Created attachment 135524 [details]
I'm using Ubuntu 17.10 (fully updated) and my system is crashing at shutdown. It's freezing until I force a reboot/shutdown using Magic SysRq.
The error is:
Nov 6 22:16:49 rene-Inspiron-5447 kernel: [ 248.294280] [drm:atom_op_jump [amdgpu]] *ERROR* atombios stuck in loop for more than 5secs aborting
Nov 6 22:16:49 rene-Inspiron-5447 kernel: [ 248.294336] [drm:amdgpu_atom_execute_table_locked [amdgpu]] *ERROR* atombios stuck executing 75A8 (len 272, WS 0, PS 4) @ 0x75F1
Nov 6 22:16:49 rene-Inspiron-5447 kernel: [ 248.294384] [drm:amdgpu_atom_execute_table_locked [amdgpu]] *ERROR* atombios stuck executing 640C (len 68, WS 0, PS 8) @ 0x6430]
This behavior is only happening when my laptop is charging. When running on battery I got a clean shutdown/reboot.
Already tried to run the latest kernel package from http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.14 and the problem is exactly the same.
DistroRelease: Ubuntu 17.10
Package: linux-image-4.13.0-16-generic 4.13.0-16.19 [modified: boot/vmlinuz-4.13.0-16-generic]
ProcVersionSignature: Ubuntu 4.13.0-16.19-generic 4.13.4
Uname: Linux 4.13.0-16-generic x86_64
USER PID ACCESS COMMAND
/dev/snd/controlC1: rene 1249 F.... pulseaudio
/dev/snd/controlC0: rene 1249 F.... pulseaudio
Date: Wed Nov 8 01:00:51 2017
InstallationDate: Installed on 2017-11-06 (1 days ago)
InstallationMedia: Ubuntu 17.10 "Artful Aardvark" - Release amd64 (20171018)
MachineType: Dell Inc. Inspiron 5447
ProcFB: 0 inteldrmfb
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-4.13.0-16-generic.efi.signed root=UUID=1237aa76-9a2f-4a06-8552-526f124914ff ro quiet splash acpi_backlight=intel_backlight
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.vendor: Dell Inc.
dmi.board.vendor: Dell Inc.
dmi.chassis.vendor: Dell Inc.
dmi.product.name: Inspiron 5447
dmi.sys.vendor: Dell Inc.
Created attachment 135525 [details]
Created attachment 135526 [details]
(In reply to Rene Barbosa from comment #0)
> Created attachment 135524 [details]
Looks like you attached the wrong file. Please attach your full dmesg output.
Created attachment 135527 [details]
Err. Sorry about that.
I've uploaded the correct file now.
An interesting information: It's only with open source amdgpu driver. With amdgpu-pro it's not happening but this driver isn't compatible with Ubuntu 17.10.
I've got a similar issue on a recently purchased Dell Inspiron 5770 with an AMD Radeon 530 graphics card with openSUSE Tumbleweed.
Reboot or attempt to switch to text console causes a hang and /var/log/messages contains the above message + kernel bug with a stack trace. Full details are here:
Are you using TLP or something similar?
I ran some tests and found that's only happening when using TLP. Tried with Ubuntu 17.10 and Fedora 27, same results.
Looks like TLP was installed on my laptop at initial installation time.
rpm -qa | grep -i tlp
The latest openSUSE Tumbleed snapshot switched to 4.14 kernel - I got the same issue with that too on first boot. The bug report in openSUSE's bugzilla contains the /var/log/messages for that first boot with 4.14.
OTOH, TLP services are not enabled according to tlp-stat -s, so does that mean my laptop is not using TLP after all?
--- TLP 1.0 --------------------------------------------
+++ System Info
System = Dell Inc. Inspiron 5770
BIOS = 1.0.5
Release = "openSUSE Tumbleweed"
Kernel = 4.14.0-1-default #1 SMP PREEMPT Mon Nov 13 21:02:46 UTC 2017 (ab9e909) x86_64
/proc/cmdline = BOOT_IMAGE=/boot/vmlinuz-4.14.0-1-default root=UUID=d58e8e85-bfdd-4a74-a694-5149ccd85110 quiet resume=/dev/sda3 splash=silent quiet showopts
Init system = systemd v234
Boot mode = UEFI
+++ TLP Status
State = enabled
Last run = 11:02:47, 14 sec(s) ago
Mode = AC
Power source = AC
Notice: tlp.service is not enabled -- invoke "systemctl enable tlp.service" to correct this!
Notice: tlp-sleep.service is not enabled -- invoke "systemctl enable tlp-sleep.service" to correct this!
Notice: systemd-rfkill.service is not masked -- invoke "systemctl mask systemd-rfkill.service" to correct this!
These systemd services are not running but according to tlp-stat, it's enabled.
Can you try to uninstall these packages and see if the error persists?
In my laptop, the problem is only happening when I install/enable them. Is it a power management bug?
Uninstalling tlp and tlp-rdw packages eliminates the error.
What I don't get is that runtime pm for amggpu driver is supposed to be disabled by default according to TLP docs.
I'll try reinstalling TLP and follow the TLP troubleshooting guide to see exactly what setting is causing this.
Is your system using 'pcieport' module?
I've added it to RUNTIME_PM_DRIVER_BLACKLIST in TLP's configuration and now the problem is fixed!
Not sure why it's happening in Linux 4.13+ and not in 4.4 with AMDGPU-Pro installed.
Sorry for the delay in responding. Thanks for your suggestion regarding the pcieport module. Adding it to RUNTIME_PM_DRIVER_BLACKLIST does fix the problem while TLP is installed.