On Fri, Jan 21, 2011 at 09:23:12AM +0100, Wolfgang Denk wrote: [...] > > In uec_phy: > > drivers/qe/uec_phy.h:34:#define LPA_1000FULL 0x0400 > > drivers/qe/uec_phy.h:35:#define LPA_1000HALF 0x0200 > > > > In Linux: > > include/linux/mii.h:133:#define LPA_1000FULL 0x0800 /* Link > > partner 1000BASE-T full duplex */ > > include/linux/mii.h:134:#define LPA_1000HALF 0x0400 /* Link > > partner 1000BASE-T half duplex */ > > > > Do you know if this is a wrong defintion on uec_phy ? > > I'm afraid the fix will lead some mis-behavior for boards has uec_phy. > > These definitions were added by commit 91cdaa3a: > > commit 91cdaa3a9d7562b869d96774e9c9ddf142c0848d > Author: Anton Vorontsov <avoront...@ru.mvista.com> > Date: Mon Mar 24 20:46:24 2008 +0300 > > uec: add support for gbit mii status readings > > Signed-off-by: Anton Vorontsov <avoront...@ru.mvista.com> > > > Anton, can you please comment?
I think that was just a shift-by-one mistake, Linux code is correct. It seems that the only way to trigger this is to force half-duplex on a gigabit link (and u-boot will think that it's a full duplex link). I believe it's safe to fix the wrong definition. Thanks, -- Anton Vorontsov Email: cbouatmai...@gmail.com _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
