From mboxrd@z Thu Jan 1 00:00:00 1970 From: Guenter Roeck Date: Wed, 21 Dec 2011 22:26:00 +0000 Subject: Re: [lm-sensors] NCT6776 global registers 0x1c=0 and 0x24=0x5c Message-Id: <20111221222600.GA18985@ericsson.com> List-Id: References: <20110410160433.GA22075@localhost> In-Reply-To: <20110410160433.GA22075@localhost> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: lm-sensors@vger.kernel.org On Wed, Dec 21, 2011 at 03:55:55PM -0500, Ira W. Snyder wrote: > On Wed, Dec 21, 2011 at 06:55:31PM +0100, Ian Dobson wrote: > > Hi, > > -------------------------------------------------- > > From: "Guenter Roeck" > > Sent: Wednesday, December 21, 2011 6:01 PM > > To: "Harald Judt" > > Cc: > > 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" > > >> >> Sent: Sunday, April 10, 2011 6:04 PM > > >> >> To: > > >> >> 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@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. > Possibly, and it might be what SpeedFan is doing. That should be done in drivers/platform/x86, though. Guenter _______________________________________________ lm-sensors mailing list lm-sensors@lm-sensors.org http://lists.lm-sensors.org/mailman/listinfo/lm-sensors