simple .o parser and loader using BPF syscall.
.o is a standard ELF generated by LLVM backend

It parses elf file compiled by llvm .c->.o
- parses 'maps' section and creates maps via BPF syscall
- parses 'license' section and passes it to syscall
- parses elf relocations for BPF maps and adjusts BPF_LD_IMM64 insns
  by storing map_fd into insn->imm and marking such insns as BPF_PSEUDO_MAP_FD
- loads eBPF programs via BPF syscall

One ELF file can contain multiple BPF programs.

int load_bpf_file(char *path);
populates prog_fd[] and map_fd[] with FDs received from bpf syscall

bpf_helpers.h - helper functions available to eBPF programs written in C
Signed-off-by: NAlexei Starovoitov <ast@plumgrid.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

this socket filter example does:
- creates arraymap in kernel with key 4 bytes and value 8 bytes

- loads eBPF program which assumes that packet is IPv4 and loads one byte of
  IP->proto from the packet and uses it as a key in a map

  r0 = skb->data[ETH_HLEN + offsetof(struct iphdr, protocol)];
  *(u32*)(fp - 4) = r0;
  value = bpf_map_lookup_elem(map_fd, fp - 4);
  if (value)
       (*(u64*)value) += 1;

- attaches this program to raw socket

- every second user space reads map[IPPROTO_TCP], map[IPPROTO_UDP], map[IPPROTO_ICMP]
  to see how many packets of given protocol were seen on loopback interface

Usage:
$sudo samples/bpf/sock_example
TCP 0 UDP 0 ICMP 0 packets
TCP 187600 UDP 0 ICMP 4 packets
TCP 376504 UDP 0 ICMP 8 packets
TCP 563116 UDP 0 ICMP 12 packets
TCP 753144 UDP 0 ICMP 16 packets
Signed-off-by: NAlexei Starovoitov <ast@plumgrid.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

introduce new setsockopt() command:

setsockopt(sock, SOL_SOCKET, SO_ATTACH_BPF, &prog_fd, sizeof(prog_fd))

where prog_fd was received from syscall bpf(BPF_PROG_LOAD, attr, ...)
and attr->prog_type == BPF_PROG_TYPE_SOCKET_FILTER

setsockopt() calls bpf_prog_get() which increments refcnt of the program,
so it doesn't get unloaded while socket is using the program.

The same eBPF program can be attached to multiple sockets.

User task exit automatically closes socket which calls sk_filter_uncharge()
which decrements refcnt of eBPF program
Signed-off-by: NAlexei Starovoitov <ast@plumgrid.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

introduce program type BPF_PROG_TYPE_SOCKET_FILTER that is used
for attaching programs to sockets where ctx == skb.

add verifier checks for ABS/IND instructions which can only be seen
in socket filters, therefore the check:
  if (env->prog->aux->prog_type != BPF_PROG_TYPE_SOCKET_FILTER)
    verbose("BPF_LD_ABS|IND instructions are only allowed in socket filters\n");
Signed-off-by: NAlexei Starovoitov <ast@plumgrid.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

To be more friendly with drop monitor, we should only call kfree_skb() when
the packets were dropped and use consume_skb() in other cases.

Cc: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NJason Wang <jasowang@redhat.com>
Acked-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The pci_dev_put() function tests whether its argument is NULL
and then returns immediately. Thus the test around the call
is not needed.

This issue was detected by using the Coccinelle software.
Signed-off-by: NMarkus Elfring <elfring@users.sourceforge.net>
Acked-by: NOlof Johansson <olof@lixom.net>
Acked-by: NLuis R. Rodriguez <mcgrof@suse.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The free_percpu() function tests whether its argument is NULL and then
returns immediately. Thus the test around the call is not needed.

This issue was detected by using the Coccinelle software.
Signed-off-by: NMarkus Elfring <elfring@users.sourceforge.net>
Acked-by: NMahesh Bandewar <maheshb@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The vfree() function performs also input parameter validation.
Thus the test around the call is not needed.

This issue was detected by using the Coccinelle software.
Signed-off-by: NMarkus Elfring <elfring@users.sourceforge.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Instead of using global variables we are going to use dynamically allocated
memory. It allows to append a support of more than one ethernet adapter which
might have different settings simultaniously.
Signed-off-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Pablo Neira Ayuso says:

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

The following batch contains netfilter updates for net-next. Basically,
enhancements for xt_recent, skip zeroing of timer in conntrack, fix
linking problem with recent redirect support for nf_tables, ipset
updates and a couple of cleanups. More specifically, they are:

1) Rise maximum number per IP address to be remembered in xt_recent
   while retaining backward compatibility, from Florian Westphal.

2) Skip zeroing timer area in nf_conn objects, also from Florian.

3) Inspect IPv4 and IPv6 traffic from the bridge to allow filtering using
   using meta l4proto and transport layer header, from Alvaro Neira.

4) Fix linking problems in the new redirect support when CONFIG_IPV6=n
   and IP6_NF_IPTABLES=n.

And ipset updates from Jozsef Kadlecsik:

5) Support updating element extensions when the set is full (fixes
   netfilter bugzilla id 880).

6) Fix set match with 32-bits userspace / 64-bits kernel.

7) Indicate explicitly when /0 networks are supported in ipset.

8) Simplify cidr handling for hash:*net* types.

