Building virtio_net driver without CONFIG_XPS fails with:

    drivers/net/virtio_net.c: In function ‘virtnet_set_affinity’:
    drivers/net/virtio_net.c:1910:3: error: implicit declaration of function ‘__netif_set_xps_queue’ [-Werror=implicit-function-declaration]
       __netif_set_xps_queue(vi->dev, mask, i, false);
       ^
Fixes: 4d99f660 ("net: allow to call netif_reset_xps_queues() under cpus_read_lock")
Signed-off-by: NKrzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Currently drivers have to check if they already have a umem
installed for a given queue and return an error if so.  Make
better use of XDP_QUERY_XSK_UMEM and move this functionality
to the core.

We need to keep rtnl across the calls now.
Signed-off-by: NJakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: NQuentin Monnet <quentin.monnet@netronome.com>
Acked-by: NBjörn Töpel <bjorn.topel@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We used to depend on real_num_rx_queues as a upper bound for sanity
checks.  For AF_XDP socket validation it's useful if the check behaves
the same regardless of CONFIG_SYSFS setting.
Signed-off-by: NJakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: NQuentin Monnet <quentin.monnet@netronome.com>
Acked-by: NJesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If an invalid MTU value is set through rtnetlink return extra error
information instead of putting message in kernel log. For other cases
where there is no visible API, keep the error report in the log.

Example:
	# ip li set dev enp12s0 mtu 10000
	Error: mtu greater than device maximum.

	# ifconfig enp12s0 mtu 10000
	SIOCSIFMTU: Invalid argument
	# dmesg | tail -1
	[ 2047.795467] enp12s0: mtu greater than device maximum
Signed-off-by: NStephen Hemminger <stephen@networkplumber.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

gro_hash size is 192 bytes, and uses 3 cache lines, if there is few
flows, gro_hash may be not fully used, so it is unnecessary to iterate
all gro_hash in napi_gro_flush(), to occupy unnecessary cacheline.

convert gro_count to a bitmask, and rename it as gro_bitmask, each bit
represents a element of gro_hash, only flush a gro_hash element if the
related bit is set, to speed up napi_gro_flush().

and update gro_bitmask only if it will be changed, to reduce cache
update
Suggested-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NLi RongQing <lirongqing@baidu.com>
Cc: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add new netdev tls op for resynchronizing HW tls context
Signed-off-by: NBoris Pismenny <borisp@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Split the query of HW-attached program from the software one.
Introduce new .ndo_bpf command to query HW-attached program.
This will allow drivers to install different programs in HW
and SW at the same time.  Netlink can now also carry multiple
programs on dump (in which case mode will be set to
XDP_ATTACHED_MULTI and user has to check per-attachment point
attributes, IFLA_XDP_PROG_ID will not be present).  We reuse
IFLA_XDP_PROG_ID skb space for second mode, so rtnl_xdp_size()
doesn't need to be updated.

Note that the installation side is still not there, since all
drivers currently reject installing more than one program at
the time.
Signed-off-by: NJakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: NQuentin Monnet <quentin.monnet@netronome.com>
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>

prog_attached of struct netdev_bpf should have been superseded
by simply setting prog_id long time ago, but we kept it around
to allow offloading drivers to communicate attachment mode (drv
vs hw).  Subsequently drivers were also allowed to report back
attachment flags (prog_flags), and since nowadays only programs
attached will XDP_FLAGS_HW_MODE can get offloaded, we can tell
the attachment mode from the flags driver reports.  Remove
prog_attached member.
Signed-off-by: NJakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: NQuentin Monnet <quentin.monnet@netronome.com>
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>

For most of these calls we can just pass NULL through to the fallback
function as the sb_dev. The only cases where we cannot are the cases where
we might be dealing with either an upper device or a driver that would
have configured things to support an sb_dev itself.

The only driver that has any significant change in this patch set should be
ixgbe as we can drop the redundant functionality that existed in both the
ndo_select_queue function and the fallback function that was passed through
to us.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch makes it so that instead of passing a void pointer as the
accel_priv we instead pass a net_device pointer as sb_dev. Making this
change allows us to pass the subordinate device through to the fallback
function eventually so that we can keep the actual code in the
ndo_select_queue call as focused on possible on the exception cases.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch adds a generic version of the ndo_select_queue functions for
either returning 0 or selecting a queue based on the processor ID. This is
generally meant to just reduce the number of functions we have to change
in the future when we have to deal with ndo_select_queue changes.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This change makes it so that we can support the concept of subordinate
device traffic classes to the core networking code. In doing this we can
start pulling out the driver specific bits needed to support selecting a
queue based on an upper device.

