Re: [lm-sensors] looking for help with W83795ADG

[Guenter Roeck <guenter.roeck@ericsson.com>]

On Fri, Nov 25, 2011 at 05:22:41AM -0500, Eddi De Pieri wrote:
> Hi Karsten,
> 
> about the complete power cycle issue:
> 
> While testing I had same problem like you... I had to remove power
> cord to get my server back...
> 
> Now I tried a soft "reboot" and I get any issue..
> 
> So my patch seems to be fully working and I'd like to know how to
> commit into mainstream.
> 

1) Don't top-post. Your e-mail may otherwise be ignored.
2) There is conflicting information if your patch is causing trouble 
   or not. Please clarify if and when a reboot fails, and what you changed
   to solve the problem.
3) This is an I2C multiplexer. Provide a patch which introduces an I2C
   multiplexer driver.
4) Read and follow Documentation/hwmon/submitting-patches for submitting
   your patch.

Thanks,
Guenter

> I think may be useful for other pc based on same chipset like ours.
> 
> Regards,
> 
> Eddi
> 
> On Mon, Nov 21, 2011 at 12:03 AM, Karsten de Freese <karsten@maxi-dsl.de> wrote:
> > hello Eddi,
> >
> > a less-than-welcome update:
> >
> > at least when the three modules i2c-piix4.ko, i2c-piix4-n36l.ko, w83795.ko
> > are loaded,  a restart (sudo shutdown -r now) of my N36L fails, the system
> > does not come up until I do a complete power cycle. I'll have to look more
> > closely at this, once I have physical access to the box again.
> >
> > regards -
> >
> > Karsten
> >
> > Am 20.11.2011 14:38, schrieb Karsten de Freese:
> >>
> >> hello Eddi,
> >>
> >> Success!
> >>
> >> Thanks for the detailed steps. This all works as described, at least as
> >> far as the 'sensors' output is concerned. With the respective modules, I get
> >> all the measurements you're listing! (See attached.)
> >>
> >> 'i2cdetect -l' will however not produce any output for whatsoever reason.
> >>
> >> The next thing I'd like to find out is how to load the modules
> >> automatically in a clean / kernel-upgrade-proof way..
> >>
> >> thanks and regards -
> >>
> >> Karsten
> >>
> >>
> >> Am 18.11.2011 11:09, schrieb Eddi De Pieri:
> >>>
> >>> I suggest you:
> >>>
> >>>
> >>> mkdir ~/piix4-n36l
> >>> cp
> >>> [yourlinuxtree]/usr/src/linux-source-2.6.32/drivers/i2c/busses/i2c-piix4.c
> >>> ~/piix4-n36l/
> >>> cp ~/Downloads/Makefile ~/piix4-n36l/
> >>> cd  ~/piix4-n36l/
> >>>
> >>> patch -p1<  proliant-microserver-n36l.diff
> >>> make
> >>> you will have your new module inside ~/piix4-n36l/
> >>>
> >>> rmmod i2c-piix4
> >>> insmod ./i2c-piix4.ko
> >>> insmod ./i2c-piix4-n36l.ko
> >>>
> >>> insmod w83795 (you should already build separately)
> >>>
> >>> sensors
> >>>
> >>> regards Eddi.
> >>>
> >>> I hope an answer from Jean... I'd like if this patch may be inserted
> >>> in mainstream linux with the necessary adjustment..
> >>>
> >>> On Thu, Nov 17, 2011 at 11:39 PM, Karsten de Freese<karsten@maxi-dsl.de>
> >>>  wrote:
> >>>>
> >>>> hello Eddi,
> >>>>
> >>>> I'm ready for two questions ;-)
> >>>>
> >>>> .diff / patch looks simple, and I also found i2c-amd756-s4882.c -
> >>>> however:
> >>>> - the diff file actually describes the patch to be on top of
> >>>> i2c-piix4.c?
> >>>> - if I'd compile the patched file, would that give me a standalone
> >>>> driver?
