System (keyboard & touchpad) freezes

Bug #1656155 reported by Jeffry R. Fisher
10
This bug affects 2 people
Affects Status Importance Assigned to Milestone
libdrm (Ubuntu)
Confirmed
Undecided
Unassigned

Bug Description

I have a 1-month old Acer Aspire F5-573-58VX that came with Win-10 installed. It relies on its Intel i5-7200u for its graphics (HD 620). It runs Win-10 with no problems (other than Win-10 itself), so I do not suspect the hardware.

I succeeded in dual-booting Ubuntu 16.04 with EFI and secure-boot both enabled, but was suffering random freezes at 3-30 minutes of use, even after setting the c-state limit in grub. It hangs both with and without Intel microcode. So I upgraded to 16.10 hoping that a newer kernel would have a fix in it.

It's still freezing at random (many apps, and sometimes using only the desktop controls). It still has the c-state parameter in grub.

It appears that if I leave the system untouched with the cursor in an empty area, then no freeze will occur (I can see display changing) as long as I don't touch anything. Bittorrent runs reliably overnight, and then the machine can freeze within seconds of my moving the cursor to a control in the morning.

Thunderbird is also suspicious: I can work most of an hour OK as long as I never move the cursor over the scroll slider in the folder pane. However, bringing the cursor close enough to change the slider's appearance will usually freeze the machine. Update: I've now experienced a freeze while the cursor was stationary and I was merely typing into an email message window.

When it freezes, the machine goes quiet (it's not spinning). There's no message to the GUI, only a sudden halt. When frozen, soft restart and reboot (ctrl-alt-back & ctrl-alt-del) both fail. Only a forced power-off has had any effect. (I'll try your alt+sysreq combos in my next freeze).

Synaptic says I have an intel driver update utility installed, but I don't know if it has run. When I try to sudo it, I am told that the command is not found. I have been unable to locate it in the file system. If anyone knows where that goes (or where to hunt), that might be helpful.

I do not (yet) know how to ssh into the sick machine to get kernel logs or crash dumps. I'm a user-space programmer, not a sysadmin, so I have a learning curve to open the firewall enough to allow an intrusion, and more learning curves for the cryptic debugging tools.

var/log/kern.log has this:
Jan 14 10:43:26 prop6 kernel: [ 1.002400] [drm:parse_csr_fw [i915_bpo]] *ERROR* Unknown stepping info, firmware loading failed
Jan 14 10:43:26 prop6 kernel: [ 1.002405] i915_bpo 0000:00:02.0: Failed to load DMC firmware [https://01.org/linuxgraphics/intel-linux-graphics-firmwares], disabling runtime power management.
Jan 14 10:43:26 prop6 kernel: [ 1.006089] vgaarb: device changed decodes: PCI:0000:00:02.0,olddecodes=io+mem,decodes=io+mem:owns=io+mem
Jan 14 10:43:26 prop6 kernel: [ 1.018209] ACPI: Video Device [GFX0] (multi-head: yes rom: no post: no)
Jan 14 10:43:26 prop6 kernel: [ 1.018522] input: Video Bus as /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/LNXVIDEO:00/input/input5
Jan 14 10:43:26 prop6 kernel: [ 1.018601] [drm] Initialized i915_bpo 1.6.0 20160229 for 0000:00:02.0 on minor 0
Jan 14 10:43:26 prop6 kernel: [ 1.064117] r8169 0000:03:00.1 enp3s0f1: renamed from eth0
Jan 14 10:43:26 prop6 kernel: [ 1.142340] fbcon: inteldrmfb (fb0) is primary device
Jan 14 10:43:26 prop6 kernel: [ 1.142437] Console: switching to colour frame buffer device 240x67
Jan 14 10:43:26 prop6 kernel: [ 1.142466] i915_bpo 0000:00:02.0: fb0: inteldrmfb frame buffer device

My current work-around is to do as much work as possible in Win-10 :(

Revision history for this message
Ubuntu Foundations Team Bug Bot (crichton) wrote :

Thank you for taking the time to report this bug and helping to make Ubuntu better. It seems that your bug report is not filed about a specific source package though, rather it is just filed against Ubuntu in general. It is important that bug reports be filed about source packages so that people interested in the package can find the bugs about it. You can find some hints about determining what package your bug might be about at https://wiki.ubuntu.com/Bugs/FindRightPackage. You might also ask for help in the #ubuntu-bugs irc channel on Freenode.

To change the source package that this bug is filed about visit https://bugs.launchpad.net/ubuntu/+bug/1656155/+editstatus and add the package name in the text box next to the word Package.

[This is an automated message. I apologize if it reached you inappropriately; please just reply to this message indicating so.]

tags: added: bot-comment
Revision history for this message
Jeffry R. Fisher (jeffryfisher) wrote :

According to FindRightPackage, "System freezes completely" is probably a fault in the kernel DRM driver (linux package). I'll attempt to set that with an update.

Also: At my latest freeze, I tried the recommended alt+sysreq combos I mentioned. Neither had any effect.

