[lm-sensors] Re: adm1027 on 2.6.9-5.EL

[lm-sensors] Re: adm1027 on 2.6.9-5.EL

[Jean Delvare]

Hi Paul,

First of all, you really shouldn't send mails with a subject line which
reads "Hello". Unless you like the idea that it'll end up in my spam
folder, of course, which is exactly what happened to this one.

Second, please send your requests for help to the lm-sensors mailing list
rather than me directly. This increases your chances to actually get
help.

> 2.6.9-5.EL kernel what is the "Can't get alarm mask data" error? and also,
> I get an alarm on in4 but I already have that ignored in the sensors.conf.
> Some of the min and max values are not looking right, maybe this is part
> of the mark data but google it and found nothing to go by. More
> importantly I downloaded i2c and lm_sensors 2.9.2 but the doc inside
> states:
>
> "FOR 2.5/2.6 KERNELS, do not attempt to compile the modules in this
> package. Use the drivers already in the 2.5/2.6 kernel tree."
>
> so.. I am confused. No need to compile then? Centos comes with 2.8.7 and
> I thought some updates will be on 2.9.2

The above warning is about kernel modules only. There still is the
user-space library and tools which both 2.4 and 2.6 kernels need. You
can compile these with "make user" and install them in /usr/local by
default with "make user_install".

If you do install it on your own, I suggest that you uninstall the
package that comes with your distribution so as to avoid library version
conflicts.

> Also, I used to be able to see /proc/sys/dev/sensors with 2.4 kernels and
> now nothing.... I am not versed on the main differences of kernels or
> their changes.

In Linux 2.6, the same data is available, although in a slightly
different form, under /sys/bus/i2c. On the latest versions you
additionally have /sys/class/hwmon, which is more or less a set of
pointers to the former. In the long run, /sys/class/hwmon will become
the right place to look at.

> My conf stuff is below.
> (...)
> adm1027-i2c-0-2e
> Adapter: SMBus I801 adapter at 1480
> ERROR: Can't get alarm mask data!
> (...)
> ***********************************
> # cat sensors.conf
>
> chip "lm85-*" "adt7463-*"

First problem, your chip is an adm1027, so you should add "adm1027-*"
here, else your configuration file is useless. This explains why your
"ignore in4" did not work.

The alarm mask error can be safely ignored. The feature is simply not
implemented in the Linux 2.6 version of the lm85 driver. Starting with
lm_sensors 2.8.8, "sensors" conveniently hides this error message. In
the meanwhile, you may try to add an "ignore alarm_mask" statement in
the configuration file, but I wouldn't swear it'll work.

--
Jean Delvare

[lm-sensors] adm1027 on 2.6.9-5.EL

[Paul Aviles]

In a previous 2.4 kernel install I had stuff under /proc/sys/dev/sensors and 
I cannot see it in 2.6.9-5. I have nothing under/sys/bus/i2c or 
/sys/class/hwmon though I do get sensors to work as the report below.

When I compared to my 2.4 working machine i2c-proc which I "guess" is normal 
when reading the doc from i2c-2.8.8 as i2c-core is supposed to provide 
this??. What I am missing is the fan stuff it used to be in the procs dir.

What am I missing?

I used the sensors.init script to load the modules and this is the result of 
my lsmod.

i2c_dev                11329  0
eeprom                  8545  0
lm85                   19433  0
i2c_sensor              3521  2 eeprom,lm85
i2c_i801                7757  0
i2c_core               22081  5 i2c_dev,eeprom,lm85,i2c_sensor,i2c_i801

and sensors

adm1027-i2c-0-2e
Adapter: SMBus I801 adapter at 1480
+2.6V:     +2.591 V  (min =  +2.47 V, max =  +2.73 V)
CPU Volt:  +1.359 V  (min =  +1.25 V, max =  +1.52 V)
+3.3V:     +3.386 V  (min =  +3.13 V, max =  +3.47 V)
+5V:      +5.104 V  (min =  +4.74 V, max =  +5.26 V)
Fan2:      8083 RPM  (min = 4800 RPM)
Fan3:      6129 RPM  (min = 2200 RPM)
Fan1:      8095 RPM  (min = 4800 RPM)
CPU Temp: +40.00?C  (low  =    +5?C, high =   +68?C)
Loc Temp: +34.00?C  (low  =    +5?C, high =   +40?C)
Sys Temp: +36.00?C  (low  =    +5?C, high =   +50?C)
pwm2:       64
pwm3:       64
pwm1:       64
vid:      -3.425 V  (VRM Version 9.1)