> >>>>
> >>>> thanks for your help -
> >>>>
> >>>> Karsten
> >>>>
> >>>> On 17.11.2011 22:25, Karsten de Freese wrote:
> >>>>>
> >>>>> hi Eddi,
> >>>>>
> >>>>> thanks very much - I had almost forgotten about the topic, but stumbled
> >>>>> across it again a few days ago, so that's very interesting.
> >>>>> I'm not yet familiar with .diff patches, and may welcome a little help.
> >>>>> However, let me try to find out on my own first..
> >>>>>
> >>>>> thanks&  regards -
> >>>>>
> >>>>> Karsten
> >>>>>
> >>>>> On 16.11.2011 17:18, Eddi De Pieri wrote:
> >>>>>>
> >>>>>> Finally I get W83795ADG working on my proliant microserver
> >>>>>>
> >>>>>> The patch is based on i2c-amd756-s4882.c and adapted for sb800
> >>>>>> southbus.
> >>>>>>
> >>>>>> Karsten, please try it too...
> >>>>>>
> >>>>>> root@proliant:/usr/src/lm-sensors/eddi# i2cdetect -l
> >>>>>> i2c-0   smbus           SMBus piix4 adapter (SDA0)              SMBus
> >>>>>> adapter
> >>>>>> i2c-1   smbus           SMBus piix4 adapter (SDA2)              SMBus
> >>>>>> adapter
> >>>>>> i2c-2   smbus           SMBus piix4 adapter (SDA3)              SMBus
> >>>>>> adapter
> >>>>>> i2c-3   smbus           SMBus piix4 adapter (SDA4)              SMBus
> >>>>>> adapter
> >>>>>> root@proliant:/usr/src/lm-sensors/eddi#
> >>>>>>
> >>>>>> yes SDA1 is reserved... so i can't multiplex it
> >>>>>>
> >>>>>> root@proliant:/usr/src/lm-sensors/eddi# sensors
> >>>>>> k10temp-pci-00c3
> >>>>>> Adapter: PCI adapter
> >>>>>> temp1:       +24.5°C  (high = +70.0°C, crit = +100.0°C)
> >>>>>>
> >>>>>> w83795adg-i2c-1-2f
> >>>>>> Adapter: SMBus piix4 adapter (SDA2)
> >>>>>> in0:         +1.02 V  (min =  +0.00 V, max =  +2.05 V)
> >>>>>> in1:         +1.52 V  (min =  +0.00 V, max =  +2.05 V)
> >>>>>> in2:         +1.10 V  (min =  +0.00 V, max =  +2.05 V)
> >>>>>> in3:         +0.89 V  (min =  +0.00 V, max =  +2.05 V)
> >>>>>> in12:        +3.35 V  (min =  +0.00 V, max =  +6.14 V)
> >>>>>> in13:        +3.28 V  (min =  +0.00 V, max =  +6.14 V)
> >>>>>> fan1:        703 RPM  (min =  329 RPM)
> >>>>>> temp1:       +23.0°C  (high = +109.0°C, hyst = +109.0°C)
> >>>>>>                       (crit = +109.0°C, hyst = +109.0°C)  sensor =
> >>>>>> thermal diode
> >>>>>> temp2:       +33.2°C  (high = +105.0°C, hyst = +105.0°C)
> >>>>>>                       (crit = +105.0°C, hyst = +105.0°C)  sensor =
> >>>>>> thermal diode
> >>>>>> temp5:       +14.0°C  (high = +39.0°C, hyst = +39.0°C)
> >>>>>>                       (crit = +44.0°C, hyst = +44.0°C)  sensor =
> >>>>>> thermistor
> >>>>>> beep_enable:disabled
> >>>>>>
> >>>>>> jc42-i2c-0-18
> >>>>>> Adapter: SMBus piix4 adapter (SDA0)
> >>>>>> temp1:       +20.5°C  (low  =  +0.0°C, high =  +0.0°C)  ALARM
> >>>>>>                       (crit =  +0.0°C, hyst =  +0.0°C)  ALARM
> >>>>>>
> >>>>>>
> >>>>>> root@proliant:/usr/src/lm-sensors/eddi# i2cdetect -y 0
> >>>>>>      0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
> >>>>>> 00:          -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 10: -- -- -- -- -- -- -- -- UU -- -- -- -- -- -- --
> >>>>>> 20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 50: 50 -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 70: -- -- -- -- -- -- -- --
