On Tue, 20 Oct 2009 21:50:31 +1000, Adam Nielsen wrote: > > 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 > > > > and see if the 3rd value would make sense as a VID value for your CPU. > > Ah interesting. It doesn't seem to for me though: > > 0 1 2 3 4 5 6 7 8 9 a b c d e f > 0800: ff 01 cf fa f7 07 ff ff ff ff ff ff ff ff ff ff > > VRM v11 (in hwmon-vid) says values can't be > 0xb2, although I don't know how > many bits are wired up to the VID. > > So is 0x802 the GPIO input? I wonder what it's wired up to...? See my answer to Simon. 6 VID bits -> 0x0f -> 1.519 V -> definitely not the right voltage for your CPU. Dead end. > > 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? > > I can test an it87 patch if you need a guinea pig - especially if it'll > compile against 2.6.30 so I don't have to reboot ;-) I'll prepare a patch as my time and health permits. -- Jean Delvare http://khali.linux-fr.org/wishlist.html _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
