On Wed, Jul 20, 2016 at 10:24:11PM -0300, Daniel Bolgheroni wrote: > On Wed, Jul 13, 2016 at 11:39:54AM +0200, Mark Kettenis wrote: > > The diff below removes the last board ID dependent code for the i.MX6 > > platform. It affects all boards that use a Micrel PHY, which would > > show up as "ukphy0" in your dmesg. If you have such a board (udoo, > > novena and sabrelight, there may be others), I'd apprciate if you > > could test this diff and check whether the network interface still > > works. > > > > I'll probably commit the diff in a couple of days if I don't get any > > test reports. > > Maybe not related to this commit, since Wandboard uses Atheros PHYs (AR8031, > same family as AR8035 as in if_fec.c) instead of Micrel, but I'm running the > latest snapshot and fec(4) does not show up.
That turns out to be different interrupt data in the fdt to implement a workaround for hardware errata. Patrick found the following two commits related to it and last I heard was looking to find a way to handle it on his sabre lite board that also has it. https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=9fc77821b17155c6e0ab50b1e1dd80c2b0e63e98 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=6261c4c8f13eb91f733e8ba6d67c409a2e841667 The following doesn't do the workaround but should at least get things working again. Index: if_fec.c =================================================================== RCS file: /cvs/src/sys/arch/armv7/imx/if_fec.c,v retrieving revision 1.8 diff -u -p -r1.8 if_fec.c --- if_fec.c 21 Jul 2016 02:32:23 -0000 1.8 +++ if_fec.c 21 Jul 2016 02:34:02 -0000 @@ -303,7 +303,7 @@ fec_attach(struct device *parent, struct if (OF_getpropintarray(faa->fa_node, "interrupts-extended", intr, sizeof(intr)) < sizeof(intr)) - return; + intr[2] = 0x76; sc->sc_node = faa->fa_node; sc->sc_iot = faa->fa_iot;