GT 650M in Acer Aspire v3 with 325.15/Kernel 3.10.5 not working

I’m still having problems with kernel 3.10.
But at least the card no longer dropps of the PCI bus…

lspci -nn

01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GK107M [GeForce GT 650M] [10de:0fd1] (rev a1)

dmesg

[36393.757065] bbswitch: enabling discrete graphics
[36395.034809] [drm:__gen6_gt_force_wake_mt_get] *ERROR* Timed out waiting for forcewake to ack request.
[36395.227630] [drm:__gen6_gt_force_wake_mt_get] *ERROR* Timed out waiting for forcewake to ack request.
[36395.770190] pci 0000:01:00.0: power state changed by ACPI to D0
[36398.592219] vgaarb: device changed decodes: PCI:0000:01:00.0,olddecodes=none,decodes=none:owns=none
[36398.640463] NVRM: The NVIDIA GPU 0000:01:00.0 (PCI ID: 10de:0fd1)
NVRM: installed in this system is not supported by the 325.15
NVRM: NVIDIA Linux driver release.  Please see 'Appendix
NVRM: A - Supported NVIDIA GPU Products' in this release's
NVRM: README, available on the Linux driver download page
NVRM: at www.nvidia.com.
[36398.688674] nvidia: probe of 0000:01:00.0 failed with error -1
[36398.688918] NVRM: The NVIDIA probe routine failed for 1 device(s).
[36398.688920] NVRM: None of the NVIDIA graphics adapters were initialized!
[36398.688921] [drm] Module unloaded

xorg.8.log

[  2636.854] (==) NVIDIA(0): Depth 24, (==) framebuffer bpp 32
[  2636.854] (==) NVIDIA(0): RGB weight 888
[  2636.854] (==) NVIDIA(0): Default visual is TrueColor
[  2636.854] (==) NVIDIA(0): Using gamma correction (1.0, 1.0, 1.0)
[  2636.855] (**) NVIDIA(0): Option "NoLogo" "true"
[  2636.855] (**) NVIDIA(0): Option "ProbeAllGpus" "false"
[  2636.855] (**) NVIDIA(0): Option "UseEDID" "false"
[  2636.855] (**) NVIDIA(0): Option "UseDisplayDevice" "none"
[  2636.855] (**) NVIDIA(0): Enabling 2D acceleration
[  2636.855] (**) NVIDIA(0): Ignoring EDIDs
[  2636.855] (**) NVIDIA(0): Option "UseDisplayDevice" set to "none"; enabling NoScanout
[  2636.855] (**) NVIDIA(0):     mode
[  2641.917] (EE) NVIDIA(0): Failed to initialize the NVIDIA GPU at PCI:1:0:0.  Please
[  2641.917] (EE) NVIDIA(0):     check your system's kernel log for additional error
[  2641.917] (EE) NVIDIA(0):     messages and refer to Chapter 8: Common Problems in the
[  2641.917] (EE) NVIDIA(0):     README for additional information.
[  2641.917] (EE) NVIDIA(0): Failed to initialize the NVIDIA graphics device!
[  2641.917] (EE) NVIDIA(0): Failing initialization of X screen 0
[  2641.917] (II) UnloadModule: "nvidia"
[  2641.917] (II) UnloadSubModule: "shadow"
[  2641.917] (II) UnloadSubModule: "wfb"
[  2641.917] (II) UnloadSubModule: "fb"
[  2641.917] (EE) Screen(s) found, but none have a usable configuration.
[  2641.917] (EE) 
Fatal server error:
[  2641.917] (EE) no screens found(EE)

Is this problem happening after a suspend/resume cycle? It sounds like the GPU is being powered off before the driver is loaded.

This problem can be reproduced 100% of the time on my machine. It doesn’t matter if re/started, suspend/resume, clean shutdown, …

I can also reproduce it with the card in a “warm” state:

tee /proc/acpi/bbswitch <<< ON

After the first failed attempt however I can’t turn the card back on. It completely ignores acpi calls.
The other problem I have is the rather restrictive bios. I can’t activate/deactivate or choose optimus from the bios.

The card woks fine with kernel 3.9.9 and nvidia 319.32

No luck at all with 3.10.x and nvidia 325.15.

Does the problem still occur if the bbswitch module is blacklisted?

