Hi Gang!
Beautiful flow-chart on your troubleshooting page.
I was working on a support request, around Item 93, but then I says
to myself, after all this modprobe rmmod modprobe rmmod, maybe I
could reboot, just to see. So now this is a feel-good positive
feedback success story.
I was getting all the pretty modules loaded for a number of
"chips", incl. lm75, lm80 and others, but lots of sensors output
with all zero values, and with never a line about the lm78 itself.
Machine is an old graphics workstation, an Intergraph TD-425 (dual
pII) that I made it into an intranet server. It's a mid-90s model.
The BIOS manual said it's got a "hardware monitor" on (default) I/O
port 0x110, and I should see the on-line Help of the Win NT monitor
app for details. I got the thing with no software whatsoever, so
that's no Help.
I had stuffed the box full of IBM SCSI disks, including a drive
cage over the ram banks that wasn't there before, and I was a bit
surprised to see in the first place that those PIIs have just big
passive heat sinks, so I thought, if there's a thermometer in
there, it wouldn't be a bad idea to read it.
It's running Debian testing/unstable and Linux 2.4.22, and I got
the Debianized sources of i2c and lm_sensors-- that's 2.8.0.
Making, installing and even modprobing went fine with Debian's make-
kpkg (after a make-kpkg clean modules_clean).
Oh: To get sensors-detect to find anything, I had to change 0x290
to 0x110 (based on the Intergraph manual, which is still available
as a PDF on their website). (That was easier than changing the
LM78's I/O addresss in the BIOS, because the box is in a closet
with no monitor or keyboard on it. One question I was about to ask
you guys was whether I was likely to have better luck if I did
change that in the BIOS to 0x290.) Since 0x110 made sensors-detect
happy, I went ahead and patched lm78.c to say
static unsigned int normal_isa[] = { 0x110, SENSORS_ISA_END };
too.
An isadump of 0x295 0x296 was all FFs, but 0x115 0x116 was a
kaleidoscope.
Anyway, since the reboot I have:
luggage:~# lsmod
Module Size Used by Not tainted
lm78 7320 0 (unused)
smbus-arp 4948 0 (unused)
ltc1710 2344 0 (unused)
eeprom 3636 0 (unused)
lm75 2908 0 (unused)
lm80 5504 0 (unused)
adm1021 5976 0 (unused)
i2c-proc 6868 0 [lm78 smbus-arp ltc1710 eeprom
lm75 lm80 adm1021]
i2c-isa 756 0 (unused)
i2c-core 15780 0 [lm78 smbus-arp ltc1710 eeprom
lm75 lm80 adm1021 i2c-proc i2c-isa]
af_packet 15016 0 (unused)
unix 17356 12 (autoclean)
luggage:~# sensors
lm78-isa-0110
Adapter: ISA adapter
Algorithm: ISA algorithm
VCore 1: +2.83 V (min = +2.52 V, max = +3.08 V)
VCore 2: +2.89 V (min = +2.52 V, max = +3.08 V)
+3.3V: +3.36 V (min = +2.97 V, max = +3.63 V)
+5V: +5.12 V (min = +4.50 V, max = +5.48 V)
+12V: +11.48 V (min = +10.79 V, max = +13.11 V)
-12V: -10.67 V (min = -13.18 V, max = -10.78 V) ALARM
-5V: -3.73 V (min = -5.48 V, max = -4.50 V) ALARM
fan1: 0 RPM (min = 3000 RPM, div = 2) ALARM
fan2: 0 RPM (min = 3000 RPM, div = 2) ALARM
fan3: 0 RPM (min = 3000 RPM, div = 2) ALARM
temp: +25.0?C (limit = +60?C, hysteresis = +50?C)
vid: +2.80 V
alarms: Chassis intrusion detection ALARM
and the voltages fluctuate, and the temperature goes up to 26? if I
block the air vents for a couple of minutes. (Did I mention that
the airflow in this box bears witness to an ample engineering
budget?)
So thanks very much! It's very cool, literally. Now I'm going to
fill in my "ignore" statements and go hunt up a funky front-end for
this info.
Oh, and maybe look into those alarms :-(. That would be the
"excess" disk drives pulling down the voltage, I imagine.
Tony
--
-- Tony Crawford
-- +49-3341-309999
--
