The is an issue where setting Relaxed Ordering (RO) bit
(in a PCI-E write transaction) on 82598 causing the chipset
to drop DCA hints.  This patch forces RO not to be set for
descriptors as well as payload.  This will only be in effect
while DCA is enabled and no performance difference was
noticed in testing.
Signed-off-by: NDon Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: NPeter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

As the assigned value is being overwritten shortly after, it can be
dropped and so the whole variable definition moved to the start of the
function.
Signed-off-by: NPhil Sutter <n0-1@freewrt.org>
Acked-by: NFlorian Fainelli <florian@openwrt.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This is copy and paste from the original driver. As skb_reserve() is
also called within korina_alloc_ring() when initially allocating the
receive descriptors, the same should be done when allocating new space
after passing an skb to upper layers.
Signed-off-by: NPhil Sutter <n0-1@freewrt.org>
Acked-by: NFlorian Fainelli <florian@openwrt.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

After the last loop iteration, i has the value RC32434_NUM_RDS and
therefore leads to an index overflow when used afterwards to address the
last element. This is yet another another bug introduced when rewriting
parts of the driver for upstream preparation, as the original driver
used 'RC32434_NUM_RDS - 1' instead.
Signed-off-by: NPhil Sutter <n0-1@freewrt.org>
Acked-by: NFlorian Fainelli <florian@openwrt.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The lro manager's frag_align_pad setting was missing,
leading to misaligned access to the skb passed up
to the stack.
Tested-by: NRick Jones <rick.jones2@hp.com>
Signed-off-by: NDivy Le Ray <divy@chelsio.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In rare cases when an underrun occur, all macb buffers where consumed
and the netif_queue was stopped infinitely. This happens then the TGO
(transfer ongoing) bit in the TSR is set (and UND). It seems like
clening up after the underrun makes the driver and the macb hardware
end up in an inconsistent state. The result of this is that in the
following calls to macb_tx no TX buffers are released -> the
netif_queue was stopped, and never woken up again.

The solution is to disable the transmitter, if TGO is set, before
clening up after the underrun, and re-enable the transmitter when the
cleaning up is done.
Signed-off-by: NRichard Röjfors <richard.rojfors@endian.se>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Updating the version and the year of updated files
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

To make sure no swapping are made by the compiler, changed HAS_WORK to inline
functions and added all the necessary barriers
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The Driver supports the 57711 and 57711E as well but the description was out of
date
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If the default PHY version (0x4321) is read - the PHY FW load failed
Signed-off-by: NYaniv Rosner <yanivr@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

10M/100M autoneg was not establishing link.
Signed-off-by: NYaniv Rosner <yanivr@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Setting loopback requires time to take effect
Signed-off-by: NYaniv Rosner <yanivr@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The HW should be configured so fast toggling between 1G and 10G will not be
missed. Make sure that the HW is re-configured in full
Signed-off-by: NYaniv Rosner <yanivr@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If the page size is > 8KB this violation happens
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Don't dump eeprom when bnx2x adapter is down.  Running ethtool -e causes an eeh
without it when the device is down
Signed-off-by: NPaul Larson <pl@linux.vnet.ibm.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Has a negative side effect when sending MAC update with no content (as done in
the self-test)
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The lock was release too soon. Make sure the HW is marked as locked until the
boot driver was unloaded from FW perspective
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Failures in the probe not handled correctly - separate the flow to handle

different failures
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Error check could result with not freeing the IRQ
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fixes a build error in absence of CONFIG_IPV6:

drivers/net/netxen/netxen_nic_main.c:1189: error: implicit declaration of function 'ipv6_hdr'
drivers/net/netxen/netxen_nic_main.c:1189: error: invalid type argument of '->'
Reported-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NDhananjay Phadke <dhananjay@netxen.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

For NX3031 only one I/O range is mapped, so unmapping other
two which are used by older chips, causes this warning on
ppc64.

"Attempt to iounmap early bolted mapping at 0x0000000000000000"
Signed-off-by: NDhananjay Phadke <dhananjay@netxen.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fix

    drivers/net/mlx4/profile.c: In function `mlx4_make_profile':
    drivers/net/mlx4/profile.c:110: warning: comparison of distinct pointer types lacks a cast

This happened because num_possible_cpus() was secretly changed by
commit ae7a47e7 ("cpumask: make cpumask.h eat its own dogfood.") from
returning "int" to (now) returning "unsigned int".  I think that was a
good change, so we should just swallow the fallout.

