[HP Compaq dc7800 Convertible Minitower PC] 14.04 freezes on suspend resume with nvidia-331

Bug #1341270 reported by TEN on 2014-07-13
This bug affects 1 person
Affects Status Importance Assigned to Milestone
nvidia-graphics-drivers-331 (Ubuntu)

Bug Description

Sometimes the screen shows a blinking text console mode prompt and the NumLock key can still turn its LED on and off for a few seconds under nivida-331 before the machine freezes with fans running after resume, while nouveau would even permit Ctrl-Alt-F1 occasionally (but then has other log-flooding issues as indicated at the prior link). This is a recurring issue already exhibited by 12.04.? LTS.

ProblemType: Bug
DistroRelease: Ubuntu 14.04
Package: linux-image-3.13.0-30-generic 3.13.0-30.55
ProcVersionSignature: Ubuntu 3.13.0-30.55-generic
Uname: Linux 3.13.0-30-generic x86_64
NonfreeKernelModules: nvidia
ApportVersion: 2.14.1-0ubuntu3.2
Architecture: amd64
 /dev/snd/controlC1: user 2417 F.... pulseaudio
 /dev/snd/controlC2: user 2417 F.... pulseaudio
 /dev/snd/controlC3: user 2417 F.... pulseaudio
 /dev/snd/controlC0: user 2417 F.... pulseaudio
CurrentDesktop: Unity
Date: Sun Jul 13 12:20:23 2014
HibernationDevice: RESUME=UUID=268e28b5-d1d2-41ba-96cd-dbc77d8ec2ad
InstallationDate: Installed on 2014-04-22 (81 days ago)
InstallationMedia: Ubuntu 14.04 LTS "Trusty Tahr" - Release amd64 (20140417)
 eth0 no wireless extensions.

 lo no wireless extensions.
MachineType: Hewlett-Packard HP Compaq dc7800p Convertible Minitower

ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-3.13.0-30-generic root=UUID=dbc867b7-24fc-40ab-93bd-b34f10eb7044 ro
 linux-restricted-modules-3.13.0-30-generic N/A
 linux-backports-modules-3.13.0-30-generic N/A
 linux-firmware 1.127.4
 1: hci0: Bluetooth
  Soft blocked: no
  Hard blocked: no
SourcePackage: linux
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 07/21/2011
dmi.bios.vendor: Hewlett-Packard
dmi.bios.version: 786F1 v01.32
dmi.board.asset.tag: CZC8277VRK
dmi.board.name: 0AACh
dmi.board.vendor: Hewlett-Packard
dmi.chassis.asset.tag: CZC8277VRK
dmi.chassis.type: 6
dmi.chassis.vendor: Hewlett-Packard
dmi.modalias: dmi:bvnHewlett-Packard:bvr786F1v01.32:bd07/21/2011:svnHewlett-Packard:pnHPCompaqdc7800pConvertibleMinitower:pvr:rvnHewlett-Packard:rn0AACh:rvr:cvnHewlett-Packard:ct6:cvr:
dmi.product.name: HP Compaq dc7800p Convertible Minitower
dmi.sys.vendor: Hewlett-Packard

TEN (launchpad-20-ten) wrote :
TEN (launchpad-20-ten) on 2014-07-13
description: updated
TEN (launchpad-20-ten) on 2014-07-13
description: updated
description: updated

This change was made by a bot.

Changed in linux (Ubuntu):
status: New → Confirmed
tags: added: precise

Curiously I can actually resume from a manually typed pm-suspend as root to the Ctrl-Alt-F2 text mode console without immediate crash.

However, even chvt 7;xrandr --auto will freeze the machine (including keyboard LEDs) afterwards, i.e. I cannot get back to anything running on the graphical desktop.

dmesg just before shows the attached entries, interestingly hinting at a recurrent problem of reading the display descriptor:
HDMI: invalid ELD buf size -1

Now I wonder how I can force-feed an overriding EDID file to the new driver(s) - as used to be possible through the old /etc/X11/xorg.conf when there still was one (http://wiki.xbmc.org/index.php?title=Creating_and_using_edid.bin_via_xorg.conf#Configure_xorg_to_read_from_EDID) , and whether that might work around the issue.

