You are not logged in.
Pages: 1
Hello,
Is there one version of virtualbox in AUR or somewhere installable for kermel 2.6.25 and x86_64 ?
Or if someone could point me whereto find the PKGBUILD for virtualbox-ose, the link to CVS from the AUR package file does not work.
Thanks,
Olivier
who had a virtualbox runnning nicely with 2.6.22 few week ago...
Offline
Hello,
Is there one version of virtualbox in AUR or somewhere installable for kermel 2.6.25 and x86_64 ?
The archlinuxfr repo [0] has VirtualBox 1.6.2 x86_64. I'm running it here on the x86_64 2.6.25 kernel.
Niels
MacBook4.1 - Arch64 - OpenBox WM
Offline
thanks for your help.
I am looking for a clean ARCH package if possible, not a tar.gz.
If someone else knows something, thanks!
Olivier
Offline
[archlinuxfr]
Server = http://repo.archlinux.fr/x86_64
add that to your pacman.conf and pacman -Sy virtualbox_bin
Offline
I am looking for a clean ARCH package if possible, not a tar.gz.
@ my64: I'm sorry, I should have been more explicit in my reply above. In my reply I wrote that the archlinuxfr repo had VirtualBox, and I gave you the URL to that repo. The url is not to download a tar.gz from, but what you enter in your pacman.conf file so pacman will pick up the repo and you can get the "clean ARCH package" when you do pacman -Sy virtualbox.
Actually noneus wrote it much better than myself, so just follow what noneus wrote and you'll be able to install the VirtualBox package without any problems.
Niels
MacBook4.1 - Arch64 - OpenBox WM
Offline
Thank you both for your clear explanations.
I have now been able to install Virtualbox 1.6.2.1, and I just realized that the "virtualbox_bin_additions" are not needed anymore.
Btw, version 1.6.2.2 works fine also, if you get youself manually the binary file from SUN site.
Olivier
Last edited by my64 (2008-07-20 14:47:22)
Offline
Thank you for providing a nice repos now I can use virtualbox on my x86_64 box
Last edited by kyanh (2008-07-31 13:32:54)
Offline
I read in the virtualbox 1.6.4 release notes that kernel 2.6.27 is supposed to be supported, but I didn't succeed in building the vbox modules against kernel 2.6.27-rc1. Is there a particular kernel option that I need to set prior to building the modules?
The error information is as follows:
In file included from /tmp/vbox.2/linux/SUPDrv-linux.c:35:
/tmp/vbox.2/SUPDRV.h:104:30: error: asm/semaphore.h: No such file or directory
/tmp/vbox.2/linux/SUPDrv-linux.c: In function 'supdrvOSGipResume':
/tmp/vbox.2/linux/SUPDrv-linux.c:1331: error: too many arguments to function 'smp_call_function'
make[2]: *** [/tmp/vbox.2/linux/SUPDrv-linux.o] Error 1
make[1]: *** [_module_/tmp/vbox.2] Error 2
make[1]: Leaving directory `/usr/src/linux-2.6.27-rc1-zenmm0-20080805'
make: *** [vboxdrv] Error 2
Note that I'm not running vanilla 2.6.27-rc1, but rather the zenmm mod.
Any ideas, anyone? Thanks.
Offline
I read in the virtualbox 1.6.4 release notes that kernel 2.6.27 is supposed to be supported, but I didn't succeed in building the vbox modules against kernel 2.6.27-rc1. Is there a particular kernel option that I need to set prior to building the modules?
The error information is as follows:
In file included from /tmp/vbox.2/linux/SUPDrv-linux.c:35: /tmp/vbox.2/SUPDRV.h:104:30: error: asm/semaphore.h: No such file or directory /tmp/vbox.2/linux/SUPDrv-linux.c: In function 'supdrvOSGipResume': /tmp/vbox.2/linux/SUPDrv-linux.c:1331: error: too many arguments to function 'smp_call_function' make[2]: *** [/tmp/vbox.2/linux/SUPDrv-linux.o] Error 1 make[1]: *** [_module_/tmp/vbox.2] Error 2 make[1]: Leaving directory `/usr/src/linux-2.6.27-rc1-zenmm0-20080805' make: *** [vboxdrv] Error 2
Note that I'm not running vanilla 2.6.27-rc1, but rather the zenmm mod.
Any ideas, anyone? Thanks.
Yes fixed this one, with some hacking
Change SUPDRV.h
Remove all the kernel version check stuff and just leave this(line 99):
# include <linux/semaphore.h>
edit linux/SUPDrv-linux.c
Line 1331
smp_call_function(VBoxDrvLinuxGipResumePerCpu, pDevExt, 0 /* retry */ );
This is the new api I found in a recent commit
Offline
Thank you, Inferno. I'm up and running now.
Offline
edit linux/SUPDrv-linux.c
Line 1331
smp_call_function(VBoxDrvLinuxGipResumePerCpu, pDevExt, 0 /* retry */ );This is the new api I found in a recent commit
Like you, I had edited this line after reading the error message about too many arguments and looking at the recent commits, but instead of setting the final argument to "0", I put it at "1".
This was just a shot in the dark; I had no idea what I was doing apart from eliminating one of the two arguments, but the result was that the module compiled. (I know, that was probably stupid, but my regular life is too safe and I wanted a little excitement.)
My question: does it matter whether line end "0" or "1"? I've been running vbox for two days now without problems, but I don't want to have laid a snare for myself. Thanks.
Offline
See here for the new version of the function smp_call_function:
http://zen-sources.org/cgi-bin/gitweb.c … 9ca39#l338
Comment is: fix: "smp_call_function: get rid of the unused nonatomic/retry argument"
int smp_call_function(void (*func)(void *), void *info, int wait)
@wait: If true, wait (atomically) until function has completed on other CPUs.
The original version from vbox is like this
smp_call_function(VBoxDrvLinuxGipResumePerCpu, pDevExt, 0 /* retry */, 1 /* wait */);
So I think you are right in changing the line to:
smp_call_function(VBoxDrvLinuxGipResumePerCpu, pDevExt, 1 /* wait */ );
So I made the function less safe, since I removed the waiting time. I will change my version to use a 1 for the wait option.
Last edited by Inferno (2008-08-08 07:30:07)
Offline
Pages: 1