On Tue, Feb 18, 2014 at 10:31:20PM +0000, Grant Likely wrote: > Adds a selftest function for the of_match_node function. of_match_node > is supposed to handle precedence for the compatible property as well as > the name and device_type values. This patch adds some test case data and > a function that makes sure each test node matches against the correct > entry of an of_device_id table. > > This code was written to verify the new of_match_node() implementation > that is an earlier part of this series. > > Currently all but one test passes. There is one scenario where the empty > "b/name2" node is getting matched against an entry without any > device_type property at all. It is unknown why this is, but it needs to > be solved before this patch can be committed. (However, this is testing > the new of_match_table implementation, which still does far better than > the old implementation which gets the precedence completely wrong.) We can drop the above paragraph now since all the test cases passed on the new version of of_match_node(). > > Signed-off-by: Grant Likely <grant.likely@xxxxxxxxxx> > Cc: Kevin Hau <haokexin@xxxxxxxxx> s/Hau/Hao/ Thanks, Kevin
Attachment:
pgpaFlObdMiG9.pgp
Description: PGP signature
