Hi Axel! On Thu, 03 Dec 2009 10:41:27 +0100, "Axel Dörfler" <axeld@xxxxxxxxxxxxxxxx> wrote: >> From other side I can simply drop unfinished support of ali5451 and >> Trident DX/NX chipsets from my code and save my part of work that was >> already completed for sis7018 chipset. In such case your driver will >> support ali5451/Trident DX/NX but my driver will be optimized only >> for >> sis7018. What do you think? >> >> 2 All: Are there any objections? > I'm not familiar enough with the chipsets to be able to judge the > situation, but if they are pretty equal, I don't really see the point > in maintaining two different drivers. I think almost all of sound drivers are pretty equal in theirs basic "divided DMA buffer" concept. ;-) But the Evil, as usual, in registers. AFAIR, they are a bit different between ali and sis chips. At least almost every HW operation procedure of the old sis7018 driver has the "switch" statement to select different registers to operate with different chips. But the difference between Trident DX/NX and sis7018 is more dramatic - especially in recording support. And, AFAIK, Linux has separate driver for ali5451. And as I have already noted - my one already works for sis7018 chips and was postponed just because of lack ali5451, Trident TX/DX support. :-\ > At least not in the long term - for now, I wouldn't mind if you work > separately for now, if the long term goal is to merge them (if > feasible, as I said, I can't really tell). My sis7018 driver uses "Michael's style" C++ implementation. Ali5451 looks like a "classical Jerome's style" implemented in C. I suspect it can be difficult to merge them. ;-) And, finally, I'm a "big SiS-fan" and would like to implement the whole family of audio, network and, probably, video drivers. Pleeeease... :-D -- Kind Regards, S.Zharski