On 5/9/05, Mark Studebaker <mds at mds.gotdns.com> wrote: > 2.4 bmcsensors doesn't have write support at all. > In fact I didn't realize that 2.6 _did_ have write support. > Nice Yani! > That is obviously a requirement for setting anything. Well I added it for setting max/min/hysteresis limits when porting the driver but I couldn't actually test it (my bmc doesn't offer any write support) .. so..err I'm surprised if it works :-). I haven't heard about it one way or the other from anyone yet. > I've never seen an IPMI SDR that looked like PWM control or fan divisor. > IPMI systems tend to want to control their own fan speeds, in my experience, so taking manual control > of PWM doesn't really fit the IPMI paradigm too well. That's my (limited) experience too, Aj what specific bmc/motherboard are you using? I'd be interested to take a look at any documentation. > For example, with my IPMI system I was testing the fan reading so I stuck a pencil in the > CPU fan. It tried to restart the fan 3 times in about 10 seconds and then the system > did an instantaneous shutdown. Nice I guess but made testing harder :) lol... mine just blinks red leds at me which seems very useful in the case of a meltdown with me not around. I used my finger I think though not a pencil - ouch ;-). It tends to run the fans at very low speeds though. > I'm not familiar with how a general-purpose output (like PWM) would be implemented > in IPMI. bmcsensors is all about sensor _inputs_ and their limits. > > So I don't think PWM control via bmcsensors makes a lot of sense. Thanks, yes if this is only supported by one or two strange bmcs then it probably wouldn't make sense trying to add support. Yani
