We prefer static_branch_unlikely() over static_key_false() these days.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Acked-by: NSoheil Hassas Yeganeh <soheil@google.com>
Acked-by: NWillem de Bruijn <willemb@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

With commit c5ad119f ("net: sched: pfifo_fast use skb_array")
pfifo_fast no longer benefit from the TCQ_F_CAN_BYPASS optimization.
Due to retpolines the cost of the enqueue()/dequeue() pair has become
relevant and we observe measurable regression for the uncontended
scenario when the packet-rate is below line rate.

After commit 46b1c18f ("net: sched: put back q.qlen into a
single location") we can check for empty qdisc with a reasonably
fast operation even for nolock qdiscs.

This change extends TCQ_F_CAN_BYPASS support to nolock qdisc.
The new chunk of code mirrors closely the existing one for traditional
qdisc, leveraging a newly introduced helper to read atomically the
qdisc length.

Tested with pktgen in queue xmit mode, with pfifo_fast, a MQ
device, and MQ root qdisc:

threads         vanilla         patched
                kpps            kpps
1               2465            2889
2               4304            5188
4               7898            9589

Same as above, but with a single queue device:

threads         vanilla         patched
                kpps            kpps
1               2556            2827
2               2900            2900
4               5000            5000
8               4700            4700

No mesaurable changes in the contended scenarios, and more 10%
improvement in the uncontended ones.

 v1 -> v2:
  - rebased after flag name change
Signed-off-by: NPaolo Abeni <pabeni@redhat.com>
Tested-by: NIvan Vecera <ivecera@redhat.com>
Reviewed-by: NEric Dumazet <edumazet@google.com>
Reviewed-by: NIvan Vecera <ivecera@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

After the previous patch, all the callers of ndo_select_queue()
provide as a 'fallback' argument netdev_pick_tx.
The only exceptions are nested calls to ndo_select_queue(),
which pass down the 'fallback' available in the current scope
- still netdev_pick_tx.

We can drop such argument and replace fallback() invocation with
netdev_pick_tx(). This avoids an indirect call per xmit packet
in some scenarios (TCP syn, UDP unconnected, XDP generic, pktgen)
with device drivers implementing such ndo. It also clean the code
a bit.

Tested with ixgbe and CONFIG_FCOE=m

With pktgen using queue xmit:
threads		vanilla 	patched
		(kpps)		(kpps)
1		2334		2428
2		4166		4278
4		7895		8100

 v1 -> v2:
 - rebased after helper's name change
Signed-off-by: NPaolo Abeni <pabeni@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Currently packet_pick_tx_queue() is the only caller of
ndo_select_queue() using a fallback argument other than
netdev_pick_tx.

Leveraging rx queue, we can obtain a similar queue selection
behavior using core helpers. After this change, ndo_select_queue()
is always invoked with netdev_pick_tx() as fallback.
We can change ndo_select_queue() signature in a followup patch,
dropping an indirect call per transmitted packet in some scenarios
(e.g. TCP syn and XDP generic xmit)

This changes slightly how af packet queue selection happens when
PACKET_QDISC_BYPASS is set. It's now more similar to plan dev_queue_xmit()
tacking in account both XPS and TC mapping.

 v1  -> v2:
  - rebased after helper name change
 RFC -> v1:
  - initialize sender_cpu to the expected value
Signed-off-by: NPaolo Abeni <pabeni@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

With the following patches, we are going to use __netdev_pick_tx() in
many modules. Rename it to netdev_pick_tx(), to make it clear is
a public API.

Also rename the existing netdev_pick_tx() to netdev_core_pick_tx(),
to avoid name clashes.
Suggested-by: NEric Dumazet <edumazet@google.com>
Suggested-by: NDavid Miller <davem@davemloft.net>
Signed-off-by: NPaolo Abeni <pabeni@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Introduce dev_change_proto_down_generic, a generic ndo_change_proto_down
implementation, which sets the netdev carrier state according to proto_down.

