On Thu, Dec 01, 2016 at 03:41:20PM -0500, Vivien Didelot wrote: > Hi Andrew, > > Andrew Lunn <and...@lunn.ch> writes: > > >> diff --git a/drivers/net/dsa/mv88e6xxx/mv88e6xxx.h > >> b/drivers/net/dsa/mv88e6xxx/mv88e6xxx.h > >> index ab52c37..9e51405 100644 > >> --- a/drivers/net/dsa/mv88e6xxx/mv88e6xxx.h > >> +++ b/drivers/net/dsa/mv88e6xxx/mv88e6xxx.h > >> @@ -765,6 +765,9 @@ struct mv88e6xxx_ops { > >> int (*phy_write)(struct mv88e6xxx_chip *chip, int addr, int reg, > >> u16 val); > >> > >> + /* Switch Software Reset */ > >> + int (*reset)(struct mv88e6xxx_chip *chip); > >> + > > > > Hi Vivien > > > > In my huge patch series of 6390, i've been using a g1_ prefix for > > functionality which is in global 1, g2_ for global 2, etc. This has > > worked for everything so far with the exception of setting which > > reserved MAC addresses should be sent to the CPU. Most devices have it > > in g2, but 6390 has it in g1. > > > > Please could you add the prefix. > > I don't understand. It looks like you are talking about the second part > of the comment I made on your RFC patchset, about the Rsvd2CPU feature:
Hi Vivien I mean + /* Switch Software Reset */ + int (*g1_reset)(struct mv88e6xxx_chip *chip); + We have a collection of function pointers with port_ prefix, another collection with stats_, and a third with ppu_, etc. And then we have some which do not fit a specific category. Those i have prefixed with g1_ or g2_. I think we should have some prefix, and that is my suggestion. Andrew