Cc: "David S. Miller" <davem@davemloft.net>
Cc: Roland Dreier <rolandd@cisco.com>
Cc: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Cc: Jack Morgenstein <jackm@dev.mellanox.co.il>
Cc: Vladimir Sokolovsky <vlad@mellanox.co.il>
Cc: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

b44 chip has some hardware limitations, that need GFP_DMA bounce
buffers in some situations.

In order to not deplete DMA zone, we should keep allocated GFP_DMA skb
only for driver use. At rx time, we copy such skb to newly allocated
skb, reusing existing copybreak infrastructure.

On machines with low amount of memory, all skb meet the hardware limitation,
so no copy is needed. We detect this situation using a new device flag, set
to one if one GFP_DMA skb was ever allocated by b44_alloc_rx_skb().

Previously allocated skb, even outside from DMA zone will then be recycled,
to have minimal impact on DMA zone use.
Signed-off-by: NEric Dumazet <dada1@cosmosbay.com>
Tested-by: NIonut Leonte <ionut.leonte@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch is for Alan Cox as it related to the tty layer.
Hopefully the hso driver is again relatively stable with this fix.
Signed-off-by: NDenis Joseph Barrow <D.Barow@option.com>
Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The HSO changes for kref introduced a recursive spinlock take. All
functions which call put_rxbuf_data already have serial->serial_lock
grabbed.

[Comment to code added-AC]
Signed-off-by: NDenis Joseph Barrow <D.Barrow@option.com>
Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

As the kernel warning states: "IRQF_DISABLED is not guaranteed on shared
IRQs". Since these IRQs' values are hardcoded and my test system doesn't
show any shared use of IRQs at all, rather make them non-shared than
non-disabled.
Signed-off-by: NPhil Sutter <n0-1@freewrt.org>
Acked-by: NFlorian Fainelli <florian@openwrt.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Apparently this doesn't make sense. Otherwise the queue gets disabled as
soon as it's getting empty and can only be resurrected by a driver
restart.
Signed-off-by: NPhil Sutter <n0-1@freewrt.org>
Acked-by: NFlorian Fainelli <florian@openwrt.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Originally this must have been a rewrite error when introducing
'chain_index'. But the original driver did not use the previous chain
item everywhere: when altering the address tx_chain_tail points to, it
should move forward, not backwards.
Also this is not an "index" but rather the penultimate element in the
chain, so rename it accordingly.
Signed-off-by: NPhil Sutter <n0-1@freewrt.org>
Acked-by: NFlorian Fainelli <florian@openwrt.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Triggering TX before the write to the DMA status mask register leads to
transferring packets with maximum payload no matter what the actual
packet size is.
While here, also trigger RX scheduling after writing the DMA status mask
register, like it was in the original driver before it was sent
upstream.
Signed-off-by: NPhil Sutter <n0-1@freewrt.org>
Acked-by: NFlorian Fainelli <florian@openwrt.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The called netif_rx_schedule() does all the work for us:
- it checks the return value of netif_rx_schedule_prep() and
- if everything is ok calls __netif_rx_schedule().
Before this change, the driver received absolutely nothing.
Signed-off-by: NPhil Sutter <n0-1@freewrt.org>
Acked-by: NFlorian Fainelli <florian@openwrt.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This function needs an early exit condition to function properly, or
else caller assumes napi workload wasn't enough to handle all received
packets and korina_rx is called again (and again and again and ...).
Signed-off-by: NPhil Sutter <n0-1@freewrt.org>
Acked-by: NFlorian Fainelli <florian@openwrt.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Without this the driver will crash when the NIC is being restarted.
Signed-off-by: NPhil Sutter <n0-1@freewrt.org>
Acked-by: NFlorian Fainelli <florian@openwrt.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The new value is the one used in the external patch before and allows at
least a standard MTU of 1500 to be handled correctly. Impact of this
change gets visible when bigger packets are to be received, issuing:
| ping -s 492 <IP>
and bigger payload sized led to 100% packet loss.
Signed-off-by: NPhil Sutter <n0-1@freewrt.org>
Acked-by: Florian Fainelli <florian@openwrt.org
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Using platform_set_drvdata() here makes no sense, since the driver_data
field has already been filled with valuable data (i.e. the MAC address).
Also having driver_data point to the net_device is rather pointless
since struct korina_device contains an apropriate field for it.
Signed-off-by: NPhil Sutter <n0-1@freewrt.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The "read for interrupts" flag must be set before enabling slow-path
interrupts as well (and not just before fast-path interrupts)
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Too big packets could pass due to wrong filter size
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Wrong initialization of the multi-queue indirection table - it should
be using the function and not the port index
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Calculation bug due to missing brackets
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>