The solution at is currently stands is only partially implemented. I have
the start of some XPS bits in here, but I would still need to allow for
configuration of the XPS maps on the queues reserved for the subordinate
devices. For now I am using the reference to the sb_dev XPS map as just a
way to skip the lookup of the lower device XPS map for now as that would
result in the wrong queue being picked.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch is meant to provide the basic tools needed to allow us to create
subordinate device traffic classes. The general idea here is to allow
subdividing the queues of a device into queue groups accessible through an
upper device such as a macvlan.

The idea here is to enforce the idea that an upper device has to be a
single queue device, ideally with IFF_NO_QUQUE set. With that being the
case we can pretty much guarantee that the tc_to_txq mappings and XPS maps
for the upper device are unused. As such we could reuse those in order to
support subdividing the lower device and distributing those queues between
the subordinate devices.

In order to distinguish between a regular set of traffic classes and if a
device is carrying subordinate traffic classes I changed num_tc from a u8
to a s16 value and use the negative values to represent the subordinate
pool values. So starting at -1 and running to -32768 we can encode those as
pool values, and the existing values of 0 to 15 can be maintained.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NAndrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

After commit 07d78363 ("net: Convert NAPI gro list into a small hash
table.")' there is 8 hash buckets, which allows more flows to be held for
merging.  but MAX_GRO_SKBS, the total held skb for merging, is 8 skb still,
limit the hash table performance.

keep MAX_GRO_SKBS as 8 skb, but limit each hash list length to 8 skb, not
the total 8 skb
Signed-off-by: NLi RongQing <lirongqing@baidu.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The ETF (Earliest TxTime First) qdisc uses the information added
earlier in this series (the socket option SO_TXTIME and the new
role of sk_buff->tstamp) to schedule packets transmission based
on absolute time.

For some workloads, just bandwidth enforcement is not enough, and
precise control of the transmission of packets is necessary.

Example:

$ tc qdisc replace dev enp2s0 parent root handle 100 mqprio num_tc 3 \
           map 2 2 1 0 2 2 2 2 2 2 2 2 2 2 2 2 queues 1@0 1@1 2@2 hw 0

$ tc qdisc add dev enp2s0 parent 100:1 etf delta 100000 \
           clockid CLOCK_TAI

In this example, the Qdisc will provide SW best-effort for the control
of the transmission time to the network adapter, the time stamp in the
socket will be in reference to the clockid CLOCK_TAI and packets
will leave the qdisc "delta" (100000) nanoseconds before its transmission
time.

The ETF qdisc will buffer packets sorted by their txtime. It will drop
packets on enqueue() if their skbuff clockid does not match the clock
reference of the Qdisc. Moreover, on dequeue(), a packet will be dropped
if it expires while being enqueued.

The qdisc also supports the SO_TXTIME deadline mode. For this mode, it
will dequeue a packet as soon as possible and change the skb timestamp
to 'now' during etf_dequeue().

Note that both the qdisc's and the SO_TXTIME ABIs allow for a clockid
to be configured, but it's been decided that usage of CLOCK_TAI should
be enforced until we decide to allow for other clockids to be used.
The rationale here is that PTP times are usually in the TAI scale, thus
no other clocks should be necessary. For now, the qdisc will return
EINVAL if any clocks other than CLOCK_TAI are used.
Signed-off-by: NJesus Sanchez-Palencia <jesus.sanchez-palencia@intel.com>
Signed-off-by: NVinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Also involved adding a way to run a netfilter hook over a list of packets.
 Rather than attempting to make netfilter know about lists (which would be
 a major project in itself) we just let it call the regular okfn (in this
 case ip_rcv_finish()) for any packets it steals, and have it give us back
 a list of packets it's synchronously accepted (which normally NF_HOOK
 would automatically call okfn() on, but we want to be able to potentially
 pass the list to a listified version of okfn().)
The netfilter hooks themselves are indirect calls that still happen per-
 packet (see nf_hook_entry_hookfn()), but again, changing that can be left
 for future work.

There is potential for out-of-order receives if the netfilter hook ends up
 synchronously stealing packets, as they will be processed before any
 accepts earlier in the list.  However, it was already possible for an
 asynchronous accept to cause out-of-order receives, so presumably this is
 considered OK.
Signed-off-by: NEdward Cree <ecree@solarflare.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Just calls netif_receive_skb() in a loop.
Signed-off-by: NEdward Cree <ecree@solarflare.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since the addition of GRO for ESP, gro_receive can consume the skb and
return -EINPROGRESS. In that case, the lower layer GRO handler cannot
touch the skb anymore.

