pm-suspend becomes catatonic - screen blanks; system never halts

pAtchSavage87 · 2019-11-28 00:10:15

Hi Everyone,

I'm a recent convert to Devuan, after several years of otherwise happy times as an Ubuntu weenie. I thought I had a pretty good understanding of linux system management but this has been an education otherwise! Let me preface everything by saying I am otherwise extremely happy with Devuan - I'll even say there seem to be a few key improvements right out of the box!

The last hangup for me is that this system is a lenovo ideapad 310 - a laptop - and I can't for the life of me get it to suspend. The essential problem statement is this: On issuing pm-suspend or pm-hibernate via CLI, ACPI Hooks, or the mate power manager (at least, I think it's the Mate power Manager), I get a nice black screen and... nothing.

To give you a precise idea what I'm working with here...

patches@nostromo:~$ inxi -F
System:    Host: nostromo Kernel: 4.9.0-11-amd64 x86_64 (64 bit)
           Desktop: MATE 1.16.2  Distro: Devuan GNU/Linux ascii
Machine:   Device: laptop System: LENOVO product: 80ST v: Lenovo ideapad 310-15ABR
           Mobo: LENOVO model: Toronto 5B1 v: SDK0J40700WIN
           UEFI: LENOVO v: 1LCN32WW date: 09/05/2016
Battery    BAT0: charge: 20.3 Wh 98.0% condition: 20.7/28.5 Wh (73%)
CPU:       Quad core AMD A12-9700P RADEON R7 10 COMPUTE CORES 4C+6G (-MCP-) cache: 4096 KB 
           clock speeds: max: 2500 MHz 1: 1600 MHz 2: 2100 MHz 3: 1600 MHz
           4: 1300 MHz
Graphics:  Card: Advanced Micro Devices [AMD/ATI] Carrizo
           Display Server: X.Org 1.19.2 drivers: ati,amdgpu (unloaded: modesetting,fbdev,vesa,radeon)
           Resolution: 1366x768@60.00hz
           GLX Renderer: Gallium 0.4 on AMD CARRIZO (DRM 3.8.0 / 4.9.0-11-amd64, LLVM 3.9.1)
           GLX Version: 3.0 Mesa 13.0.6
Audio:     Card-1 Advanced Micro Devices [AMD] Device 157a
           driver: snd_hda_intel
           Card-2 Advanced Micro Devices [AMD/ATI] Kabini HDMI/DP Audio
           driver: snd_hda_intel
           Sound: Advanced Linux Sound Architecture v: k4.9.0-11-amd64
Network:   Card-1: Realtek RTL8821AE 802.11ac PCIe Wireless Network Adapter
           driver: rtl8821ae
           IF: wlan0 state: up mac: 78:45:61:cb:84:c3
           Card-2: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller
           driver: r8169
           IF: eth0 state: down mac: c8:5b:76:76:eb:be
Drives:    HDD Total Size: 1000.2GB (14.4% used)
           ID-1: /dev/sda model: WDC_WD10JPVX size: 1000.2GB
Partition: ID-1: / size: 904G used: 124G (15%) fs: ext4 dev: /dev/dm-1
           ID-2: /boot size: 237M used: 66M (30%) fs: ext2 dev: /dev/sda2
           ID-3: swap-1 size: 12.23GB used: 0.00GB (0%) fs: swap dev: /dev/dm-2
Sensors:   System Temperatures: cpu: 0.0C mobo: N/A gpu: 0.0
           Fan Speeds (in rpm): cpu: N/A
Info:      Processes: 222 Uptime: 21 min Memory: 1641.2/11403.5MB
           Client: Shell (bash) inxi: 2.3.5

I've been on Devuan for about a month now, and trying to debug this suspend issue off and on throughout. I've tried a few things (made sure the right permissions were given to my user to actually trigger it, for a start - added "rtl8821ae and xhci_hd" to MODULES_SUSPEND, etc - but no love. pm-hibernate produces the same results, and I've noticed when working through the pm_tests progression that I tend to fail at the `devices` test - freeze works just fine. I suspect that fixing the problem with suspend will likely also allow proper hibernation.

I'm assuming there's a bad module somewhere but I'm at wits end trying to figure out which it is, so any pointers in that regard would be helpful.

