[Impact]
System lockup on vt switch or resume from suspend
[Development Fix]
Fixed in upstream revision r838. Maverick includes up through r916 and natty through r920.
[Stable Fix]
Backport of r838 (omitting the change to Changelog) has been verified to build and uploaded to lucid-proposed.
[Test Case]
On affected hardware do a vt switch or suspend/resume cycle. Bug should repro 100%.
[Regression Potential]
Moderate. There may be some hardware for which a full memcpy worked and copied registers that aren't copied by this patch, however the severity of such issues will be less than this bug. Looking through the changelog subsequent to this fix, there is only one bug I see that this patch could expose - graphical corruption to the cursor following a suspend for VX855 and VX900; possibly these two chips aren't functioning post-resume anyway due to this bug (I can't tell without access to the given hw), so maybe that cosmetic issue is worth the risk.
[Original Report]
The openchrome driver locks up the system when switching VT on Via VX800 based systems (including the Samsung NC20 I'm using). This was upstream bug 321 (http://www.openchrome.org/trac/ticket/321) and was fixed in upstream SVN revision 838, which unfortunately Ubuntu hasn't yet packaged. Any chance of the package being updated in Lucid, or even just that one commit being included? This bug means anyone affected cannot switch back to X after switching to another VT without the system locking up completely, and this in turn breaks suspend, or rather resume, as well.
Accepted xserver-