VMWare + Latest Kernel

[Robert Locke]

On Wed, 2005-11-30 at 11:27 +0800, Edward Dekkers wrote:
> Robert Locke wrote:
> > On Wed, 2005-11-30 at 09:04 +0800, Edward Dekkers wrote:
> > 
> >>Below a log of vmware building the vmnet module and failing. This has 
> >>worked up until my latest update. Don't bother with the URL links, as 
> >>I've tried them and one is a page not found the other has no 
> >>troubleshooting. Does one of the programmers on this list know what this 
> >>is? The newsgroup link from the page seems to be broken also.
> >>
> >>---
> >>
> >>Building the vmnet module.
> >>
> >>Using 2.6.x kernel build system.
> >>make: Entering directory `/tmp/vmware-config0/vmnet-only'
> >>make -C /lib/modules/2.6.14-1.1637_FC4smp/build/include/.. SUBDIRS=$PWD 
> >>SRCROOT=$PWD/. modules
> >>make[1]: Entering directory `/usr/src/kernels/2.6.14-1.1637_FC4-smp-i686'
> >>   CC [M]  /tmp/vmware-config0/vmnet-only/driver.o
> >>In file included from /tmp/vmware-config0/vmnet-only/vnet.h:14,
> >>                  from /tmp/vmware-config0/vmnet-only/vnetInt.h:10,
> >>                  from /tmp/vmware-config0/vmnet-only/driver.c:40:
> >>/tmp/vmware-config0/vmnet-only/vm_atomic.h:54:5: warning: "_MSC_VER" is 
> >>not defined
> >>In file included from /tmp/vmware-config0/vmnet-only/vm_oui.h:13,
> >>                  from /tmp/vmware-config0/vmnet-only/vnetInt.h:11,
> >>                  from /tmp/vmware-config0/vmnet-only/driver.c:40:
> >>/tmp/vmware-config0/vmnet-only/vm_basic_asm.h:48:5: warning: "_MSC_VER" 
> >>is not defined
> >>/tmp/vmware-config0/vmnet-only/driver.c: In function ˜VNetProcessOwnsPort:
> >>/tmp/vmware-config0/vmnet-only/driver.c:1698: error: struct files_struct 
> >>has no member named max_fds
> >>make[2]: *** [/tmp/vmware-config0/vmnet-only/driver.o] Error 1
> >>make[1]: *** [_module_/tmp/vmware-config0/vmnet-only] Error 2
> >>make[1]: Leaving directory `/usr/src/kernels/2.6.14-1.1637_FC4-smp-i686'
> >>make: *** [vmnet.ko] Error 2
> >>make: Leaving directory `/tmp/vmware-config0/vmnet-only'
> >>Unable to build the vmnet module.
> >>
> >>For more information on how to troubleshoot module-related problems, please
> >>visit our Web site at 
> >>"http://www.vmware.com/download/modules/modules.html" and
> >>"http://www.vmware.com/support/reference/linux/prebuilt_modules_linux.html".
> >>
> >>Execution aborted.
> >>
> > 
> > 
> > OK, I'll ask the obvious questions...
> > 
> > What version of VMWare?  I had no problem building VMWare Workstation
> > 5.0 on both that kernel 1637 and the newer 1644.
> 
> 5.0.0
> 
> > 
> > So next obvious check, are you using the latest vmware-any-any-update96?
> > Appears to be needed particularly if you are using VMWare 4.5.2....
> 
> Nope - didn't even know about it. Is this still required for version 5? 
> If so do I grab that from vmware?

Not sure if it is necessary for 5.0, though I have applied it.  Though
usually it comes back and tells me there was nothing to do.  It is
hosted on this guys site at:

http://knihovny.cvut.cz/ftp/pub/vmware/

Apparently the guy is a developer for VMWare, I guess, cause the stuff
does work it's way back to them, but is unofficial except that everyone
points to it even on the VMWare Forums....

> 
> Understand my company bought vmware, threw it at me to install on the 
> workstation here. I managed to get it working every time before when 
> there were updates and I had to re-compile, just not this time. I don't 
> know a hell of a lot about it, I just used the perl script and left the 
> defaults as they were initially set up, it would simply compile the 
> vmnet.ko module.
> 
> Not this time though :(
> 
> > 
> > If using the released today 5.5, I have no idea yet as I haven't had a
> > chance to build those yet.
> 
> No way - this is at least a couple of months old.
> 
> > 
> > The only difference in my environment that I can see is that I am not
> > smp - using the UP kernel...
> > 
> > Other thought as to a difference.  I noticed a gcc update announcement
> > today that I have yet to get so perhaps that is different - I am still
> > running: gcc-4.0.1-4.fc4....
> > 
> > --Rob
> 
> Well it broke after these updates (from what I can tell):
> 
> Package Installed:
>      dlm-kernel.i686 2.6.11.5-20050601.152643.FC4.16
>      gnbd-kernel.i686 2.6.11.2-20050420.133124.FC4.52
>      GFS-kernel.i686 2.6.11.8-20050601.152643.FC4.19
>      cman-kernel.i686 2.6.11.5-20050601.152643.FC4.17
> 
>   Package Updated:
>      dlm-kernheaders.i686 2.6.11.5-20050601.152643.FC4.16
>      openswan-doc.i386 2.4.4-1.0.FC4.1
>      GFS-kernheaders.i686 2.6.11.8-20050601.152643.FC4.19
>      openswan.i386 2.4.4-1.0.FC4.1
>      perl.i386 3:5.8.6-16.fc4
>      gnbd-kernheaders.i686 2.6.11.2-20050420.133124.FC4.52
>      cman-kernheaders.i686 2.6.11.5-20050601.152643.FC4.17
>      perl-suidperl.i386 3:5.8.6-16.fc4
> 
> Thanks so far.

Have to admit that I don't have much of that installed.  Are you really
running the clustering stuff?  Or did you just do an "Everything"
install?  You might want to consider paring down some of the "packages
installed" to see if you can eliminate the problem....  The kernheader
updates might be conflicting???  Also, the kernel is not in the list?
Of course, you should only need to re-run vmware-config.pl when the
kernel itself and kernel-devel packages (or in your case, kernel-smp and
kernel-smp-devel) get installed....

Also, do you have the UP kernel installed and kernel-devel for that
version?  You might want to try booting in to the UP kernel and seeing
if VMWare will compile for that.  If it works there but not in
kernel-smp with kernel-smp-devel, then perhaps there is a regression in
the smp kernel and we have narrowed it down to just that...

I would try the vmware-any-any-update tarball from the above site first.
If that doesn't resolve it, then I would try compiling it with the UP
kernel, and if that doesn't resolve it, I would begin attacking some of
the other kernheader related packages that are depositing modules that I
don't use on to the system....

Heading to bed, but good luck....

--Rob