I've modified i2c-viapro and via686a so that the via686a driver will register the i2c bus driver as well, so that i2c-viapro is not required if via686a is loaded. The problem is making it pretty and making it work for all combinations of built-in or module. This works around the PCI conflict of via686a and viapro. This doesn't apply to vt1211 because it isn't a PCI module. So I agree there isn't any conflict. Agreed there's no reason to combine vt1211 and viapro. BARRE Francois wrote: > Well, it was off-topic, let's make it another topic ;-p > > In fact no, i2c-viapro and vt1211 are not in conflict. As far as I > understand it, vt1211 (I mean the lm_sensors module which deals with the i2c > bus dedicated to temps, fans, etc. on the via nehemiah) is connected to the > (fake ?) isa bus, whereas the i2c-viapro and via-ircc are both connected to > a pci bus. > > I don't think it would be aethetic (and even possible) to integrate both > fonctionnalities in the same driver, and I don't know yet how to make two > drivers share the same device, but I guess that's the way to do it. Or maybe > to push "shared" code in a kind of (fake) bus driver and specific code in a > chip. If Mark has some idea/any kind of experience dealing with it, I shall > be interrested. I didn't manage to find any example implemented in the > kernel yet. > > Anyway I didn't dig into the i2c-viapro code yet, but I shall do it in order > to plug an PCF8591 in it (the smbus shall be handled by i2c-viapro, am I > right ?), together with an irda transceiver (with the help of via-ircc)... > > >>-----Message d'origine----- >>De : Jean Delvare [mailto:khali at linux-fr.org] >>Envoy? : mercredi 17 mars 2004 22:21 >>? : BARRE Francois >>Cc : sensors at Stimpy.netroedge.com >>Objet : Re: [RFC][2.6] Additional i2c adapter flags for i2c >>client isolation >> >> >>>This topic makes me think of something _interresting_... >>> >>>According to the tests I made on my VIA Nehemiah M10k, the vt1211 >>>can be handled by i2c-viapro as an smbus device (via the vt596 it >>>says) and via-ircc as an irda device, BUT this device can not be >>>binded to both drivers at the same time : when one is loaded (and >>>activated), the other finds no device (ENODEV I presume). (I wonder >>>how I should use both smbus*and* irda device)... >>> >>>Is this totally out of topic ? >> >>"Off-topic" even, yes ;) >> >> >>>Could a hack prevent these two drivers to be mutually exclusive? >> >>We knew of a similar problem between i2c-viapro and via686a. I didn't >>know that i2c-viapro and via-ircc had the same problem. And I >>guess that >>i2c-viapro and vt1211 would conflict too? >> >>Mark D. Studebaker is working on a solution for i2c-viapro >>and via686a. >>I don't know if the same solution will apply in the other cases. Since >>i2c-viapro is involved in all three conflicts, it promises to be damn >>complex... >> >>-- >>Jean Delvare >>http://www.ensicaen.ismra.fr/~delvare/ >> > > >
