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.