Thanks guys, particularly to Jean.

Paul 

[lm-sensors] adm1027 on 2.6.9-5.EL

[Jean Delvare]

Hi Paul,

On 2005-11-08, Paul Aviles wrote:
> In a previous 2.4 kernel install I had stuff under /proc/sys/dev/sensors
> and I cannot see it in 2.6.9-5. I have nothing under /sys/bus/i2c or
> /sys/class/hwmon though I do get sensors to work as the report below.

Sorry but this just can't be true. In 2.6.9, hardware monitoring data is
found under /sys/bus/i2c (or more exactly $sysfs/bus/i2c where $sysfs is
your sysfs mount point). If "sensors" work, then /sys/bus/i2c must be
populated. Check again.

> When I compared to my 2.4 working machine i2c-proc which I "guess" is
> normal when reading the doc from i2c-2.8.8 as i2c-core is supposed to
> provide this??

Sorry, I can't make any sense of this sentence. Please rephrase.

> What I am missing is the fan stuff it used to be in the procs dir.
>
> What am I missing?

Do you refer to the fan control features? They should be under
/sys/bus/i2c/devices/0-002e, search for pwm* files.

> I used the sensors.init script to load the modules and this is the result
> of my lsmod.
>
> i2c_dev                11329  0
> eeprom                  8545  0
> lm85                   19433  0
> i2c_sensor              3521  2 eeprom,lm85
> i2c_i801                7757  0
> i2c_core               22081  5 i2c_dev,eeprom,lm85,i2c_sensor,i2c_i801

Looks OK to me. You usually don't even need i2c-dev, it's only used by
sensors-detect and for bus scanning/chip dumping.

> adm1027-i2c-0-2e
> Adapter: SMBus I801 adapter at 1480
> +2.6V:     +2.591 V  (min =  +2.47 V, max =  +2.73 V)

+2.6V isn't a usual VCC value, how did you come up with this label?

> vid:      -3.425 V  (VRM Version 9.1)

Wow. Not good. What value does /sys/bus/i2c/devices/0-002e/cpu0_vid
contain?

> Thanks guys, particularly to Jean.

You're welcome.

--
Jean Delvare

[lm-sensors] adm1027 on 2.6.9-5.EL

[Jean Delvare]

Hi again Paul,

> > vid:      -3.425 V  (VRM Version 9.1)
>
> Wow. Not good. What value does /sys/bus/i2c/devices/0-002e/cpu0_vid
> contain?

That and the missing +12V reading make me suspicious about the fact that
you actually have an ADM1027 chip. I suspect that you actually have an
ADT7463 chip, misdetected (which is plausible when I look at the driver
code in 2.6.9). Do you confirm this? If you do not know which chip you
do have, please provide the output of "i2cdump 0 0x2e b", this should
answer my question.

According to the datasheet, the ADT7463 can use VRM 10 (i.e. 6-bit VID
values), while other chips in the family couldn't (5-bit VID values).
This is done by connecting VID5 to the +12V input, and changing a
configuration bit in the chip. This would explain the lack of +12V
reading. This also explains the bogus VID reading, as the driver code
didn't expect the new configuration bits in the VID register.

We'll need to update the lm85 driver to support that case.

Thanks,
--
Jean Delvare

[lm-sensors] adm1027 on 2.6.9-5.EL

[Paul Aviles]

This is my output i get.

