From: Guenter Roeck <linux@roeck-us.net>
To: lm-sensors@vger.kernel.org
Subject: Re: [lm-sensors] SuperMicro H8DG6-F-O IPMI BMC and lm_sensors
Date: Fri, 02 Jan 2015 08:53:56 +0000 [thread overview]
Message-ID: <20150102085356.GA29009@roeck-us.net> (raw)
In-Reply-To: <CADw2B2PqoMeXkr0yVPbn_R=z17yM6F1QgefYwBmvaqHHqtCDKA@mail.gmail.com>
On Thu, Jan 01, 2015 at 03:07:25PM -0800, G. Richard Bellamy wrote:
> I need help configuring lm_sensors for a SuperMicro H8DG6-F-O motherboard.
>
> I have an interesting challenge... my motherboard is a SuperMicro
> H8DG6-F-O motherboard with an IPMI BMC Chipset [1], and I'm trying to
> collect fan speed (among other things) via lm-sensors. I've been in
> ongoing discussions with SuperMicro, and I need some help - I can get
> what I need via the IPMI interface:
> --------------------------------------------------------------------------
> 2014-12-28 10:35:58
> rbellamy@x i ~ % sudo ipmitool sdr
> CPU1 Temp | 0x00 | ok
> CPU2 Temp | 0x00 | ok
> System Temp | 40 degrees C | ok
> CPU1 Vcore | 1.04 Volts | ok
> CPU2 Vcore | 0.92 Volts | ok
> CPU1 DIMM | 1.52 Volts | ok
> CPU2 DIMM | 1.52 Volts | ok
> +1.1 V | 1.10 Volts | ok
> +1.8 V | 1.83 Volts | ok
> +5 V | 4.99 Volts | ok
> +12 V | 12.14 Volts | ok
> VTT | 1.14 Volts | ok
> +3.3 V | 3.26 Volts | ok
> +3.3VSB | 3.34 Volts | ok
> VBAT | 3.14 Volts | ok
> FAN 1 | no reading | ns
> FAN 2 | no reading | ns
> FAN 3 | no reading | ns
> FAN 4 | 441 RPM | cr
> FAN 5 | no reading | ns
> FAN 6 | no reading | ns
> FAN 7 | 961 RPM | ok
> FAN 8 | 1024 RPM | ok
> Intrusion | 0x01 | ok
> PS Status | 0x01 | ok
> --------------------------------------------------------------------------
>
> But nothing comes through for sensors:
> --------------------------------------------------------------------------
> 2014-12-28 10:36:00
> rbellamy@eanna i ~ % sudo sensors
> fam15h_power-pci-00c4
> Adapter: PCI adapter
> power1: 72.18 W (crit = 115.21 W)
>
> fam15h_power-pci-00d4
> Adapter: PCI adapter
> power1: 55.81 W (crit = 115.21 W)
>
> k10temp-pci-00c3
> Adapter: PCI adapter
> temp1: +24.2°C (high = +70.0°C)
> (crit = +70.0°C, hyst = +67.0°C)
>
> k10temp-pci-00cb
> Adapter: PCI adapter
> temp1: +24.2°C (high = +70.0°C)
>
> k10temp-pci-00d3
> Adapter: PCI adapter
> temp1: +15.2°C (high = +70.0°C)
> (crit = +70.0°C, hyst = +67.0°C)
>
> k10temp-pci-00db
> Adapter: PCI adapter
> temp1: +15.2°C (high = +70.0°C)
>
> radeon-pci-4200
> Adapter: PCI adapter
> temp1: +74.5°C (crit = +120.0°C, hyst = +90.0°C)
>
> jc42-i2c-0-1d
> Adapter: SMBus PIIX4 adapter at 0b00
> temp1: +39.4°C (low = +0.0°C) ALARM (HIGH, CRIT)
> (high = +0.0°C, hyst = +0.0°C)
> (crit = +0.0°C, hyst = +0.0°C)
