Re: mv88e6240 configuration broken for B850v3

On Mon, Dec 06, 2021 at 08:07:45PM +0000, Russell King (Oracle) wrote:

My conclusion from having read this thread is the CPU port is using PPU
polling, meaning that in mac_link_up():

        if ((!mv88e6xxx_phy_is_internal(ds, port) &&
             !mv88e6xxx_port_ppu_updates(chip, port)) ||
            mode == MLO_AN_FIXED) {

is false - because mv88e6xxx_port_ppu_updates() returns true, and
consequently we never undo this force-down.

We know that
1. A == mv88e6xxx_phy_is_internal(ds, port), B == mv88e6xxx_port_ppu_updates(chip, port), C == mode == MLO_AN_FIXED
2. (!A && !B) || C == false. This is due to the effect we observe: link is not forced up
2. C == false. This is due to the device tree.
3. !A && !B == false. This is due to statement (2), plus the rule that if X || Y == false and Y == false, then X must also be false.
4. We know that A is true, again due to device tree: port 4 < .num_internal_phys for MV88E6240 which is 5.
5. !A is false, due to 4.

So we have:

false && !B == false.

Therefore "!B" is "don't care". In other words we don't know whether
mv88e6xxx_port_ppu_updates() is true or not.