This adds the ability to set protodown on vxlan and macvlan devices in a
generic way for use by control protocols like VRRPD.
Signed-off-by: NAndy Roulin <aroulin@cumulusnetworks.com>
Acked-by: NRoopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

qdisc_pkt_len_init expects transport_header to be set for GSO packets.
Patch [1] skips transport_header validation for GSO packets that don't
have network_header set at the moment of calling virtio_net_hdr_to_skb,
and allows them to pass into the stack. After patch [2] no placeholder
value is assigned to transport_header if dissection fails, so this patch
adds a check to the place where the value of transport_header is used.

[1] https://patchwork.ozlabs.org/patch/1044429/
[2] https://patchwork.ozlabs.org/patch/1046122/Signed-off-by: NMaxim Mikityanskiy <maximmi@mellanox.com>
Acked-by: NWillem de Bruijn <willemb@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The features attribute is of type u64 and stored in the native endianes on
the system. The for_each_set_bit() macro takes a pointer to a 32 bit array
and goes over the bits in this area. On little Endian systems this also
works with an u64 as the most significant bit is on the highest address,
but on big endian the words are swapped. When we expect bit 15 here we get
bit 47 (15 + 32).

This patch converts it more or less to its own for_each_set_bit()
implementation which works on 64 bit integers directly. This is then
completely in host endianness and should work like expected.

Fixes: fd867d51 ("net/core: generic support for disabling netdev features down stack")
Signed-off-by: NHauke Mehrtens <hauke.mehrtens@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In preparation for getting rid of switchdev_ops, create a dedicated NDO
operation for getting the port's parent identifier. There are
essentially two classes of drivers that need to implement getting the
port's parent ID which are VF/PF drivers with a built-in switch, and
pure switchdev drivers such as mlxsw, ocelot, dsa etc.

We introduce a helper function: dev_get_port_parent_id() which supports
recursion into the lower devices to obtain the first port's parent ID.

Convert the bridge, core and ipv4 multicast routing code to check for
such ndo_get_port_parent_id() and call the helper function when valid
before falling back to switchdev_port_attr_get(). This will allow us to
convert all relevant drivers in one go instead of having to implement
both switchdev_port_attr_get() and ndo_get_port_parent_id() operations,
then get rid of switchdev_port_attr_get().
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NFlorian Fainelli <f.fainelli@gmail.com>
Reviewed-by: NIdo Schimmel <idosch@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since commit a25717d2 ("xdp: support simultaneous driver and
hw XDP attachment") users can load an XDP program for offload and
in native driver mode simultaneously.  Allow a similar mix of
offload and SKB mode/generic XDP.
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>

In order to provide more meaningful messages to user when the process of
loading xdp program onto network interface failed, let's add extack
messages within dev_change_xdp_fd.
Suggested-by: NJakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: NMaciej Fijalkowski <maciej.fijalkowski@intel.com>
Acked-by: NJesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>

Assign a default net namespace to netdevs created by init_dummy_netdev().
Fixes a NULL pointer dereference caused by busy-polling a socket bound to
an iwlwifi wireless device, which bumps the per-net BUSYPOLLRXPACKETS stat
if napi_poll() received packets:

  BUG: unable to handle kernel NULL pointer dereference at 0000000000000190
  IP: napi_busy_loop+0xd6/0x200
  Call Trace:
    sock_poll+0x5e/0x80
    do_sys_poll+0x324/0x5a0
    SyS_poll+0x6c/0xf0
    do_syscall_64+0x6b/0x1f0
    entry_SYSCALL_64_after_hwframe+0x3d/0xa2

Fixes: 7db6b048 ("net: Commonize busy polling code to focus on napi_id instead of socket")
Signed-off-by: NJosh Elsasser <jelsasser@appneta.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Currently, CONFIG_JUMP_LABEL just means "I _want_ to use jump label".

The jump label is controlled by HAVE_JUMP_LABEL, which is defined
like this:

  #if defined(CC_HAVE_ASM_GOTO) && defined(CONFIG_JUMP_LABEL)
  # define HAVE_JUMP_LABEL
  #endif

We can improve this by testing 'asm goto' support in Kconfig, then
make JUMP_LABEL depend on CC_HAS_ASM_GOTO.

