 [details]
dmesg screenshot

Screenshot of dmesg output while desktop was freezed by compiz process at 100% (more or less).

Created attachment 119172 [details]
Dmesg during freeze

Dmesg command launched during freeze.

(In reply to Carla sella from comment #6)
> (In reply to Ilia Mirkin from comment #5)
> > Please give the patch at
> > http://lists.freedesktop.org/archives/nouveau/2015-October/022732.html a
> > shot -- it will hopefully help with the suspend/resume issues, and who knows
> > -- perhaps even fix these -- you only have 256MB of vram, so similar
> > situations will occur as with suspend/resume, with buffers getting swapped
> > out.
> 
> 
> 
> Ok I will give it a shot and let you know.
> Meanwhile it has just happened again, I will attach two screenshots with top
> command and one with dmesg, just for you info.



I am waiting for Ubuntu to have the 4.3-rc7 kernel and install it:


http://kernel.ubuntu.com/~kernel-ppa/mainline/?C=N;O=D

so it will be easier for me to test your patch.

I will let you know what happens.

Thanks.

(In reply to Carla sella from comment #11)
> 
> I am waiting for Ubuntu to have the 4.3-rc7 kernel and install it:
> 
> 
> http://kernel.ubuntu.com/~kernel-ppa/mainline/?C=N;O=D
> 
> so it will be easier for me to test your patch.
> 
> I will let you know what happens.
> 
> Thanks.

You reported in bug 92504 that the suspend/resume issue was gone... is this one as well?

(In reply to Ilia Mirkin from comment #12)
> (In reply to Carla sella from comment #11)
> > 
> > I am waiting for Ubuntu to have the 4.3-rc7 kernel and install it:
> > 
> > 
> > http://kernel.ubuntu.com/~kernel-ppa/mainline/?C=N;O=D
> > 
> > so it will be easier for me to test your patch.
> > 
> > I will let you know what happens.
> > 
> > Thanks.
> 
> You reported in bug 92504 that the suspend/resume issue was gone... is this
> one as well?


I was waiting some time before answering to see if this issue occurred, but it hasn't occurred since I installed Ubuntu 4.3.0-040300rc7-generic kernel, so I can affirm that the issue is solved.