>
> jc42-i2c-0-1f
> Adapter: SMBus PIIX4 adapter at 0b00
> temp1: +37.5°C (low = +0.0°C) ALARM (HIGH, CRIT)
> (high = +0.0°C, hyst = +0.0°C)
> (crit = +0.0°C, hyst = +0.0°C)
>
> w83627dhg-isa-0a10
> Adapter: ISA adapter
> Vcore: +2.04 V (min = +2.04 V, max = +2.04 V) ALARM
> in1: +2.04 V (min = +2.04 V, max = +2.04 V) ALARM
> AVCC: +4.08 V (min = +4.08 V, max = +4.08 V) ALARM
> +3.3V: +4.08 V (min = +4.08 V, max = +4.08 V) ALARM
> in4: +2.04 V (min = +2.04 V, max = +2.04 V) ALARM
> in5: +2.04 V (min = +2.04 V, max = +2.04 V) ALARM
> in6: +2.04 V (min = +2.04 V, max = +2.04 V) ALARM
> 3VSB: +4.08 V (min = +4.08 V, max = +4.08 V) ALARM
> Vbat: +4.08 V (min = +4.08 V, max = +4.08 V) ALARM
> fan1: 0 RPM (min = 0 RPM, div = 128) ALARM
> fan2: 0 RPM (min = 0 RPM, div = 128) ALARM
> fan3: 0 RPM (min = 0 RPM, div = 128) ALARM
> fan4: 0 RPM (min = 0 RPM, div = 128) ALARM
> temp1: -1.0°C (high = -1.0°C, hyst = -1.0°C) ALARM sensor
> = CPU diode
> temp2: +0.0°C (high = +0.0°C, hyst = +0.0°C) ALARM sensor
> = CPU diode
> temp3: +0.0°C (high = +0.0°C, hyst = +0.0°C) ALARM sensor
> = CPU diode
> cpu0_vid: +0.000 V
> intrusion0: ALARM
> --------------------------------------------------------------------------
>
> I've seen mention that I shouldn't be using the Windbond w83627dhg
> sensor, as the temps and fans aren't connected. Is this true, or am I
> just missing the correct SMBus offsets?
>
You can see from the values above that the sensors are not connected (noe of
them as far as I can see).
> SuperMicro recommends against installing lm_sensors:
> --------------------------------------------------------------------------
> SuperMicro: We are not recommend user install lmssenors or any 3th
> party sensor monitor utility to monitor sensor reading on motherboard
> with IPMI (BMC) chipset. It could cause system to crash due to 2
> sensors monitor probing sensor reading in the same time.
> Me: So you are saying NOT to install lm-sensors with a BMC-equipped motherboard?
> SuperMicro: Yes. Our LAB able the duplicate issue regarding system
> crash when installing lm-sensor on motherboard with BMC-equipped
> motherboard. System crash could happen in random time.
> --------------------------------------------------------------------------
>
> Does that make sense to the list?
>
Yes, that is correct.
> Their advice is confusing, since at the same time as they were telling
> me to NOT install lm_sensors, they gave me the offsets necessary to
> make the WinBond work with lm_sensors:
>
> SuperMicro:
> --------------------------------------------------------------------------
> Please check the offset address for lm-sensor to match information below.