Ugly #ifdef HAVE_JUMP_LABEL will go away, and CONFIG_JUMP_LABEL will
match to the real kernel capability.
Signed-off-by: NMasahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Michael Ellerman <mpe@ellerman.id.au> (powerpc)
Tested-by: NSedat Dilek <sedat.dilek@gmail.com>

This avoids an indirect calls for L3 GRO receive path, both
for ipv4 and ipv6, if the latter is not compiled as a module.

Note that when IPv6 is compiled as builtin, it will be checked first,
so we have a single additional compare for the more common path.

v1 -> v2:
 - adapted to INDIRECT_CALL_ changes
Signed-off-by: NPaolo Abeni <pabeni@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When a device address is about to be changed, or an address added to the
list of device HW addresses, it is necessary to ensure that all
interested parties can support the address. Therefore, send the
NETDEV_PRE_CHANGEADDR notification, and if anyone bails on it, do not
change the address.
Signed-off-by: NPetr Machata <petrm@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Reviewed-by: NIdo Schimmel <idosch@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The NETDEV_CHANGEADDR notification is emitted after a device address
changes. Extending this message to allow vetoing is certainly possible,
but several other notification types have instead adopted a simple
two-stage approach: first a "pre" notification is sent to make sure all
interested parties are OK with a change that's about to be done. Then
the change is done, and afterwards a "post" notification is sent.

This dual approach is easier to use: when the change is vetoed, nothing
has changed yet, and it's therefore unnecessary to roll anything back.
Therefore adopt it for NETDEV_CHANGEADDR as well.

To that end, add NETDEV_PRE_CHANGEADDR and an info structure to go along
with it.
Signed-off-by: NPetr Machata <petrm@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Reviewed-by: NIdo Schimmel <idosch@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

A follow-up patch will add a notifier type NETDEV_PRE_CHANGEADDR, which
allows vetoing of MAC address changes. One prominent path to that
notification is through dev_set_mac_address(). Therefore give this
function an extack argument, so that it can be packed together with the
notification. Thus a textual reason for rejection (or a warning) can be
communicated back to the user.
Signed-off-by: NPetr Machata <petrm@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Reviewed-by: NIdo Schimmel <idosch@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Drivers may need to validate configuration of a device that's about to
be upped. Should the validation fail, there's currently no way to
communicate details of the failure to the user, beyond an error number.

To mend that, change __dev_open() to take an extack argument and pass it
from __dev_change_flags() and dev_open(), where it was propagated in the
previous patches.

Change __dev_open() to call call_netdevice_notifiers_extack() so that
the passed-in extack is attached to the NETDEV_PRE_UP notifier.
Signed-off-by: NPetr Machata <petrm@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Reviewed-by: NIdo Schimmel <idosch@mellanox.com>
Reviewed-by: NDavid Ahern <dsahern@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In order to propagate extack through NETDEV_PRE_UP, add a new function
call_netdevice_notifiers_extack() that primes the extack field of the
notifier info. Convert call_netdevice_notifiers() to a simple wrapper
around the new function that passes NULL for extack.
Signed-off-by: NPetr Machata <petrm@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Reviewed-by: NIdo Schimmel <idosch@mellanox.com>
Reviewed-by: NDavid Ahern <dsahern@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In order to pass extack together with NETDEV_PRE_UP notifications, it's
necessary to route the extack to __dev_open() from diverse (possibly
indirect) callers. The last missing API is __dev_change_flags().

Therefore extend __dev_change_flags() with and extra extack argument and
update the two existing users.

Since the function declaration line is changed anyway, name the struct
net_device argument to placate checkpatch.
Signed-off-by: NPetr Machata <petrm@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Reviewed-by: NIdo Schimmel <idosch@mellanox.com>
Reviewed-by: NDavid Ahern <dsahern@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In order to pass extack together with NETDEV_PRE_UP notifications, it's
necessary to route the extack to __dev_open() from diverse (possibly
indirect) callers. One prominent API through which the notification is
invoked is dev_change_flags().

