legacy nvidia drivers fail to work with latest kernel

[Darko Veberic]

i have a MacBookPro3,1 with a NVIDIA Corporation G84M [GeForce 8600M GT] (rev a1) video card. while in ubuntu 15.10 everything worked fine, after the upgrade in 16.04 the legacy nvidia-304 drivers do not. the X11 fails to start, the relevant lines in the /var/log/Xorg.0.log file are

[ 3021.745] (EE) NVIDIA(0): Failed to initialize the NVIDIA GPU at PCI:1:0:0. Please
[ 3021.745] (EE) NVIDIA(0): check your system's kernel log for additional error
[ 3021.745] (EE) NVIDIA(0): messages and refer to Chapter 8: Common Problems in the
[ 3021.745] (EE) NVIDIA(0): README for additional information.
[ 3021.745] (EE) NVIDIA(0): Failed to initialize the NVIDIA graphics device!
[ 3021.745] (EE) NVIDIA(0): Failing initialization of X screen 0
[ 3021.745] (II) UnloadModule: "nvidia"

and the journalctl --dmesg gives

kernel: NVRM: failed to copy vbios to system memory.
kernel: NVRM: RmInitAdapter failed! (0x30:0xffffffff:867)
kernel: NVRM: rm_init_adapter(0) failed

nevertheless, the nvidia kernel module is loaded without a problem.

i also tried more recent versions of nvidia-X but without success. they either fail with the same messages as above or they do not support the video card.

ProblemType: Bug
DistroRelease: Ubuntu 16.04
Package: linux-image-4.4.0-6-generic 4.4.0-6.21
ProcVersionSignature: Ubuntu 4.4.0-6.21-generic 4.4.1
Uname: Linux 4.4.0-6-generic x86_64
NonfreeKernelModules: nvidia
ApportVersion: 2.20-0ubuntu3
Architecture: amd64
AudioDevicesInUse: Error: command ['fuser', '-v', '/dev/snd/by-path', '/dev/snd/hwC0D0', '/dev/snd/pcmC0D2c', '/dev/snd/pcmC0D1c', '/dev/snd/pcmC0D1p', '/dev/snd/pcmC0D0c', '/dev/snd/pcmC0D0p', '/dev/snd/controlC0', '/dev/snd/seq', '/dev/snd/timer'] failed with exit code 1:
CRDA:
 country AW: DFS-ETSI
  (2402 - 2482 @ 40), (N/A, 20), (N/A)
  (5170 - 5250 @ 80), (N/A, 20), (N/A)
  (5250 - 5330 @ 80), (N/A, 20), (0 ms), DFS
  (5490 - 5710 @ 160), (N/A, 27), (0 ms), DFS
Date: Fri Feb 19 18:05:07 2016
HibernationDevice: RESUME=UUID=484d04a2-0891-47bb-b890-ee3ab0189ef7
InstallationDate: Installed on 2016-02-17 (2 days ago)
InstallationMedia: Ubuntu-Server 15.10 "Wily Werewolf" - Release amd64 (20151021)
MachineType: Apple Inc. MacBookPro3,1
ProcFB: 0 EFI VGA
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-4.4.0-6-generic.efi.signed root=UUID=fddff8cb-9428-47b7-a215-034c4b79c0fe ro
PulseList: Error: command ['pacmd', 'list'] failed with exit code 1: No PulseAudio daemon running, or not running as session daemon.
RelatedPackageVersions:
 linux-restricted-modules-4.4.0-6-generic N/A
 linux-backports-modules-4.4.0-6-generic N/A
 linux-firmware 1.156
SourcePackage: linux
UpgradeStatus: Upgraded to xenial on 2016-02-17 (1 days ago)
dmi.bios.date: 03/05/08
dmi.bios.vendor: Apple Inc.
dmi.bios.version: MBP31.88Z.0070.B07.0803051658
dmi.board.asset.tag: Base Board Asset Tag
dmi.board.name: Mac-F4238BC8
dmi.board.vendor: Apple Inc.
dmi.board.version: PVT
dmi.chassis.asset.tag: Asset Tag#
dmi.chassis.type: 2
dmi.chassis.vendor: Apple Inc.
dmi.chassis.version: Mac-F4238BC8
dmi.modalias: dmi:bvnAppleInc.:bvrMBP31.88Z.0070.B07.0803051658:bd03/05/08:svnAppleInc.:pnMacBookPro3,1:pvr1.0:rvnAppleInc.:rnMac-F4238BC8:rvrPVT:cvnAppleInc.:ct2:cvrMac-F4238BC8:
dmi.product.name: MacBookPro3,1
dmi.product.version: 1.0
dmi.sys.vendor: Apple Inc.

[Brad Figg (brad-figg)]

This change was made by a bot.

[penalvch (penalvch)]

Darko Veberic, thank you for reporting this and helping make Ubuntu better.