# i2cdump 0 0x2e b
WARNING! This program can confuse your I2C bus, cause data loss and worse!
I will probe file /dev/i2c-0, address 0x2e, mode byte
Continue? [Y/n] y
     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f    0123456789abcdef
00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
20: c7 74 c5 c4 00 29 23 25 a1 02 6e 03 a0 02 00 00    ?t??.)#%??n???..
30: 40 40 40 64 64 64 00 00 00 00 00 00 00 27 41 6a    @@@ddd.......'Aj
40: 05 80 01 d3 be d2 6b 82 b6 ca b6 ca 00 ff 05 44    ??????k?????..?D
50: 05 28 05 32 65 04 96 09 65 04 ff ff 02 42 02 74    ?(?2e???e?..?B?t
60: c7 74 e0 00 40 40 40 3c 5a 28 42 64 2e 44 40 00    ?t?.@@@<Z(Bd.D@.
70: 00 00 00 00 00 00 7e 40 06 00 00 55 00 00 00 02    ......~@?..U...?
80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
90: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
a0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
b0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
c0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
d0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
e0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
f0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
[root@andeya1 /]#



On Tue, 8 Nov 2005 16:11:06 +0100 (CET), Jean Delvare wrote
> Hi again Paul,
> 
> > > vid:      -3.425 V  (VRM Version 9.1)
> >
> > Wow. Not good. What value does /sys/bus/i2c/devices/0-002e/cpu0_vid
> > contain?
> 
> That and the missing +12V reading make me suspicious about the fact that
> you actually have an ADM1027 chip. I suspect that you actually have 
> an ADT7463 chip, misdetected (which is plausible when I look at the driver
> code in 2.6.9). Do you confirm this? If you do not know which chip 
> you do have, please provide the output of "i2cdump 0 0x2e b", this should
> answer my question.
> 
> According to the datasheet, the ADT7463 can use VRM 10 (i.e. 6-bit 
> VID values), while other chips in the family couldn't (5-bit VID 
> values). This is done by connecting VID5 to the +12V input, and 
> changing a configuration bit in the chip. This would explain the 
> lack of +12V reading. This also explains the bogus VID reading, as 
> the driver code didn't expect the new configuration bits in the VID 
register.
> 
> We'll need to update the lm85 driver to support that case.
> 
> Thanks,
> --
> Jean Delvare


--
Open WebMail Project (http://openwebmail.org)

[lm-sensors] adm1027 on 2.6.9-5.EL

[Paul Aviles]

On Tue, 8 Nov 2005 15:02:17 +0100 (CET), Jean Delvare wrote
> Hi Paul,
> 
> Sorry but this just can't be true. In 2.6.9, hardware monitoring 
> data is found under /sys/bus/i2c (or more exactly $sysfs/bus/i2c 
> where $sysfs is your sysfs mount point). If "sensors" work, then 
> /sys/bus/i2c must be populated. Check again.

OK, my bad, I was looking under /proc instead. Yes, there are there.
> 
> Do you refer to the fan control features? They should be under
> /sys/bus/i2c/devices/0-002e, search for pwm* files.

I used to run this to lower the speeds of the fans and I guess the format is 
now different. 

echo "0.2 32.0 98.0 1.0 .0.0" > /proc/sys/dev/sensors/`/bin/ls -
1 /proc/sys/dev/sensors |  grep -v eeprom | grep -v chips | tail -1 `/pwm2_cfg

Under the new location, the file I think is related is fanX_pwm. When I do a:
cat fan1_pwm
64

and I tried with echo "32" > fan1_pwm but it does not change the value even 
logged in as root.

When I try pwmconfig I get.

