On Sun, 25 Sep 2005 20:08:59 +0200, Jean Delvare <khali at linux-fr.org> wrote: >Hi Grant, > >Sorry for the delay, That's okay... >> # isadump -y -f 0x290 16 >> 0 1 2 3 4 5 6 7 8 9 a b c d e f >> 00: 01 ff ff 00 00 00 1a 80 ff ff ff ff 00 00 00 00 > >Hm, that's odd. The device is supposed not to decode offsets 0-4 and 7. >I checked on my own LM78 and it actually doesn't (they all return >0xff). I suspect that you have another device mapped in the same I/O >region, maybe in such a way that there is no major conflict, but it >still makes the LM78 very hard to detect. Hmm, so an ISA NIC at 0x280 is probably a bad idea? Wasn't showing in /proc/ioports as the driver wasn't loaded. Sorry about that, me not think of it. Removing the offending NIC, we get: root at silly:~# isadump -y -f 0x290 16 0 1 2 3 4 5 6 7 8 9 a b c d e f 00: ff ff ff ff ff 40 01 ff ff ff ff ff ff ff ff ff root at silly:~# sensors-detect [...] Probing for `National Semiconductor LM78' Trying address 0x0290... Failed! Probing for `National Semiconductor LM78-J' Trying address 0x0290... Success! (confidence 6, driver `lm78') [...] Okay, since I eyeballed an LM78-J I think we now know where I goofed :o) Dunno how many people using NICs from a dozen years ago with 0x280 and irq 3, but it is a standard setting on older ISA NICs, only today tracking different issue on another box do I see NIC using I/O offsets 0..0x1f. Cheers, Grant.