dkms unknow symbol mount (err0)

[vigilian]

problem at the installer, impossible to use dkms with my kernel apparently and I c'ant downgrade my kernel because of compatibility issues. Os: ubuntu 13.10 saucy kernel 3.11.0-14 ubuntu saucy version

I have tried with all version of virtualbox

[Frank Mehnert]

The information you provided is not sufficient. I don't see any log file, e.g. the file /var/log/vbox-install.log.

[vigilian]

i've didn't attached it before because it is useless in this case, it seems to be well installed but I've still have an error... ans still the system didn't recognize it as well installed it blocks me to run any VM...

[Frank Mehnert]

Then I would like to know what your actual problem is. You say that it is 'impossible to use dkms with your kernel' but you didn't mention what the actual problem is.

Also: What is the output of 'uname -a' on your host?

[vigilian]

not possible to run anything with virtualbox because it says that the vboxdrv is not setup and that's because the installation return an error when I'm trying to install virtualbox at the line "trying to register the virtualbox kernel modules using dkms" it failed everytime... and because of that I can't run any VM in virtualbox because then it tells me that the drivers aren't installed and I have to run vboxdrv again but it always the same....

And the error in dmesg is what I've put in the title...

You don't have to make me fool, if you put the title in google you will have a lot of hits and it's a pretty common problem for years now and we nver had a good answer to that problem. That's why people go see the product of vmware or qemu.

[vigilian]

the uname -a response is Linux hostname 3.11.0-14-generic #21-Ubuntu

[vigilian]

and if you don't believe it I can put you a screenshot

[Frank Mehnert]

I don't like your tone. If you want help then provide facts and don't insult people which try to help you. The first step is to provide an exact description of your problem and not just a few words without any log file.

We had a number of reports in the past from people where the VirtualBox kernel modules could not be installed. In many cases I saw it was a mistake done by the users. With VBox 4.3.4 I'm not aware of any problem with DKMS, therefore I'm asking. Googling "VirtualBox dkms 4.3.4" does not show me any relevant hit.

And actually the symbol 'mount' is not part of the VirtualBox host kernel modules. Therefore I would really like to see a screenshot from your problem, or better the complete output of 'dmesg' on your host.

[Frank Mehnert]

The output of the 'uname -a' command you posted in comment 5 was not complete.

I took my time and set up an Ubuntu 13.10 guest locally. The output of 'uname -a' is:

Linux vbox-VirtualBox 3.11.0-14-generic #21-Ubuntu SMP Tue Nov 12 17:04:55 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux

On this host I installed dkms and after that the VirtualBox 4.3.4 package for 64-Bit Ubuntu 13.04/13.10 (virtualbox-4.3_4.3.4-91027~Ubuntu~raring_amd64.deb). It installed without any issue, the kernel modules were automatically compiled using DKMS and I could start VMs without any problem.

When you mention the strange 'mount (err0)' failure I get more and more the feeling that this error comes from compiling kernel modules unrelated to VirtualBox. But I hope the output of 'dmesg' will provide the necessary information.

[vigilian]

Linux mnserver 3.11.0-14-generic #21-Ubuntu SMP TUe Nov 12 17:04:55 UTC 2013 x86_64 x86_64 x86_64 GNU/linux

[Frank Mehnert]

The provided output of 'dmesg' does not contain any error message. And there is no sign of any 'mount (err0)' message at all.

[vigilian]

sorry my bad wrong one

[vigilian]

sorry i can't get the dmesg updated. it show it on the command "dmesg" but not on the file in /var/log/dmesg. So i f you have another way to do it i'm open to all suggestions

[vigilian]

anyway you have all the screenshots to believe me . I could even open you a ssh access

[Frank Mehnert]

You can create a file containing the output of 'dmesg' by

dmesg > ~/dmesg.txt

After that a file dmesg.txt will be located in your home directory.

Looking at your last screenshot shows that the actual error message is

vboxdrv: Unknown symbol mcount (err 0)

And this is completely different from your original description! This makes a difference because 'mcount' hints to some profiling option enabled in your kernel.

Next please attach the vboxdrv.ko kernel module which does not load to this ticket and also attach the file /var/lib/dkms/vboxhost/4.3.4/3.11.0.-14-generic/x86_64/log/make.log file.

[vigilian]

okey no problem then i'm fully capable to understand that I've done a mistake by being upset after you. I will explain me after we've solved this thing. give me 3 minutes

[vigilian]

done

[Frank Mehnert]

Interesting. The make.log already shows that there are undefined 'mcount' symbols. And I see this symbol in your module while I don't see this symbol in the module on my local system. Unfortunately the actual command line for compiling the modules is not recorded. Please could you also provide the complete output of

gcc -v

Best attach it as file (gcc -v > ~/gcc.txt).

[vigilian]

it appear to not working (this command I mean, to put it in a file. it shouws me on screen nothing in the file) will try something else

[Frank Mehnert]

Sorry, please try

gcc -v > ~/gcc.txt

[vigilian]

I was meaning the gcc -v > name_of_thefile

well tee doesn't work either... I will put you the output here then :

vigilian@mnserver:/var/log$ gcc -v
Using built-in specs.
Target: x86_64-linux-gnu
Configured with: ../src/configure -v --with-pkgversion='Ubuntu/Linaro 4.4.7-2ubuntu2' --with-bugurl=file:///usr/share/doc/gcc-4.4/README.Bugs --enable-languages=c,c++,fortran --prefix=/usr --program-suffix=-4.4 --enable-shared --enable-linker-build-id --with-system-zlib --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --with-gxx-include-dir=/usr/include/c++/4.4 --libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug --disable-werror --with-arch-32=i686 --with-tune=generic --enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu
Thread model: posix
gcc version 4.4.7 (Ubuntu/Linaro 4.4.7-2ubuntu2)

[Frank Mehnert]

Ok, that's the reason for your problem. Your gcc version is outdated. Your kernel was compiled with gcc-4.8.1-10ubuntu9 while you are now trying to compile the modules with gcc-4.4.7. This will not work.

dpkg -l gcc

should show 4:4.8.1-2ubuntu3 installed but it will most likely show a 4.4.7 version installed. If the command shows 4.8.1 on your box as well then you have a 2nd gcc version installed which takes preference over 4.8.1.

I don't see an easy way how the VirtualBox installer could check that both gcc versions match.

[vigilian]

okey I understand. Now sorry to have been impolite earlier but for the record it's not the first time that I have this problem. some years ago again with different kernel and different ubuntu of course. And if you put this error in google (I mean the mcount error) you'll see a lot of post in forums with different kernel and different os sometimes. So I don't say that the users are a bit stupid like me but maybe there isn't enough basic documentation like a "must have before installing" section which precise maybe all the required component like gcc and the version needed and maybe also the command line to check it jsut in case. Especially, since the Linux Os are opening to the market to consumer users like Steam platform introduction. Anyway thank you for your assistance, I will do a wiki on the subject with the error message as title, Think it will be a good help to a lot of people.

[vigilian]

I'm confirming that it was the solution of the problem, just solved it

[Frank Mehnert]

Great, closing this ticket.