./pwmconfig: line 116: 0-002e/fan3_pwm_enable: Permission denied
  0-002e/fan1_input ... speed was 7952 now 7952
./pwmconfig: line 101: 0-002e/fan3_pwm_enable: Permission denied
    no correlation
  0-002e/fan2_input ... speed was 6150 now 6122
./pwmconfig: line 101: 0-002e/fan3_pwm_enable: Permission denied
    no correlation
  0-002e/fan3_input ... speed was 8095 now 8181
./pwmconfig: line 101: 0-002e/fan3_pwm_enable: Permission denied
    no correlation

> 
> > adm1027-i2c-0-2e
> > Adapter: SMBus I801 adapter at 1480
> > +2.6V:     +2.591 V  (min =  +2.47 V, max =  +2.73 V)
> 
> +2.6V isn't a usual VCC value, how did you come up with this label?
> 
> > vid:      -3.425 V  (VRM Version 9.1)
> 
> Wow. Not good. What value does /sys/bus/i2c/devices/0-002e/cpu0_vid
> contain?

cat /sys/bus/i2c/devices/0-002e/cpu0_vid
-3425

I don't remember where I got that value. I think what I did was to compare 
the values I got back from sensors and tried to compared with what the BIOS 
shows for the voltages so this may be completly off.  I need to recheck my 
stuff later today.


Thanks again.

Paul

[lm-sensors] adm1027 on 2.6.9-5.EL

[Jean Delvare]

Hi Paul,

On 2005-11-08, Paul Aviles wrote:
> # i2cdump 0 0x2e b
> (...)
>      0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f    0123456789abcdef
> 00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
> 10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
> 20: c7 74 c5 c4 00 29 23 25 a1 02 6e 03 a0 02 00 00    ?t??.)#%??n???..
> 30: 40 40 40 64 64 64 00 00 00 00 00 00 00 27 41 6a    @@@ddd.......'Aj
> 40: 05 80 01 d3 be d2 6b 82 b6 ca b6 ca 00 ff 05 44    ??????k?????..?D

This confirms that this is an ADT7463. The detection code is already
fixed in later versions of the driver. In the meantime you may want to
add the following line to /etc/modprobe.conf:

options lm85 force_adt7463=0,0x2e

One mystery remains. The VRM detection code said your CPU used VRM 9.1,
while the chip is configured in VRM 10 mode. What's your CPU? Please
paste /proc/cpuinfo. Maybe your CPU is too recent and was not yet
supported by our code in 2.6.9.

Thanks,
--
Jean Delvare

[lm-sensors] adm1027 on 2.6.9-5.EL

[Paul Aviles]

almost positive this is my issue not the cpu or board.

cat /proc/cpuinfo
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 15
model           : 4
model name      : Intel(R) Pentium(R) 4 CPU 3.00GHz
stepping        : 1
cpu MHz         : 2994.321
cache size      : 1024 KB
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 5
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov 
pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe pni monitor ds_cpl 
cid xtpr
bogomips        : 5914.62

Also, what about the pwm for the fans

On Tue, 8 Nov 2005 20:00:35 +0100 (CET), Jean Delvare wrote
> Hi Paul,
> 
> On 2005-11-08, Paul Aviles wrote:
> > # i2cdump 0 0x2e b
> > (...)
> >      0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f    0123456789abcdef
> > 00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
> > 10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00    ................
> > 20: c7 74 c5 c4 00 29 23 25 a1 02 6e 03 a0 02 00 00    ?t??.)#%??n???..
> > 30: 40 40 40 64 64 64 00 00 00 00 00 00 00 27 41 6a    @@@ddd.......'Aj
> > 40: 05 80 01 d3 be d2 6b 82 b6 ca b6 ca 00 ff 05 44    ??????k?????..?D
> 
> This confirms that this is an ADT7463. The detection code is already
> fixed in later versions of the driver. In the meantime you may want 
> to add the following line to /etc/modprobe.conf:
> 
> options lm85 force_adt7463=0,0x2e
> 
> One mystery remains. The VRM detection code said your CPU used VRM 
> 9.1, while the chip is configured in VRM 10 mode. What's your CPU? Please
> paste /proc/cpuinfo. Maybe your CPU is too recent and was not yet
> supported by our code in 2.6.9.
> 
> Thanks,
> --
> Jean Delvare