[lm-sensors] NCT6776 global registers 0x1c=0 and 0x24=0x5c onASRock Extreme4

Ira W. Snyder iws at ovro.caltech.edu
Wed Dec 21 21:55:55 CET 2011


On Wed, Dec 21, 2011 at 06:55:31PM +0100, Ian Dobson wrote:
> Hi,
> --------------------------------------------------
> From: "Guenter Roeck" <guenter.roeck at ericsson.com>
> Sent: Wednesday, December 21, 2011 6:01 PM
> To: "Harald Judt" <h.judt at gmx.at>
> Cc: <lm-sensors at lm-sensors.org>
> Subject: Re: [lm-sensors] NCT6776 global registers 0x1c=0 and 0x24=0x5c 
> onASRock Extreme4
> 
> > Hi,
> > On Wed, Dec 21, 2011 at 09:50:31AM -0500, Harald Judt wrote:
> >> Hi,
> >>
> >> > On Tue, 12 Apr 2011 19:02:19 +0200, Ian Dobson wrote:
> >> >> Hi,
> >> >>
> >> >> --------------------------------------------------
> >> >> From: "Mike Campin" <lm_sensors at ootsa.homelinux.net>
> >> >> Sent: Sunday, April 10, 2011 6:04 PM
> >> >> To: <lm-sensors at lm-sensors.org>
> >> >> Subject: [lm-sensors] NCT6776 global registers 0x1c=0 and 0x24=0x5c 
> >> >> onASRock
> >> >> Extreme4
> >> >>
> >> >> > Hi,
> >> >> >
> >> >> > Only 2 of the 5 fans are showing up on my ASRock P67 Extreme4 
> >> >> > motherboard.
> >> >> > It due to the NCT6776 global configuration register 0x1c and 0x24 
> >> >> > values.
> >> >> > The fans work if I modify these registers prior to loading the 
> >> >> > w83627ehf
> >> >> > driver.
> >> >> >
> >> >> > My question is where are these registers initialized? Should I ask 
> >> >> > ASRock
> >> >> > to fix the BIOS?
> >> >> >
> >> >> >  isadump -y -k 0x87,0x87 0x2e 0x2f
> >> >> >
> >> >> >          0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
> >> >> >     00: ff ff 00 ff ff ff ff 02 ff ff ff ff ff ff ff ff
> >> >> >     10: ff ff ff ff ff ff ff ff ff ff f8 0e 00 00 ff ff
> >> >> >     20: c3 33 ff 00 5c 00 00 98 00 ff 20 00 80 00 00 01
> >> >> >     30: 00 ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
> >> >> >
> >> >> >  isaset -y -f 0x2e 0x87
> >> >> >  isaset -y -f 0x2e 0x87
> >> >> >  isaset -y 0x2e 0x2f 0x1c 0x3
> >> >> >  isaset -y 0x2e 0x2f 0x24 0x1c
> >> >> >
> >> >> > Thanks, Mike
> >> >> >
> >> >> > --
> >> >> > Mike Campin
> >> >> > lm_sensors at ootsa.homelinux.net
> >> >> >
> >> >> > _______________________________________________
> >> >> > lm-sensors mailing list
> >> >> > lm-sensors at lm-sensors.org
> >> >> > http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
> >> >>
> >> >> ASRock designed the board, so they should setup the hardware 
> >> >> monitoring chip
> >> >> correctly. Allowing users to change the chip configuration is a really 
> >> >> bad
> >> >> idea, as many of the pins can have different functionality/electrical
> >> >> spcifications depending on the config. Setting up a pin incorrectly 
> >> >> could
> >> >> blow the chip (Configuring a pin so that it's a source that's 
> >> >> connected to
> >> >> an unprotected/limited sink for example).
> >> >>
> >> >> So ASRock should fix the BIOS.
> >> >
> >> > Seconded. But before flaming Asrock, two things worth checking:
> >> > * Availability of a BIOS update fixing the issue.
> >> > * Options in the BIOS to enable/disable the monitoring of specific
> >> >   fans. I can imagine that the BIOS skips the configuration steps for
> >> >   fans for which monitoring was disabled (although that would probably
> >> >   mean more code than just doing it unconditionally...)
> >> >
> >> > --
> >> > Jean Delvare
> >>
> >> I own an Asrock P67 Extreme4 too and have the same problem. It can be
> >> solved by following Mike Campin's instructions, which I successfully
> >> incorporated into /etc/init.d/lm_sensors. There is no BIOS available
> >> that fixes this issue, but I noticed that SpeedFan for Windows
> >> (www.almico.com/speedfan.php) seems to do something similar, otherwise
> >> it would show only 2 fans too. So maybe we could get that into
> >> lm_sensors, even if the BIOS is at fault here?
> >>
> > See above ... worst case this can result in defective hardware if 
> > misconfigured.
> >
> > We could possibly introduce some kind of module parameter, but I don't 
> > feel
> > too comfortable with that - wrong values may still blow up the hardware.
> > Any opinions, anyone ?
> >
> > Guenter
> >
> > _______________________________________________
> > lm-sensors mailing list
> > lm-sensors at lm-sensors.org
> > http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
> 
> I wouldn't recommend a module parameter, leave the driver as it is.
> If there is a hack (using isaset) just use that. It's too easy for a user to 
> set a module parameter hoping that it'll fix some random problem, and if the 
> module kills the hardware, then the module is too blame.
> 
> Asrock should fix the BIOS, that's the correct option.
> 

I wonder if you could set these register values only on this specific
motherboard by checking the DMI table to determine the motherboard
vendor and model number.

At least then you can work around a broken BIOS until the vendor
provides a fix.

Ira




More information about the lm-sensors mailing list