On Wed, Nov 05, 2014 at 11:42:26PM +0100, Richard Cochran wrote: > Here is my current series. The first patch is the improved Tx time > stamp polling, and patches 2-5 are trivial clean ups.
Those look good to me. > Patches 6-10 represent my second attempt at the "jbod" BC. With any > luck, this should work. Compile tested only! In a test with two ports where the slave/master states are swapped, it seems to be working as expected now. I noticed another problem, however. When the two ports are both in a non-slave state, phc2sys doesn't keep them close together. Should phc2sys pick the synchronization source randomly? > (We still would need to change the max_adj for the servo, and maybe > there are a few other things I have forgotten.) Hm, yes. Maybe also the clock check should be disabled. The problem is that the clock struct represents a single physical clock and now it would be a collection of clocks pretending it's one clock. -- Miroslav Lichvar ------------------------------------------------------------------------------ _______________________________________________ Linuxptp-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/linuxptp-devel
