Don't do the num_tx_queues based masking on calculating tx queue
index.

 1) num_tx_queues is not always power-of-2, because it also depends on
    the online cpu numbers. So the masking could be a performance bug
    on a 6 cpu system.
 2) queue_mapping will be limited by real_num_tx_queues=num_tx_queues
    in the generic netdev function set_cur_queue_map(). So the bound
    limiting here is not necessary.
Signed-off-by: NWu Fengguang <fengguang.wu@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Replace all DMA_32BIT_MASK macro with DMA_BIT_MASK(32)

Signed-off-by: Yang Hongyang<yanghy@cn.fujitsu.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Replace all DMA_64BIT_MASK macro with DMA_BIT_MASK(64)

Signed-off-by: Yang Hongyang<yanghy@cn.fujitsu.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

We were basicly ignoring ethtool users request for FC autoneg
and replying to queries with a "best guess".  This patch
enables the driver to store if we want to enable/disable
autoneg FC and do the correct behavior.
Signed-off-by: NDon Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When network interface is made active we were not handling the error
scenarios properly to clean up rx & tx resources which might result in
a driver panic.
Signed-off-by: NMallikarjuna R Chilakala <mallikarjuna.chilakala@intel.com>
Acked-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>

This patch resolves an issue with map single being used to map a buffer and
then unmap page being used to unmap it.  In addition it handles any error
conditions that may be detected using skb_dma_map.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Acked-by: NMallikarjuna R Chilakala <mallikarjuna.chilakala@intel.com>
Acked-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>

MSI-X allocation broke after the 82599 merge on systems with more than 8
CPU cores.  82598 drops back into MSI mode, which isn't sufficient to run
full, efficient 10G line rate.
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>

In the past flow control wasn't enabled by default under the
incorrect assumption that this opened up us to a denial of
service attack.  However since any switch that forwarded flow
control would be extremely msiconfigured and/or buggy, this
concern no longer out weighs the preformance gains from
having FC enabled.
Signed-off-by: NDon Skidmore <donald.c.skidmore@intel.com>
Acked-by: NMallikarjuna R Chilakala <mallikarjuna.chilakala@intel.com>
Acked-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>

The ixgbe driver had issues when DEBUG was defined because the hw_dbg macro
was incomplete.  This patch completes the code based off of the code that
already existed in the igb module.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Acked-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>

Convert ixgbe to use net_device_ops properly.
Rather than changing the select_queue function pointer
just check the flag.
Signed-off-by: NStephen Hemminger <shemminger@vyatta.com>
Acked-by: NPeter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The watchdog will schedule an interrupt to help make sure queues are
cleaned in the case when an interrupt is missed, most likely due to very
high load.  On 82599, there are extra interrupt registers to account for
the larger number of MSI-X vectors (64 total for 82599 vs. 18 total for
82598).  These must be taken into account when performing this operation in
the watchdog.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@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>

Cleanup a bit of whitespace, add some function header comments, and fix a
few comments around the driver.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: NPeter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Acked-by: NMallikarjuna R Chilakala <mallikarjuna.chilakala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The Tx DMA unit should be disabled when bringing the device down.  Also,
the KX4 device with 82599 supports WoL, so we should clear the Wake Up
Status (WUS) after a PCIe slot reset.
Signed-off-by: NPeter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Acked-by: NMallikarjuna R Chilakala <mallikarjuna.chilakala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

There are possible times that a driver may fail to completely initialize,
due to a buggy platform or a buggy kernel.  In those cases, we'd rather
fail gracefully instead of a panic.  Add a few safety checks to some
critical paths to try and prevent a panic in these corner-case situations.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@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>

This cleans up the following pieces of the Rx initialization path:

- Enable the ECC memory fault interrupt in OTHER causes.

- Fix an 82598 initialization of RDRXCTL when depending on RSS and VMDq to
be enabled.  We don't need these features enabled to safely set the MVMEN
bit to allow multiple SRRCTL register mappings into the RXDCTL registers.

- Fix the RSS initialization path to not stomp on DCB accidentally.  When
configuring the MRQC (multiple Rx queue contol) register, we want to make
sure we only OR in features as necessary, instead of full assignment.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@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>

The Tx accounting when cleaning during NAPI was not completely properly.
We should use the work_limit to determine when to finish cleaning, and
use the same to return the cleaned status.  The impact of running like this
causes the NAPI clean for this Tx to get stuck in a scheduling loop, and
can result in Tx not getting cleaned, ending with a Tx hang and device
reset.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@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>

Occasionally if the driver was loaded in a system that
didn't support MSI-X or MSI and was on a shared interrupt,
the driver would then panic in NAPI on the first shared
interrupt because we hadn't called napi_add yet.