Therefore extend dev_change_flags() with and extra extack argument and
update all users. Most of the calls end up just encoding NULL, but
several sites (VLAN, ipvlan, VRF, rtnetlink) do have extack available.

Since the function declaration line is changed anyway, name the other
function arguments to placate checkpatch.
Signed-off-by: NPetr Machata <petrm@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Reviewed-by: NIdo Schimmel <idosch@mellanox.com>
Reviewed-by: NDavid Ahern <dsahern@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In order to pass extack together with NETDEV_PRE_UP notifications, it's
necessary to route the extack to __dev_open() from diverse (possibly
indirect) callers. One prominent API through which the notification is
invoked is dev_open().

Therefore extend dev_open() with and extra extack argument and update
all users. Most of the calls end up just encoding NULL, but bond and
team drivers have the extack readily available.
Signed-off-by: NPetr Machata <petrm@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Reviewed-by: NIdo Schimmel <idosch@mellanox.com>
Reviewed-by: NDavid Ahern <dsahern@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

list_del() leaves the skb->next pointer poisoned, which can then lead to
 a crash in e.g. OVS forwarding.  For example, setting up an OVS VXLAN
 forwarding bridge on sfc as per:

========
$ ovs-vsctl show
5dfd9c47-f04b-4aaa-aa96-4fbb0a522a30
    Bridge "br0"
        Port "br0"
            Interface "br0"
                type: internal
        Port "enp6s0f0"
            Interface "enp6s0f0"
        Port "vxlan0"
            Interface "vxlan0"
                type: vxlan
                options: {key="1", local_ip="10.0.0.5", remote_ip="10.0.0.4"}
    ovs_version: "2.5.0"
========
(where 10.0.0.5 is an address on enp6s0f1)
and sending traffic across it will lead to the following panic:
========
general protection fault: 0000 [#1] SMP PTI
CPU: 5 PID: 0 Comm: swapper/5 Not tainted 4.20.0-rc3-ehc+ #701
Hardware name: Dell Inc. PowerEdge R710/0M233H, BIOS 6.4.0 07/23/2013
RIP: 0010:dev_hard_start_xmit+0x38/0x200
Code: 53 48 89 fb 48 83 ec 20 48 85 ff 48 89 54 24 08 48 89 4c 24 18 0f 84 ab 01 00 00 48 8d 86 90 00 00 00 48 89 f5 48 89 44 24 10 <4c> 8b 33 48 c7 03 00 00 00 00 48 8b 05 c7 d1 b3 00 4d 85 f6 0f 95
RSP: 0018:ffff888627b437e0 EFLAGS: 00010202
RAX: 0000000000000000 RBX: dead000000000100 RCX: ffff88862279c000
RDX: ffff888614a342c0 RSI: 0000000000000000 RDI: 0000000000000000
RBP: ffff888618a88000 R08: 0000000000000001 R09: 00000000000003e8
R10: 0000000000000000 R11: ffff888614a34140 R12: 0000000000000000
R13: 0000000000000062 R14: dead000000000100 R15: ffff888616430000
FS:  0000000000000000(0000) GS:ffff888627b40000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007f6d2bc6d000 CR3: 000000000200a000 CR4: 00000000000006e0
Call Trace:
 <IRQ>
 __dev_queue_xmit+0x623/0x870
 ? masked_flow_lookup+0xf7/0x220 [openvswitch]
 ? ep_poll_callback+0x101/0x310
 do_execute_actions+0xaba/0xaf0 [openvswitch]
 ? __wake_up_common+0x8a/0x150
 ? __wake_up_common_lock+0x87/0xc0
 ? queue_userspace_packet+0x31c/0x5b0 [openvswitch]
 ovs_execute_actions+0x47/0x120 [openvswitch]
 ovs_dp_process_packet+0x7d/0x110 [openvswitch]
 ovs_vport_receive+0x6e/0xd0 [openvswitch]
 ? dst_alloc+0x64/0x90
 ? rt_dst_alloc+0x50/0xd0
 ? ip_route_input_slow+0x19a/0x9a0
 ? __udp_enqueue_schedule_skb+0x198/0x1b0
 ? __udp4_lib_rcv+0x856/0xa30
 ? __udp4_lib_rcv+0x856/0xa30
 ? cpumask_next_and+0x19/0x20
 ? find_busiest_group+0x12d/0xcd0
 netdev_frame_hook+0xce/0x150 [openvswitch]
 __netif_receive_skb_core+0x205/0xae0
 __netif_receive_skb_list_core+0x11e/0x220
 netif_receive_skb_list+0x203/0x460
 ? __efx_rx_packet+0x335/0x5e0 [sfc]
 efx_poll+0x182/0x320 [sfc]
 net_rx_action+0x294/0x3c0
 __do_softirq+0xca/0x297
 irq_exit+0xa6/0xb0
 do_IRQ+0x54/0xd0
 common_interrupt+0xf/0xf
 </IRQ>
========
So, in all listified-receive handling, instead pull skbs off the lists with
 skb_list_del_init().

