Open NoMore201 opened 7 years ago
Did this occur with older kernel versions too?
Have you tried whether the nouveau driver with runtime PM works for you? That is the recommended approach if you don't use the binary nvidia driver.
Yes, it occurs with older kernel versions, unfortunately I cannot tell exactly from which version. I use the official nvidia driver, but I can check with the nouveau one.
I have exactly the same issue as NoMore201. After waking up the laptop Nvidia card is enabled and there's no way to disable it. There were no any issues with this before.
I'm using Arch Linux with the latest kernel (4.10.4) & mesa/Nvidia drivers (17.0.1 & 378.13, respectively). Intel HD4000 + Geforce GTX 760M.
Edit: Nothing to do with power-off, just forgot to recompile bbswitch after upgrading the kernel. I'm sorry about misinformation.
However, the main issue is still valid (bbswitch not working after suspend).
I have a quite stupid workaround for this. Basically a systemd service file which starts after resuming from suspend, which simply unloads and reloads the bbsiwtch module, like this
[Unit]
Description=<your description>
After=suspend.target
[Service]
Type=oneshot
ExecStart=rmmod bbswitch && modprobe bbswitch
As concerns the issue, I still didn't test with the nouveau driver, I will report back after I do (and maybe do some debugging).
Not sure if related but worth noting:
Bumblebee + Nvidia 378.13 driver: RmInitAdapter failed
Partial output of dmesg | grep -C 10 bbswitch:
command on my Arch Linux:
[ 376.216768] ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20160930/nsarguments-95) [ 376.419570] ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20160930/nsarguments-95) [ 377.744949] bbswitch: Unloaded. Discrete card 0000:01:00.0 is on [ 377.773648] bbswitch: version 0.8 [ 377.773653] bbswitch: Found integrated VGA device 0000:00:02.0: \_SB_.PCI0.GFX0 [ 377.773658] bbswitch: Found discrete VGA device 0000:01:00.0: \_SB_.PCI0.PEG0.PEGP [ 377.773669] ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20160930/nsarguments-95) [ 377.773745] bbswitch: detected an Optimus _DSM function [ 377.773752] bbswitch: device 0000:01:00.0 is in use by driver 'nvidia', refusing OFF [ 377.773753] bbswitch: Succesfully loaded. Discrete card 0000:01:00.0 is on [ 379.141629] ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20160930/nsarguments-95) [ 379.141683] ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20160930/nsarguments-95)
Kernel: 4.10.4-1-ARCH
Isn't this basically the same of the closed-non-closed #142?
@mirh Not sure, @NoMore201 can you post your info according to https://bugs.launchpad.net/lpbugreporter/+bug/752542?
You could find guides that arrived to contemplate tables are now exposed under /sys/firmware/acpi/tables/
and there's no need to install additional programs 😥
@Lekensteyn here is the output of the script ASUSTeK_COMPUTER_INC.-S551LN.tar.gz
FWIW, I'm experiencing the same issue here on 4.10.10 "disabling already-disabled device". Unloading and reloading the bbswitch module on resume works. ACPI DSDT info: Timi-TM1613.tar.gz.
Any updates on this issue?
pcie_port_pm=off
didn't fix the issue for me.
This occurs also using the nouveau module. To be more precise, after resuming from suspend
when using nouveau, /sys/kernel/debug/vgaswitcheroo/switch
shows
0:IGD:+:Pwr:0000:00:02.0
1:DIS: :DynOff:0000:04:00.0
which seems ok, but if I check power consumption using powertop, I get ~11W battery discharge rate against the ~6W on a fresh boot. Could it be an ACPI bug? Where should I report?
Distro: Archlinux Machine: ASUS S551LN Kernel: Linux hostname 4.10.1-1-ARCH SMP PREEMPT Sun Feb 26 21:08:53 UTC 2017 x86_64 GNU/Linux
The module is working fine, but after resuming the notebook from suspend the nvidia card stays on, there's no way I can disable it again (unless rebooting the system) and I get the following messages: