News Archive
02:21, 18 October 2011 (CEST)
No News is Good News?
Its been a while since the last news post, but nothing in particular really jumps out as having occurred around here ... unless, of course, you count the general day by day activities such as: more new drivers, work on the utilities, some ongoing refinement of the build config, the creation of a tvtime development repo here on LinuxTV, changes to the patchwork system (see the main page), and etc., etc., etc.
Although this stands as short notice to the fact, please take note of the Media Subsystem Workshop being held Oct 23-25, happening in along side the 2011 Linux Kernel Summit (see the main page for more details).
03:43, 13 February 2011 (UTC)
Free at Last, Free at Last: A Final Fairwell to V4L1 ... And Just What the Heck is xf86-video-v4l Anyway ?
Upon release of the forthcoming 2.6.38 kernel, all but a very few vestiges of the antiquated V4L1 API will have been removed forever in favour of its successor, the V4L2 API (which, itself, now resides within the larger Linux Media Infrastructure API).
One of the more obvious ramifications of this significant change is that older V4L1 based userspace applications will have to be brought up to speed, so to speak, to V4L2 spec. Devin and the guys at Kernellabs, for instance, have already tackled the problem with the very popular, but essentially previously abandoned, Tvtime. Likewise, reworking of the also popular 3.9x branch of xawtv has taken place.
Further still, the removal of V4L1 has even necessitated a change in the X server code base; specifically, to the xf86-video-v4l module. video-v4l is a relatively older Xorg DDX driver that had, up until just very recently, suffered in the way of development neglect and, as a consequence of “bit rot”, hasn't provided much in the way of utility for the end user for a number of years. In addition, existing documentation (e.g. have a look at its manpage by typing “man v4l” in a console) is probably a bit vague or, at first glance, not fully instructive on the driver's purpose and usage. That state of affairs, in turn, has essentially lead to video-v4l becoming a bit of a curiosity – at least, for those end users who are even aware of its existence (questions about it have surfaced from time to time on the past and present LMML and related #irc channels).
video-v4l now, however, has managed to find its way into the spotlight for 15mins of fame after Mauro submitted a major rewrite of the driver to Xorg this past week and this happened to be picked up as a news brief in Phoronix. Though, despite the article's explanation, as well as that provided by Mauro in his submittal message, it was evident from several Phoronix reader responses that many were still a little unclear about the drivers purpose. Fortunately, Alex Deucher, no stranger to V4L or the Phoronix forums, was able to provide those readers a nice succinct summary.
While it might be left at that, it might be more instructive to parse together a number of online sources to provide an even more thorough explanation (much will be verbatim, just rearranged for clairity; sources will be listed at end). So, without further adu, lets see if we can pull all the pieces together and answer the age old head scratcher: "just what the heck is xf86-video-v4l anyway?" (Note: its a “bit” technical and assumes some basic knowledge of X)
09:22, 9 January 2011 (UTC)
Introducing the "Media_build" System:
If you hadn't noticed, over the course of the past year, there has been some monumental changes in the revision control systems employed for V4L-DVB development. These changes have had an impact not only upon developers, but also on end users seeking to use the latest drivers set.
If you recall, back in January 2010, GIT was, for better or worse, adopted as the mainstay for future V4L-DVB development, while the former Mercurial (Hg) build system was relegated to the Backport tree. [1] With the backport repo in place, end users would still be able to easily adopt current V4L-DVB driver snapshots with their currently installed kernel; something that cannot be equally said in the case with GIT. However, as time progressed, the Hg backport tree unfortunately saw less and less "backporting" activity, and the Hg snapshots soon became stale and error prone during module compilation. In October, the tree became maintainer-less [2]. Since then, Hg has all but become formally abandoned around these parts.
Alas, all was not lost for the end user seeking to easily adopt a new driver snapshot! An experimental new build system, the "media_build" GIT tree (which was originally entitled "new_build" before being renamed), was developed in hopes of overcoming the formal restrictions imposed by GIT and to maintain the benefits the former Hg system provided the V4L-DVB user community.
For much fuller background information regarding the "media_build" tree, see the following:
- http://www.mail-archive.com/linux-media@vger.kernel.org/msg23394.html
- http://www.mail-archive.com/linux-media@vger.kernel.org/msg23764.html
- http://www.mail-archive.com/linux-media@vger.kernel.org/msg23784.html
- http://www.mail-archive.com/linux-media@vger.kernel.org/msg23881.html
- http://www.mail-archive.com/linux-media@vger.kernel.org/msg24353.html
- http://www.mail-archive.com/linux-media@vger.kernel.org/msg26724.html
- http://www.mail-archive.com/linux-media@vger.kernel.org/msg26733.html
- http://www.mail-archive.com/linux-media@vger.kernel.org/msg26750.html
- http://www.mail-archive.com/linux-media@vger.kernel.org/msg26773.html
- http://git.linuxtv.org/media_build.git?a=tree;h=a452eb4a9d742e13987697da8778c0cf22855fa0;hb=a452eb4a9d742e13987697da8778c0cf22855fa0
23:10, 27 June 2010 (UTC)
Merger of Wikis Comes to a Close:
After running some $47Million dollars over budget and 18 months late, the merger of the V4L and DVB wikis is finally complete! When asked about the glaring cost and time overruns, a well tanned project coordinator candidly replied that "there were, ummm, some issues", before hurrying away in a private limousine, escorted in arm by two, equally as tanned, attractive women, whom this reporter can only assume were the project's creative advisors.