This patch does s/rebuid/rebuild/.

Change-ID: I77ff0c26db3d23ca5dc971b82d71b4ed8c6bcc14
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Fix some badly formatted lines, long lines and a mis-formatted else.

Change-ID: Iac2eef064ae27c55a0c3d9c15c525bf8fed8ab6f
Signed-off-by: NMitch Williams <mitch.a.williams@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The alloc_rx_buff_failed and alloc_rx_page_failed variables
are both part of an rx specific structure so just remove
the _rx part of the name.  No functional changes.

Change-ID: Icffa2f5d13c6f2b1e09cf45b9472b83c9dae8fc6
Signed-off-by: NMitch Williams <mitch.a.williams@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The admin queue can be used to send messages among the physical
function interfaces.  This adds the code to handle that case.

Change-ID: I0700fcc47e41433131a381f0eb72fc7b01b6bd87
Signed-off-by: NShannon Nelson <shannon.nelson@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Since nearly everything we do is synchronous, using the interrupt-on-error bit
is unnecessary and causing unneeded interrupts.  If anyone wants to use the bit
they can turn it on in individual AQ requests using the cmd_details parameter.

Change-ID: I4690a9c561d3e0836aeadb4f88f8a8702b1d1366
Signed-off-by: NShannon Nelson <shannon.nelson@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Call AQ to release any reservation held by this PF on the NVM resource
lock on startup, in order to clear anything that might have been left over from
a previous run.  The lock is only cleared by the requestor calling for it to be
cleared, on power-on, or firmware reset.  This should help limit the need for
rebooting a customer machine if something goes wrong on a firmware update or
some other action.

Change-ID: I8c8473e601d4ef512dda7baa77a6e75f2e5fea49
Signed-off-by: NShannon Nelson <shannon.nelson@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

RSS initialization was doing some extra work, remove the extra
work and any bugs it created when managing number of queues.

Change-ID: Iea75b04a70d73ce76947b6a177ce89ab4899d4c6
Signed-off-by: NAnjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The driver and hardware support splitting packets on headers
but with the use of GRO we don't need the extra bus
overhead, so make this driver more like igb and ixgbe and
disable packet split.

Change-ID: Id42f2c3736baa9d5bdfe1f72d64226e7d8ebd737
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The memory barrier used in maybe_stop_tx can use a comment.

Also add checks to VSI->rx_rings to ensure a kernel panic is not induced.

Change-ID: I48cc1bf1d6cf301818155b737edeef77c0d790c7
Change-ID: I1363a8445fbf521a26267849966296ed55f43ad8
Signed-off-by: NGreg Rose <gregory.v.rose@intel.com>
Signed-off-by: NMitch Williams <mitch.a.williams@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The MAC address format expected by the hardware is in a very specific
format, and the driver was filling in the data incorrectly.

Change-ID: I7bc66505ef459ee347dd3bda68051004c141c689
Signed-off-by: NKamil Krawczyk <kamil.krawczyk@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The GPL header included in each file in the i40e driver doesn't
need to include the "this program" text since this driver
is already part of the larger kernel.
Signed-off-by: NGreg Rose <gregory.v.rose@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Replace uses of I40E_LENGTH_OF_ADDRESS with ETH_ALEN.
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The hardware can occasionally give an interrupt on the misc
queue for which there is no driver work to do.  In that case
the driver was not re-enabling interrupts even though they
were auto masked by hardware.  This left interrupts disabled
on this queue.

Re-enable the interrupt whenever leaving this function.
Signed-off-by: NAnjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Pablo Neira Ayuso says:

====================
Netfilter updates for net-next

The following patchset contains three Netfilter updates, they are:

* Fix wrong usage of skb_header_pointer in the DCCP protocol helper that
  has been there for quite some time. It was resulting in copying the dccp
  header to a pointer allocated in the stack. Fortunately, this pointer
  provides room for the dccp header is 4 bytes long, so no crashes have been
  reported so far. From Daniel Borkmann.

* Use format string to print in the invocation of nf_log_packet(), again
  in the DCCP helper. Also from Daniel Borkmann.