TEN (launchpad-20-ten) wrote :

The following from a resumed-to Ctrl-Alt-F2 text console results in the same sort of crashes, so the cause seems to be something other than (only) the display auto(re)detection:

chvt 7
xrandr --newmode "1920x1080_60.00" 173.00 1920 2048 2248 2576 1080 1083 1088 1120 -hsync +vsync
xrandr --addmode DFP-1 "1920x1080_60.00"

The final crash hit so hard it corrupted all of /dev/sda's file systems (requiring to press F on each to even reboot).

Looks a lot like https://bugs.freedesktop.org/show_bug.cgi?id=35648 back with a vengeance.

description: updated
tags: added: latest-bios-1.32
removed: precise
summary: - Ubuntu 14.04 freezes on resume from suspend: GeForce GT 520 GPU with
- nouveau or proprietary nvidia-331 drivers, even when booted nosplash
- noquiet
+ [HP Compaq dc7800 Convertible Minitower PC] 14.04 freezes on suspend
+ resume with nvidia-331

TEN, using the nvidia drivers narrows the scope of people who may help you on graphics driver issues to nvidia only.

As well, please do not mix two different drivers (nvidia and nouveau) into one report. Instead, please create one per driver, not per symptom.

affects: linux (Ubuntu) → nvidia-graphics-drivers-331 (Ubuntu)
Changed in nvidia-graphics-drivers-331 (Ubuntu):
importance: Undecided → Low
status: Confirmed → New
TEN (launchpad-20-ten) wrote :

Christopher, why is this supposed to be a "graphics driver issue" that would need to be split up into isolated bugs per GPU and per system (so that none shows the "big picture" and aggregate Bug Heat of likely Mode Setting or similar hardware re-initialization troubles on resume) ?

With the references of the original report https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1341270/comments/0 to possibly related bugs edited out, individualized instances may be made appear as of Low importance and remain Unassigned for long (to what end though?), but freeze-on-resume means a serious total loss of all unsaved data, potential file system corruption and severe risk recurrent from earlier LTS releases where the circumstances triggering it and their "solution" (some kernels simply unaffected) never seem to have been fully understood either.

Guidance as to what contributions may be provided for the purpose of debugging is much appreciated.

Per Wahlström at https://bugs.launchpad.net/ubuntu/+source/pm-utils/+bug/999638 #18 et seq. had already made every conceivable effort more than 2 years ago but there still seems to be no indication who picked up how from there.

TEN (launchpad-20-ten) wrote :

Freezes at resume also occur on Kernel 3.13.0-32-generic #57-Ubuntu SMP Tue Jul 15 03:51:08 UTC 2014.
While the /usr/share/apport/apportcheckresume reporting window pops up 1-2 reboots later, it is not traceable how https://wiki.ubuntu.com/Apport "uploads the collected information to the bug tracking system. After that it opens the packages' bug filing page with a sensible default bug title and leaves the rest of bug filing process to the web UI."
Neither Firefox opens nor anything in it.
If there is any way to find the uploads and verify whether they have been completed, please advise.

TEN (launchpad-20-ten) wrote :

Except for some erratic flickering on resume, the mainline kernel just out returns from suspend without crash after this & a reboot into it:

cd /usr/local;mkdir v3.16-utopic;cd v3.16-utopic;wget http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.16-utopic/linux-headers-3.16.0-031600-generic_3.16.0-031600.201408031935_amd64.deb;wget http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.16-utopic/linux-headers-3.16.0-031600_3.16.0-031600.201408031935_all.deb;wget http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.16-utopic/linux-image-3.16.0-031600-generic_3.16.0-031600.201408031935_amd64.deb;dpkg -i *.deb

Will have to test for a while before I trust it my production environment, and of course this should not preclude further investigation to avoid regressions.

Changed in nvidia-graphics-drivers-331 (Ubuntu):
status: New → Confirmed
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers

Remote bug watches

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