There are multiple places in our device nvm where the version is stored.
This update fixes some output errors with some types of images and
refactors the way the version data is gathered and stored for ethtool output.
Signed-off-by: NCarolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: NAaron Brown <aaron.f.brown@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Newer devices supported by igb can support auto-crossover detection in
forced operation modes. Enable this in the driver, rather than clobbering
this functionality in forced operation.
Signed-off-by: NMatthew Vick <matthew.vick@intel.com>
Tested-by: NAaron Brown <aaron.f.brown@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Ignoring the return value from a call to the kernel dma_map API functions
can cause data corruption and system instability.  Check the return value
and take appropriate action.
Signed-off-by: NGreg Rose <gregory.v.rose@intel.com>
Tested-by: NSibai Li <sibai.li@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The driver should not forward LLDP type frames.  Inspect the ether type and
do not send if it is an LLDP ethertype frame.
Signed-off-by: NGreg Rose <gregory.v.rose@intel.com>
Tested-by: NSibai Li <sibai.li@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Hw timestamping code caused performance regression in ixgbe driver when the
timestamping is not enabled. The culprit is IXGBE_READ_REG call in the Rx
path which is executed for every received skb. This call is not needed when
the timestamping is disabled or for non-ptp packets.

netperf results:

The ixgbe side of the connection was acting as a server, the netperf command
line on the other side was:
netperf -H 192.168.1.23 -T0,0 -t UDP_STREAM -l 20

The values below mean throughput as reported by netperf (local/remote), for
3 runs, with timestamping not enabled.

3.7.0-rc1+ with CONFIG_IXGBE_PTP off:
5373.83 / 3329.32
5721.88 / 3033.89
5653.42 / 3112.38

3.7.0-rc1+ with CONFIG_IXGBE_PTP on:
5233.64 / 1226.85
5448.67 / 1039.32
5421.36 / 1095.66

Patched 3.7.0-rc1+ with CONFIG_IXGBE_PTP on:
5594.72 / 2942.53
5428.95 / 3110.16
5343.56 / 3200.48
Reported-by: NJesper Brouer <jbrouer@redhat.com>
Signed-off-by: NJiri Benc <jbenc@redhat.com>
Signed-off-by: NAndy Gospodarek <gospo@redhat.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Adds/updates ASCII descriptor maps for 82598 and 82599 Tx/Rx descriptors.
Current descriptor maps were out of date for 82598 and incorrect for
82599.
Signed-off-by: NJosh Hay <joshua.a.hay@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This change makes it so that compare the total_rx_packets cleaned to budget
instead of decrementing budget.  The advantage to this approach is that budget
can now be const and we only end up modifying total_rx_packets instead of
modifying both it and budget.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

When setting a MAC filter for the VF the function should return a success
or failure code, not the index of the new filter.  It causes spurious NACK
returns to the VF driver.
Signed-off-by: NGreg Rose <gregory.v.rose@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Tested-by: NSibai Li <sibai.li@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch simplifies the check for calling en/disable_tx_laser() function
pointer. The pointer is only set on parts that can use it.
Signed-off-by: NEmil Tantilov <emil.s.tantilov@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

In SR-IOV mode the PF driver acts as the uplink port and is
used to send control packets e.g. lldpad, stp, etc.

   eth0.1     eth0.2     eth0
   VF         VF         PF
   |          |          |   <-- stand-in for uplink
   |          |          |
  --------------------------
  |  Embedded Switch       |
  --------------------------
              |
             MAC   <-- uplink

But the embedded switch is setup to forward multicast addresses
to all interfaces both VFs and PF and onto the physical link.
This results in reserved MAC addresses used by control protocols
to be forwarded over the switch onto the VF.

In the LLDP case the PF sends an LLDPDU and it is currently
being forwarded to all the VFs who then see the PF as a peer.
This is incorrect.