Solution: call napi_add before calling request_irq
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@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>

The interrupt models using EITR have changed in 82599.  The way the register
is laid out, the change is transparent to some of the existing code.
However, some of it isn't.  This patch fixes all the cases where EITR
handling is different than 82598.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@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>

82599 mistakenly enabled drop on Rx queues in the packet buffer.  The
default mode should be store-and-forward from the FIFO.
Signed-off-by: NPeter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Acked-by: NMallikarjuna R Chilakala <mallikarjuna.chilakala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The rx_no_dma_resources counter reported by ethtool -S ethX is not
counting correctly.  In 82599, the queue mappings for the counters need
to be mapped properly, and accounted for properly.
Signed-off-by: NPeter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Acked-by: NMallikarjuna R Chilakala <mallikarjuna.chilakala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Multiple unicast address support appears to have been broken with the
change to support net_device_ops.  This a regression from 2.6.28 to 2.6.29.

I'm not 100% on whether ndo_set_multicast_list can be NULL after this
or not.  If ndo_set_rx_mode is set everything _should_ be using it.
Signed-off-by: NChris Leech <christopher.leech@intel.com>
Acked-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>

With the hardware-specific code in place, add all supported device id's,
along with base driver changes to enable 82599 devices.  The devices
being enabled are:

8086:10f7: 82599EB 10 Gigabit KX4 Network Connection
8086:10fb: 82599EB 10 Gigabit Network Connection

The device 8086:10fb is a fully-pluggable SFP+ NIC.
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>

it was pointed out on the list that ixgbe was failing when using 64kB pages
and large 16kB MTU.

since with a 64kB PAGE_SIZE MAX_SKB_FRAGS = 3, the way the driver was
configuring page usage was assuming 2kB is half a page, and was only
ever dmaing that much data to a half page.

(16kB - header size) / 2048 = 7 or 8 pages, which would far exceed 3

adjust the driver to account for these large pages, the hardware can
support DMA to up to 16kB for each descriptor.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The failure happens when an interrupt occurs and the driver is reading
EICR.  This read will cause a clear-by-read which leads to two TLP
being inserted in the PCIe retry buffer leading to an overflow of the
buffer and corruption of TLPs.

The solution is different depending where the reading of EICR takes place.

For ixgbe_msix_lsc() since we are in MSIX mode and know OCD is enabled a
clear-by-write is done instead of the normal clear-by-read.

For ixgbe_intr() 0xffffffff is written to EIMC before the read, masking the
interrupts.
Signed-off-by: NDon Skidmore <donald.c.skidmore@intel.com>
Acked-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>

Removing the unused PAGE_USE_COUNT() macro. Also removing the unused
pages variable at ixgbe_configure_rx() function.
Signed-off-by: NBreno Leitão <leitao@linux.vnet.ibm.com>
Signed-off-by: NJay Vosburgh <fubar@us.ibm.com>
Acked-by: NPeter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Now that our set_num_queues() routines for each feature are re-entrant, and
can be called at any point, they shouldn't zero out the feature's indices
or mask bits.  Subsequent calls into those routines for those features can
result in zero Rx and Tx queues being assigned, causing a panic later in
driver reinitialization.
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>

Tx Head writeback is causing multi-microsecond stalls on PCIe chipsets, due
to partial cacheline writebacks.  Removing this feature removes these
issues.
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>

The flow control handling is overly complicated and difficult to maintain.
This patch cleans up the flow control handling and makes it much more
explicit.  It also adds 1G flow control autonegotiation, for 1G copper
links, 1G KX links, and 1G fiber links.
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>

Currently ixgbe does not display the EEPROM version in ethtool -i, where
other drivers do.  The EEPROM version is located at offset 0x29.  This
patch adds support to display it.
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>

The current code has some flaws in it when performing autonegotiation,
especially on KX/KX4 links.  This patch updates the code to better handle
the autonegotiation states on link setup.  The patch also removes a redundant
link configuration call on driver load, and moves link configuration to
the ->open() path.
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>

The current code to determine the number of queues the device will want
on driver initialization is ugly and difficult to maintain.  It also
doesn't allow for easy expansion for future features or future hardware.
This patch refactors these routines, and make them easier to deal with.
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>

New year, new copyright date ranges.  Also bump the driver version
number to reflect many of the recent changes.
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>

Our current MSI-X allocation mechanism does not support new hardware
at all.  It also isn't getting the actual number of supported MSI-X vectors
from the device.

This patch allows the number of MSI-X vectors to be specific to a device,
plus it gets the number of MSI-X vectors available from PCIe configuration
space.
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>

Add the device ID for BX devices using the 82598 MAC.
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>

Base versions handle constant folding now.
Signed-off-by: NHarvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

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

And support for the KX/KX4 mezzanine card.  Device id 0x10B6.
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>

Load times for NICs that use i2c to communicate with the phy were taking
~4.5 sec per port.  This fix first checks to see if the link is already
up before calling get_link_capabilities, since if it is we don't need
query the phy for link state.
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>

This patch makes ixgbe invoke the GRO hooks instead of LRO.  As
GRO has a compatible external interface to LRO this is a very
straightforward replacement.

As GRO uses the napi structure to track the held packets, I've
modified the code paths involved to pass that along.
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Following the removal of the unused struct net_device * parameter from
the NAPI functions named *netif_rx_* in commit 908a7a16, they are
exactly equivalent to the corresponding *napi_* functions and are
therefore redundant.
Signed-off-by: NBen Hutchings <bhutchings@solarflare.com>
Acked-by: NNeil Horman <nhorman@tuxdriver.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>