Hi Hans, On Mon, 15 Dec 2008 13:51:13 +0100, Hans de Goede wrote: > I just realised there is an issue with the > hwmon-f71882fg-11-separate-max-crit-alarm-and-beep.patch > patch. > > It removes the temp#_alarm attributes (replacing them by temp#_max_alarm), > which is fine for lm_sensors v3, but will cause a problem with lm_sensors v2, > the "sensors" command from lm_sensors v2 will now give failed to read data for > temp# for all 3 temps :( Indeed. I admit I didn't think about lm_sensors v2 when reviewing your patch. This is probably a good reason to _not_ add support for new chips to lm_sensors v2. Owners of such recent hardware should really use lm_sensors v3 anyway. > So it looks like we are sorta stuck with our past mistake of not doing separate > alarms. > > We could do separate alarms like this: > temp#_alarm > temp#_crit_alarm > > This will work with both libsensors and with "sensors" from both v2 and v3. It's a bit confusing due to the asymmetry, but I admit it should work fine in practice. Feel free to send an updated patch doing this. Including an explanation of why we do this, so that someone reading the code in 3 or 4 years can understand the idea behind it. -- Jean Delvare