> >>>>>>
> >>>>>> root@proliant:/usr/src/lm-sensors/eddi# i2cdetect -y 1
> >>>>>>      0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
> >>>>>> 00:          -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- UU
> >>>>>> 30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 50: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 60: -- 61 -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 70: -- -- -- -- -- -- -- --
> >>>>>>
> >>>>>> root@proliant:/usr/src/lm-sensors/eddi# i2cdetect -y 2
> >>>>>>      0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
> >>>>>> 00:          -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 40: -- -- -- -- -- -- -- -- -- -- -- -- 4c -- -- --
> >>>>>> 50: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 70: -- -- -- -- -- -- -- --
> >>>>>>
> >>>>>> root@proliant:/usr/src/lm-sensors/eddi# i2cdetect -y 3
> >>>>>>      0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
> >>>>>> 00:          -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 50: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> >>>>>> 70: -- -- -- -- -- -- -- --
> >>>>>>
> >>>>>> pay attention that the msleep seems to be really needed...
> >>>>>>
> >>>>>> Regards, Eddi
> >>>>>>
> >>>>>> On Tue, Aug 16, 2011 at 9:26 AM, Eddi De Pieri<eddi@depieri.net>
> >>>>>>  wrote:
> >>>>>>>
> >>>>>>> Hi Jean...
> >>>>>>>
> >>>>>>> The bios show fan rpm and temperature. Since my own proliant
> >>>>>>> microserver
> >>>>>>> lacks of ipmi card, I doubt the bios give me the results using ipmi
> >>>>>>> call.
> >>>>>>>
> >>>>>>> The HP Proliant Microserver chipset (SB820M) have a multiplexed smbus
> >>>>>>> like your.
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>> http://support.amd.com/it/Embedded_TechDocs/47283_sb820m_ds_pub_2.00.pdf
> >>>>>>>
> >>>>>>> Here is technical document from amd...
> >>>>>>>
> >>>>>>>
> >>>>>>> at p52/53 they say:
> >>>>>>>
> >>>>>>> (1) The SDA1 and SCL1 SMBus interface is dedicated for ASF devices
> >>>>>>> only.
> >>>>>>> It should not be used to
> >>>>>>> connect to any other devices.
> >>>>>>> (2) There are only two SMBus controllers. The SCL1/SDA1 pair is
> >>>>>>> controlled by SMBus controller 1.
> >>>>>>>
> >>>>>>>
> >>>>>>> SCL0/SDA0, SCL2/SDA2, SCL3/SDA3 and SCL4/SDA4 are multiplexed pins
> >>>>>>> that
> >>>>>>> are all controlled by
> >>>>>>> SMBus controller 0, and only 1 pair of those pins can be active at
> >>>>>>> any
> >>>>>>> time.
> >>>>>>>
> >>>>>>> I can't understand how to enable 2, 3 and 4 muxed bus..
> >>>>>>>
> >>>>>>>
> >>>>>>> If you consolidate a solution for i2c-i801.c I hope we can be
> >>>>>>> replicated
> >>>>>>> for i2c-piix4.c too..
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>> Hi Karsten,
> >>>>>>>
> >>>>>>> On Fri, 22 Oct 2010 16:04:10 +0200, karsten at maxi-dsl.de wrote:
> >>>>>>>>
> >>>>>>>> hi Jean,
> >>>>>>>>>
> >>>>>>>>> The w83795 driver uses the same detection mechanism used in
> >>>>>>>>> sensors-detect. So if sensors-detect doesn't see your chip, I am
> >>>>>>>>> not
> >>>>>>>>> surprised that the driver didn't either.
