diff options
author | Luc Verhaegen <libv@skynet.be> | 2009-10-28 16:07:12 +0100 |
---|---|---|
committer | Luc Verhaegen <libv@skynet.be> | 2009-11-04 15:13:03 +0100 |
commit | 89807ffc284b677bfe987ea2bab695bdf34c8b33 (patch) | |
tree | b4ba507bbc0b3ece8da2d146c1be2a4eea09cbcc | |
parent | 89b85ba5d1da8a309fd7ab9fdf2df2a55f7ec8d5 (diff) |
README: Provide intro, add package building and XvMC section.
-rw-r--r-- | README | 126 |
1 files changed, 112 insertions, 14 deletions
@@ -1,18 +1,50 @@ -Building The Unichrome Projects drivers for your X version: ------------------------------------------------------------ +xf86-video-unichrome: +--------------------- -WARNING: The content of this file is about building and installing this -driver yourself. This will not make your package manager happy. You might -also want to back up via_drv.o/so and the manpage first. +xf86-video-unichrome is a free software graphics driver for VIA Technologies +Integrated graphics of the CastleRock, Unichrome and Chrome9 denominations. + +It currently supports the following chipsets: + + VT3122: CLE266 (CastleRock). + VT7205: KM400, KM400A, KN400, P4M800 (Unichrome). + VT3108: K8M800, K8N800, K8N800A (Unichrome Pro, aka Pro B). + VT3118: CN400, PM800, PM880, PN800 (Unichrome Pro A). + VT3344: VN800, P4M800Pro, CN700, P4M800CE (Unichrome Pro). + VT3157: CX700, CX700M, VX700, CX700-M2 (Unichrome Pro). + VT3230: K8M890 (Chrome9 IGP). + VT3343: P4M890, VN890 (Unichrome Pro). + VT3371: P4M900, VN896, CN896 (Chrome 9 HC IGP). + VT1122: VX800, VX820 (Chrome 9 HC3 IGP). + +The level of support for these chipsets varies. + +The canonical source for this driver is: + http://cgit.freedesktop.org/~libv/xf86-video-unichrome/ + +You can get a copy of this repository by running: + git clone git://anongit.freedesktop.org/~libv/xf86-video-unichrome + +Its main project website is: + http://unichrome.sf.net If you run into any difficulties, please contact the unichrome-users -mailinglist at unichrome-users@lists.sourceforge.net . Subscribe at -http://lists.sourceforge.net/mailman/listinfo/unichrome-users first. +mailinglist at unichrome-devel@lists.sourceforge.net . Subscribe at +http://lists.sourceforge.net/mailman/listinfo/unichrome-devel first. +1) Building this driver manually: +--------------------------------- -This driver is suited for Xorg 7.0 through 7.5. +WARNING: This section is about building and installing this driver yourself. +This will not make your package manager happy, and you might want to check +the next section instead. You might also want to back up via_drv.o/so and the +manpage first. -2.1) Requirements: +This driver is suited for Xorg 7.0 through 7.5. While it used to support +older X versions, such support has now been removed, but can be re-added upon +popular request. Such wide range of support is unique in today's X world. + +1.1) Requirements: - An X version more recent than 7.0 RC2 (there was a macro difference before that). @@ -24,7 +56,7 @@ This driver is suited for Xorg 7.0 through 7.5. Contact me if you would like to add to this list. -2.2) Building: +1.2) Building: Building against modular is a breeze. All it takes is: @@ -37,9 +69,9 @@ This driver is suited for Xorg 7.0 through 7.5. and # make install -2.3) Common problems: +1.3) Common problems: -2.3.1) aclocal is unable to find X macros. +1.3.1) aclocal is unable to find X macros. You will see configure complain with: ./configure: line xxxxx: XORG_MANPAGE_SECTIONS: command not found @@ -55,7 +87,7 @@ This driver is suited for Xorg 7.0 through 7.5. - temporary solution: run # export ACLOCAL_LOCALDIR="<prefix>/share/aclocal/" -2.3.2) XORG_DRIVER_CHECK_EXT missing from macros. +1.3.2) XORG_DRIVER_CHECK_EXT missing from macros. Configure fails with following message: ./configure: line xxxxx: syntax error near unexpected token `RANDR,' @@ -71,7 +103,7 @@ This driver is suited for Xorg 7.0 through 7.5. for the existence of individual macros at the autoheader stage, please don't hesitate to contact me. -2.3.3) Automake too old. +1.3.3) Automake too old. autogen.sh fails with: configure.ac: xxx: required file `./[config.h].in' not found @@ -79,4 +111,70 @@ This driver is suited for Xorg 7.0 through 7.5. Make sure that you have at least automake version 1.7 +2) Building this driver as a package: +------------------------------------- + +This driver comes with some support for being built as a package directly. +This is, once again, a rather unique feature of this driver. + +2.1) Debian based systems: + + This driver comes with built in debian package scripts in the debian/ + directory. It is suited for at least debian and ubuntu, and even handles + the early ubuntu days where they used completely different package names + from debian. + + All you have to do is run dpkg-buildpackage in the drivers top level + directory. dpkg-buildpackage is a part of the dpkg-dev package. + +2.2) RPM based systems: + + No spec file is provided at this time, i should poke sndirsch (SUSE Xorg + package/bugmaster) to see whether i am allowed to just steal his. It works + for most rpm based distributions directly. + + +3) Using XvMC with this driver. +------------------------------- + +This driver provides Slice decoder support for xine and MythTV, even +though there are some things that you should be aware of. + +2.1) Xine issues: + + The Xine "XXMC" video output driver has a rather appalling amount of + issues. + +2.1.1) video_out_xxmc requires subpicture support. + + This for no reason whatsoever, the xxmc video out module depends on + subpictures. When running this driver, you can notice the following when + xine dies: + + abort: video_out_xxmc.c:622: create_ximage: Aborting. + + To alleviate this, add the following to your device section in your + xorg.conf: + + Option "XvMCBrokenXine" "True" + + This fools xine in that the server advertises subpictures, while not + actually implementing them. You will notice these messages: + + XvMCCreateSubpicture is not implemented. + XvMCBlendSubpicture2 is not implemented. + + These are rather harmless. The first will have the xxmc code fall back to + software drawing just fine. And the second is actually just a horrible + example of abusing an API for other purposes. + +2.1.2) Xine splash is not always showing correctly. + + This is an issue with the xxmc code as well: it tends to feed the + dimensions of the mpeg2 stream as the dimensions of the xine splash, which + the standard Xv interface then feeds to the driver. You can notice a + strangely squated splash, with some green edges with sometimes flickering + areas. This is not a problem, as it is just showing random (empty) FB + content. + Luc Verhaegen, <libv@skynet.be> |