It's winter, my laptop is freezing (again!)

[Simon]

Hello everyone,

Following the latest Qubes update the Qubes seems again to randomly
freeze every 1-2 day making it unusable.

After the last freeze I fetched the following error from dmesg (I'm
using a Thinkpad T500):

--------------- 8< --------------------

[18899.015916] audit: type=1110 audit(1481304606.026:978): pid=7687
uid=0 auid=4294967295 ses=4294967295 msg='op=PAM:setcred
grantors=pam_env,pam_unix acct="root" exe="/usr/bin/sudo" hostname=?
addr=? terminal=? res=success'
[18971.119105] radeon 0000:01:00.0: ring 0 stalled for more than
10372msec
[18971.119122] radeon 0000:01:00.0: GPU lockup (current fence id
0x000000000001950d last fence id 0x000000000001962b on ring 0)
[18971.354376] radeon 0000:01:00.0: Saved 9145 dwords of commands on
ring 0.
[18971.354394] radeon 0000:01:00.0: GPU softreset: 0x00000008
[18971.354398] radeon 0000:01:00.0: R_008010_GRBM_STATUS =
0xA0003030
[18971.354401] radeon 0000:01:00.0: R_008014_GRBM_STATUS2 =
0x00000003
[18971.354404] radeon 0000:01:00.0: R_000E50_SRBM_STATUS =
0x200010C0
[18971.354408] radeon 0000:01:00.0: R_008674_CP_STALLED_STAT1 =
0x00000000
[18971.354411] radeon 0000:01:00.0: R_008678_CP_STALLED_STAT2 =
0x00000000
[18971.354414] radeon 0000:01:00.0: R_00867C_CP_BUSY_STAT =
0x00020186
[18971.354418] radeon 0000:01:00.0: R_008680_CP_STAT =
0x80028645
[18971.354421] radeon 0000:01:00.0: R_00D034_DMA_STATUS_REG =
0x44C83D57
[18971.410243] radeon 0000:01:00.0: R_008020_GRBM_SOFT_RESET=0x00004001
[18971.410299] radeon 0000:01:00.0: SRBM_SOFT_RESET=0x00000100
[18971.412443] radeon 0000:01:00.0: R_008010_GRBM_STATUS =
0xA0003030
[18971.412446] radeon 0000:01:00.0: R_008014_GRBM_STATUS2 =
0x00000003
[18971.412449] radeon 0000:01:00.0: R_000E50_SRBM_STATUS =
0x200080C0
[18971.412453] radeon 0000:01:00.0: R_008674_CP_STALLED_STAT1 =
0x00000000
[18971.412456] radeon 0000:01:00.0: R_008678_CP_STALLED_STAT2 =
0x00000000
[18971.412460] radeon 0000:01:00.0: R_00867C_CP_BUSY_STAT =
0x00000000
[18971.412463] radeon 0000:01:00.0: R_008680_CP_STAT =
0x80100000
[18971.412467] radeon 0000:01:00.0: R_00D034_DMA_STATUS_REG =
0x44C83D57
[18971.412481] radeon 0000:01:00.0: GPU reset succeeded, trying to
resume
[18971.430177] [drm] PCIE GART of 512M enabled (table at
0x0000000000254000).
[18971.430200] radeon 0000:01:00.0: WB enabled
[18971.430203] radeon 0000:01:00.0: fence driver on ring 0 use gpu addr
0x0000000010000c00 and cpu addr 0xffff88013a777c00
[18971.430981] radeon 0000:01:00.0: fence driver on ring 5 use gpu addr
0x00000000000521d0 and cpu addr 0xffffc900014121d0
[18971.462438] [drm] ring test on 0 succeeded in 1 usecs
[18971.638803] [drm] ring test on 5 succeeded in 1 usecs
[18971.638811] [drm] UVD initialized successfully.
[18971.639046] switching from power state:
[18971.639048] ui class: none
[18971.639050] internal class: boot
[18971.639052] caps: video
[18971.639055] uvd vclk: 0 dclk: 0
[18971.639057] power level 0 sclk: 60000 mclk: 70000 vddc: 1100
[18971.639058] power level 1 sclk: 60000 mclk: 70000 vddc: 1100
[18971.639060] power level 2 sclk: 60000 mclk: 70000 vddc: 1100
[18971.639061] status: c b
[18971.639063] switching to power state:
[18971.639064] ui class: performance
[18971.639066] internal class: none
[18971.639068] caps: single_disp video
[18971.639071] uvd vclk: 0 dclk: 0
[18971.639072] power level 0 sclk: 11000 mclk: 40500 vddc: 900
[18971.639074] power level 1 sclk: 30000 mclk: 70000 vddc: 1100
[18971.639075] power level 2 sclk: 60000 mclk: 70000 vddc: 1100
[18971.639076] status: r
[18981.658102] radeon 0000:01:00.0: ring 0 stalled for more than
10020msec
[18981.658119] radeon 0000:01:00.0: GPU lockup (current fence id
0x000000000001950e last fence id 0x000000000001962c on ring 0)
[18981.658250] [drm:r600_ib_test [radeon]] *ERROR* radeon: fence wait
failed (-35).
[18981.658330] [drm:radeon_ib_ring_tests [radeon]] *ERROR* radeon:
failed testing IB on GFX ring (-35).
[18982.811925] audit_printk_skb: 15 callbacks suppressed
[18982.811929] audit: type=1130 audit(1481304689.822:984): pid=1 uid=0
auid=4294967295 ses=4294967295 msg='unit=getty@tty2 comm="systemd"
exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'

--------------- 8< --------------------


It really seems I'm out-of-luck with Qubes freezes and, frankly, I feel
a little depressed at the idea I may have to through it away to switch
back to a more traditional Linux distro :'(.

For the record, I already encountered what appears to be the same
freezes with Qubes R2, but setting KWin to use XRender instead of the
OpenGL as the compositing engine definitively solved the issue with no
noticeable drawback and gave me a rock solid Qubes (with full screen
Youtube videos and all the fancy, good old times!).

Since Qubes R3.2 and its switch to XFCE, there is not such option
available anymore. I tried to completely disable compositing from XFCE
settings, it kinda worked by reducing the freezes to about one per month
(which means I encountered 2 or 3 freezes since September, not the ideal
but acceptable), but since a recent update (a rather large one, bringing
a new kernel and xorg packages) now the system is again unusable and
makes me loose my work every 1-2 days.

For the record, my laptop also has an Intel integrated graphic card, it
worked great with Qubes R2 but does not work at all with Qubes R3.2 (the
booting process behaves as normal, but the screen stays black when the X
server should start, setting the Grub option i915.enable_rc6=0 does not
help).

Here is my previous message on the topic:

https://groups.google.com/forum/#!topic/qubes-users/O4UO9CjO4TM

Is there anything I can do or try?

I hope someone has some good news for me...

Thank you by advance for your help on this :) !

Simon.