Pull spi updates from Mark Brown:
 "Quite a lot of activity in SPI this cycle, almost all of it in drivers
  with a few minor improvements and tweaks in the core.

   - Updates to pxa2xx to support Intel Broxton and multiple chip selects.
   - Support for big endian in the bcm63xx driver.
   - Multiple slave support for the mt8173
   - New driver for the auxiliary SPI controller in bcm2835 SoCs.
   - Support for Layerscale SoCs in the Freescale DSPI driver"

* tag 'spi-v4.4' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi: (87 commits)
  spi: pxa2xx: Rework self-initiated platform data creation for non-ACPI
  spi: pxa2xx: Add support for Intel Broxton
  spi: pxa2xx: Detect number of enabled Intel LPSS SPI chip select signals
  spi: pxa2xx: Add output control for multiple Intel LPSS chip selects
  spi: pxa2xx: Use LPSS prefix for defines that are Intel LPSS specific
  spi: Add DSPI support for layerscape family
  spi: ti-qspi: improve ->remove() callback
  spi/spi-xilinx: Fix race condition on last word read
  spi: Drop owner assignment from spi_drivers
  spi: Add THIS_MODULE to spi_driver in SPI core
  spi: Setup the master controller driver before setting the chipselect
  spi: dw: replace magic constant by DW_SPI_DR
  spi: mediatek: mt8173 spi multiple devices support
  spi: mediatek: handle controller_data in mtk_spi_setup
  spi: mediatek: remove mtk_spi_config
  spi: mediatek: Update document devicetree bindings to support multiple devices
  spi: fix kernel-doc warnings about missing return desc in spi.c
  spi: fix kernel-doc warnings about missing return desc in spi.h
  spi: pxa2xx: Align a few defines
  spi: pxa2xx: Save other reg_cs_ctrl bits when configuring chip select
  ...

Minor overlapping changes in net/ipv4/ipmr.c, in 'net' we were
fixing the "BH-ness" of the counter bumps whilst in 'net-next'
the functions were modified to take an explicit 'net' parameter.

Signed-off-by: David S. Miller <davem@davemloft.net>

In the packet timestamping function, check that the ptp version and
protocol of the packet matches what we have configured the hardware to
actually generate timestamps for, before looking/waiting for a timestamp.

Signed-off-by: Stefan Sørensen <stefan.sorensen@spectralink.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The list of rx timestamps are currently only pruned of old entries when a
new entry is inserted. If no new entries are added, old timestamps may
survive beyond their lifetime, possible causing them to be attached to
packets with the same sequence number after a rollover.

Signed-off-by: Stefan Sørensen <stefan.sorensen@spectralink.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently rx_timestamp_work reschedules itself as a regular workqueue item,
effectively causing it run constantly as long as there are packets left in
the queue. Fix by using delayed workqueue items, limiting it to run only
every two jiffies.

Signed-off-by: Stefan Sørensen <stefan.sorensen@spectralink.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Only using the message type and sequence id for matching timestamps
with packets is error prone, as multiple clients may very well be
sending packets with the same messagetype and timestamp at the same
time. Fix by extending the check to include the hash of bytes 20-29
(source id in PTPv2) that is provided with the timestamps.

Signed-off-by: Stefan Sørensen <stefan.sorensen@spectralink.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When probing devices-in-package for a c45 phy, device zero is the last
device to probe, however, if driver reads 0 from device zero,
c45_ids->devices_in_package is set to '0', the loop condition of probing
will be matched again, see codes below:

for (i = 1;i < num_ids && c45_ids->devices_in_package == 0;i++)

driver will run in a dead loop.

This patch restructures the bug and confusing loop, it provides a helper
function get_phy_c45_devs_in_pkg which to read devices-in-package registers
of a MMD, and rewrites the loop with using the helper function.

Signed-off-by: Shaohui Xie <Shaohui.Xie@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

An spi_driver does not need to set an owner, it will be populated by the
driver core.

Signed-off-by: Andrew F. Davis <afd@ti.com>
Acked-by: Jonathan Cameron <jic23@kernel.org>
Signed-off-by: Mark Brown <broonie@kernel.org>