Linux UI frequently freezes for <1 second

Question

The Linux Desktop (Xubuntu 16.04) freezes frequently (multiple times a minute) for less than 1-2 seconds and then returns operating normally.

During the freeze only the mouse cursor is movable.

The error occurs with and without a connected ac adapter.

The Laptop is a Thinkpad E550 with an Intel i7-5500U, 16GB RAM and a Samsung SSD 850 EVO 120GB. The average Load is between 0.5 and 0.7.

According to smartctl the SSD has no errors and gets trimmed before each shutdown.

Even powering the the second graphic adapter off, using vgaswitcheroo does not help:

# sudo cat /sys/kernel/debug/vgaswitcheroo/switch 
0:DIS: :DynPwr:0000:05:00.0
1:IGD:+:Pwr:0000:00:02.0

The power management is already set to performance in the ThinkPad BIOS.

After each micro freeze the following messages occur int the syslog:

Jan 18 09:37:54 localhost kernel: [ 2868.686075] [drm] probing gen 2 caps for device 8086:9c98 = 5323c42/0
Jan 18 09:37:54 localhost kernel: [ 2868.686082] [drm] PCIE gen 2 link speeds already enabled
Jan 18 09:37:54 localhost kernel: [ 2868.693583] [drm] PCIE GART of 2048M enabled (table at 0x00000000002E8000).
Jan 18 09:37:54 localhost kernel: [ 2868.693688] radeon 0000:05:00.0: WB enabled
Jan 18 09:37:54 localhost kernel: [ 2868.693690] radeon 0000:05:00.0: fence driver on ring 0 use gpu addr 0x0000000080000c00 and cpu addr 0xffff8800355ffc00
Jan 18 09:37:54 localhost kernel: [ 2868.693692] radeon 0000:05:00.0: fence driver on ring 1 use gpu addr 0x0000000080000c04 and cpu addr 0xffff8800355ffc04
Jan 18 09:37:54 localhost kernel: [ 2868.693693] radeon 0000:05:00.0: fence driver on ring 2 use gpu addr 0x0000000080000c08 and cpu addr 0xffff8800355ffc08
Jan 18 09:37:54 localhost kernel: [ 2868.693694] radeon 0000:05:00.0: fence driver on ring 3 use gpu addr 0x0000000080000c0c and cpu addr 0xffff8800355ffc0c
Jan 18 09:37:54 localhost kernel: [ 2868.693695] radeon 0000:05:00.0: fence driver on ring 4 use gpu addr 0x0000000080000c10 and cpu addr 0xffff8800355ffc10
Jan 18 09:37:54 localhost kernel: [ 2868.695260] radeon 0000:05:00.0: fence driver on ring 5 use gpu addr 0x0000000000075a18 and cpu addr 0xffffc90002835a18
Jan 18 09:37:54 localhost kernel: [ 2868.795722] radeon 0000:05:00.0: VCE init error (-110).
Jan 18 09:37:54 localhost kernel: [ 2868.989381] [drm] ring test on 0 succeeded in 1 usecs
Jan 18 09:37:54 localhost kernel: [ 2868.989388] [drm] ring test on 1 succeeded in 1 usecs
Jan 18 09:37:54 localhost kernel: [ 2868.989392] [drm] ring test on 2 succeeded in 1 usecs
Jan 18 09:37:54 localhost kernel: [ 2868.989410] [drm] ring test on 3 succeeded in 2 usecs
Jan 18 09:37:54 localhost kernel: [ 2868.989417] [drm] ring test on 4 succeeded in 4 usecs
Jan 18 09:37:54 localhost kernel: [ 2869.165112] [drm] ring test on 5 succeeded in 2 usecs
Jan 18 09:37:54 localhost kernel: [ 2869.165118] [drm] UVD initialized successfully.
Jan 18 09:37:54 localhost kernel: [ 2869.165153] [drm] ib test on ring 0 succeeded in 0 usecs
Jan 18 09:37:54 localhost kernel: [ 2869.165183] [drm] ib test on ring 1 succeeded in 0 usecs
Jan 18 09:37:54 localhost kernel: [ 2869.165213] [drm] ib test on ring 2 succeeded in 0 usecs
Jan 18 09:37:54 localhost kernel: [ 2869.165241] [drm] ib test on ring 3 succeeded in 0 usecs
Jan 18 09:37:54 localhost kernel: [ 2869.165269] [drm] ib test on ring 4 succeeded in 0 usecs
Jan 18 09:37:55 localhost kernel: [ 2869.812290] [drm] ib test on ring 5 succeeded

Answer 1

It can be due to power management.

setting power management performance does not necessarily limit CPU c-state and p-state. You may want to manual set these parameters in BIOS (max c-state to c1, and disable pstate).

In addition, there are new intel drivers which overrides these settings, which need to be disabled using kernel commandline like

intel_idle.max_cstate=0,
intel_pstate=disable, processor.max_cstate=0

But not sure if these are applicable for Ubuntu. please check.

Answer 2

This thread suggests that the issue may come from Intel Graphics drivers. The issue may be solved (to a degree) by switching to UXA acceleration. This can be done by adding

Option      "AccelMethod"  "uxa"

to the relevant Device section of xorg.conf