RE: [Unichrome-devel] Need help with v4l
Brought to you by:
dwdeath
From: Warren M. <wa...@ca...> - 2004-10-28 12:04:20
|
Luc, The reason for doing so is this: This group is claiming that they are the better solution for VIA CLE266 drivers. VIA themselves are pushing their chipset to be used in this type of solution, where I2C input is used. The boards that do this are appearing now and will be mass produced within the month. One of the main applications that will run on a board like this is MythTV, which as of the next release will exclusively run using your drivers. So, given that you want to be the de-facto standard for VIA/CLE266 drivers and have convinced other software providers to exclusively use your software, there becomes a bit of a burden upon you to support the boards being made that use this chipset. Also, if you go back in the VIA arena forums (which until very recently I thought was the only place to ask questions about this stuff), I was asking questions about I2C (V4L) input a year ago. I apologize for not asking here earlier, but as I stated, I did not know that this was the place to ask. Put simply: HELP!!!!!!! Regards, Warren -----Original Message----- From: uni...@li... [mailto:uni...@li...] On Behalf Of Luc Verhaegen Sent: Thursday, October 28, 2004 7:43 AM To: uni...@li... Subject: Re: [Unichrome-devel] Need help with v4l On Wed, Oct 27, 2004 at 09:29:56AM -0400, Warren Melnick wrote: > HI all, >=20 > =20 >=20 > I am new to this list. I asked this message on the VIA Arena forums and > was asked to submit it here instead. >=20 > =20 >=20 > I am CTO of a software company that is writing a suite of programs for a > board based on the EPIA-M10000 with some modifications. The board in > question is pretty much the same as the VIA board, but has swapped the > network chip and added a few things such as a pcmcia slot, smart card > reader, component video out, lots more USB and firewire connectors and > most importantly for this, video inputs and a hardware mpeg-4 > processor. These inputs are connected through an SAA7114 chip and were > previously accessed through /dev/video0 and /dev/video1 using the > VIA_V4L_DRV. >=20 > =20 >=20 > We had been using kernel 2.4.26-epia2 and XWindows 4.3.99.903 with all > the appropriate patches to make everything work to get the video to a > full screen overlay. >=20 > =20 >=20 > Now the hardware vendor wants mythtv to be used. I have gotten it to > work beautifully using 2.4.26 and 2.4.27 with the drm patch and using > either X-4.4 or X.org. I am quite happy with the results and would like > to move our other software to this platform. The problem is that I no > longer know how to access the incoming video from the 2 devices, > formerly /dev/video0 and /dev/video1. >=20 > =20 >=20 > Could someone here possibly point me in the proper direction? >=20 > =20 >=20 > Regards and Thanks, >=20 > Warren >=20 Please check out=20 http://sourceforge.net/mailarchive/forum.php?thread_id=3D5828259&forum_id= =3D 38837 The SAA7114 here is wired straight to an input bus on the CLE266,=20 controlled through i2c (or io pins emulating it). Since it is connected=20 straight to the unichrome, and X should have full and complete control=20 over the unichrome (in the current driver model), X should also=20 exclusively handle the SAA7114 and the necessary HQV work. This is how=20 Alan Cox saw the situation when he added via_tuner.c, this is how i=20 see it too. Now, since most of this code is dead, since no-one before reported that=20 sort of hardware, since no-one is around to write, test, maintain this=20 code, the only real option would be to remove it. As outlined in that thread: all support for i2c connected devices will=20 eventually be moved out of the drivers themselves, and they will move=20 to xfree86/drivers/i2c. Currently the gatos framegrabbers/tv tuners are=20 sitting in there, but it will, if i ever get at that sate, be expanded=20 to hold TV/LVDS/TMDS encoders too. I find the infrastructure done for=20 gatos too simplistic and too dependant on the actual driver handling=20 most of it, but that will be changed, in cooperation with vladimir and=20 whoever else is seperating out i2c devices from drivers. I do not see how, with the i2c and hqv code firmly in the X driver, one=20 would even think about involving V4L into the lot. Luc Verhaegen. ------------------------------------------------------- This SF.Net email is sponsored by: Sybase ASE Linux Express Edition - download now for FREE LinuxWorld Reader's Choice Award Winner for best database on Linux. http://ads.osdn.com/?ad_id=3D5588&alloc_id=3D12065&op=3Dclick _______________________________________________ Unichrome-devel mailing list Uni...@li... https://lists.sourceforge.net/lists/listinfo/unichrome-devel |