This patch adds the multicast addresses to the RAR table in the
hardware to prevent this behavior.
Signed-off-by: NJohn Fastabend <john.r.fastabend@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Tested-by: NSibai Li <sibai.li@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The function to set the macvlan filter should return success or failure
instead of the index of the filter.  The message processing function was
misinterpreting the index as a non-zero return code indicating failure and
NACKing MAC filter set messages that actually succeeded.
Signed-off-by: NGreg Rose <gregory.v.rose@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Calling the ixgbe_reset_pipeline_82599 function will ensure a full pipeline
reset on all 82599 devices.  This is necessary to avoid possible link issues.
Since this patch accomplishes this by modifying AUTOC.LMS we need to wrap
all AUTOC writes when LESM is enabled.

v2- fix LMS behaviour based on feedback by Martin Josefsson

CC: Martin Josefsson <gandalf@mjufs.se>
Signed-off-by: NDon Skidmore <donald.c.skidmore@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Signed-off-by: NGreg Rose <gregory.v.rose@intel.com>
Tested-by: NSibai Li <sibai.li@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The lockdep splat below identifies a case where irq safe to unsafe
lock order is detected. Resolved by making mbx_lock bh.

======================================================
[ INFO: SOFTIRQ-safe -> SOFTIRQ-unsafe lock order detected ]
3.6.0-rc5jk-net-next+ #119 Not tainted
------------------------------------------------------
ip/2608 [HC0[0]:SC0[2]:HE1:SE0] is trying to acquire:
 (&(&adapter->mbx_lock)->rlock){+.+...}, at: [<ffffffffa008114e>] ixgbevf_set_rx_mode+0x36/0xd2 [ixgbevf]

and this task is already holding:
 (_xmit_ETHER){+.....}, at: [<ffffffff814097c8>] dev_set_rx_mode+0x1e/0x33
which would create a new lock dependency:
 (_xmit_ETHER){+.....} -> (&(&adapter->mbx_lock)->rlock){+.+...}

but this new dependency connects a SOFTIRQ-irq-safe lock:
 (&(&mc->mca_lock)->rlock){+.-...}
... which became SOFTIRQ-irq-safe at:
  [<ffffffff81092ee5>] __lock_acquire+0x2f2/0xdf3
  [<ffffffff81093b11>] lock_acquire+0x12b/0x158
  [<ffffffff814bdbcd>] _raw_spin_lock_bh+0x4a/0x7d
  [<ffffffffa011a740>] mld_ifc_timer_expire+0x1b2/0x282 [ipv6]
  [<ffffffff81054580>] run_timer_softirq+0x2a2/0x3ee
  [<ffffffff8104cc42>] __do_softirq+0x161/0x2b9
  [<ffffffff814c6a7c>] call_softirq+0x1c/0x30
  [<ffffffff81011bc7>] do_softirq+0x4b/0xa3
  [<ffffffff8104c8d5>] irq_exit+0x53/0xd7
  [<ffffffff814c734d>] do_IRQ+0x9d/0xb4
  [<ffffffff814be56f>] ret_from_intr+0x0/0x1a
  [<ffffffff813de21c>] cpuidle_enter+0x12/0x14
  [<ffffffff813de235>] cpuidle_enter_state+0x17/0x3f
  [<ffffffff813deb6c>] cpuidle_idle_call+0x140/0x21c
  [<ffffffff8101764c>] cpu_idle+0x79/0xcd
  [<ffffffff814a59f5>] rest_init+0x149/0x150
  [<ffffffff81ca2cbc>] start_kernel+0x37c/0x389
  [<ffffffff81ca22dd>] x86_64_start_reservations+0xb8/0xbd
  [<ffffffff81ca23e3>] x86_64_start_kernel+0x101/0x110

