|Subject: [NYLXS - HANGOUT] Virtualbox (was: Apple crimes)
Quoting Ruben Safir (mrbrklyn-at-panix.com):
> No - it is about the technology. I'm compiling kernels after
> tinkering with the scheduler and other kernel internals. So I set up a
> virtualbox to do this. When I do, I followed the Manjaro instructions
> located here.
> No some of this I'm not understanding. Virtualization is built into the
> hardware, so why do I have to use these OS and Kernel specific packages.
We discussed this briefly when we chatted yesterday on the 'phone,
though I didn't have the page in front of me at the time. I've now
pulled it up. Some packages are integration addons that make the guest
OS able to better deal with various services that cross the VM
interface, such as host-side printer services and physical USB devices.
It's common to have such 'extension packs' for VM environments (VMware,
VirtualBox, etc.) to improve the cross-divide functionality that you get
by default. Clipboard integration, as one prime example, is really nice
and helps you not go insane from the frustration of not being able to
copy/paste between host and guest.
Then there are also (always) host-side kernel modules you need to run
the VM at all. Those add things like routing for virtual networking.
The host-side kernel modules are obligatory. The guest-side 'extension
packs' aren't, but you'll want them.
> And more immediate to my problem, I'm HACKING the kernel. So if I make
> Kernel changes, will that break the virtualbox, regardless of any
> changes in the kernel.
True, that. Definitely a risk.
Consider what the VM must handle to provide an entire emulated network
router to an imaginary machine. There's really no way around having
some obligatory kernel modules on the host side, other than doing
without the VM of course.