The xorg modular video drivers have the wrong Xv ColorKey (plain black) set and don't support the "VideoKey" config option for example. The reason is that -DXvExtension that is #ifdef'ed in the source is never set by the build system.
*** Bug 4912 has been marked as a duplicate of this bug. ***
*** Bug 4224 has been marked as a duplicate of this bug. ***
Can confirm this bug ist still valid for the savage driver and Version RC3
The way i worked around this in modular unichrome.sf.net code is to define this locally in the config.h through configure.ac. Look at http://cvs.sourceforge.net/viewcvs.py/unichrome/xf86-video-unichrome/configure.ac?rev=1.1&view=log Lines 58 and 100 are the workaround for this. I'm not sure wether this is the correct solution though. My current configure.ac for a start doesn't do a PKG_CHECK_MODULES and i'm not sure wether i dug this out when i was looking into this. Twini (who's sitting next to me right now) said he has his driver ignore XvExtension completely after a certain Xfree86 version. Apparently after 4.3 alanh removed this #ifdeffing from the sis driver and thomas followed that (while remaining backward compatible). Ajax, maybe you have a better idea about what should happen here. Should drivers just depend on Xv being enabled and present?
it is actually optional in the build; XvExtension should probably go in xorg-server.h.in. either that, or we admit that no-one who wants a light DDX actually uses Xorg and no-one actually bothers not compiling Xv, and make it mandatory.
*** Bug 5067 has been marked as a duplicate of this bug. ***
Moving Severity from Minor to Major. Not having Video on Via boxes, which are typically used for media centers, is not a minor issue :) Bug 5067 attachment 3835 [details] [review] (first chunk) is a minimum patch for the i386 case, but I agree this is not 100%. Let's fix that one way or another though!
The following configure.ac snippet should probably do the trick, unless I've missed something obvious? Also makes sure that videoproto is in. SAVED_RM="$REQUIRED_MODULES" XORG_DRIVER_CHECK_EXT(XV, videoproto) if test "x$REQUIRED_MODULES" != "x"; then AC_DEFINE(XvExtension,1,[Enable X Video support]) fi REQUIRED_MODULES="$SAVED_RM $REQUIRED_MODULES"
(In reply to comment #8) > The following configure.ac snippet should probably do the trick, unless I've > missed something obvious? Also makes sure that videoproto is in. > > SAVED_RM="$REQUIRED_MODULES" > XORG_DRIVER_CHECK_EXT(XV, videoproto) > if test "x$REQUIRED_MODULES" != "x"; then > AC_DEFINE(XvExtension,1,[Enable X Video support]) > fi > REQUIRED_MODULES="$SAVED_RM $REQUIRED_MODULES" > Of course I forgot a REQUIRED_MODULES="" between the first and second line. >
AH_BOTTOM([ #ifdef XV #define XvExtension 1 #endif]) Xv is not an absolute requirement in most cases.
Is there a real patch I could test? RC3 is getting wrapped up soon, correct?
Created attachment 3961 [details] [review] AH_BOTTOM for all 3 drivers that require XvExtension Trivial really. But preferred above an xorg-server.h, as this somewhat forces drivers to be adapted (post-release), and only a rather meek 3 drivers apparently are capable of building without Xv support.
I agree that this is a good solution. IMHO we should get this in before RC3. /Thomas
Comment on attachment 3961 [details] [review] AH_BOTTOM for all 3 drivers that require XvExtension I've checked in a different fix that will enable both Xv and XvMC support for the server and drivers.
Fixed in CVS.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.