to a SOFTIRQ-irq-unsafe lock:
 (&(&adapter->mbx_lock)->rlock){+.+...}
... which became SOFTIRQ-irq-unsafe at:
...  [<ffffffff81092f59>] __lock_acquire+0x366/0xdf3
  [<ffffffff81093b11>] lock_acquire+0x12b/0x158
  [<ffffffff814bd862>] _raw_spin_lock+0x45/0x7a
  [<ffffffffa0080fde>] ixgbevf_negotiate_api+0x3d/0x6d [ixgbevf]
  [<ffffffffa008404b>] ixgbevf_open+0x6c/0x43e [ixgbevf]
  [<ffffffff8140b2c1>] __dev_open+0xa0/0xe6
  [<ffffffff814099b6>] __dev_change_flags+0xbe/0x142
  [<ffffffff8140b1eb>] dev_change_flags+0x21/0x57
  [<ffffffff8141a523>] do_setlink+0x2e2/0x7f4
  [<ffffffff8141ad8c>] rtnl_newlink+0x277/0x4bb
  [<ffffffff81419c08>] rtnetlink_rcv_msg+0x236/0x253
  [<ffffffff8142f92d>] netlink_rcv_skb+0x43/0x94
  [<ffffffff814199cb>] rtnetlink_rcv+0x26/0x2d
  [<ffffffff8142f6dc>] netlink_unicast+0xee/0x174
  [<ffffffff8142ff12>] netlink_sendmsg+0x26a/0x288
  [<ffffffff813f5a0d>] __sock_sendmsg_nosec+0x58/0x61
  [<ffffffff813f7d57>] __sock_sendmsg+0x3d/0x48
  [<ffffffff813f7ed9>] sock_sendmsg+0x6e/0x87
  [<ffffffff813f93d4>] __sys_sendmsg+0x206/0x288
  [<ffffffff813f95ce>] sys_sendmsg+0x42/0x60
  [<ffffffff814c57a9>] system_call_fastpath+0x16/0x1b

other info that might help us debug this:

Chain exists of:
  &(&mc->mca_lock)->rlock --> _xmit_ETHER --> &(&adapter->mbx_lock)->rlock

 Possible interrupt unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(&(&adapter->mbx_lock)->rlock);
                               local_irq_disable();
                               lock(&(&mc->mca_lock)->rlock);
                               lock(_xmit_ETHER);
  <Interrupt>
    lock(&(&mc->mca_lock)->rlock);

 *** DEADLOCK ***
Signed-off-by: NJohn Fastabend <john.r.fastabend@intel.com>
Acked-by: NGreg Rose <gregory.v.rose@intel.com>
Tested-by: NSibai Li <sibai.li@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Ignoring the return value from a call to the kernel dma_map API functions
can cause data corruption and system instability.  Check the return value
and take appropriate action.
Signed-off-by: NGreg Rose <gregory.v.rose@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Tested-by: NSibai Li <sibai.li@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

ixgbevf_alloc_q_vectors() calls netif_napi_add for each qvector
where qvectors is determined by the number of msix vectors. This
makes perfect sense.

However on cleanup when ixgbevf_free_q_vectors() is called and
for each qvector we should call netif_napi_del there is some
extra logic to add a dependency on RX queues. This patch makes
the add/del operations symmetric by removing the RX queues
dependency.

Without this if  free_netdev() is called we see the general
protection fault below in netif_napi_del when list_del_init()
is called.

