Question about address 0x69

Philip Edelbrock phil at netroedge.com
Thu Feb 20 00:08:59 CET 2003


Yes, I'm very familiar with that.  It's an I2C chip which is on an 
SMBus.  The two are somewhat compatible, but not completely.  The I2C 
chip in this case is locking up and taking over the bus and not letting 
go because it thinks the transaction is still pending (ironicly, the 
clock chip has no time out).  The only way to make it stop is to power 
it down and back up (complete system power down).  If you stick with 
using SMBus commands which are compatible with this chip, you should be 
OK.  But, using commands like a byte-read when the chip expects a 
word-read (this is a hypothetical example and may not hold true) could 
cause bad things like this to happen.


Phil


Maik Broemme wrote:

>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>Hi,
>
>i have a small question about smbus block read/write because i start to write a utility to set the FSB. I have an ASUS TUSL2-C with an ICS94201 I²C and a ASUS A7V Mainboard. For both Mainboards the Clock Chip Generator is at address 0x69, but if i use i2cdump 0 0x69 s to read data from an error occurs.
>
>Feb 19 23:50:48 homer kernel: i2c-dev.o: opened i2c-1
>Feb 19 23:50:48 homer kernel: i2c-dev.o: i2c-1 ioctl, cmd: 0x705, arg: bffff618.
>Feb 19 23:50:48 homer kernel: i2c-dev.o: i2c-1 ioctl, cmd: 0x706, arg: 69.
>Feb 19 23:50:53 homer kernel: i2c-dev.o: i2c-1 ioctl, cmd: 0x720, arg: bffff720.
>Feb 19 23:50:53 homer kernel: i2c-i801.o: Block (pre 1): CNT=14, CMD=00, ADD=d3, DAT0=71, BLKDAT=7f
>Feb 19 23:50:54 homer kernel: i2c-i801.o: SMBus Timeout!
>Feb 19 23:50:54 homer kernel: i2c-i801.o: Error: no response!
>Feb 19 23:50:54 homer kernel: i2c-i801.o: Block (post 1): CNT=14, CMD=00, ADD=d3, DAT0=71, BLKDAT=7f
>Feb 19 23:50:54 homer kernel: i2c-dev.o: Closed: i2c-1
>
>Thats the debug output from kernel log. It comes on both mainboards with different chipsets, does anyone knows this problem? The BUS is not busy or such things, also after coldstart of both systems it does not work. Don't know exactly if this is a bug in the i2c code. :)
>
>- --
>Best regards
>
>Maik Broemme
>UNIX/Linux Developer
>
>WOLK Kernel Developer - http://wolk.sourceforge.net/
>TuxBox Developer - http://tuxbox.berlios.de/ 
>-----BEGIN PGP SIGNATURE-----
>Version: GnuPG v1.0.7 (GNU/Linux)
>
>iD8DBQE+VAtUj7mZcU7rMfERArQfAKDCAl0AvUXFBcaQzHz4qAjN/y7qpACdFLTM
>07+fZhhyvEWSAXkBQYJwn2k=
>=Q7sN
>-----END PGP SIGNATURE-----
>
>  
>



More information about the lm-sensors mailing list