Hello Chunhao: * Huang0 at Winbond.com.tw <Huang0 at Winbond.com.tw> [2004-12-23 09:51:24 +0800]: > I have some questions about VRM. > > I find such codes in w83781d_init_client() of w83781d.c: > data->vrm = (type == w83791d) ? 90 : 82; > That's to say, if the chip is w83791d, the driver w83781d fix > the value of VRM into 90. > In my driver of 792, I also fix the value of VRM into 90, > Is that right? > > Since I'm not very familiar with VRM, Could you give me a general > introduction to it when you are free? When we (lm_sensors people) talk about VRM, we are talking specifically about the interpretation of the VID lines; i.e. the formula for converting from 4-6 binary inputs to a voltage. Intel has many standards for the processor VRM, the latest of which can be found here: http://developer.intel.com/design/Xeon/guides/302731.htm The vrm value for w83791d in the w83781d driver is not *fixed* to 90, it is merely *initialized* to that. The user is free to change this value. You may wonder why the user is allowed to change it at all, when the value obviously depends on what type of processor is present. Well, in the 2.6 kernel we made some moves in that direction: see drivers/i2c/i2c-sensor-vid.c. So to answer your immediate question ... from the recently posted 792 datasheet, it looks like VRM 9.0 and 10.0 are supported. You'll want to pick one of these as a default, but still allow the user to choose the other one. Wait a sec... just read the 792 datasheet a little more closely. That part also supports some sort of VID *outputs* as well. That could be downright dangerous to fool around with. I wonder what you're planning to do with that? I'll have a look at your driver at some point. Now for my own question to you: I have a w83791sd on my mainboard. Will you publish a better datasheet for that one? The datasheet that is currently available on your website doesn't have enough info to create a driver. Regards, -- Mark M. Hoffman mhoffman at lightlink.com
