Jean We have 3 monitored fans. They all showed the same behaviour. I make absolutely no claims that rev 8 is correct - it just gets us out of a hole :-) I'll resubmit the patch if needs be...would be good to get to the bottom of this first. If I get chance I'll have a look at the specs. Best regards John -----Original Message----- From: Jean Delvare [mailto:khali at linux-fr.org] Sent: 15 September 2008 15:29 To: John Gumb; Andrew Paprocki Cc: lm-sensors at lm-sensors.org Subject: Re: it8712 rev 7; fan speeds reported incorrectly. Hi John, Le lundi 15 septembre 2008, John Gumb a ?crit?: > I've just picked up linux-2.6.27-rc5 and it would appear fan speeds are > reported incorrectly now. We have it8712 rev7 on our systems. We see > speeds of around 40K RPM (!!). > > The following hack fixes it - it just punts 16 bit fan support out to > rev8. It would appear our it8172 does not have 16 bit tacho support. Or > if it does, the driver isn't enabling it properly. Thanks for reporting. I don't have access to my collection of IT87xx datasheets and personal notes on chip revisions, so I can't check now. I will check when I am back home. Andrew, what do you think? John, how many fans are monitored on your systems, and were they all screwed up in 2.6.27-rc5, or only some of them? > --- drivers/hwmon/it87.c.orig 2008-09-05 15:51:11.000000000 +0100 > > +++ drivers/hwmon/it87.c 2008-09-05 15:51:53.000000000 +0100 > > @@ -283,7 +283,7 @@ > > IT8712F Datasheet 0.9.1, section 8.3.5 indicates 7h == > Version I. > > These are the first revisions with 16bit tachometer support. > */ > > return (data->type == it87 && data->revision >= 0x03) > > - || (data->type == it8712 && data->revision >= 0x07) > > + || (data->type == it8712 && data->revision >= 0x08) > > || data->type == it8716 > > || data->type == it8718; > Patch was totally destroyed by your mailer, so we can't apply it. Please resend with HTML mails disabled or as a text attachment, or whatever is needed for the patch to reach us in its original format. We will also need a Signed-off-by statement from you as described in section 12 of Documentation/SubmittingPatches. Thanks, -- Jean Delvare
