[vbox-dev] linux kernel taint?

Bernd Petrovitsch bernd at tuxoid.at
Wed Oct 12 10:14:22 GMT 2011

On Tue, 11 Oct 2011 13:59:04 -0500, Perry Halbert wrote:
> Stone throwing seems to be the way of things these days.  Some figure if 
> they throw enough they will actually hit an issue.  I would be more 
> interested in see the actual tickets myself and seeing what the problem 
> really is, but here again this was not given so how could anyone expect 
> to take it seriously.
> I myself suspect this has more to do with politics than actual issues, 
> but that would be speculation without proof which is again lacking.  
> Anyway how could you take someone seriously that plainly puts something 
> like this is crap, or garbage in a code field.  Sounds more like sour 
> grapes to me.

The naming - "crap" - is very bad. Better would have been "out of tree"
or the like which is what it actually means.

> Besides kernel_taint just seems to be a way to not have to deal with 
> issues.

Yes, with the issues which do not belong there but somewhere else. Or
do you expect people to work on bugs for free in some closed-source
driver from some company making money?

>          Look at Debian kern.log after you install the non-free nVidia 
> drivers.  Debian does not want to support it so they taint it.  This 
> does not necessarily mean that the drivers are bad, just that the 
> maintainer does not want to deal with it.

The "taint" stuff comes from the Linux kernel as such. LKML got (and gets)
regularly bug reports with close-source drivers are loaded (and the
nvidia driver is the most prominent there) and many developers refuse to
look for bugs in the kernel if some closed-source part could have done
evil things.
These bug reporters gets redirected to either their distribution "vendor"
since usually they decide which packages to include or the vendor of the
closed-source thingy. Or - especially if the bug is quite easily
reproducible - they are asked to reproduce it without any tainting
module loaded and send that bug report (which will be handled).

Fedora/RedHat and SuSE/Novell (and perhaps others) have paid people to
deal with the bug reports (and some of them read the LKML too) but
Debian can't do much (unless they find someone who does it for
free as in beer).


"I dislike type abstraction if it has no real reason. And saving
on typing is not a good reason - if your typing speed is the main
issue when you're coding, you're doing something seriously wrong."
    - Linus Torvalds

More information about the vbox-dev mailing list