Quoting Jean Delvare <khali@xxxxxxxxxxxx>:
Hi Simon,
The raw VID value is at 0xfc, value is 0x3f, which means all VID pins
are high. Which means they are not wired to the CPU, otherwise at least
one of them would be low.
Now, looking at register 0x27, it appears that all VID input pins have
been configured for their alternative function (GPIO.) This is not the
default value for this register, which means that the motherboard
vendor decided to not use these pins for VID input but for another
function.
BTW... the pins in question are in input mode. So who knows... maybe
they ARE used for VID monitoring, and GPIO was preferred over true VID
for obscure reasons (incompatible voltage levels?) You may want to run,
as root:
isadump -f 0x800 16
WARNING! Running this program can cause system crashes, data loss and worse!
I will probe address range 0x800 to 0x80f.
Continue? [Y/n] y
0 1 2 3 4 5 6 7 8 9 a b c d e f
0800: ff 21 cf fa f5 07 ff ff ff ff ff ff ff ff ff ff
and see if the 3rd value would make sense as a VID value for your CPU.
Does it? CF hex = 207 dec... surely not 2.07V?!? According to Intel
the e6850 should have a VID of 1.3 or 1.35V.
Ask the manufacturer to confirm this, and if they do, blame it on them.
I can ask them in my IT job capacity - they may help... can you help
me with the gist of what I need to ask them? :)
One thing we could do is teach the it87 driver to not export the VID
value if any of VID pins 0-3 are configured for GPIO function. That
wouldn't be too difficult, but would you be able to test a kernel patch?
Happy to help - as long as you provide idiot-proof instructions LOL...
the closest I've ever got to kernel stuff I've ever done is loading
the kmod drivers, which isn't exactly difficult. Bear in mind I'm
running a xen kernel:
2.6.18-164.el5xen
I "believe" that in0 actually provides a VCore reading on this setup
anyway - it varies between 1.04V and 1.2V, which would seem to me to
be about right. Unless I'm off on the wrong track there too and that
needs a modifying calcluation for something else.
Thanks for your help Jean, much appreciated. It's not critical, but I
like having things finished and working well.
Simon
--
Simon Wilson
www.simonandkate.net
_______________________________________________
lm-sensors mailing list
lm-sensors@xxxxxxxxxxxxxx
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