Commit 5f114163 ("net: Add a skb_gro_flush_final helper.") converted
some of the gro_receive handlers that can lead to ESP's gro_receive so
that they wouldn't access the skb when -EINPROGRESS is returned, but
missed other spots, mainly in tunneling protocols.

This patch finishes the conversion to using skb_gro_flush_final(), and
adds a new helper, skb_gro_flush_final_remcsum(), used in VXLAN and
GUE.

Fixes: 5f114163 ("net: Add a skb_gro_flush_final helper.")
Signed-off-by: NSabrina Dubroca <sd@queasysnail.net>
Reviewed-by: NStefano Brivio <sbrivio@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Refactor XPS code to support Tx queue selection based on
CPU(s) map or Rx queue(s) map.
Signed-off-by: NAmritha Nambiar <amritha.nambiar@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Improve the performance of GRO receive by splitting flows into
multiple hash chains.
Suggested-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Manage pending per-NAPI GRO packets via list_head.

Return an SKB pointer from the GRO receive handlers.  When GRO receive
handlers return non-NULL, it means that this SKB needs to be completed
at this time and removed from the NAPI queue.

Several operations are greatly simplified by this transformation,
especially timing out the oldest SKB in the list when gro_count
exceeds MAX_GRO_SKBS, and napi_gro_flush() which walks the queue
in reverse order.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Added ndo_xsk_async_xmit. This ndo "kicks" the netdev to start to pull
userland AF_XDP Tx frames from a NAPI context.
Signed-off-by: NMagnus Karlsson <magnus.karlsson@intel.com>
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>

Extend ndo_bpf with two new commands used for query zero-copy support
and register an UMEM to a queue_id of a netdev.
Signed-off-by: NBjörn Töpel <bjorn.topel@intel.com>
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>

All drivers are cleaned up and no references to ndo_xdp_flush
are left in drivers, it is time to remove the net_device_ops
operation ndo_xdp_flush.
Signed-off-by: NJesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>

This patch only change the API and reject any use of flags. This is an
intermediate step that allows us to implement the flush flag operation
later, for each individual driver in a separate patch.

The plan is to implement flush operation via XDP_XMIT_FLUSH flag
and then remove XDP_XMIT_FLAGS_NONE when done.
Signed-off-by: NJesper Dangaard Brouer <brouer@redhat.com>
Acked-by: NSong Liu <songliubraving@fb.com>
Signed-off-by: NAlexei Starovoitov <ast@kernel.org>

mq offload is trivial, we just need to let the device know
that the root qdisc is mq.  Alternative approach would be
to export qdisc_lookup() and make drivers check the root
type themselves, but notification via ndo_setup_tc is more
in line with other qdiscs.

Note that mq doesn't hold any stats on it's own, it just
adds up stats of its children.
Signed-off-by: NJakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The failover module provides a generic interface for paravirtual drivers
to register a netdev and a set of ops with a failover instance. The ops
are used as event handlers that get called to handle netdev register/
unregister/link change/name change events on slave pci ethernet devices
with the same mac address as the failover netdev.

This enables paravirtual drivers to use a VF as an accelerated low latency
datapath. It also allows migration of VMs with direct attached VFs by
failing over to the paravirtual datapath when the VF is unplugged.
Signed-off-by: NSridhar Samudrala <sridhar.samudrala@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

LAG upper event notifiers contain the tx type used by the LAG device.
Extend this to also include the hash policy used for tx types that
utilize hashing.
Signed-off-by: NJohn Hurley <john.hurley@netronome.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch change the API for ndo_xdp_xmit to support bulking
xdp_frames.

When kernel is compiled with CONFIG_RETPOLINE, XDP sees a huge slowdown.
Most of the slowdown is caused by DMA API indirect function calls, but
also the net_device->ndo_xdp_xmit() call.

Benchmarked patch with CONFIG_RETPOLINE, using xdp_redirect_map with
single flow/core test (CPU E5-1650 v4 @ 3.60GHz), showed
performance improved:
 for driver ixgbe: 6,042,682 pps -> 6,853,768 pps = +811,086 pps
 for driver i40e : 6,187,169 pps -> 6,724,519 pps = +537,350 pps

With frames avail as a bulk inside the driver ndo_xdp_xmit call,
further optimizations are possible, like bulk DMA-mapping for TX.

Testing without CONFIG_RETPOLINE show the same performance for
physical NIC drivers.

