Hi Jean, Comments below: > -----Original Message----- > From: Jean Delvare [mailto:khali at linux-fr.org] > Sent: 15 November 2005 12:02 > To: roger at planbit.co.uk > Cc: 'Mark Studebaker'; LM Sensors > Subject: RE: RE: vt8231.c > > > Hi Roger, > > On 2005-11-15, Roger Lucas wrote: > > We have another sample of the board here now, so we will be repeating > the > > measurements to ensure that the VIA results are predictable from board > to > > board. > > OK, great. I've tried some mathematics on the formulae to find out if > anything obvious with regards to the physical meaning of the constants > would spring out, but did not find anything. > > > I agree that whatever we do, the VIA users must apply some kind of > > equation to get their correct temperature, but I feel that the safest > > approach is "first do no wrong". > > Fair enough. > > > I am extremely uncomfortable with the situation where we release a > > driver that we know will give incorrect information under some > > conditions. I feel that it is much better to return the result from > > the driver as a register value and put the text below: > > The problem is that the standard sysfs interface sets some contraints on > what we can or cannot do. In particular, it dictates that the magnitude > for temperature values is 3. This suggests that the value we would be > returning would be REG * 1000. As it seems that there are an additional > 2 bits of resolution, that would be (REG << 2 + ADDREG) * 250. > >From the above, if the driver returns the result ((REG << 2 + ADDREG) * 250) to the sensors user-space application, then you are OK with this? The SENSORS.CONF file would then apply one of the two lines below to give user the temperature in degrees Centigrade. > > # If you have an Intel CPU, then uncomment the line below > > # temp0_input = (@-65)/0.9686, (@*0.9686)+65 > > > > # If you have an VIA EPIA CPU, then uncomment the line below > > # temp0_input = (@-45)/0.7809, (@*0.7809)+45 > > s/temp0_input =/compute temp1/, but yes. > > BTW, does this suggest that you decided that the diode temperature would > be temp1, and thermistor-based ones are temp2+? I have no objection a > priori, just curious. Nope. I don't care which is which. Really. If there is a general trend for the CPU-0 temperature to be on a specific sensor then let me know and I'll make the driver match. > > > This way, an incorrect value is never returned and the user must > > correctly set the board configuration before a temperature is > > returned. > > By commenting out both compute lines, I'd rather say that an incorrect > value will *always* be returned. I'd prefer to keep the Intel line by > default so as to preserve backward compatibility as much as we can. > > I think you will end up with a dedicated configuration file for your > board anyway. The best we can do to help out VIA EPIA users is to make > this file publicly available. Fair enough. > > Thanks, > -- > Jean Delvare If you can send me the results for the code review of the driver then I'll wrap these changes up into it and re-submit it. Hopefully then it is complete. Best regards, Roger