> >>>>>>>>> The full output of sensors-detect is needed to diagnose your
> >>>>>>>>> problem
> >>>>>>>>> further. It could be that your SMBus controller isn't supported, or
> >>>>>>>>> maybe you are supposed to access the chip through IPMI.
> >>>>>>>>
> >>>>>>>> IPMI is probably the key..
> >>>>>>>>   From the sensors-detect output:
> >>>>>>>> "Warning: the required module ipmisensors is not currently installed
> >>>>>>>> on your system. If it is built into the kernel then it's OK.
> >>>>>>>> Otherwise, checkhttp://www.lm-sensors.org/wiki/Devices  for
> >>>>>>>> driver availability" (full ouput attached)
> >>>>>>>>
> >>>>>>>> I did see the output before, but didn't really understand the
> >>>>>>>> implications (IPMI seems to be an alternative way of accessing the
> >>>>>>>> monitoring chips?)
> >>>>>>>
> >>>>>>> Yes. The strength of IPMI access is that it can be used remotely too,
> >>>>>>> for example from a BMC plugged into your system.
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>>> The suggested ipmisensors driver however does not seem to be
> >>>>>>>> available,
> >>>>>>>> right?
> >>>>>>>> (http://lm-sensors.org/wiki/Devices  says "(2004-12-12)  Port to
> >>>>>>>> Linux
> >>>>>>>> 2.6 in progress by Yani Ioannou")
> >>>>>>>
> >>>>>>> That project is essentially dead. At the time being, the best way to
> >>>>>>> retrieve IPMI-based sensor values is using "ipmitool sensor". The
> >>>>>>> drawback is that this isn't integrated with libsensors, so the
> >>>>>>> usually
> >>>>>>>
> >>>>>>>
> >>>>>>> monitoring applications won't work.
> >>>>>>>
> >>>>>>> If you have a BMC or are otherwise already using IPMI on your
> >>>>>>> machine,
> >>>>>>> just use ipmitool and you're done.
> >>>>>>>
> >>>>>>> If not, then it might be worth trying ipmitool once, and then try
> >>>>>>>
> >>>>>>>
> >>>>>>> sensors-detect again. There is a small chance that, once ipmitool
> >>>>>>> managed to access the monitoring device, sensors-detect will see it.
> >>>>>>> My
> >>>>>>> own mainboard which has a W83795ADG chip has a multiplexed SMBus, and
> >>>>>>> I
> >>>>>>> can only see the W83795ADG in sensors-detect when the multiplexer is
> >>>>>>>
> >>>>>>>
> >>>>>>> set properly. Otherwise I would see the memory modules, which _do_
> >>>>>>> appear in your sensors-detect output. So maybe your board has a
> >>>>>>> multiplexed SMBus too.
> >>>>>>>
> >>>>>>> If the trick doesn't work, you may want to try the new jc42 driver
> >>>>>>>
> >>>>>>>
> >>>>>>> which can monitor the temperature of some DDR3 memory modules. I
> >>>>>>> suspect your memory modules have such sensors. A more recent version
> >>>>>>> of
> >>>>>>> sensors-detect would confirm that:
> >>>>>>>   http://dl.lm-sensors.org/lm-sensors/files/sensors-detect
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>> I also see that sensors-detect suggested to use the k10temp driver,
> >>>>>>> which your kernel doesn't have. You can try the standalone flavor of
> >>>>>>> the drivers which I maintain at:
> >>>>>>>   http://khali.linux-fr.org/devel/misc/k10temp/
> >>>>>>>
> >>>>>>>
> >>>>>>> The accuracy of the CPU internal sensors is limited, but that's still
> >>>>>>> better than nothing.
> >>>>>>>
> >>>>>>> --
> >>>>>>> Jean Delvare
> >>>>>>> http://khali.linux-fr.org/wishlist.html
> >>>>>>>
> >>>>>>>
> >>>>>
> >>>>
> >>
> >
> >
> 
> _______________________________________________
> lm-sensors mailing list
> lm-sensors@lm-sensors.org
> http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors