* [lm-sensors] HP Compaq ProLiant DL380 detect failure
2006-11-25 22:34 [lm-sensors] HP Compaq ProLiant DL380 detect failure Thomas Brown
@ 2006-12-22 21:48 ` Jean Delvare
2006-12-22 22:11 ` Juerg Haefliger
` (6 subsequent siblings)
7 siblings, 0 replies; 9+ messages in thread
From: Jean Delvare @ 2006-12-22 21:48 UTC (permalink / raw)
To: lm-sensors
Hi Thomas,
On Sat, 25 Nov 2006 17:34:42 -0500, Thomas Brown wrote:
> Anyone have any luck getting lm_sensors to work with the HP/Compaq
> ProLiant DL series (i.e. DL380)? With the latest 'sensors-detect'
> script I get "Sorry, no sensors were detected."
>
> Here is some data for your perusal:
>
>
> dmesg output:
> piix4_smbus 0000:00:0f.0: Found 0000:00:0f.0 device
> piix4_smbus 0000:00:0f.0: SMB base address uninitialized - upgrade BIOS
> or use force_addr=0xaddr
This is a fatal error. If there are sensor chips connected to the
SMBus, you cannot reach them until the SMBus itself works. Try
upgrading the BIOS as suggested. Or try using force_addr as instructed,
just make sure you pick an unused I/O area (check in /proc/ioports.)
But it is also possible that the sensor chips are hidden elsewhere, and
the SMBus isn't used, and that's why it is disabled. I can't tell.
You'd need to ask the manufacturer for technical information.
> HP has a utility to read data from the Internal Lights Out Management board:
> # hplog -t
> ID TYPE LOCATION STATUS CURRENT THRESHOLD
> 1 ADM1022 Processor Zone Nominal 113F/ 45C 143F/ 62C
> 2 ADM1022 CPU (1) Nominal 113F/ 45C 156F/ 69C
> 3 ADM1022 I/O Zone Nominal 113F/ 45C 145F/ 63C
> 4 ADM1022 CPU (2) Nominal 111F/ 44C 156F/ 69C
> 5 ADM1022 Pwr. Supply Bay Nominal 100F/ 38C 129F/ 54C
The driver for the ADM1022 (thmc50) wasn't ported to Linux 2.6 yet,
only Linux 2.4 is supported. Not that it really matters until you can
actually reach these chips.
>
> # hplog -f
> ID TYPE LOCATION STATUS REDUNDANT FAN SPEED
> 1 Var. Speed Processor Zone Nominal Yes Low ( 10)
> 2 Var. Speed Processor Zone Nominal Yes Low ( 10)
> 3 Var. Speed I/O Zone Nominal Yes Low ( 20)
> 4 Var. Speed I/O Zone Nominal Yes Low ( 20)
> 5 Var. Speed Processor Zone Nominal Yes Low ( 10)
> 6 Var. Speed Processor Zone Nominal Yes Low ( 10)
> 7 Var. Speed Pwr. Supply Bay Nominal Yes Low ( 20)
> 8 Var. Speed Pwr. Supply Bay Nominal Yes Low ( 20)
>
> # hplog -p
> ID TYPE LOCATION STATUS REDUNDANT
> 1 Standard Pwr. Supply Bay Nominal No
> 2 Standard Pwr. Supply Bay Absent No
Do you have the source code of that tool? If you do, just check where
it gets the information from.
I also invite you to bug HP themselves about it. Ask them why
lm_sensors doesn't work with their hardware, see what they answer.
--
Jean Delvare
^ permalink raw reply [flat|nested] 9+ messages in thread* [lm-sensors] HP Compaq ProLiant DL380 detect failure
2006-11-25 22:34 [lm-sensors] HP Compaq ProLiant DL380 detect failure Thomas Brown
2006-12-22 21:48 ` Jean Delvare
@ 2006-12-22 22:11 ` Juerg Haefliger
2006-12-24 9:06 ` Jean Delvare
` (5 subsequent siblings)
7 siblings, 0 replies; 9+ messages in thread
From: Juerg Haefliger @ 2006-12-22 22:11 UTC (permalink / raw)
To: lm-sensors
Well, I happen to work for HP, different organization though. But
we're about to be using the DL385 out of ISS to integrate into our
product line. We'll be running Debian and we need to be able to read
temp and fan information but we'll rely on hp tools to get there (not
my choice). Anyways, we might have a business need to get the source
and I might be able to pass some information along. No promise though
since we're pretty uptight about those things (not my choice either).
Anyways, let me know what you want to know and I'll see what I can do...
...juerg
On 12/22/06, Jean Delvare <khali at linux-fr.org> wrote:
> Hi Thomas,
>
> On Sat, 25 Nov 2006 17:34:42 -0500, Thomas Brown wrote:
> > Anyone have any luck getting lm_sensors to work with the HP/Compaq
> > ProLiant DL series (i.e. DL380)? With the latest 'sensors-detect'
> > script I get "Sorry, no sensors were detected."
> >
> > Here is some data for your perusal:
> >
> >
> > dmesg output:
> > piix4_smbus 0000:00:0f.0: Found 0000:00:0f.0 device
> > piix4_smbus 0000:00:0f.0: SMB base address uninitialized - upgrade BIOS
> > or use force_addr=0xaddr
>
> This is a fatal error. If there are sensor chips connected to the
> SMBus, you cannot reach them until the SMBus itself works. Try
> upgrading the BIOS as suggested. Or try using force_addr as instructed,
> just make sure you pick an unused I/O area (check in /proc/ioports.)
>
> But it is also possible that the sensor chips are hidden elsewhere, and
> the SMBus isn't used, and that's why it is disabled. I can't tell.
> You'd need to ask the manufacturer for technical information.
>
> > HP has a utility to read data from the Internal Lights Out Management board:
> > # hplog -t
> > ID TYPE LOCATION STATUS CURRENT THRESHOLD
> > 1 ADM1022 Processor Zone Nominal 113F/ 45C 143F/ 62C
> > 2 ADM1022 CPU (1) Nominal 113F/ 45C 156F/ 69C
> > 3 ADM1022 I/O Zone Nominal 113F/ 45C 145F/ 63C
> > 4 ADM1022 CPU (2) Nominal 111F/ 44C 156F/ 69C
> > 5 ADM1022 Pwr. Supply Bay Nominal 100F/ 38C 129F/ 54C
>
> The driver for the ADM1022 (thmc50) wasn't ported to Linux 2.6 yet,
> only Linux 2.4 is supported. Not that it really matters until you can
> actually reach these chips.
>
> >
> > # hplog -f
> > ID TYPE LOCATION STATUS REDUNDANT FAN SPEED
> > 1 Var. Speed Processor Zone Nominal Yes Low ( 10)
> > 2 Var. Speed Processor Zone Nominal Yes Low ( 10)
> > 3 Var. Speed I/O Zone Nominal Yes Low ( 20)
> > 4 Var. Speed I/O Zone Nominal Yes Low ( 20)
> > 5 Var. Speed Processor Zone Nominal Yes Low ( 10)
> > 6 Var. Speed Processor Zone Nominal Yes Low ( 10)
> > 7 Var. Speed Pwr. Supply Bay Nominal Yes Low ( 20)
> > 8 Var. Speed Pwr. Supply Bay Nominal Yes Low ( 20)
> >
> > # hplog -p
> > ID TYPE LOCATION STATUS REDUNDANT
> > 1 Standard Pwr. Supply Bay Nominal No
> > 2 Standard Pwr. Supply Bay Absent No
>
> Do you have the source code of that tool? If you do, just check where
> it gets the information from.
>
> I also invite you to bug HP themselves about it. Ask them why
> lm_sensors doesn't work with their hardware, see what they answer.
>
> --
> Jean Delvare
>
> _______________________________________________
> lm-sensors mailing list
> lm-sensors at lm-sensors.org
> http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
>
^ permalink raw reply [flat|nested] 9+ messages in thread* [lm-sensors] HP Compaq ProLiant DL380 detect failure
2006-11-25 22:34 [lm-sensors] HP Compaq ProLiant DL380 detect failure Thomas Brown
2006-12-22 21:48 ` Jean Delvare
2006-12-22 22:11 ` Juerg Haefliger
@ 2006-12-24 9:06 ` Jean Delvare
2006-12-26 14:25 ` Juerg Haefliger
` (4 subsequent siblings)
7 siblings, 0 replies; 9+ messages in thread
From: Jean Delvare @ 2006-12-24 9:06 UTC (permalink / raw)
To: lm-sensors
Hi Juerg,
On Fri, 22 Dec 2006 14:11:40 -0800, Juerg Haefliger wrote:
> Well, I happen to work for HP, different organization though. But
> we're about to be using the DL385 out of ISS to integrate into our
> product line. We'll be running Debian and we need to be able to read
> temp and fan information but we'll rely on hp tools to get there (not
> my choice). Anyways, we might have a business need to get the source
> and I might be able to pass some information along. No promise though
> since we're pretty uptight about those things (not my choice either).
>
> Anyways, let me know what you want to know and I'll see what I can do...
Thanks for the proposal. Basically we want to know what hardware
monitoring chips are used in the system, and how they are wired. I
suspect that HP uses a non-standard I2C bus, possibly GPIO-driven.
Thomas, I've thought a bit more about your problem, and I'm now almost
certain that you won't find anything attached to your motherboard's
SMBus, even if you upgrade the BIOS or force the SMBus base address.
You said that the "hplog" utility was printing temperature values, so
it must get them from somewhere, and it definitely can't get them from
anything connected to the currently disabled SMBus. So there has to be
another, non-standard I2C bus somewhere on the board.
--
Jean Delvare
^ permalink raw reply [flat|nested] 9+ messages in thread* [lm-sensors] HP Compaq ProLiant DL380 detect failure
2006-11-25 22:34 [lm-sensors] HP Compaq ProLiant DL380 detect failure Thomas Brown
` (2 preceding siblings ...)
2006-12-24 9:06 ` Jean Delvare
@ 2006-12-26 14:25 ` Juerg Haefliger
2006-12-26 22:30 ` Thomas Brown
` (3 subsequent siblings)
7 siblings, 0 replies; 9+ messages in thread
From: Juerg Haefliger @ 2006-12-26 14:25 UTC (permalink / raw)
To: lm-sensors
Hi Jean,
> Thanks for the proposal. Basically we want to know what hardware
> monitoring chips are used in the system, and how they are wired. I
> suspect that HP uses a non-standard I2C bus, possibly GPIO-driven.
If I'm not totaly mistaken it's all controlled by the proprietary iLO
ASIC which shows up as a PCI device. I don't know the interface to
that beast and how temp and sensors are connected to it. I'll check
the schematics as soon as I have established connection to the HP
intranet. I'm currently slacking away in Switzerland but am supposed
to work remotely... :-)
Not a big deal until it starts snowing...
...juerg
> Thomas, I've thought a bit more about your problem, and I'm now almost
> certain that you won't find anything attached to your motherboard's
> SMBus, even if you upgrade the BIOS or force the SMBus base address.
> You said that the "hplog" utility was printing temperature values, so
> it must get them from somewhere, and it definitely can't get them from
> anything connected to the currently disabled SMBus. So there has to be
> another, non-standard I2C bus somewhere on the board.
>
> --
> Jean Delvare
>
^ permalink raw reply [flat|nested] 9+ messages in thread* [lm-sensors] HP Compaq ProLiant DL380 detect failure
2006-11-25 22:34 [lm-sensors] HP Compaq ProLiant DL380 detect failure Thomas Brown
` (3 preceding siblings ...)
2006-12-26 14:25 ` Juerg Haefliger
@ 2006-12-26 22:30 ` Thomas Brown
2007-01-09 13:28 ` Juerg Haefliger
` (2 subsequent siblings)
7 siblings, 0 replies; 9+ messages in thread
From: Thomas Brown @ 2006-12-26 22:30 UTC (permalink / raw)
To: lm-sensors
Thanks all--if we can get this working there is a large community of HP
ProLiant server users who will benefit.
Upon tracing the "hplog" process behavior, it seems to read the data is
by using message queues attached to the Proliant System Health Monitor
(hpasmd) process which, in turn, may also rely on data from the HP
Lights-Out Drivers and Agents (hprsm) daemon.
There are two routes that might be taken here:
1. Reverse-engineer the hpasmd process and read values directly from
the HP/Compaq hardware.
or
2. Write a shim to work in concert with an existing hpasmd process and
feed data back using existing drivers (i.e. /dev/cpqhealth/proc,
/dev/cpqhealth/cevt, etc.)
-Tom
^ permalink raw reply [flat|nested] 9+ messages in thread* [lm-sensors] HP Compaq ProLiant DL380 detect failure
2006-11-25 22:34 [lm-sensors] HP Compaq ProLiant DL380 detect failure Thomas Brown
` (4 preceding siblings ...)
2006-12-26 22:30 ` Thomas Brown
@ 2007-01-09 13:28 ` Juerg Haefliger
2007-01-14 11:37 ` Jean Delvare
2007-01-17 16:13 ` Juerg Haefliger
7 siblings, 0 replies; 9+ messages in thread
From: Juerg Haefliger @ 2007-01-09 13:28 UTC (permalink / raw)
To: lm-sensors
I contacted the iLO ASIC guys and they are checking the request for
providing technical information. I also checked the schematics of the
DL385G2. The iLO ASIC provides 4 I2C buses to which the temp sensors
(SMSC EMC1043) are attached to.
I already wrote the driver for the EMC1043 and have a framework in
place for the iLO driver. More later...
...juerg
On 12/26/06, Thomas Brown <twb0 at lymenet.org> wrote:
> Thanks all--if we can get this working there is a large community of HP
> ProLiant server users who will benefit.
>
> Upon tracing the "hplog" process behavior, it seems to read the data is
> by using message queues attached to the Proliant System Health Monitor
> (hpasmd) process which, in turn, may also rely on data from the HP
> Lights-Out Drivers and Agents (hprsm) daemon.
>
> There are two routes that might be taken here:
>
> 1. Reverse-engineer the hpasmd process and read values directly from
> the HP/Compaq hardware.
> or
> 2. Write a shim to work in concert with an existing hpasmd process and
> feed data back using existing drivers (i.e. /dev/cpqhealth/proc,
> /dev/cpqhealth/cevt, etc.)
>
> -Tom
>
^ permalink raw reply [flat|nested] 9+ messages in thread* [lm-sensors] HP Compaq ProLiant DL380 detect failure
2006-11-25 22:34 [lm-sensors] HP Compaq ProLiant DL380 detect failure Thomas Brown
` (5 preceding siblings ...)
2007-01-09 13:28 ` Juerg Haefliger
@ 2007-01-14 11:37 ` Jean Delvare
2007-01-17 16:13 ` Juerg Haefliger
7 siblings, 0 replies; 9+ messages in thread
From: Jean Delvare @ 2007-01-14 11:37 UTC (permalink / raw)
To: lm-sensors
Hi Juerg,
On Tue, 9 Jan 2007 14:28:25 +0100, Juerg Haefliger wrote:
> I contacted the iLO ASIC guys and they are checking the request for
> providing technical information. I also checked the schematics of the
> DL385G2. The iLO ASIC provides 4 I2C buses to which the temp sensors
> (SMSC EMC1043) are attached to.
> I already wrote the driver for the EMC1043 and have a framework in
> place for the iLO driver. More later...
Hey, it's great to hear you're making progress :) Please keep informed,
and good luck!
--
Jean Delvare
^ permalink raw reply [flat|nested] 9+ messages in thread* [lm-sensors] HP Compaq ProLiant DL380 detect failure
2006-11-25 22:34 [lm-sensors] HP Compaq ProLiant DL380 detect failure Thomas Brown
` (6 preceding siblings ...)
2007-01-14 11:37 ` Jean Delvare
@ 2007-01-17 16:13 ` Juerg Haefliger
7 siblings, 0 replies; 9+ messages in thread
From: Juerg Haefliger @ 2007-01-17 16:13 UTC (permalink / raw)
To: lm-sensors
Hi Jean,
> Hey, it's great to hear you're making progress :) Please keep informed,
> and good luck!
The iLO people came back to me and they are a little bit reluctant at
this point. They are concerned that we screw up the HW or interfere
with the BMC if we mock around with fans and I2C registers :-)
They suggested to use IPMI, so I started to look into it and it does
seem promising. I noticed that Yani has a patch to interface IPMI to
lm-sensors (see
http://lists.lm-sensors.org/pipermail/lm-sensors/2006-August/017071.html).
What is the state of this patch and is there any chance it'll make it
into mainline? Anything I can help with, Yani?
I have a DL385 G2 at my hand that I can use and I'll play around with
the patch probably next week after returning from vacation.
...juerg
^ permalink raw reply [flat|nested] 9+ messages in thread