The doorbell writes may be seen out of order by the firmware if they
are in WC memory since the tx spin(un)lock does not flush WC writes.
Hence if the "stop" is written on a different CPU than the "go", it
is possible that the stop will arrive after the go unless we add an
explicit memory barrier (and mmiowb() is not enough).

It fixes transmit hangs in multi tx queue mode.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Use mmiowb() to ensure "stop" and "go" commands are sent in order on ia64.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Disable NAPI if a failure occurs when setting up the interface. Leaving
it enabled could cause a BUG the next time an ifconfig up is issued.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

We cannot select INTEL_IOATDMA in Kconfig as soon as MYRI10GE or IXGBE
is enabled since the former is not available on all architectures.

Just use a Kconfig bool {IXGBE,MYRI10GE}_DCA set to =y when DCA
support can actually be built.

[myri10ge portion written and signed-off-by] Brice Goglin <brice@myri.com>

Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Update myri10ge driver version number to 1.4.3-1.369.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

No need to modulus the queue number in ->hard_start_xmit() since the
core is going to do that for you modulus ->real_num_tx_queues.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add multiqueue TX support to myri10ge.

[ Removed reference to deprecated CONFIG_NETDEVICES_MULTIQUEUE and
  NETIF_F_MULTI_QUEUE ]

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Use DCA in myri10ge when CONFIG_DCA_MODULE is set as well.

And thus force INTEL_IOATDMA to =y so that DCA=y if we are =y.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Rename the cryptic "dca_capable" to "dca_capable_firmware"
and "dca_enabled" to "dca_device_present" in the firmware
counters.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Stop scaring people with what looks like a fatal message when DCA support
is compiled into their kernel, but the DCA device is not present.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>