# addr2line -e ./vmlinux ffffffff8140810c
net-next/include/linux/list.h:88

general protection fault: 0000 [#1] SMP
Modules linked in: bonding ixgbevf ixgbe(-) mdio libfc scsi_transport_fc scsi_tgt 8021q garp stp llc cpufreq_ondemand acpi_cpufreq freq_table mperf ipv6 uinput coretemp lpc_ich i2c_i801 shpchp hwmon i2c_core serio_raw crc32c_intel mfd_core joydev pcspkr microcode ioatdma igb dca pata_acpi ata_generic usb_storage pata_jmicron [last unloaded: bonding]
CPU 10
Pid: 4174, comm: rmmod Tainted: G        W    3.6.0-rc3jk-net-next+ #104 Supermicro X8DTN/X8DTN
RIP: 0010:[<ffffffff8140810c>]  [<ffffffff8140810c>] netif_napi_del+0x24/0x87
RSP: 0018:ffff88027f5e9b48  EFLAGS: 00010293
RAX: ffff8806224b4768 RBX: ffff8806224b46e8 RCX: 6b6b6b6b6b6b6b6b
RDX: 6b6b6b6b6b6b6b6b RSI: ffffffff810bf6c5 RDI: ffff8806224b46e8
RBP: ffff88027f5e9b58 R08: ffff88033200b180 R09: ffff88027f5e98a8
R10: ffff88033320b000 R11: ffff88027f5e9ae8 R12: 6b6b6b6b6b6b6aeb
R13: ffff8806221d11c0 R14: 0000000000000000 R15: ffff88027f5e9cf8
FS:  00007f5e58b9b700(0000) GS:ffff880333200000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 00000000010ef2b8 CR3: 0000000281fff000 CR4: 00000000000007e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process rmmod (pid: 4174, threadinfo ffff88027f5e8000, task ffff88032f888000)
Stack:
 ffff8806221d1160 6b6b6b6b6b6b6aeb ffff88027f5e9b88 ffffffff81408e46
 ffff8806221d1160 ffff8806221d1160 ffff8806221d1ae0 ffff8806221d5668
 ffff88027f5e9bb8 ffffffffa009153c ffffffffa0092a30 ffff8806221d5700
Call Trace:
 [<ffffffff81408e46>] free_netdev+0x64/0xd7
 [<ffffffffa009153c>] ixgbevf_remove+0xa6/0xbc [ixgbevf]
 [<ffffffff8127a7a1>] pci_device_remove+0x2d/0x51
 [<ffffffff8131f503>] __device_release_driver+0x6c/0xc2
 [<ffffffff8131f640>] device_release_driver+0x25/0x32
 [<ffffffff8131e821>] bus_remove_device+0x148/0x15d
 [<ffffffff8131cb6b>] device_del+0x130/0x1a4
 [<ffffffff8131cc2a>] device_unregister+0x4b/0x57
 [<ffffffff81275c27>] pci_stop_bus_device+0x63/0x85
 [...]
Signed-off-by: NJohn Fastabend <john.r.fastabend@intel.com>
Acked-by: NGreg Rose <gregory.v.rose@intel.com>
Acked-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NSibai Li <sibai.li@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch updates the igb driver version to 4.0.17.
Signed-off-by: NCarolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: NAaron Brown <aaron.f.brown@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

There was a problem in the initial implementation of the get cable length
function for i210 and it did not work properly.  This patch fixes that
problem for i210/i211 devices.
Signed-off-by: NCarolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: NAaron Brown <aaron.f.brown@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This is a HW requirement. Although a buffer as short as 1 byte is allowed,
the total length of packet before, padding and CRC insertion, must be at
least 17 bytes.  So pad all small packets manually up to 17 bytes before
delivering them to HW.
Signed-off-by: NTushar Dave <tushar.n.dave@intel.com>
Tested-by: NJeff Pieper <jeffrey.e.pieper@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

We were not correctly freeing the temporary rings on error in
ixgbe_set_ring_param.  In order to correct this I am unwinding a number of
changes that were made in order to get things back to the original working
form with modification for the current ring layouts.

This approach has multiple advantages including a smaller memory footprint,
and the fact that the interface is stopped while we are allocating the rings
meaning that there is less potential for some sort of memory corruption on the
ring.

The only disadvantage I see with this approach is that on a Rx allocation
failure we will report an error and only update the Tx rings.  However the
adapter should be fully functional in this state and the likelihood of such
an error is very low.  In addition it is not unreasonable to expect the
user to need to recheck the ring configuration should they experience an
error setting the ring sizes.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch adds a function that forces a full pipeline reset.  This
function will be used in following patches to completely reset the PHY
during resets.
Signed-off-by: NDon Skidmore <donald.c.skidmore@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

We still had some code floating around from the old single buffer receive
path.  As a result we were adding VLAN_HLEN to max_frame although the
resultant value was never used.  Since that is the case we can drop this from
the function.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Driver pad skb up to 17 bytes because of the HW requirement. However, that code
implementation mess up the skb tail pointer after padding. This patch sets
skb->tail correctly.
Signed-off-by: NTushar Dave <tushar.n.dave@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Using is_zero_ether_addr() to simplify the code.
Signed-off-by: NWei Yongjun <yongjun_wei@trendmicro.com.cn>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch modifies when and where PTP registers and data are set. Previously
a work-around was used inside cyclecounter_start in order to reset some of the
time registers. This patch creates a new ixgbe_ptp_reset specifically for this
purpose. The cyclecounter configuration has trimmed down to only modify what
is necessary. Due to hardware conditions after probe and before open, PTP init
has now moved into the ixgbe_open call. This allows the ptp device name in the
sysfs to be the ethernet device name instead of the MAC address.