* Revert "netfilter: avoid get_random_bytes call" as prandom32 does not
  guarantee enough entropy when being calling this at boot time, that may
  happen when reloading the rule.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Jeff Kirsher says:

====================
Intel Wired LAN Driver Updates

This series contains updates to i40e only.

Anjali adds more functionality to debugfs to assist development and
testing of admin queue commands.

Greg makes sure broadcast promiscuous is disabled by default, otherwise
VLAN tagged packets out of the assigned VLAN domain are received.  Also
provides a fix when the 8021q driver is loaded, so that VLAN 0 tagged
packets are accepted so that upper layers can interpret the priority
bits. Then provides a fix to let the VF to request the PF to set its
already assigned MAC address without generating an error.  Greg also
adds helper functions to enable or disable internal switch loopback
when VFs are created or destroyed via the sysfs interface.

Shannon provides most of the changes, where he adds code to ensure
that the hardware waits to make sure that the firmware is ready as well
after reset.  Also updates the code to use the new features in the
firmware.  Provides a fix while in MFP mode where resources are
reduced, so use a smaller range of test registers than when in SFP mode.
Moves the PF ID initialization code to earlier in the driver
initialization function since a few operations need the information
before the first PF reset is called.  Shannon adds a check for MAC
type before reading anything from the registers to ensure we dealing
with the correct MAC type.  Then reworks Shadow RAM read word/buffer
functions as to not block whole NVM resources for SR read operations.

Mitch lastly provides a fix to correctly setup ARQ descriptors in
the cleanup code.

Catherine bumps the driver version due to all the recent changes.

v2:
 - Added blank lines after local variable declarations to patch 01
   as suggested by David Miller
 - Used the suggested memset() line in patch 15 as suggested by David
   Miller
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When cleaning descriptors, we must set up ALL fields, not just the DMA
address. The initial setup does this correctly, but not the cleanup
code, so the firmware would process the ring exactly once and then fail.

Change-ID: I2930b83c76194b3016a8ac0fa693f9a573995640
Signed-off-by: NMitch Williams <mitch.a.williams@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The admin queue length register is updated in
config_a<sq|rq>_regs functions.  We should not update it again,
as we will trigger firmware to init the AQ again. In this case
firmware will lose the information about the AQ Rx tail position
and will see Rx queue as full (no free desc for FW to use).
Signed-off-by: NKamil Krawczyk <kamil.krawczyk@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The PF VSI was never updated to enable or disable internal switch loopback
when VFs were created or destroyed via the sysfs interface.  Add some
helper functions to take care of that.
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>

Addresses a few code format issues that have crept in over time.

Change-Id: I1a62cbd16b29a218a933b0f7176abe748f9615e8
Signed-off-by: NMitch Williams <mitch.a.williams@intel.com>
Signed-off-by: NShannon Nelson <shannon.nelson@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Rework Shadow RAM read word/buffer functions to not use AQ Request
Resource commands.  Requesting resource is not needed for SR read
operations which are done through SRCTL register.  Access to SR through
register is controlled through DONE bit within SRCTL.  With this change
we do not block whole NVM resource for SR read operations.

Change-Id: I73e96cdea39a45ee7b5bdf038e527308de2d9efe
Signed-off-by: NKamil Krawczyk <kamil.krawczyk@intel.com>
Signed-off-by: NShannon Nelson <shannon.nelson@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

We need to check if we are dealing with the correct MAC type before we
try to read anything from the registers.

Change-Id: I3989516999d06c3009e87d6a2eafc20af305c5c2
Signed-off-by: NKamil Krawczyk <kamil.krawczyk@intel.com>
Signed-off-by: NShannon Nelson <shannon.nelson@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Move PF ID initialization code from PF reset routine to earlier driver
initialization function.  There are a few operations which need the
PF ID before the first PF reset is called.

Change-Id: I7e971f7556b46a837149850ec05ce115c35db575
Signed-off-by: NKamil Krawczyk <kamil.krawczyk@intel.com>
Signed-off-by: NShannon Nelson <shannon.nelson@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Use a smaller range of test registers in MFP mode as there are fewer
resources than when in SFP mode.

