On Wed, 6 Jun 2007 17:47:36 +0200, Ivo Manca wrote: > Yeah, it seems like this was the problem. > Strange though, since it did work in my 2.6.19 kernel. Depends on the order in which the modules are loaded, and a number of different things can change that. The good news is that my patch seems to work. Please try unloaded and reloading the sis5595 driver to make sure (without i2c-sis5595 nor i2c-sis630 loaded, of course.) The conflict between sis5595 and i2c-sis5595/i2c-sis630 is of the same nature as the conflict between via686a/vt8231 and i2c-viapro. It _was_ solved for the VIA drivers (in 2.6.10), but seemingly not for the SiS drivers. I'm surprised nobody ever reported it, I guess SiS hardware is way less popular than VIA hardware. It's not too difficult to fix, I could give it a try if you want. > Loading them in the way sensors-detect suggest (first i2c-sis5595, then > sis5595), gives me the following output: > > Jun 6 15:23:01 localhost kernel: i2c-dev: adapter [SMBus SIS5595 adapter at > 0438] registered as minor 0 > Jun 6 15:23:02 localhost kernel: i2c-adapter i2c-0: ioctl, cmd=0x705, > arg=0xbf886b54 > Jun 6 15:23:03 localhost kernel: i2c-adapter i2c-0: ioctl, cmd=0x5401, > arg=0xbfa9e818 > Jun 6 15:23:03 localhost kernel: i2c-adapter i2c-0: i2c ioctl, cmd: 0x5401, > arg: 0xbfa9e818 > Jun 6 15:23:03 localhost kernel: i2c-adapter i2c-0: ioctl, cmd=0x705, > arg=0x82ff3c0 > Jun 6 15:23:03 localhost kernel: i2c-adapter i2c-0: ioctl, cmd=0x703, > arg=0x09 > Jun 6 15:23:03 localhost kernel: i2c-adapter i2c-0: ioctl, cmd=0x720, > arg=0x82ff550 > Jun 6 15:23:03 localhost kernel: i2c-adapter i2c-0: SMBus busy (0400). > Resetting... > Jun 6 15:23:03 localhost kernel: i2c-adapter i2c-0: Failed! (400) > > Then it repeats: > Jun 6 15:23:07 localhost kernel: i2c-adapter i2c-0: ioctl, cmd=0x703, > arg=0x77 > Jun 6 15:23:07 localhost kernel: i2c-adapter i2c-0: ioctl, cmd=0x720, > arg=0x82ff550 > Jun 6 15:23:07 localhost kernel: i2c-adapter i2c-0: SMBus busy (0400). > Resetting... > Jun 6 15:23:08 localhost kernel: i2c-adapter i2c-0: Failed! (400) > > Where it changes the 0x77 from 0x09 to 0x77 > Bug or feature? You see these messages only because you have CONFIG_I2C_DEBUG_BUS=y. I'm a bit surprised that the reset is said to have failed, but other than that, these messages are caused by failed probes, which sensors-detect and i2c-detect do a lot by design, so you can just ignore them. -- Jean Delvare