Yes also if bbswitch is blacklisted.
Now on 4.10.7:

$ cat /etc/modprobe.d/modprobe.conf

blacklist nouveau
install bbswitch /usr/bin/false
options nvidia NVreg_EnableMSI=1 (with or without)

$ lspci

01:00.0 VGA compatible controller: NVIDIA Corporation GK107M [GeForce GT 650M] (rev a1)

$ dmesg (after trying to run glxsphears)

[  469.132238] nvidia: module license 'NVIDIA' taints kernel.
[  469.132242] Disabling lock debugging due to kernel taint
[  469.136974] nvidia 0000:01:00.0: enabling device (0006 -> 0007)
[  469.137034] vgaarb: device changed decodes: PCI:0000:01:00.0,olddecodes=io+mem,decodes=none:owns=none
[  469.137226] [drm] Initialized nvidia-drm 0.0.0 20130102 for 0000:01:00.0 on minor 1
[  469.137233] NVRM: loading NVIDIA UNIX x86_64 Kernel Module  325.15  Wed Jul 31 18:50:56 PDT 2013
[  470.119922] nvidia 0000:01:00.0: irq 48 for MSI/MSI-X
[  474.527152] NVRM: GPU at 0000:01:00.0 has fallen off the bus.
[  474.527166] NVRM: os_pci_init_handle: invalid context!
[  474.527169] NVRM: os_pci_init_handle: invalid context!
[  474.527178] NVRM: GPU at 0000:01:00.0 has fallen off the bus.
[  474.527185] NVRM: os_pci_init_handle: invalid context!
[  474.527187] NVRM: os_pci_init_handle: invalid context!
[  474.554858] NVRM: RmInitAdapter failed! (0x25:0x28:1157)
[  474.554867] NVRM: rm_init_adapter(0) failed

$ lspci (after trying to run glxsphears)

01:00.0 VGA compatible controller: NVIDIA Corporation GK107M [GeForce GT 650M] (rev ff)

Problem still present on linux 3.11.0 and linux 3.10.x, nvidia 325.15 and nvidia 319.49
tested with bbswitch enabled/disabled

dmesg after a test:

[   59.721840] bbswitch: enabling discrete graphics
[   60.142619] pci 0000:01:00.0: power state changed by ACPI to D0
[   63.287960] nvidia: module license 'NVIDIA' taints kernel.
[   63.287964] Disabling lock debugging due to kernel taint
[   63.293077] vgaarb: device changed decodes: PCI:0000:01:00.0,olddecodes=io+mem,decodes=none:owns=none
[   63.293268] [drm] Initialized nvidia-drm 0.0.0 20130102 for 0000:01:00.0 on minor 1
[   63.293275] NVRM: loading NVIDIA UNIX x86_64 Kernel Module  325.15  Wed Jul 31 18:50:56 PDT 2013
[   64.270094] nvidia 0000:01:00.0: irq 48 for MSI/MSI-X
[   64.275131] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20130517/nsarguments-95)
[   64.275165] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20130517/nsarguments-95)
[   64.275181] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20130517/nsarguments-95)
[   64.275195] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20130517/nsarguments-95)
[   64.275209] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20130517/nsarguments-95)
[   64.275223] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20130517/nsarguments-95)
[   64.275269] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20130517/nsarguments-95)
[   64.275284] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20130517/nsarguments-95)
[   68.336754] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20130517/nsarguments-95)
[   68.362027] NVRM: GPU at 0000:01:00.0 has fallen off the bus.
[   68.362035] NVRM: os_pci_init_handle: invalid context!
[   68.362037] NVRM: os_pci_init_handle: invalid context!
[   68.362042] NVRM: GPU at 0000:01:00.0 has fallen off the bus.
[   68.362046] NVRM: os_pci_init_handle: invalid context!
[   68.362048] NVRM: os_pci_init_handle: invalid context!
[   68.389746] NVRM: RmInitAdapter failed! (0x25:0x28:1157)
[   68.389756] NVRM: rm_init_adapter(0) failed

faither, Please provide nvidia-bug-report with reproduction steps.

Sent in several bug reports via e-mail.
Hope this helps resolving the problem.

filed bug 1372956 to track this issue …

Just a little update:
Card is still falling off the bus with linux 3.11.4 and nvidia-beta 331.13

We reproduced issue and investigated on: Dell Audi Turbo A7 Inspiron 7720 + Debian GNU/Linux sid . And found this bug in linux kernel ACPI driver.