Change-Id: I08424890c3f57b5dde5ee99e99724ce252e0875a
Signed-off-by: NKamil Krawczyk <kamil.krawczyk@intel.com>
Signed-off-by: NShannon Nelson <shannon.nelson@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The firmware's AdminQ interface has matured a little, so update the
code to use the new fields and values.

Change-Id: I8fcd7b443f268dcf9346bd6a9e940fe9c2958891
Signed-off-by: NShannon Nelson <shannon.nelson@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The RSTAT comes back with DEVICE READY to indicate that the HW is ready,
but we still have to wait for the FW to indicate that the Core and Global
modules are back up again.  This needs a read of the NVM_ULD register.

Change-Id: I88276165f9cd446d2f166fb4b8cff00521af4bec
Signed-off-by: NAnjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: NShannon Nelson <shannon.nelson@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Version update to 0.3.25-k
Signed-off-by: NCatherine Sullivan <catherine.sullivan@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The VF is allowed to request the PF to set its already assigned
MAC address without generating an error.

Change-Id: I8dfdf353396995dbbb26cafab4e42b451911da3d
Signed-off-by: NGreg Rose <gregory.v.rose@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Acked-by: NShannon Nelson <shannon.nelson@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

When the 8021q driver is loaded it sets VLAN 0 filters on all devices.
This does not mean that any VLAN tagged packet should be accepted.
Instead accept only VLAN 0 tagged packets so that upper layers can
interpret the priority bits.

Change-Id: I17274a540b613749612ffe23a3aef2b8ee6ff6a4
Signed-off-by: NGreg Rose <gregory.v.rose@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NSibai Li <sibai.li@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Broadcast promiscuous should only be turned on when general
promiscuous mode is turned on, otherwise VLAN tagged packets out of
the assigned VLAN domain are received.

Add a broadcast MAC filter in order to continue to receive
broadcast traffic on VLANs, MAIN or VMDQ VSI.

Change-Id: I99d8e382a082ee51201228f1226af3b46452ac55
Signed-off-by: NGreg Rose <gregory.v.rose@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NSibai Li <sibai.li@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Add more functionality to debugfs to assist development
and testing of AQ commands.

adds:
send aq_cmd
send indirect aq_cmd

Change-Id: I01710cddd33110a6c1e1aabf84cb6e93cda4c42a
Signed-off-by: NAnjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: NKavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Fix below compiler warning:

net/xfrm/xfrm_policy.c:1644:12: warning: ‘xfrm_dst_alloc_copy’ defined but not used [-Wunused-function]
Signed-off-by: NYing Xue <ying.xue@windriver.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Jon Maloy says:

====================
tipc: link setup and failover improvements

This series consists of four unrelated commits with different purposes.

- Commit #1 is purely cosmetic and pedagogic, hopefully making the
  failover/tunneling logics slightly easier to understand.
- Commit #2 fixes a bug that has always been in the code, but was not
  discovered until very recently.
- Commit #3 fixes a non-fatal race issue in the neighbour discovery
  code.
- Commit #4 removes an unnecessary indirection step during link
  startup.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When a link is created we delay the start event by launching it
to be executed later in a tasklet. As we hold all the
necessary locks at the moment of creation, and there is no risk
of deadlock or contention, this delay serves no purpose in the
current code.

We remove this obsolete indirection step, and the associated function
link_start(). At the same time, we rename the function tipc_link_stop()
to the more appropriate tipc_link_purge_queues().
Signed-off-by: NJon Maloy <jon.maloy@ericsson.com>
Reviewed-by: NPaul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Currently, only 'bearer_lock' is used to protect struct link_req in
the function disc_timeout(). This is unsafe, since the member fields
'num_nodes' and 'timer_intv' might be accessed by below three different
threads simultaneously, none of them grabbing bearer_lock in the
critical region:

link_activate()
  tipc_bearer_add_dest()
    tipc_disc_add_dest()
      req->num_nodes++;

tipc_link_reset()
  tipc_bearer_remove_dest()
    tipc_disc_remove_dest()
      req->num_nodes--
      disc_update()
        read req->num_nodes
	write req->timer_intv