Fixes: 9af86f93 ("net: core: fix use-after-free in __netif_receive_skb_list_core")
Fixes: 7da517a3 ("net: core: Another step of skb receive list processing")
Fixes: a4ca8b7d ("net: ipv4: fix drop handling in ip_list_rcv() and ip_list_rcv_finish()")
Fixes: d8269e2c ("net: ipv6: listify ipv6_rcv() and ip6_rcv_finish()")
Signed-off-by: NEdward Cree <ecree@solarflare.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

netif_napi_add() could report an error like this below due to it allows
to pass a format string for wildcarding before calling
dev_get_valid_name(),

"netif_napi_add() called with weight 256 on device eth%d"

For example, hns_enet_drv module does this.

hns_nic_try_get_ae
  hns_nic_init_ring_data
    netif_napi_add
  register_netdev
    dev_get_valid_name

Hence, make it a bit more human-readable by using netdev_err_once()
instead.
Signed-off-by: NQian Cai <cai@gmx.us>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Trace events are already present for the receive entry points, to indicate
how the reception entered the stack.

This patch adds the corresponding exit trace events that will bound the
reception such that all events occurring between the entry and the exit
can be considered as part of the reception context. This greatly helps
for dependency and root cause analyses.

Without this, it is not possible with tracepoint instrumentation to
determine whether a sched_wakeup event following a netif_receive_skb
event is the result of the packet reception or a simple coincidence after
further processing by the thread. It is possible using other mechanisms
like kretprobes, but considering the "entry" points are already present,
it would be good to add the matching exit events.

In addition to linking packets with wakeups, the entry/exit event pair
can also be used to perform network stack latency analyses.
Signed-off-by: NGeneviève Bastien <gbastien@versatic.net>
CC: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
CC: Steven Rostedt <rostedt@goodmis.org>
CC: Ingo Molnar <mingo@redhat.com>
CC: David S. Miller <davem@davemloft.net>
Reviewed-by: Steven Rostedt (VMware) <rostedt@goodmis.org> (tracing side)
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: NCong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit 04157469 ("net: Use static_key for XPS maps") introduced a
static key for XPS, but the increments/decrements don't match.

First, the static key's counter is incremented once for each queue, but
only decremented once for a whole batch of queues, leading to large
unbalances.

Second, the xps_rxqs_needed key is decremented whenever we reset a batch
of queues, whether they had any rxqs mapping or not, so that if we setup
cpu-XPS on em1 and RXQS-XPS on em2, resetting the queues on em1 would
decrement the xps_rxqs_needed key.

This reworks the accounting scheme so that the xps_needed key is
incremented only once for each type of XPS for all the queues on a
device, and the xps_rxqs_needed key is incremented only once for all
queues. This is sufficient to let us retrieve queues via
get_xps_queue().

