Hi Justin, > > Granted it's not part of the standard yet, but you would have three > > files temp[1-3]_crit_enable if we stick to our chip-indenpendent > > interface logic. Either make 1 read-write and [2-3] read-only, or make > > all read-write and each one changes the three values. > > Any reason not to simply provide 3 sysfs files pointing at the same > variable/register bit? Maintaining separate variables for a single, > uncomplicated value seems rather overkill. That's pretty much what I was proposing, actually ;) I never meant three different variables. The discussion was about having all files read-write or not. I usually have only the first file read-write and others read-only mirrors thereof, but sometimes people don't like that arbitrary symmetry breakage and go for read-write mirrors. You choose. > (...) > I would agree that it is a bit confusing that there are essentially 2 > temperature-motivated mechanisms for forcing fan speeds to full (automatic > PWM fan control, and critical temperature monitoring), but I think that the > utility of providing a critical temperature fail-safe is worth the > minor amount of confusion. Agreed. > No problem. Sorry that it's taking so much revision to get the kinks > worked out. This will undoubtedly become less trouble as I get more > familiar with i2c/lm_sensors/kernel issues. Well, the adm1026 driver is a complex driver which never received testing and reviewing so far, because it is so rarely found on motherboards. It's perfectly normal that it takes a couple revisions and discussion to get it in good shape. Thanks for the good work so far. -- Jean Delvare