disc_timeout()
  read req->num_nodes
  read/write req->timer_intv

Without lock protection, the only symptom of a race is that discovery
messages occasionally may not be sent out. This is not fatal, since such
messages are best-effort anyway. On the other hand, since discovery
messages are not time critical, adding a protecting lock brings no
serious overhead either. So we add a new, dedicated spinlock in
order to guarantee absolute data consistency in link_req objects.
This also helps reduce the overall role of the bearer_lock, which
we want to remove completely in a later commit series.
Signed-off-by: NYing Xue <ying.xue@windriver.com>
Reviewed-by: NPaul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: NJon Maloy <jon.maloy@ericsson.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The flag 'has_redundant_link' is defined only in RESET and ACTIVATE
protocol messages. Due to an ambiguity in the protocol specification it
is currently also transferred in STATE messages. Its value is used to
initialize a link state variable, 'permit_changeover', which is used
to inhibit futile link failover attempts when it is known that the
peer node has no working links at the moment, although the local node
may still think it has one.

The fact that 'has_redundant_link' incorrectly is read from STATE
messages has the effect that 'permit_changeover' sometimes gets a wrong
value, and permanently blocks any links from being re-established. Such
failures can only occur in in dual-link systems, and are extremely rare.
This bug seems to have always been present in the code.

Furthermore, since commit b4b56102
("tipc: Ensure both nodes recognize loss of contact between them"),
the 'permit_changeover' field serves no purpose any more. The task of
enforcing 'lost contact' cycles at both peer endpoints is now taken
by a new mechanism, using the flags WAIT_NODE_DOWN and WAIT_PEER_DOWN
in struct tipc_node to abort unnecessary failover attempts.

We therefore remove the 'has_redundant_link' flag from STATE messages,
as well as the now redundant 'permit_changeover' variable.
Signed-off-by: NJon Maloy <jon.maloy@ericsson.com>
Reviewed-by: NYing Xue <ying.xue@windriver.com>
Reviewed-by: NPaul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The functionality related to link addition and failover is unnecessarily
hard to understand and maintain. We try to improve this by renaming
some of the functions, at the same time adding or improving the
explanatory comments around them. Names such as "tipc_rcv()" etc. also
align better with what is used in other networking components.

The changes in this commit are purely cosmetic, no functional changes
are made.
Signed-off-by: NJon Maloy <jon.maloy@ericsson.com>
Reviewed-by: NYing Xue <ying.xue@windriver.com>
Reviewed-by: NPaul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We should const-ify comparisons on skb_queue_* inline helper
functions as their parameters are const as well, so lets not
drop that.
Suggested-by: NBrad Spengler <spender@grsecurity.net>
Signed-off-by: NDaniel Borkmann <dborkman@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fix three warnings related to:

  net/xfrm/xfrm_policy.c:1644:1: warning: 'inline' is not at beginning of declaration [-Wold-style-declaration]
  net/xfrm/xfrm_policy.c:1656:1: warning: 'inline' is not at beginning of declaration [-Wold-style-declaration]
  net/xfrm/xfrm_policy.c:1668:1: warning: 'inline' is not at beginning of declaration [-Wold-style-declaration]

Just removing the inline keyword is sufficient as the compiler will
decide on its own about inlining or not.
Signed-off-by: NDaniel Borkmann <dborkman@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Now that mlx4_en includes a PHC driver it must select PTP_1588_CLOCK.

   drivers/built-in.o: In function `mlx4_en_get_ts_info':
>> en_ethtool.c:(.text+0x391a11): undefined reference to `ptp_clock_index'
   drivers/built-in.o: In function `mlx4_en_remove_timestamp':
>> (.text+0x397913): undefined reference to `ptp_clock_unregister'
   drivers/built-in.o: In function `mlx4_en_init_timestamp':
>> (.text+0x397b20): undefined reference to `ptp_clock_register'

Fixes: ad7d4eae ("mlx4_en: Add PTP hardware clock")
Signed-off-by: NShawn Bohrer <sbohrer@rgmadvisors.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch built on top of Commit 299603e8
("net-gro: Prepare GRO stack for the upcoming tunneling support") to add
the support of the standard GRE (RFC1701/RFC2784/RFC2890) to the GRO
stack. It also serves as an example for supporting other encapsulation
protocols in the GRO stack in the future.

