Voodoo3 3500 TV / News: Recent posts

This was a tough job, but it would probably be necessary sooner or later anyway.

The .c-files of the main driver module are now compiled separately, and then linked together to make "v3tv.o" (I think this is a better name, than just voodoo). This means that you need to update your /etc/modules.conf to load "v3tv" in stead of "voodoo" if you use autoloading. It might also be a good idea to remove the old "voodoo.o" module from your /lib/modules/x.x.x/misc directory.... read more

I've move the init code around some. This may solve the pink.green stuff. Also change the v4l open/close functions. My test app allows glide after closing. xawtv and kwintv being programmed by college smucks prolly, don't call close on the v4l device. So the registers never get set back. Anyone want to write a patch?

New install documentation is on the way for the 2.2.x series. I'm also working on the bug that 3D (glide) will not fire up correctly after running xawtv. I have a partial fix. Restart X :-/ That solves the problem, but isn't the fix. I'll let you know what I find.

I've checked in and update to voodoo-v4l.c and README. The audio ioctls are now in. The README has more info on installation, FAQs, and a list of known bugs/workarounds. This is in the 2.2.x series code. Next up is video capture and the video out stuff.

I've checked in the code that works under the 2.2.x kernels. Its been tested on 2.2.16 on a RedHat 6.2 box. Please see CVS for the latest. Its in the "v3tv for 2.2.x kernels" module.

I've taking the time out and back ported the current code to the 2.2.16 kernel. Found a couple pointer bugs that affected the msp3400 and input switching code. Everything seems to be working. I'll post a tarball up later tonight as soon as I have finished the build readme.

Currently sitting in an internet cafe in London... Flight to us of a departs in abt.4 hours so I better get going soon. I read some of my email while here and will reply to most of it by the end of weekend(only 179 unread msgs - mostly mailing lists, though). Maybe someone can give me a summary what has been going on in past two weeks or so?

Watching the Soldier while working on the audio changing code on xawtv. Color is great in 16 and 24 bpp. No cpu usage also. This is how it should be. Back to the code...

I'm off to Europe in about 3 hours. I'll be back on 6th of May.

So during the next week and a half there will be no development from my part. You're free to examine the code and submit ideas and patches but I will not be able to respond until after I come back..

Request:
If someone could come up with a YCbCr->RGB conversion routine using a lookup table and a some quantization routines during my absence? I have a conversion routine included in the code that does the calculations real-time(not a good thing) and currently i have no quantization.. I know that this should not be needed since programs dealing with v4l-devices should use the native color-format of the driver but since it is nicer to get rgb data out of the device we could still implement it..

Everything went pretty smoothly today and I can now say that the code in cvs repository is pretty much functional.

This means that you can watch tv with your voodoo3500-card while using linux and in stereo audio too.. (xawtv worked for me)

Our current Video4Linux-interface covers only the basic subset necessary to show the image on screen(sorry.. no capture yet) but that's a pretty good start..

The code is still very much pre-alpha quality but since it has very little functionality it probably will run stable(hasn't failed me so far). Beware that you will need a late development kernel-source installed and that you need to update i2c-drivers separately. ... read more

I've uploaded an initial version of the source code to the repository.

What Works:
- tuner
- audio
- composite input

What Doesn't:
- v4l-interface
- radio
- capture

Also all configuration is currently only possibly by changing the sources..

I've gotten some request to put the sources up. The code at the moment does very little. I'm still working out the I2C and VMI bus communication registers. If someone wants to help out on this, I can put what I have up. If not, I'll upload when the code actually does something useful.

I found the Voodoo3 register specs. I can now continue programming the drivers. I have the support chipset code in place, but could not get to them through the main chip. Now I can. :-) Sooooonnnnnnn........

The doc for the chipset is on the home page.

I got this far. I'm updating the web-pages and cvs now. Stop by later and check them out. I'll have linkes to the chipset documentation from the web pages.