I'm including the pm-suspend.log file from the most recent round of debugging attempts. There's one in there where PM_Debug was set to true, but I'm at a loss as to how to interpret it. It's long, so I threw it on pastebin.

Log: https://pastebin.com/GJf6Ehh7

Apologies if this is an often-asked question - I did my best to grep around the forum looking for similar instances but I haven't been able to find a solution. I like Devuan and I'll probably keep it for my fileserver, but without these power management features working it won't be practical for me as a daily driver.

Last edited by pAtchSavage87 (2019-11-28 03:11:58)

pAtchSavage87 · 2019-11-29 14:53:27

One thing I neglected to mention was this line of testing:

# echo devices > /sys/power/pm_test
# echo platform > /sys/power/disk
# echo disk > /sys/power/state

Produces the same result, while echo freezer > /sys/power/pm_test completes successfully.

RJVB · 2020-08-09 09:06:20

I suppose the power LED doesn't go off (if you have one)?

As far as I can tell there's a regression that was introduced in the 4.19 kernel sometime after 4.19.118 (aka 4.19.0-9 in Debian speak). See https://unix.stackexchange.com/question … 19-0-9-ker

Phil · 2020-08-15 03:42:59

I see that you are running with and AMD processor, I also have an AMD A12 and I have run into several suspend issues over the last few years. Twice the kernel folks broke systems with AMD processors causing suspend issues when someone was updating some kernel stuff. I was running Fedora when this happened. My system would suspend but the screen would never come back on wake up. The screen was blank, black and the system would not respond to the keyboard.

I had the same issue when changing to Devuan Chimera. I fixed my problem by using synaptic and searching for amd. I installed the AMD firmware package, along with the firmware and graphics package which fixed my problem. My system suspends and hibernates properly now.

I hope this helps.

Phil

Phil · 2020-08-15 20:56:55

If you are using Beowolf, I would recommend that yuotry and older kernel. I do remember that problem with one version of the 4. something kernel and then again with a version of the 5 kernel.

I am running a Dell laptop with an AMD A12 processor and I remember that other folks had the same problem.

Good Luck fixing things.

Phil

Marjorie · 2020-08-15 23:22:03

Just to check:

As it's been suggested it may be a kernel issue have you tried using either of the two userspace, i.e. non-kernel, sleep methods (uspswap, tuxonice) instead? I recall I had to use uspswap on ascii but could use the kernel method again when I upgraded to Beowulf.

For suspend a black screen is the usual result of calling suspend and its almost instant. if you hibernate then you'll usually see some evidence of disk activity before it shuts down. How are to trying to resume? Pressing the power button is the usual method. Some laptops will suspend/resume when you close/open the lid but whether this happens depends on the laptop.

On my computer pm-utils is set up to call a screensaver on resume so that there's some superficial security (log in required).

Also (but it only affect hibernate so can't be your present problem) have you got a resume=/dev/dm-2 parameter in the linux /boot/vmlinuz... line(s) of your /boot/grub/grub.cfg? It should point to your swap partition, You need this for the system to know where to save/reload its memory state when hibernating.

Head_on_a_Stick · 2020-08-16 12:29:48

That inxi output is too opaque — which kernel driver is loaded for the video card?

lspci -k | grep -iA3 vga

I think it should be using amdgpu, which you should probably also add to /etc/initramfs-tools/modules (and then rebuild the initramfs).

The officially official Devuan Forum!

#1 2019-11-28 00:10:15

pm-suspend becomes catatonic - screen blanks; system never halts

#2 2019-11-29 14:53:27

Re: pm-suspend becomes catatonic - screen blanks; system never halts

#3 2020-08-09 09:06:20

Re: pm-suspend becomes catatonic - screen blanks; system never halts

#4 2020-08-15 03:42:59

Re: pm-suspend becomes catatonic - screen blanks; system never halts

#5 2020-08-15 20:56:55

Re: pm-suspend becomes catatonic - screen blanks; system never halts

#6 2020-08-15 23:22:03

Re: pm-suspend becomes catatonic - screen blanks; system never halts

#7 2020-08-16 12:29:48

Re: pm-suspend becomes catatonic - screen blanks; system never halts

Board footer