On Wed, 20 Nov 2002 v4l@xxxxxxxxxxxxxxx wrote: > It would be so cool if these drivers > (http://www.shspvr.com/download/iTVC15_Linux_Drivers.zip) worked, but > I don't have a WinTV PVR-250 and no use for one (I don't use Windows > in any usable form here) if they don't. Does anyone have one they > want to try out? > > I would love to hear success/failure stories. Anyone? Here's the sales flyer for that set: http://www.conexant.com/servlets/DownloadServlet/102068a.pdf?FileId=1022 Interesting reading. I'm still trying to figure out why they specify a proprietary network stack. ::shrugs:: Unfortunately it didn't seem to be a full set: the datasheet is missing. Yeah, using that set with a handful of tweaks, I can get hardware encoded MPEG (vcd,svcd,dvd) from the card. But no audio. It's either a misconfig of the audio chip or the gpio bits aren't getting frobbed right. I'm (obsessively) trying to work thru this, but quite frankly, I'm about to the point where one buys board#2 and starts removing components to follow the traces. (Donations gladly accepted :) http://www.shspvr.com/forum/viewtopic.php?t=1305 Also look at the 2 topics in the Linux Driver section. Word of warning. It looks like Provideo screwed up big-time. Notice that they don't seem to sell an iTVC15 based board anymore, but that could be me not being able to find things. It's a crying shame that they decided to go down the closed route. They've spent considerable effort rolling the functionality (NOT CODE!) of tuner.o, msp34xx.o, i2c-bit-algo, and video4linux into one monstrous driver set that gloms 4 majors. If you have any delusions of writing a real set of drivers for it, DON'T look at that code. Unless somebody tells me otherwise, I consider myself tainted w.r.t that kind of effort. The best I'll be able to offer is ... The i2c registers may be found at: mmio+0x02007000 (setscl) mmio+0x02007004 (setsda) mmio+0x02007008 (getscl) mmio+0x0200700c (getsda) Use the LSBit. Looks like the sense is reversed (0->1, 1->0) Note, I don't have this quite working with i2c-algo-bit yet, so I've either got a bug, or I've misread something.