To advise, as per nvidia.com the latest driver version for your chipset is https://launchpad.net/ubuntu/+source/nvidia-graphics-drivers-340 . Hence, could you please uninstall the legacy package, install the 340 version and advise if this works?

[Darko Veberic (darko-veberic-kit)]

ok, i purged all nvidia related packages and installed your request. the packages now are

libcuda1-340 340.96-0ubuntu2
nvidia-340 340.96-0ubuntu2
nvidia-opencl-icd-340 340.96-0ubuntu2
nvidia-prime 0.8.2
nvidia-settings 361.28-0ubuntu1

unfortunately, the x11 still does not start. the Xorg.0.log is the same as above but the journalctl --dmesg is slightly different:

kernel: NVRM: failed to copy vbios to system memory.
kernel: NVRM: RmInitAdapter failed! (0x30:0xffffffff:747)
kernel: NVRM: rm_init_adapter failed for device bearing minor number 0
kernel: NVRM: nvidia_frontend_open: minor 0, module->open() failed, error -5

i hope this helps. it was really nice to revive these old macs with ubuntu...

[Darko Veberic (darko-veberic-kit)]

i forgot to mention that these macs do not work with nouveau so that the legacy nvidia drivers are the only hope.

[penalvch (penalvch)]

Darko Veberic, to clarify, only on the MacBookPro3,1 (not other hardware) does using the nouveau driver allow x to start?

If not, you would want to file a new report about that not working by ensuring you have the package xdiagnose installed, and that you click the Yes button for attaching additional debugging information running the following from a terminal:
ubuntu-bug xorg

Also, please feel free to subscribe me to it.

For more on why this is helpful, please see https://wiki.ubuntu.com/ReportingBugs.

[Darko Veberic (darko-veberic-kit)]

this bug post is only about MacBookPro3,1.

in ubuntu 15.10 the nouveau driver did not work but nvidia drivers did. now, in ubuntu 16.04, the nvidia drivers also do not work.

just to see why nouveau fails i uninstalled all nvidia packages and purged dependencies. while booting without splash the output in the console stops at the message "fb: switching to nouveaufb from EFI VGA" and the dmesg has some more:

fb: switching to nouveaufb from EFI VGA
Console: switching to colour dummy device 80x25
nouveau 0000:01:00.0: enabling device (0002 -> 0003)
nouveau 0000:01:00.0: NVIDIA G84 (084700a2)
nouveau 0000:01:00.0: Invalid ROM contents
nouveau 0000:01:00.0: bios: unable to locate usable image
nouveau 0000:01:00.0: bios ctor failed, -22
nouveau: probe of 0000:01:00.0 failed with error -22

i never really tried to get nouveau working with 15.10 so i cannot tell whether this is a regression or not.

do you want me to file a separate bug report on nouveau on MacBookPro3,1?

[Darko Veberic (darko-veberic-kit)]

after some more digging the problems might be related to the fact that the disk partition table was left in the hybrid GPT/MBR format and thus the mac is booting in the efi mode. according to http://www.rodsbooks.com/ubuntu-efi/ this might cause limited support for nvidia graphics card and related bios rom problems. in other words, the mac is not booting in the bios compatibility mode.

since ubuntu is the only operating system on the machine, i will reinstall it by wiping the disk and using a dos partition table. this should force the mac to boot into the bios mode. unfortunately, since ubuntu 16.04 desktop dvd does not boot via the optical drive, i have to use ubuntu 15.10 server cd and then do-release-upgrade since 16.04 desktop does not fit into a standard cd.

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[MensVeritas (mensveritas)]

Also experiencing this issue, on a Sony Vaio laptop. The card is an Nvidia GeForce 8400M GT. It was working under 15.10.

The latest driver in the graphics-drivers PPA, nvidia-364, does not support the video card:

[ 243.591384] NVRM: The NVIDIA GeForce 8400M GT GPU installed in this system is
               NVRM: supported through the NVIDIA 340.xx Legacy drivers. Please
               NVRM: visit http://www.nvidia.com/object/unix.html for more
               NVRM: information. The 364.19 NVIDIA driver will ignore
               NVRM: this GPU. Continuing probe...

The current recommended driver in the graphics-drivers PPA, nvidia-352, also does not support the card:

[ 183.787047] NVRM: The NVIDIA GeForce 8400M GT GPU installed in this system is
               NVRM: supported through the NVIDIA 340.xx Legacy drivers. Please
               NVRM: visit http://www.nvidia.com/object/unix.html for more
               NVRM: information. The 361.42 NVIDIA driver will ignore
               NVRM: this GPU. Continuing probe...

For some reason installing the package nvidia-352 also installs the package nvidia-361. The one option in the graphics-drivers PPA after the driver-recommended version 340 but before the PPA-recommended version 352 is 346. Installing nvidia-346 also installs nvidia-352 and nvidia-361 and so gives the same message as nvidia-352.

