Date: Wed, 17 Aug 2005 02:01:27 +0200 From: Hans Petter Selasky <hselasky@c2i.net> To: freebsd-isdn@freebsd.org Subject: Re: New ISDN driver Message-ID: <200508170201.28540.hselasky@c2i.net> In-Reply-To: <200508161814.j7GIESTv010770@peedub.jennejohn.org> References: <200508161814.j7GIESTv010770@peedub.jennejohn.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tuesday 16 August 2005 20:14, Gary Jennejohn wrote: > Michael Reifenberger writes: > > On Tue, 16 Aug 2005, Hans Petter Selasky wrote: > > ... > > > > >> Well, doesn't "i4l_" indicate that this came from "ISDN4LINUX" ? > > > > > > If you "cat" all the files in "/sys/i4b" and search for the word > > > "linux" or "Linux" you will get some hits. > > > > > > So I suspect that some parts of the drivers in "/sys/i4b/layer1" are > > > derivatives of Linux. And I have got some hints about that too. So > > > maybe we should just drop support for some hardware, and stick with the > > > chip-drivers > > > > I > > > > > have written from scratch ? Else we might end up getting a bad rumor > > > about stealing code from Linux. > > > > The occurance of [Ll]inux or l_* doesnt imply that these files are from > > linux . > > /sys/compat/linux/* files are stuffed with these letters. > > Usually they are for compatibility reasons. > > So its not reasonable to suspect any I4B files to be linux tainted. > > Several drivers were developed after examining the Linux code, which > in these cases was the only form of documentation available. AVM in > particular is very bad about providing documentation to any developer > outside the Linux world! Been there, done that. They might be afraid that someone will find problems about their products. For example their B-channel FIFO interface ill fits my system, hence you have to pre-write data lengths, which is a bad idea. Because then one has to pre-walk mbuf chains. But it might be interesting to dissassemble that AVM firmware, and see if it is really solid. I think that we won't have much use for those active AVM controllers in the future. Modern passive E1 chips have many useful features. For example DTMF detection, support for telephony conferences, interconnection of channels, and very large buffers. The only thing these chips haven't got, is a modem emulator, which I think is not very difficult to have in software. Detecting sine waves is not very difficult. One simply puts the waves one wants to detect, into a linear equation set, solves it, and finds the unknown amplitudes. This is something like Fast Fourier Transform, though we cannot use FFT, hence these modem frequencies doesn't fit exactly into "n" samples. So it is not very difficult. There also exist other algorithms to demodulate sine waves, but I think one will get best results with linear equation sets? --HPS
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200508170201.28540.hselasky>