>
> Windbond W83795AG, Slave Address=0x2f (0x5E in 8-Bit format)
> =============================================================
> Fan1 Fan Speed, Offset 0x2e RPM = 84375/Data
> Fan2 Fan Speed, Offset 0x2f RPM = 84375/Data
> Fan3 Fan Speed, Offset 0x30 RPM = 84375/Data
> Fan4 Fan Speed, Offset 0x31 RPM = 84375/Data
> Fan5 Fan Speed, Offset 0x32 RPM = 84375/Data
> Fan6 Fan Speed, Offset 0x33 RPM = 84375/Data
> Fan7 Fan Speed, Offset 0x34 RPM = 84375/Data
> Fan8 Fan Speed, Offset 0x35 RPM = 84375/Data
> CPU1 Voltage, Offset 0x10 Voltage = Data* 0.008
> CPU2 Voltage, Offset 0x11 Voltage = Data* 0.008
> CPU1 DIMM Voltage, Offset 0x12 Voltage = Data* 0.008
> CPU2 DIMM Voltage, Offset 0x13 Voltage = Data* 0.008
> CPU1 DIMM VTT Voltage, Offset 0x14 Voltage = Data* 0.008
> CPU2 DIMM VTT Voltage, Offset 0x15 Voltage = Data* 0.008
> +1.1V Voltage, Offset 0x16 Voltage = Data* 0.008
> +1.8V Voltage, Offset 0x17 Voltage = Data* 0.008
> +5V Voltage, Offset 0x18 Voltage = Data* 0.008 * (40/10)
> +12V Voltage, Offset 0x19 Voltage = Data* 0.008/ (10./66.2)
> -12V Voltage, Offset 0x1a Voltage =
> ((Data*0.008-2.048)/(10./242.))+2.048
> HT Voltage, Offset 0x1b Voltage = Data* 0.008
> +3.3V Voltage, Offset 0x1c Voltage = Data* 0.024
> +3.3VSB Voltage, Offset 0x1d Voltage = Data* 0.024
> VBAT Voltage, Offset 0x1e Voltage = Data* 0.024
> CPU1 Temperature, Offset 0x26 Temperature = Data
> CPU2 Temperature, Offset 0x27 Temperature = Data
> System Temperature, Offset 0x1f Temperature = Data
> Air Temperature, Offset 0x20 Temperature = Data
> VRD1 Temperature, Offset 0x21 Temperature = Data
> VRD2 Temperature, Offset 0x22 Temperature = Data
> NB Temperature, Offset 0x23 Temperature = Data
> Chassis Intrusion, Bank 0, Offset 0x46, BitMask 0x40 1 = Bad, 0 = Good
> (Clear Bit: Bank 0, Offset 0x4d, BitMask 0x80)
>
That information is not very helpful by itself. What you would need is to load
the i2c-i801 driver to connect to the SMBus (i2 bus), and the w83795 driver
to access the W83795AG chip.
>
> Power Supply Failure, GP31(From W83527HG) 1 = Good, 0 = Bad
> --------------------------------------------------------------------------
>
> I found mention of this on the List Archive [2].
>
> I also found the configuration settings for a SuperMicro H8DC8 with a
> WinBond W83627HF. [3]
>
> I am unclear about using the WinBond vs IPMI - are they the same (I
> think not)? If they're not, should I be using WinBond with this
> motherboard, or stick with IPMI? If I should stick with IPMI, there is
> no lm_sensor support for the BMC chipset, so what are my options? I
> would rather not have to abandon lm_sensors, for any number of reasons
> - not least of which is I would then have to roll my own sysstat,
> removing all lm_sensor deps.
>
You should stick with IPMI. The reason that you can not see the w83795 in the
output of the 'sensors' command is that the memory space for accessing the i2c
controller (i2c-i801 driver) is reserved by ACPI. You can override that,
and connect to the controller anyway, but then you risk the above mentioned
crashes. You would have to disable the BMC (if that is possible) if you want
to access the chip through lm-sensors.
Guenter
_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
next prev parent reply other threads:[~2015-01-02 8:53 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-01-01 23:07 [lm-sensors] SuperMicro H8DG6-F-O IPMI BMC and lm_sensors G. Richard Bellamy
2015-01-02 8:53 ` Guenter Roeck [this message]
2015-01-03 22:30 ` G. Richard Bellamy
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20150102085356.GA29009@roeck-us.net \
--to=linux@roeck-us.net \
--cc=lm-sensors@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.