The virtual NIC driver tun sees a huge performance boost, as it can
avoid doing per frame producer locking, but instead amortize the
locking cost over the bulk.

V2: Fix compile errors reported by kbuild test robot <lkp@intel.com>
V4: Isolated ndo, driver changes and callers.
Signed-off-by: NJesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: NAlexei Starovoitov <ast@kernel.org>

The new dev_direct_xmit will be used by AF_XDP in later commits.
Signed-off-by: NMagnus Karlsson <magnus.karlsson@intel.com>
Signed-off-by: NAlexei Starovoitov <ast@kernel.org>

We do not require this inline function to be used in multiple different
locations, just inline it where it gets used in register_netdevice().
Suggested-by: NDavid Miller <davem@davemloft.net>
Suggested-by: NStephen Hemminger <stephen@networkplumber.org>
Signed-off-by: NFlorian Fainelli <f.fainelli@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add new netdev ops to add and delete tls context
Signed-off-by: NIlya Lesokhin <ilyal@mellanox.com>
Signed-off-by: NBoris Pismenny <borisp@mellanox.com>
Signed-off-by: NAviad Yehezkel <aviadye@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We have about 53 netdev_features_t bits defined and counting, add a
build time check to catch when an u64 type will not be enough and we
will have to convert that to a bitmap. This is done in
register_netdevice() for convenience.
Signed-off-by: NFlorian Fainelli <f.fainelli@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

I am dropping the export of __skb_tx_hash as after my patches nobody is
using it outside of the net/core/dev.c file. In addition I am renaming and
repurposing it to just be a static declaration of skb_tx_hash since that
was the only user for it at this point. By doing this the compiler can
inline it into __netdev_pick_tx as that will improve performance.
Signed-off-by: NAlexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Virtual devices such as tunnels and bonding can handle large packets.
Only segment packets when reaching a physical or loopback device.
Signed-off-by: NWillem de Bruijn <willemb@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Changing API ndo_xdp_xmit to take a struct xdp_frame instead of struct
xdp_buff.  This brings xdp_return_frame and ndp_xdp_xmit in sync.

This builds towards changing the API further to become a bulk API,
because xdp_buff is not a queue-able object while xdp_frame is.

V4: Adjust for commit 59655a5b ("tuntap: XDP_TX can use native XDP")
V7: Adjust for commit d9314c47 ("i40e: add support for XDP_REDIRECT")
Signed-off-by: NJesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Provide a pointer to the SFP bus in struct net_device, so that the
ethtool module EEPROM methods can access the SFP directly, rather
than needing every user to provide a hook for it.
Reviewed-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NRussell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: NFlorian Fainelli <f.fainelli@gmail.com>
Reviewed-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NRussell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: NFlorian Fainelli <f.fainelli@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

NETIF_F_HW_VLAN_[CS]TAG_FILTER features require more than just a bit
flip in dev->features in order to keep the driver in a consistent state.
These features notify the driver of each added/removed vlan, but toggling
of vlan-filter does not notify the driver accordingly for each of the
existing vlans.

This patch implements a similar solution to NETIF_F_RX_UDP_TUNNEL_PORT
behavior (which notifies the driver about UDP ports in the same manner
that vids are reported).

Each toggling of the features propagates to the 8021q module, which
iterates over the vlans and call add/kill ndo accordingly.
Signed-off-by: NGal Pressman <galp@mellanox.com>
Reviewed-by: NTariq Toukan <tariqt@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Last user is gone after bdf5bd7f "rds: tcp: remove
register_netdevice_notifier infrastructure.", so we can
remove this netdevice command. This allows to delete
rtnl_lock() in netdev_run_todo(), which is hot path for
net namespace unregistration.

dev_change_net_namespace() and netdev_wait_allrefs()
have rcu_barrier() before NETDEV_UNREGISTER_FINAL call,
and the source commits say they were introduced to
delemit the call with NETDEV_UNREGISTER, but this patch
leaves them on the places, since they require additional
analysis, whether we need in them for something else.
Signed-off-by: NKirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch is preparation to drop NETDEV_UNREGISTER_FINAL.
Since the cmd is used in usnic_ib_netdev_event_to_string()
to get cmd name, after plain removing NETDEV_UNREGISTER_FINAL
from everywhere, we'd have holes in event2str[] in this
function.

Instead of that, let's make NETDEV_XXX commands names
available for everyone, and to define netdev_cmd_to_name()
in the way we won't have to shaffle names after their
numbers are changed.
Signed-off-by: NKirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>