Hello. On Sat, 2010-08-28 at 11:29 +0200, ext Hans Verkuil wrote: > On Thursday, August 26, 2010 09:39:45 Matti J. Aaltonen wrote: > > Hi. > > > > On Wed, 2010-08-25 at 23:20 +0200, ext Pavan Savoy wrote: > > > > > > > I'm sorry for not answering to you earlier. But I don't > > > > have my own > > > > public repository. But to create the whole thing is > > > > extremely simple: > > > > just take the current mainline tree and apply my patches on > > > > top of it... > > > > > > Yep, that I can do, the reason I asked for was, we've pushed a few > > > patches of our own for WL1283 over shared transport/UART (Not HCI-VS, but > > > I2C like commands, packed in a CH8 protocol format). > > > The FM register set in both chip are a match, with only transport being > > > the difference (i2c vs. UART). > > > Also we have the Tx version of driver ready too, it just needs a bit of > > > cleanup and more conformance to already existing V4L2 TX Class.. > > > > > > So I was wondering, although there is no problem with WL1273 with I2C and > > > WL1283 with UART being there on the kernel (whenever that happens), but > > > it would be way more cooler if the transport was say abstracted out .. > > > > > > what do you say? just an idea... > > > > I think it's a good idea. And the WL1273 ship can also used with a UART > > connection, we just chose I2C when the driver development started etc... > > Making a completely bus-independent driver is actually possible. It would > require > that the driver uses the subdev API (include/media/v4l2-subdev.h). Any > register > read or writes can be done by calling the v4l2_device notify() callback and > the > bridge/host driver can then translate the callback to either i2c or uart read > or writes. > > Both v4l2_device and v4l2_subdev structs are completely abstract structs (i.e. > they do not rely on any particular bus), so it should be possible to implement > this. > > I had this scenario in the back of my mind when I designed these APIs, but > this > would be the first driver where this would actually apply to. >
That sounds interesting. I think that after the driver gets accepted in its current form we can start to work according to the above scenario... Cheers, Matti -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
