Hi Andrew,
Andrew Lunn <and...@lunn.ch> writes:
> This patchset add interrupt controller support to the MV88E6xxx. This
> allows access to the interrupts the internal PHY generate. These
> interrupts can then be associated to a PHY device in the device tree
> and used by the PHY lib, rather than polling.
>
> Since interrupt handling needs to make MDIO bus accesses, threaded
> interrupts are used. The phylib needs to request the PHY interrupt
> using the threaded IRQ API. This in term allows some simplification to
> the code, in that the phylib interrupt handler can directly call
> phy_change(), rather than use a work queue. The work queue is however
> retained for the phy_mac_interrupt() call, which can be called in hard
> interrupt context.
>
> Since RFC v1:
>
> Keep phy_mac_interrupt() callable in hard IRQ context.
>
> The fix to trigger the phy state machine transitions on interrupts has
> already been submitted, so is dropped from here.
>
> Added back shared interrupts support.
>
>
> Andrew Lunn (5):
> net: dsa: mv88e6xxx: Implement interrupt support.
> net: phy: Use threaded IRQ, to allow IRQ from sleeping devices
> net: phy: Threaded interrupts allow some simplification
> net: phy: Use phy name when requesting the interrupt
> arm: vf610: zii devel b: Add support for switch interrupts
>
> .../devicetree/bindings/net/dsa/marvell.txt | 21 +-
> arch/arm/boot/dts/vf610-zii-dev-rev-b.dts | 51 +++++
> drivers/net/dsa/mv88e6xxx/chip.c | 248
> ++++++++++++++++++++-
> drivers/net/dsa/mv88e6xxx/global2.c | 139 +++++++++++-
> drivers/net/dsa/mv88e6xxx/global2.h | 11 +
> drivers/net/dsa/mv88e6xxx/mv88e6xxx.h | 31 +++
> drivers/net/phy/phy.c | 52 +++--
> drivers/net/phy/phy_device.c | 2 +-
> include/linux/phy.h | 5 +-
> 9 files changed, 522 insertions(+), 38 deletions(-)
mv88e6xxx_g1_* should've been moved to global1.c, I'll move them later.
For what it's worth:
Reviewed-by: Vivien Didelot <vivien.dide...@savoirfairelinux.com>
Thanks,
Vivien
