Hi Hans, On Fri, 06 Apr 2007 09:19:48 +0200, Hans de Goede wrote: > jk wrote: > > I have a p4pe board with the asb_100 sensor running under linux 2.6.20. > > > > I now notice that the sensors location varies between > > /sys/bus/i2c/drivers/asb100/0-002d and /sys/bus/i2c/drivers/asb100/2-002d > > between reboots. > > > > This creates problems for me since I use sensord and a crontab entry to read the > > sensors but now the location of the sensors in /sys varies with each boot. > > > > Is this a bug? > > If not is there a way to fix the location of the sensor directory in the /sys > > heirarch? (Otherwise, I will need to write some klugey shell script to try to > > find the location at boot-up and then automatically change the crontab entry > > accordingly. > > Interesting, I've been thinking about this for while, as I foresee problems > here in relation to the DMI based motherboard config project we are working on too. > > I think that the currently used scheme where busses are purely numbered instead > of named needs fixing. > > Here is what I have on my system: > /sys/bus/i2c/devices/0-0050 > /sys/bus/i2c/devices/0-0051 > > And here is what I would like to have: > /sys/bus/i2c/devices/viapro-0-0050 > /sys/bus/i2c/devices/viapro-0-0051 > > The idea here is that the 0 added here is in case one can have multiple > instances of the same i2c master driver Which directly points to a weakness in your proposal: if you have two adapters of the same type, they'll get named foo-0 and foo-1. How do you know which is which? You don't, so you only moved the problem one step further, you didn't solve it. My understanding is that this is a problem for user-space to solve, it should not assume stable device numbering accross reboots. But it used to be the case and many user-space tools still rely on this. > So if I would also have an i2c driver for my ati radeon, then it would look like: > /sys/bus/i2c/devices/viapro-0-0050 > /sys/bus/i2c/devices/viapro-0-0051 > /sys/bus/i2c/devices/radeon-0-00xx > > This way the order in which the drivers get loaded doesn't matter. We do > ofcourse need to provide compat symlinks with the old names which will still be > driver loading order dependend. -- Jean Delvare
