When will Fedora work again?

[Adam Williamson]

On Sun, 2009-03-08 at 17:10 +0000, Leon Stringer wrote:
> Al Dunsmuir wrote:
> > 
> > It seems strange (and sad) that so soon into the life of a release major
> > regressions can happen... followed by a deafening science.  There  seems  to  be
> > very  little urgency (or developer time/focus) on fixing this in F10.  Rawhide
> > may be important... but only if existing releases can actually be used in real life.
> > 
> 
> 
> Exactly. I can't test or contribute if it doesn't work and there 
> *appears* to be little interest to resolve this. (Maybe there's just a 
> bigger focus on other chipsets and a limited number of developers).
> 
> I now see that there are quite a few people with Intel controllers who 
> are in the same position and feeling pretty frustrated...

First off, as others have said, Deependra is just one guy posting his
opinion, so don't take what he says as gospel :). There's no general
policy that Fedora will only support really recent hardware. There's no
coherent policy for Fedora saying "we will support exactly X, Y, and Z
hardware", or "we will support 3 year old but not 4 year old hardware",
or anything like that. It's more or less up to developers and
maintainers for each particular driver.

As far as Intel goes, it may help to explain the situation with regards
to that driver. The Fedora maintainers for the 'intel' driver are also
the main upstream developers for the driver. Developing the driver
itself is their main focus. The driver covers a very wide range of
hardware, from i810 all the way up to the current GMA series adapters,
which is twenty or thirty different chips from a timespan of several
years, so they have a very wide range of hardware scenarios to try and
cope with. The 'intel' driver is also kinda one of the 'showpiece'
drivers for X.org, so it tends to get implementations of new
technologies like RandR 1.2 very early, and to an extent it is used as a
development base for those technologies.

So the intel driver devs, who also have to maintain the Fedora package,
have a lot on their plate - trying to support several years worth of
different hardware, and also working on cutting-edge X.org developments
(like, for instance, at present they are working on the new GEM
architecture). Because they're working on so much and because so much
experimental new stuff goes into the intel driver, that's why you will
see regressions in performance and complete bugs in support for specific
chips like this sometimes coming up, and it probably also explains why
it takes them a while to address issues sometimes.

In this specific case, as others have said, there are some workarounds
you can try. The emergency one is to use the 'vesa' driver, which will
make X work, but will be rather slow and may not support all the
resolutions you need. You can also create a /etc/X11/xorg.conf file -
running 'Xorg -configure' from runlevel 3 should do this - and then try
either disabling acceleration or switching to XAA rather than EXA
acceleration:

Option "NoAccel" "true"

or

Option "AccelMethod" "XAA"

in the Driver section of the xorg.conf file will do those things. Either
of those may get you going. If so, of course, please add this
information to your bug report(s). Thanks!
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net