This patch introduces a new reset_xps_maps(), which reinitializes and
frees the appropriate map (xps_rxqs_map or xps_cpus_map), and drops a
reference to the needed keys:
 - both xps_needed and xps_rxqs_needed, in case of rxqs maps,
 - only xps_needed, in case of CPU maps.

Now, we also need to call reset_xps_maps() at the end of
__netif_set_xps_queue() when there's no active map left, for example
when writing '00000000,00000000' to all queues' xps_rxqs setting.

Fixes: 04157469 ("net: Use static_key for XPS maps")
Signed-off-by: NSabrina Dubroca <sd@queasysnail.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Before commit 80d19669 ("net: Refactor XPS for CPUs and Rx queues"),
netif_reset_xps_queues() did netdev_queue_numa_node_write() for all the
queues being reset. Now, this is only done when the "active" variable in
clean_xps_maps() is false, ie when on all the CPUs, there's no active
XPS mapping left.

Fixes: 80d19669 ("net: Refactor XPS for CPUs and Rx queues")
Signed-off-by: NSabrina Dubroca <sd@queasysnail.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We very often have few flows/chains to look at, and we
might increase GRO_HASH_BUCKETS to 32 or 64 in the future.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Eric noted that with UDP GRO and NAPI timeout, we could keep a single
UDP packet inside the GRO hash forever, if the related NAPI instance
calls napi_gro_complete() at an higher frequency than the NAPI timeout.
Willem noted that even TCP packets could be trapped there, till the
next retransmission.
This patch tries to address the issue, flushing the old packets -
those with a NAPI_GRO_CB age before the current jiffy - before scheduling
the NAPI timeout. The rationale is that such a timeout should be
well below a jiffy and we are not flushing packets eligible for sane GRO.

v1  -> v2:
 - clarified the commit message and comment

RFC -> v1:
 - added 'Fixes tags', cleaned-up the wording.
Reported-by: NEric Dumazet <eric.dumazet@gmail.com>
Fixes: 3b47d303 ("net: gro: add a per device gro flush timer")
Signed-off-by: NPaolo Abeni <pabeni@redhat.com>
Acked-by: NWillem de Bruijn <willemb@google.com>
Acked-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

eth_type_trans() assumes initial value for skb->pkt_type
is PACKET_HOST.

This is indeed the value right after a fresh skb allocation.

However, it is possible that GRO merged a packet with a different
value (like PACKET_OTHERHOST in case macvlan is used), so
we need to make sure napi->skb will have pkt_type set back to
PACKET_HOST.

Otherwise, valid packets might be dropped by the stack because
their pkt_type is not PACKET_HOST.

napi_reuse_skb() was added in commit 96e93eab ("gro: Add
internal interfaces for VLAN"), but this bug always has
been there.

Fixes: 96e93eab ("gro: Add internal interfaces for VLAN")
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Currently netdev_rx_csum_fault() only shows a device name,
we need more information about the skb for debugging csum
failures.

Sample output:

 ens3: hw csum failure
 dev features: 0x0000000000014b89
 skb len=84 data_len=0 pkt_type=0 gso_size=0 gso_type=0 nr_frags=0 ip_summed=0 csum=0 csum_complete_sw=0 csum_valid=0 csum_level=0

Note, I use pr_err() just to be consistent with the existing one.
Signed-off-by: NCong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This removes assumptions about VLAN_TAG_PRESENT bit.
Signed-off-by: NMichał Mirosław <mirq-linux@rere.qmqm.pl>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Before calling dev_hard_start_xmit(), upper layers tried
to cook optimal skb list based on BQL budget.

Problem is that GSO packets can end up comsuming more than
the BQL budget.

Breaking the loop is not useful, since requeued packets
are ahead of any packets still in the qdisc.

It is also more expensive, since next TX completion will
push these packets later, while skbs are not in cpu caches.

It is also a behavior difference with TSO packets, that can
break the BQL limit by a large amount.

Note that drivers should use __netdev_tx_sent_queue()
in order to have optimal xmit_more support, and avoid
useless atomic operations as shown in the following patch.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Just like with normal GRO processing, we have to initialize
skb->next to NULL when we unlink overflow packets from the
GRO hash lists.

