On Wed, Oct 30, 2013 at 04:33:26PM +0100, Jean Delvare wrote: > Hi Guenter, > > On Mon, 15 Jul 2013 10:33:22 -0700, Guenter Roeck wrote: > > On Mon, Jul 15, 2013 at 06:57:27PM +0200, Jean Delvare wrote: > > > Unrelated to this patch, but Guenter, I am worried about the MAX6696 > > > handling here. I realize that I am the one who accepted your code, but > > > now it looks wrong. Specifically: > > > * We check for (status2 & 0xfe) i.e. 7 alarm bits, but the code below > > > only reports 2 alarms bits. So if any of the 5 other alarm bits in > > > STATUS2 are, we may return true (chip is tripped) but not print the > > > cause. > > > * At least bits 1 and 2 of STATUS 2 fit totally fine in the driver as > > > it currently exists, so I can't think of any reason for not handling > > > them. Why are we not? Ideally we should print a message for every > > > alarm bit so that we never return "true" without printing a message. > > > Even though OT2 limits aren't handled by the driver... > > > * If you think this piece of code shouldn't deal with OT/THERM limits > > > because they do not trigger an SMBus alarm, this can be discussed, > > > but all chips should be handled the same in this respect then. > > > * Why in the first place is max6696's data->alert_alarms set to 0x187c > > > and not 0x1c7c? Including 1OPEN but not 2OPEN makes no sense. > > > > I am about to leave for vacation, so this will have to wait for a couple of > > weeks. I'll look at it after I am back. > > Are you back now? ;-) > Yes, only I completely forgot about this :-). I'll add it to my task list. Guenter _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
