> On Tue, 17 Feb 2004, Jean Delvare wrote: > > Interesting. I will analyze the data and try to summarize some > > facts. Could you dump the chip again under high load? If it is some > > temperature sensor, we should see at least one register increase. > > Here they are, baked at 75 C :) Thanks. I have more information since yesterday. It happens that this mysterious chip is simply your via686a chip, except that you access it through the SMBus instead of ISA. Until then I didn't know that this chip could possibly be accessed that way. Now, there's a theory I would like to confirm, if you can provide one more dump. I'd like you to load i2c-isa and via686a. Then note the address of the via686a. If should be relatively high (such as 0x5800 or 0x6000). Then unload via686a and dump it using the following command: isadump -f 0x<address> This should give you a dump similar to what i2cdump gave you so far. Please send it to me :) Thanks. (BTW, your lm80 driver will be part of Linux 2.6.4, in case I didn't already told you! Congratulations.) -- Jean Delvare http://www.ensicaen.ismra.fr/~delvare/