Revision history for this message
Jeffry R. Fisher (jeffryfisher) wrote :

FindRightPackage suggests "kernel DRM driver (linux package)". Since similar bug reports seem to be clustering on Intel CPUs that have on-board GPUs, I can only guess that it means the Intel-specific libdrm. However, I wonder about the i965-va-driver on my Kabylake i5-7200u.

affects: ubuntu → libdrm (Ubuntu)
description: updated
Revision history for this message
Jeffry R. Fisher (jeffryfisher) wrote :

More info (sort of):

After booting today, I used SSH from another machine (successfully logging in). Unfortunately, when the laptop froze, so did my ssh session. The terminal was not "disconnected by host", but it was inert. The laptop has quieted, but not completely silent. Something is whirring in there, but nothing I've tried can get any response from it (except holding down the power key until it powers off).

I can't get anything from Acer; they deliver the machine with Win-10, so that's what they support.

Should I write to Intel?

As for the "Unknown stepping info" error, because I don't know whether it's related, I'm going to open a help request at Ask Ubuntu.

Revision history for this message
Launchpad Janitor (janitor) wrote :

Status changed to 'Confirmed' because the bug affects multiple users.

Changed in libdrm (Ubuntu):
status: New → Confirmed
Revision history for this message
Viacheslav (erlaveri) wrote :

I have fully the same situation with Lenovo v310-15 IKB and i5-7200u, HD620 graphic on board. I had this bug on fresh installed Ubuntu 16.04.2 and 16.10

Revision history for this message
Jeffry R. Fisher (jeffryfisher) wrote :

@V: Did you upgrade to 16.10, or did you do a fresh install? Check what kernel version you ended up with. If less than 4.6, get a fresh install image of 16.10 with kernel 4.8 (or else follow a recipe for upgrading your kernel).

Avoid the bug by choosing a command line boot level (stay out of the GUI) or using ssh to login from a different machine.

That's all I have. Being a user-space kind of guy, I'll defer to anyone who overrides anything I've said.

I'm still rock-stable since achieving kernel 4.8. Somebody else might have something to say about choosing the right install options. There was a 3rd-party option that I said yes to, and it was associated with a log message about Intel drivers later on.

Revision history for this message
Viacheslav (erlaveri) wrote :

I did fresh install.

Currently I am on Ubuntu 16.04 with `uname -r` -> 4.4.0-62-generic. I checked Ubuntu 16.10 on Feb 17 2017, sorry, but didn't check kernel version, but assure it was fresh as possible (after install I made `sudo apt upgrade`).

Also I found that, maybe, all freezes occur in Chrome browser (in my scenario), so respectively to assumptions that Intel Graphic is guilty in this bug I switched off GPU support in Chrome settings and keep watching my laptop.

Revision history for this message
Viacheslav (erlaveri) wrote :

After I switched off GPU in Chrome there was not freezing in Chrome. But I left laptop for a night and in morning I found that he was `frozen`.

Revision history for this message
Viacheslav (erlaveri) wrote :

Hi, it seems installing this `intel-graphics-update-tool_2.0.2_amd64.deb` tool helps me (Ubuntu 16.04)

https://01.org/linuxgraphics/downloads/intel-graphics-update-tool-linux-os-v2.0.2

If doesn't help, you can try install firmware for yours CPU version from here

https://01.org/linuxgraphics/downloads/firmware

Revision history for this message
Viacheslav (erlaveri) wrote :

Hi, again. That I wrote above didn't help. Super sad...

Revision history for this message
Jeffry R. Fisher (jeffryfisher) wrote :

You got some error in the boot log about an unknown stepping function, right? Then an older, faulty version of the Intel driver loaded in its place?

Go back to doing a fresh install of 16.10. Make sure you allow 3rd-party and choose the Intel microcode when offered (and have a working network connection to download and install updates during the install process).

Verify kernel 4.6+ (should be 4.8), and look at the boot log after installing (the error about stepping should be gone).

Good luck.

Revision history for this message
Viacheslav (erlaveri) wrote :

Hi. Installing Ubuntu 16.04, installing INTEL GRAPHICS UPDATE TOOL FOR LINUX* OS V2.0.4 didn't help. Except one.

> echo "options i915 enable_rc6=0" | sudo tee /etc/modprobe.d/i915.conf
> sudo update-initramfs -k `uname -r` -u

This command disable power-saving render C-state 6 for i915 driver. It helped me on Ubuntu 16.10 and 16.04.
Also there are other powersaving possibly harmful for performance and stability parameters. They are:
-enable_fbc
-enable_dc

Also you may look at `semaphores` parameter. To read all i915 parameters:
> modinfo -p i915

This is a very useful link about intel graphic that I found:
https://wiki.archlinux.org/index.php/intel_graphics

@Jeffry R. Fisher: in logs(16.04) I had the same error as you, about missing DMC firmware. I didn't fix it and apparently will not. All this power-saving stuff makes no sense if you almost always work with power supply.

Thank you and good luck.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.