Re: Purge of firmware to recommence
Herbert Xu wrote:
>In accordance with the latest directive from the RM, I shall recommence
>the removal of non-free firmware where no acceptable solution using
>user-space firmware loaders has been provided.
>This will take place with the next upstream release of the respective
>packages, i.e., 2.2.26/2.4.27/2.6.6.
>If it gets real close to the release and no upstream release is in
>sight, then I will remove the firmware from the version that we
Please be advised that I have a *working* userland firmware loading solution
for the Radeon and R128 DRM drivers (2.6 kernels only of course), which has
been submitted to the DRM maintainers. I have no idea how long it will take
to get it into the DRM tree, from there into the kernel tree, and from there
into an upstream release; there will likely be small changes for preferred
file/function naming & whatnot. If at any time you would like me to submit a
version of this patch for the Debian 2.6 kernels, please let me know.
I intend to create more such fixes; fixes for most other hardware should be
easier to get into upstream due to there being fewer layers to go through.
If anyone has a favorite, tell me and I'll try to work on it early. (tg3 is
currently on the top of my list because it was described as being 'very
Remember that I can't actually test most of these myself since I don't have
the hardware (I found a Radeon lying around, but that was luck), but I would
love to send patches to people who are
(a) willing to build their own kernels, and have done so successfully
(b) have the hardware to test.
In addition, a package in 'non-free' will be required to hold the three
firmware files for ATI's cards. I am planning to create such a package (as
an exercise in packaging skills), but if anyone wants to sponsor the upload
and/or take over maintenance of them and/or offer advice, please contact me.
Similarly, 'non-free' packages will be needed for other hardware, but I'll
cross that bridge when I come to it. Ideally, if the firmware is supplied
standalone by upstream in some particular format, that would be used
(simplifying firmware updates immensely); in the ATI case, I didn't find any
such object, so made up my own format based on the "format" in the hex blobs.