(Crowd funded) fix for SiS 671/771 video cards

[Connor Behan]

On 06/05/13 04:44 PM, pander at users.sourceforge.net wrote:
> My experience with building drivers for Xorg is relatively small. As I
> recall, this one compiles but crashed Xorg and I don't know exactly
> what to submit exactly. It is better if the people working on that
> driver or other Xorg experts give it a try and write a proper analysis
> of the specific work that needs to be done (probably restoring support
> that has been removed but that might not be trivial).
>
That's the point, there are no people working on that driver. By looking
at the git logs, xf86-video-intel received 1904 commits in 2012,
xf86-video-sis received 16. I'm sure the people who made those 16
commits would be happy to fix a bug but that requires detailed
information about what the bug is. Here is exactly what you should
submit (for starters). The file /etc/X11/xorg.conf, the file
/var/log/Xorg.0.log, a text file containing whatever gets printed out
when you run "startx" and a text file containing whatever gets printed
out when you run "dmesg".
> Well, I don't mind getting my hands dirty but I am already involved in
> too many other open source project myself at the moment. With so many
> users of this hardware out there and people getting used to chipping
> in for applications or features via small donations such as Google
> Play for Android, I think it is a nice way to show appreciation for
> their efforts. But if someone would like to start working on it
> nevertheless-simply because to support users with this kind of
> hardware-that is also fine.
>
Yes, debugging can be a big time commitment and that's why there are
only three Xorg drivers being actively developed (two of which are paid
for by the manufacturer). Don't get me wrong... I think it's a shame
that the other 20+ drivers get such little attention and that's why I
started to learn driver hacking last year.

The problem here is not that some gigantic reverse engineering effort is
needed. The problem is that a few tiny janatorial fixes were required.
And it's easy to miss one when the developer implementing the fixes
doesn't have a SiS card on which to test. Maybe you'll find someone who
has a SiS card lying around and say "I'll pay you $100 to put that card
in your computer, write an xorg.conf file for it, figure out why it
won't start and then change 2 or 3 lines to make it start again" but it
won't be worth it. Because a similar problem will just come up the next
year. Unless you can make the popularity of SiS skyrocket to the point
where it is as widely used as ATI, gaining some basic familiarity with
the code is the only viable strategy in the long run.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 555 bytes
Desc: OpenPGP digital signature
URL: <http://lists.x.org/archives/xorg-devel/attachments/20130506/7852d065/attachment.pgp>