On Tue, Sep 19, 2017 at 2:35 PM, Rob Herring <[email protected]> wrote: > On Tue, Sep 19, 2017 at 1:11 PM, Alan Tull <[email protected]> wrote: >> On Mon, Sep 18, 2017 at 4:11 PM, Rob Herring <[email protected]> wrote: >>> On Mon, Sep 11, 2017 at 12:58:57PM -0700, Guenter Roeck wrote: >>>> On Mon, Sep 11, 2017 at 02:16:49PM -0500, Alan Tull wrote: >>>> > Add new device tree binding for max1619. >>>> > >>>> > Signed-off-by: Alan Tull <[email protected]> >>>> >>>> Technically that should already work, without explicit binding, >>>> or did the i2c core change lately ? >>> >>> There was some work in that direction IIRC. >> >> Yes, after Guenter replied I tried it without this patch and it >> worked. i2c drivers can have a 'detect' function that can check part >> id, etc. If detection succeeds, it will fill in the device name in >> i2c_board_info->type and the i2c-core will enumerate using that. > > Right, I expected it still worked, but am saying we don't want to rely > on that behavior and introduce new places relying on it. Otherwise, > the driver will match on say "rob,max1619" as well. The detect > function is the really old way to probe devices IIRC. The function to > look at is i2c_of_match_device. We want OF style match, not a match > with i2c_of_match_device_sysfs which strips the vendor prefix from the > compatible.
OK, well my patch is out there if someone wants to take it. Alan > > Rob
