Radeon driver in linux libre

Alexandre Oliva lxoliva at fsfla.org
Wed Jan 28 18:50:27 UTC 2009

On Jan 28, 2009, Rubén Rodríguez Pérez <ruben at es.gnu.org> wrote:

> I'm reviewing the linux-libre list of changes, to see if any of them can
> be done by removing the non-free parts instead of the whole files.

It's been a while since we stopped removing whole files.  Unless
deblob-check is missing, that is.  Running deblob-<kver> without
deblob-check will remove files after a warning, and this seems to catch
a lot of people off guard.  I'm thinking maybe we should only run
without deblob-check if a command-line option is provided or some such.

> As many drivers cover several devices, some of them may work. The one
> I'm currently reviewing is the ATI radeon (and r128) driver.

Odd that you should mention it.  Earlier this week, I modified the way
we handle these in the SVN version of deblob-2.6.28.  With the changes I
made, it will fail to initialize and probably not work at all.  If you
have the time to experiment with *not* triggering such a failure, I'd be
glad to lend you a hand.

That said, I very much doubt radeon will actually work with any 3D
acceleration without the removed firmware or Free reimplementations
thereof.  It may be worth a try, though.  Just make sure (somehow) that
you don't have remnants of non-Free firmware in the not-quite-volatile
memory of the video card.  Like, power off the machine if you ever had a
kernel that may have loaded the non-Free firmware into the device, to
make sure it's not working just because it's still there.

Even without the kernel radeon driver, the X radeon driver actually
works for me, even with 2D acceleration, at least with the oldish radeon
card on my home firewall/gateway/server.

I hope this helps.  Let me know how it goes, and if you need any help,
you can use this list, or get ahold of me (lxo) on #linux-libre at

> Someone at this forum said that this parser can be avoided via a (less
> powerfull) MMIO method, thus providing 3D support for ATI cards in the
> linux-libre kernel, which would be great. I'll ask this person about it.

Hmm, interesting, I hadn't realized that.  This would indeed be awesome.

That said, it's probably the kind of change that would probably make
more sense to try to integrate upstream first.  The idea of
accummulating and maintaining patches for actual functionality (rather
than mere removal of non-Free stuff) is a bit scary to me, considering
that I'm not much a kernel developer.  (or at all ;-)

Thanks in advance for looking into this,

Alexandre Oliva           http://www.lsd.ic.unicamp.br/~oliva/
You must be the change you wish to see in the world. -- Gandhi
Be Free! -- http://FSFLA.org/   FSF Latin America board member
Free Software Evangelist      Red Hat Brazil Compiler Engineer

More information about the linux-libre mailing list