On Wed, 16 Mar 2011 06:51:10 -0700, Guenter Roeck wrote: > On Wed, Mar 16, 2011 at 04:26:49AM -0400, Jean Delvare wrote: > > Note that this means that this device has now in1_alarm and > > in1_crit_alarm. This is a case your proposed rewrite of the "sensors" > > code won't properly deal with, as it will only check for limit-specific > > alarm flags in the absence of a generic alarm flag (so in this case > > in1_crit_alarm will be ignored.) Not that the original code was better, > > but maybe it's the right time to get it right. > > > Yes, I know :(. I didn't want to get rid of the _crit attribute, since it is > supported and adds value. Figured I'd deal with the sensors problem later > (if I find a good solution). Just always check for all possible alarm flags (generic and limit-specific alike)? The only annoyance I can see is that when in1_crit_alarm is set, whether or not in1_alarm is also set won't affect the output, it will be "ALARM (CRIT)" in both cases. But I don't think it is a problem in practice. -- Jean Delvare _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
