Ubuntu
xserver-xorg-video-geode package

  1. Bug #214119
  2. Comment #56

Comment 56 for bug 214119

Revision history for this message
Martin-Éric Racine (q-funk) wrote :

I'm nominating this as an SRU for Hardy:

1. The bug is a failure to perform DDC polling on hardware with broken BIOS environment. It affects a number of hardware products, including some OEM hardware sold by Ubuntu-certified Canonical partners.

2. Upstream addressed this issue by replacing code fragments that are dependent upon BIOS environments with new code that leverages X.org's own DDC polling library. This allows the driver to directly poll the DDC pins itself via libDDC and to parse the EDID information in a reliable and consistent way that is independent of BIOS features.

3. Find the attached xf86-video-geode_280_to_290.diff that shows the diff between 2.8.0 currently in Hardy and this new 2.9.0 released by upstream. The upstream commit diff between 2.8.0 and 2.9.0 is attached. The only other changes seen in the upstream tarball are the ChangeLog and autoreconf results, both of which are auto-generated at release time by tarball creation scripts. Upstream produced this release using Hardy versions of the libtool/autoconf/automake on a host running Hardy.

4. Reproducing the issue requires an OEM version of the FIC ION603, as sold by vendors such as Koolu, Linutop and Sumo Systems.

TEST CASE:

a) Launch X.org with an xorg.conf configured to utilize the "geode" driver.
b) Witness the driver falling back to its 800x600 lowest default resolution.

NOTE: the same issue is randomly reported on #ltsp as affecting other Geode -based products with a BIOS from other vendors as well.

5. We cannot think of any situation where this change would introduce any regression. If there's one thing, it should suddenly make this driver usable on a much wider variety of Geode -based hardware than before.

*****
This package needs to be sync'ed from Debian/unstable to hardy-proposed.