On Sun, 04 May 2008 17:27:22 +0200, achim wrote: > Now with the 9600BE in the sapphire board there are additional chips > > X2 5000BE > --------------------------------------------------- > 0 1 2 3 4 5 6 7 8 9 a b c d e f > 00: -- -- -- -- -- -- -- -- -- -- -- -- -- > 10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > 20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- 2e -- > 30: -- -- -- -- -- -- -- -- 38 -- -- -- -- -- -- -- > 40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- 4e -- > 50: -- -- 52 53 -- -- -- 57 -- -- -- -- -- -- -- -- > 60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- 6e -- > 70: 70 -- -- -- -- -- -- -- > --------------------------------------------------- > > 9600BE > ?--------------------------------------------------- > 0 1 2 3 4 5 6 7 8 9 a b c d e f > 00: -- -- -- -- -- -- -- -- -- -- -- -- -- > 10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- > 20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- 2e -- > 30: -- -- -- -- -- -- -- -- 38 -- -- -- -- -- -- -- > 40: -- -- -- -- -- -- -- 47 -- -- -- -- 4c -- 4e -- > 50: -- -- 52 53 -- -- -- 57 -- -- -- -- -- -- -- -- > 60: -- -- -- -- -- -- -- -- -- 69 -- -- -- -- 6e -- > 70: 70 -- -- -- -- -- -- -- > ?--------------------------------------------------- > > Now the clock chips shows up at 0x69 maybe a second run with the X2 > would also have shown it. Strange... This could be related to the chip at 0x70 if it's an I2C mux as I suspect. Another possibility would be that the extra addresses are in the CPU itself, but at least for 0x69 we know it's not the case. > I spotted one thing. The output of 0x38 is identical on the M3A and the > Sapphire board and slightly different on the GBT board. > M3A and Sapphire use the same clock chip ?9LPRS477BKL the GBT > an ?9LPRS477CKL. Seems this 0x38 address also point to the cloch > generator. > > I tried all types of bios modifications but none showed up at the dumps > of 0x4e and 0x6e. > > Trying to dump 0x47 results in XX till i reboot. Never seen a chip at this address. sensors-detect does probe it, because the Maxim MAX6633 can use this address in theory. However if it is dangerous to probe this address, I think I'll just remove the probe. We're probing 0x40-0x47 only for the MAX6633 and it's a rather rare chip - not sure if we already saw a single user with it, so the risk is just not worth the benefit. > For the record I attached the dumps of 0x38 0x57 0x4c 0x69 0x6e and > 0x70 Doesn't look like anything I know. The chip at 0x70 doesn't look like a mux after all. > If I run sensors-detect and pass 0x2e,0x6e,0x47 as excludes, will it > pass? Yes, it should pass. > With the current patch applied the scanning will be completely skipped, > so i can't test the patch atm. Of course. The second patch I sent was broken anyway, I've resent it to the i2c list already but didn't Cc you, sorry. I'll send an updated patch to you now. -- Jean Delvare
