Kernel 4.14.2 = no nvidia, no bbswitch support

Support
21

With kernel 4.14.14-22 things are working now.

Building module:
cleaning build area....
make KERNELRELEASE=4.14.14-desktop-22omv KVERSION=4.14.14-desktop-22omv -j8...
/usr/sbin/dkms: linia 1326: /var/lib/dkms/bbswitch/0.8-6///module/symvers.gz: Nie ma takiego pliku ani katalogu
cleaning build area...

DKMS: build completed.

bbswitch:
Running module version sanity check.
 - Original module
   - No original module exists within this kernel
 - Installation
   - Installing to /lib/modules/4.14.14-desktop-22omv/kernel/drivers/acpi/

sed: -e wyraĹźenie #1, znak 6: nieznane polecenie: `m'
depmod................

DKMS: install completed.
22

Still it is not working.

journalctl --since=2018-01-25
led 25 11:24:23 localhost.localdomain sh[4350]: make -C /lib/modules/4.14.14-desktop-22omv/build M=/var/lib/dkms/nvidia-current/384.98-2/build clean
led 25 11:24:23 localhost.localdomain sh[4350]: make: Vstupuje se do adresáře „/usr/src/linux-4.14.14-desktop-22omv“
led 25 11:24:24 localhost.localdomain sh[4350]: make: Opouští se adresář „/usr/src/linux-4.14.14-desktop-22omv“
led 25 11:24:24 localhost.localdomain sh[4350]: { 'make' CC=gcc CXX=g++ SYSSRC=/lib/modules/4.14.14-desktop-22omv/build modules -j16; } >> /var/lib/dkms/nvidia-current/384.98-2/build/make.log 2>&1
led 25 11:24:53 localhost.localdomain sh[4350]: (bad exit status: 2)
led 25 11:24:53 localhost.localdomain sh[4350]: Error! Bad return status for module build on kernel: 4.14.14-desktop-22omv (x86_64)

For working boot I must use the older 4.13.12-desktop-2omv kernel.
I see many errors in the /var/lib/dkms/nvidia-current/384.98-2/build/make.log file.
See the small section of them bellow.
/var/lib/dkms/nvidia-current/384.98-2/build/nvidia/nv-acpi.c: In function ‘nv_acpi_remove_one_arg’:
/var/lib/dkms/nvidia-current/384.98-2/build/nvidia/nv-acpi.c:308:29: note: byref variable will be forcibly initialized
     struct acpi_object_list control_argument_list = { 0, NULL };
                             ^~~~~~~~~~~~~~~~~~~~~
/var/lib/dkms/nvidia-current/384.98-2/build/nvidia/nv-acpi.c: In function ‘nv_acpi_add’:
/var/lib/dkms/nvidia-current/384.98-2/build/nvidia/nv-acpi.c:190:29: note: byref variable will be forcibly initialized
     struct acpi_object_list control_argument_list = { 0, NULL };
                             ^~~~~~~~~~~~~~~~~~~~~
/var/lib/dkms/nvidia-current/384.98-2/build/nvidia/nv-acpi.c: In function ‘nv_acpi_nvif_method’:
/var/lib/dkms/nvidia-current/384.98-2/build/nvidia/nv-acpi.c:985:24: note: byref variable will be forcibly initialized
     struct acpi_buffer output = { ACPI_ALLOCATE_BUFFER, NULL };
                        ^~~~~~
/var/lib/dkms/nvidia-current/384.98-2/build/nvidia/nv-acpi.c:984:29: note: byref variable will be forcibly initialized
     struct acpi_object_list input;
                             ^~~~~
  CC [M]  /var/lib/dkms/nvidia-current/384.98-2/build/nvidia/nv-p2p.o
/var/lib/dkms/nvidia-current/384.98-2/build/nvidia/nv-acpi.c: In function ‘nv_acpi_mxmi_method’:
/var/lib/dkms/nvidia-current/384.98-2/build/nvidia/nv-acpi.c:821:24: note: byref variable will be forcibly initialized
     struct acpi_buffer output = { ACPI_ALLOCATE_BUFFER, NULL };
                        ^~~~~~
...
make[1]: *** [Makefile:1507: _module_/var/lib/dkms/nvidia-current/384.98-2/build] Chyba 2
make[1]: Opouští se adresář „/usr/src/linux-4.14.14-desktop-22omv“
make: *** [Makefile:81: modules] Chyba 2

Edit: Moderator took the liberty of posting code as code. < /> without the space.

23

Yes, bbswitch compiled fine with 4.14.14. But no nvidia yet.

24

The kernel modules for nVidia340 (340.106 mine version), VBoxAdditions and VirtulaBox are build with kernel 4.14.14-desktop-22

1 Like
25

Just in case it may help, it follows the make.log for

$dkms install -m nvidia-current -v 384.98-2nvidia_make.log.txt (107,0 KB)

26

I don’t have a nVidia hardware that use the current version.
But the 384.111 version build successful for kernel 4.14.14-desktop-22.

You have to wait until it’s available.

27

Update,

nVidia driver 390.25 is available.
http://www.nvidia.com/object/unix.html

28

Tried 387-34, from non-free testing, no luck. Waiting on availability of 390.25 on testing …

29

New kernel and new nvidia drivers are building

30

Installed new kernel 4.15 and nvidia 390.25 from testing repos and it seems to work. Also bbswitch builded fine.

The only thing that seems to be a problem here is that quite often after new kernel/nvidia drivers installation I get a blackscreen at reboot and have to launch XFdrake at recovery mode to fix. Notice that nvidia is not the main graphical card here, it is a dual graphic computer where the main graphical support comes from the intel graphical card. Why, then, installing nvidia or new kernel affects graphical initialization after all.

31

It seems new kernel 4.15 (and nvidia) is already available at main repos. But it does not appear to me to be installed since it has the same version as that I’ve downloaded from main/testing repos.
The package I’ve installed had Bad Signatures, so it is different from the new available one.

Why the new package has the same version? And if I reinstall it from the main?

32

Packages in testing repos are not signed by default, this is standard practice. The signature or key is added when packages are moved to updates repos.

I think you are talking about a package moved from testing repo to updates repo. The version number does not change because it is the same package. The only difference is that the signature is added. No reason to reinstall anything.

FWIW: “bad signatures” is not really true the error is “rpm: Medium without key” which means there is no signature. This is normal for testing repos.

33

Ok