The cyclecounter check flag is renamed to PTP_ENABLED and is used to prevent
PTP init from happening when PTP has not been enabled.

CC: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: NJacob Keller <jacob.e.keller@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch adds a subdevice id for new 82599 device. The define is needed
to allow enabling WOL support.
Signed-off-by: NEmil Tantilov <emil.s.tantilov@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This change adds support for DCB and SR-IOV from the VF.  With this change
in place the VF will correctly use a traffic class other than 0 in the case
that the PF is configured with the default user priority belonging to a
traffic class other than 0.

Cc: Greg Rose <gregory.v.rose@intel.com>
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NSibai Li <sibai.li@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This change switches on the last few bits for us enabling version 1.1 VF
support in the PF.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Tested-by: NRobert Garrett <RobertX.Garrett@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch addresses several issues in regards to the combination of DCB
and SR-IOV. Specifically it allows us to send information to the VF on
which queues it should be using.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Tested-by: NRoss Brattain <ross.b.brattain@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

It is necessary to track the default user priority in the PF so that we can
force it upon the VFs.  The motivation behind this is to keep the VFs from
getting access to user priorities meant for things like storage.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This change adds support for IPv6 and UDP to ixgbe_get_headlen. The
advantage to this is that we can now handle ipv4/UDP, ipv6/TCP, and
ipv6/UDP with a single memcpy instead of having to do them in multiple
pskb_may_pull calls.

A quick bit of testing shows that we increase throughput for a single
session of netperf from 8800Mpbs to about 9300Mpbs in the case of ipv6/TCP.
As such overall ipv6 performance should improve with this change.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NStephen Ko  <stephen.s.ko@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This change makes it so that igb_update_dca is broken into two halves, one
for Rx and one for Tx.  The advantage to this is primarily readability.

In addition I am enabling relaxed ordering for reads from hardware since
this is supported on all of the igb parts.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NAaron Brown <aaron.f.brown@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This change helps to address locking issues seen with
netif_set_real_num_tx_queues and netif_set_real_num_rx_queues when used in
the igb_set_interrupt_capability function.  To resolve these locking issues
I have moved the two function calls into __igb_open so that they can be
called while the RTNL lock is held.

An added advantage to this is that the number of queues is not updated
until the last possible moment so if there are any issues in allocating
MSI-X interrupts or resources for the rings we have time to change the
values prior to updating the netdev.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NAaron Brown <aaron.f.brown@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This change combines the the allocation of q_vectors and rings into a single
function.  The advantage of this is that we are guaranteed we will avoid
overlap in the L1 cache sets.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NAaron Brown <aaron.f.brown@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This change locks us in at 2K buffers even on a system that supports larger
frames.  The reason for this change is to make better use of pages and to
reduce the overall truesize of frames generated by igb.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NAaron Brown <aaron.f.brown@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

In order to try and isolate things a bit further I am moving the code
related to retrieving data from the rx_buffer_info structure into a
separate function.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NAaron Brown <aaron.f.brown@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This change makes it so that we map the entire page and just sync half of
it for the device at a time.  The advantage to this approach is that we can
avoid the locking on map/unmap seen in many IOMMU implementations.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NAaron Brown <aaron.f.brown@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This change is meant to just clean-up a number of function calls that were
made at the end of the Rx clean-up path by combining them into a single
function call.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NAaron Brown <aaron.f.brown@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This change makes it so that we no longer use header split.  The idea is to
reduce partial cache line writes by hardware when handling frames larger
then header size.  We can compensate for the extra overhead of having to
memcpy the header buffer by avoiding the cache misses seen by leaving an
full skb allocated and sitting on the ring.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NAaron Brown <aaron.f.brown@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

In order to support page based receive we will need to split up the two
different types of timestamping into two separate functions.  The first one
will handle legacy timestamps with the value in the register, and the new
one will handle timestamps in the Rx buffer itself.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Acked-by: NMatthew Vick <matthew.vick@intel.com>
Tested-by: NAaron Brown <aaron.f.brown@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>