Should drivers INIT? Was: Re: success on Asus P4S8X-X

Philip Pokorny ppokorny at penguincomputing.com
Mon Apr 14 08:05:06 CEST 2003


You know, perhaps it should be made a new recommendation/rule that drivers 
*not* initialize limits on chips by default.

I had a problem with the LM85 shutting down Margit's motherboard.  Other 
drivers have an init=0 option.

With the different VRM specs that are now out there it's even harder to set 
accurate defaults.

And if the BIOS *has* setup the chip, then when the driver "initializes" it, 
the user looses what are presumably good limits for the given motherboard.

Lastly, it would make the chip drivers smaller by a significant amount which 
seems to be a current focus of development [grin]...

:v)


> "set vrm 9.0" worked. thanks.
> now I have a better vid value :
> 
> VCore 1:   +1.58 V  (min =  +1.44 V, max =  +1.60 V)              
> [...]
> vid:      +1.525 V
> 
> I just can now ear a very short alarm signal trigered by the motherboard,
> after the module is loaded, and before "sensors -s" is launched. I can 
> live with that. I usually don't reboot very often :-)






More information about the lm-sensors mailing list