The version those later driver versions recommend, 340, seemingly fails in some way. It apparently does load but then gives the same error mentioned in the OP. This happens both when I install that version of the driver manually, either nvidia-340 or nvidia-340-updates, using `sudo apt-get install nvidia-340` or `sudo apt-get install nvidia-340-updates`:

[ 17.015956] nvidia: module license 'NVIDIA' taints kernel.
[ 17.015961] Disabling lock debugging due to kernel taint
[ 17.028907] nvidia: module verification failed: signature and/or required key missing - tainting kernel
[ 17.036538] vgaarb: device changed decodes: PCI:0000:01:00.0,olddecodes=io+mem,decodes=none:owns=io+mem
[ 17.037284] [drm] Initialized nvidia-drm 0.0.0 20150116 for 0000:01:00.0 on minor 0
[ 17.037295] NVRM: loading NVIDIA UNIX x86_64 Kernel Module 340.96 Sun Nov 8 22:33:28 PST 2015
[ 17.335519] nvidia_uvm: Loaded the UVM driver, major device number 246
[ 17.757950] NVRM: failed to copy vbios to system memory.
[ 17.758187] NVRM: RmInitAdapter failed! (0x30:0xffffffff:747)
[ 17.758215] NVRM: rm_init_adapter failed for device bearing minor number 0
[ 17.758284] NVRM: nvidia_frontend_open: minor 0, module->open() failed, error -5

And when I let Ubuntu install it automatically, using `sudo ubuntu-drivers autoinstall`:

[ 20.921331] nvidia: module license 'NVIDIA' taints kernel.
[ 20.921336] Disabling lock debugging due to kernel taint
[ 20.934363] nvidia: module verification failed: signature and/or required key missing - tainting kernel
[ 20.941739] vgaarb: device changed decodes: PCI:0000:01:00.0,olddecodes=io+mem,decodes=none:owns=io+mem
[ 20.942365] [drm] Initialized nvidia-drm 0.0.0 20150116 for 0000:01:00.0 on minor 0
[ 20.942376] NVRM: loading NVIDIA UNIX x86_64 Kernel Module ...

[MensVeritas (mensveritas)]

Have just tried using the manual install package from Nvidia, version 340.96. Same messages in the dmesg log:

[ 30.269523] NVRM: failed to copy vbios to system memory.
[ 30.269758] NVRM: RmInitAdapter failed! (0x30:0xffffffff:747)
[ 30.269781] NVRM: rm_init_adapter failed for device bearing minor number 0
[ 30.269851] NVRM: nvidia_frontend_open: minor 0, module->open() failed, error -5

[penalvch (penalvch)]

MensVeritas, it will help immensely if you filed a new report with Ubuntu by ensuring you have the package xdiagnose installed, and that you click the Yes button for attaching additional debugging information running the following from a terminal:
ubuntu-bug xorg

Also, please feel free to subscribe me to it.

For more on why this is helpful, please see https://wiki.ubuntu.com/ReportingBugs.

[MensVeritas (mensveritas)]

Thanks Christopher. Am now suspecting it may be a heat problem. Turns out the internal fan died; just happened to coincide with the update. I have a solution for that on the way & plan to try again without the overheating. Will come back to this after the repair.

[MensVeritas (mensveritas)]

I got some external cooling for this laptop & the problem went away. I've run some intensive GPGPU + CPU processing for some hours with no problems showing up. I have the 8400M GT running fine under Ubuntu 16.04. Removing myself from the "affects me" list.

[Nick Barry (itsnickbarry)]

I was able to fix this problem on a MacBook Pro 4,1 with a GeForce 8600M GT by following these instructions:
https://ubuntuforums.org/showthread.php?t=2209602

In this case, installing the current version of Ubuntu (16.04.1) via USB with a GUID partition table was not a problem.

[KeithG (grider-4)]

I get loginloop where it is trying to set the video mode but just resets. I cannot even login to a terminal as it constantly wipes it out and sends me to tty7. My card is:

  *-display
       description: VGA compatible controller
       product: G71GL [Quadro FX 3500] [10DE:29D]
       vendor: NVIDIA Corporation [10DE]
       physical id: 0
       bus info: pci@0000:01:00.0
       version: a1
       width: 64 bits
       clock: 33MHz
       capabilities: vga_controller bus_master cap_list rom
       configuration: driver=nouveau latency=0

Tried 304. 340, used the ppa and tried 331. All do the same thing. No login screen. I am running updated Ubuntu Gnome 16.04

[jiggles100 (jiggles)]

I have similar problems with nvidia-384.90, the only working solution for me is to use the 4.4 kernel instead.

linux-image-4.10 enters a loop of doom at startup
linux-image-4.13 runs for some random amount of time usually fails within 15 minutes but then hangs the system completely