The patch supports version 0 and all the flags (key, csum, seq#) but
will flush any pkt with the S (seq#) flag. This is because the S flag
is not support by GSO, and a GRO pkt may end up in the forwarding path,
thus requiring GSO support to break it up correctly.

Currently the "packet_offload" structure only contains L3 (ETH_P_IP/
ETH_P_IPV6) GRO offload support so the encapped pkts are limited to
IP pkts (i.e., w/o L2 hdr). But support for other protocol type can
be easily added, so is the support for GRE variations like NVGRE.

The patch also support csum offload. Specifically if the csum flag is on
and the h/w is capable of checksumming the payload (CHECKSUM_COMPLETE),
the code will take advantage of the csum computed by the h/w when
validating the GRE csum.

Note that commit 60769a5d "ipv4: gre:
add GRO capability" already introduces GRO capability to IPv4 GRE
tunnels, using the gro_cells infrastructure. But GRO is done after
GRE hdr has been removed (i.e., decapped). The following patch applies
GRO when pkts first come in (before hitting the GRE tunnel code). There
is some performance advantage for applying GRO as early as possible.
Also this approach is transparent to other subsystem like Open vSwitch
where GRE decap is handled outside of the IP stack hence making it
harder for the gro_cells stuff to apply. On the other hand, some NICs
are still not capable of hashing on the inner hdr of a GRE pkt (RSS).
In that case the GRO processing of pkts from the same remote host will
all happen on the same CPU and the performance may be suboptimal.

I'm including some rough preliminary performance numbers below. Note
that the performance will be highly dependent on traffic load, mix as
usual. Moreover it also depends on NIC offload features hence the
following is by no means a comprehesive study. Local testing and tuning
will be needed to decide the best setting.

All tests spawned 50 copies of netperf TCP_STREAM and ran for 30 secs.
(super_netperf 50 -H 192.168.1.18 -l 30)

An IP GRE tunnel with only the key flag on (e.g., ip tunnel add gre1
mode gre local 10.246.17.18 remote 10.246.17.17 ttl 255 key 123)
is configured.

The GRO support for pkts AFTER decap are controlled through the device
feature of the GRE device (e.g., ethtool -K gre1 gro on/off).

1.1 ethtool -K gre1 gro off; ethtool -K eth0 gro off
thruput: 9.16Gbps
CPU utilization: 19%

1.2 ethtool -K gre1 gro on; ethtool -K eth0 gro off
thruput: 5.9Gbps
CPU utilization: 15%

1.3 ethtool -K gre1 gro off; ethtool -K eth0 gro on
thruput: 9.26Gbps
CPU utilization: 12-13%

1.4 ethtool -K gre1 gro on; ethtool -K eth0 gro on
thruput: 9.26Gbps
CPU utilization: 10%

The following tests were performed on a different NIC that is capable of
csum offload. I.e., the h/w is capable of computing IP payload csum
(CHECKSUM_COMPLETE).

2.1 ethtool -K gre1 gro on (hence will use gro_cells)

2.1.1 ethtool -K eth0 gro off; csum offload disabled
thruput: 8.53Gbps
CPU utilization: 9%

2.1.2 ethtool -K eth0 gro off; csum offload enabled
thruput: 8.97Gbps
CPU utilization: 7-8%

2.1.3 ethtool -K eth0 gro on; csum offload disabled
thruput: 8.83Gbps
CPU utilization: 5-6%

2.1.4 ethtool -K eth0 gro on; csum offload enabled
thruput: 8.98Gbps
CPU utilization: 5%

2.2 ethtool -K gre1 gro off

2.2.1 ethtool -K eth0 gro off; csum offload disabled
thruput: 5.93Gbps
CPU utilization: 9%

2.2.2 ethtool -K eth0 gro off; csum offload enabled
thruput: 5.62Gbps
CPU utilization: 8%

2.2.3 ethtool -K eth0 gro on; csum offload disabled
thruput: 7.69Gbps
CPU utilization: 8%

2.2.4 ethtool -K eth0 gro on; csum offload enabled
thruput: 8.96Gbps
CPU utilization: 5-6%
Signed-off-by: NH.K. Jerry Chu <hkchu@google.com>
Reviewed-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>