Fixes: d4546c25 ("net: Convert GRO SKB handling to list_head.")
Reported-by: NOleksandr Natalenko <oleksandr@natalenko.name>
Tested-by: NOleksandr Natalenko <oleksandr@natalenko.name>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

DEC FDDIcontroller 700 (DEFZA) uses a Tx/Rx queue pair to communicate
SMT frames with adapter's firmware.  Any SMT frame received from the RMC
via the Rx queue is queued back by the driver to the SMT Rx queue for
the firmware to process.  Similarly the firmware uses the SMT Tx queue
to supply the driver with SMT frames which are queued back to the Tx
queue for the RMC to send to the ring.

When a network tap is attached to an FDDI interface handled by `defza'
any incoming SMT frames captured are queued to our usual processing of
network data received, which in turn delivers them to any listening
taps.

However the outgoing SMT frames produced by the firmware bypass our
network protocol stack and are therefore not delivered to taps.  This in
turn means that taps are missing a part of network traffic sent by the
adapter, which may make it more difficult to track down network problems
or do general traffic analysis.

Call `dev_queue_xmit_nit' then in the SMT Tx path, having checked that
a network tap is attached, with a newly-created `dev_nit_active' helper
wrapping the usual condition used in the transmit path.
Signed-off-by: NMaciej W. Rozycki <macro@linux-mips.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since commit 5aad1de5 ("ipv4: use separate genid for next hop
exceptions"), exceptions get deprecated separately from cached
routes. In particular, administrative changes don't clear PMTU anymore.

As Stefano described in commit e9fa1495 ("ipv6: Reflect MTU changes
on PMTU of exceptions for MTU-less routes"), the PMTU discovered before
the local MTU change can become stale:
 - if the local MTU is now lower than the PMTU, that PMTU is now
   incorrect
 - if the local MTU was the lowest value in the path, and is increased,
   we might discover a higher PMTU

Similarly to what commit e9fa1495 did for IPv6, update PMTU in those
cases.

If the exception was locked, the discovered PMTU was smaller than the
minimal accepted PMTU. In that case, if the new local MTU is smaller
than the current PMTU, let PMTU discovery figure out if locking of the
exception is still needed.

To do this, we need to know the old link MTU in the NETDEV_CHANGEMTU
notifier. By the time the notifier is called, dev->mtu has been
changed. This patch adds the old MTU as additional information in the
notifier structure, and a new call_netdevice_notifiers_u32() function.

Fixes: 5aad1de5 ("ipv4: use separate genid for next hop exceptions")
Signed-off-by: NSabrina Dubroca <sd@queasysnail.net>
Reviewed-by: NStefano Brivio <sbrivio@redhat.com>
Reviewed-by: NDavid Ahern <dsahern@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

XDP can modify (and resize) the Ethernet header in the packet.

There is a bug in generic-XDP, because skb->protocol and skb->pkt_type
are setup before reaching (netif_receive_)generic_xdp.

This bug was hit when XDP were popping VLAN headers (changing
eth->h_proto), as skb->protocol still contains VLAN-indication
(ETH_P_8021Q) causing invocation of skb_vlan_untag(skb), which corrupt
the packet (basically popping the VLAN again).

This patch catch if XDP changed eth header in such a way, that SKB
fields needs to be updated.

V2: on request from Song Liu, use ETH_HLEN instead of mac_len,
in __skb_push() as eth_type_trans() use ETH_HLEN in paired skb_pull_inline().

Fixes: d4455169 ("net: xdp: support xdp generic on virtual devices")
Signed-off-by: NJesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: NAlexei Starovoitov <ast@kernel.org>

It documents what is happening, and eliminates the spurious list
pointer poisoning.

In the long term, in order to get proper list head debugging, we
might want to use the list poison value as the indicator that
an SKB is a singleton and not on a list.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

An SKB is not on a list if skb->next is NULL.

Codify this convention into a helper function and use it
where we are dequeueing an SKB and need to mark it as such.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>