9) Allocate the proper size of memory when /0 networks are supported.

10) Explicitly add padding elements to hash:net,net and hash:net,port,
    because the elements must be u32 sized for the used hash function.

Jozsef is also cooking ipset RCU conversion which should land soon if
they reach the merge window in time.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Jeff Kirsher says:

====================
Intel Wired LAN Driver Updates 2014-12-05

This series contains updates to ixgbe and ixgbevf.

Alex provides a couple of patches to cleanup ixgbe.  First cleans up the
page reuse code getting it into a state where all the workarounds needed
are in place as well as cleaning up a few minor oversights such as using
__free_pages instead of put_page to drop a locally allocated page.  Then
cleans up the tail writes for the ixgbe descriptor queues.

Mark Peterson adds support to lookup MAC addresses in Open Firmware or
IDPROM.

Emil provides patches for ixgbe and ixgbevf to fix an issue on rmmod and
to add support for X550 in the VF driver.  First removes the read/write
operations to the CIAA/D registers since it can block access to the PCI
config space and make use of standard kernel functions for accessing the
PCI config space.  Then fixes an issue where the driver has logic to free
up used data in case any of the checks in ixgbe_probe() fail, however
there is a similar set of cleanups that can occur on driver unload in
ixgbe_remove() which can cause the rmmod command to crash.

Don provides the remaining patches in the series to complete the addition
of X550 support into the ixgbe driver.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch resolves couple of issues in ixgbevf_probe/remove():

1. Fix a case where adapter->state is tested after free_netdev() this is
same as the patch for ixgbe from Daniel Borkmann <dborkman@redhat.com>:
commit b5b2ffc0 ("ixgbe: fix use after free adapter->state test in ixgbe_remove/ixgbe_probe")

2. Move pci_set_drvdata() after all the error checks in ixgbevf_probe() and
then add a check in ixgbevf_probe() to avoid running the cleanup functions
twice in cases where probe failed.

CC: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: NEmil Tantilov <emil.s.tantilov@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch adds initial support for VFs on a new mac - X550.

The patch adds the basic structures and device IDs for the X550 VFs
that would allow the driver to load and pass traffic.
Signed-off-by: NEmil Tantilov <emil.s.tantilov@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The driver has logic to free up used data in case any of the checks in
ixgbe_probe() fail, however there is a similar set of cleanups that can
occur on driver unload in ixgbe_remove() which can cause the rmmod command
to crash.

This patch aims to fix the logic by moving pci_set_drvdata() after all error
checks and then adds a check in ixgbe_remove() to skip it altogether if
adapter comes up empty.
Signed-off-by: NEmil Tantilov <emil.s.tantilov@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Since we now support X550 mac's bump the version number to reflect this.
Signed-off-by: NDon Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch extends the function pointer structure to include the new
X550 class MAC types. This creates a new file ixgbe_x550.c that contains
all of the new methods.  Because of similarities to the X540 part in
some cases we just use it's methods where they can be used without any
modification.  These exported functions are now defined in the new
ixgbe_x540.h file.
Signed-off-by: NDon Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Currently the shared code checksum calculation function only
returns a u16 and cannot return an error code. Unfortunately
a variety of errors can happen that completely prevent the
calculation of a checksum. So, change the function return value
from a u16 to an s32 and return a negative value on error, or the
positive checksum value when there is no error.
Signed-off-by: NDon Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Some X550 procedures will be using CS4227 PHY and need to
perform combined read and write operations.  This patch
adds those methods.
Signed-off-by: NDon Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The X550 hardware will use more bits in the mask, so change
the prototypes to match.  This larger mask will require changes
in callers which use the higher bits. Likewise since X550 will
use different semaphore mask values and will use the lan_id
value.  So save these values in the ixgbe_phy_info struct.
Signed-off-by: NDon Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Since on X550 we use host interface commands to read,write and erase
some commands require more time to complete. So this adds a timeout
parameter to ixgbe_host_interface_command as wells as a return_data
parameter allowing us to return with any data.
Signed-off-by: NDon Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The new X550 family of MAC's will have a larger RSS hash (16 -> 64).
It will also support individual VF to have their own independent RSS
hash key.  This patch will enable this functionality
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>

Accessing the CIAA/D register can block access to the PCI config space.

This patch removes the read/write operations to the CIAA/D registers
and makes use of standard kernel functions for accessing the PCI config
space.

In addition it moves ixgbevf_check_for_bad_vf() into the watchdog subtask
which reduces the frequency of the checks.

CC: Alex Williamson <alex.williamson@redhat.com>
Reported-by: NAlex Williamson <alex.williamson@redhat.com>
Signed-off-by: NEmil Tantilov <emil.s.tantilov@intel.com>
Tested-by: NAaron Brown <aaron.f.brown@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Attempt to look up the MAC address in Open Firmware on systems that
support it. On SPARC resort to using the IDPROM if no OF address is
found.
Signed-off-by: NMartin K Petersen <martin.petersen@oracle.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This change cleans up the tail writes for the ixgbe descriptor queues.  The
current implementation had me confused as I wasn't sure if it was still
making use of the surprise remove logic or not.

It also adds the mmiowb which is needed on ia64, mips, and a couple other
architectures in order to synchronize the MMIO writes with the Tx queue
_xmit_lock spinlock.

Cc: Don Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: NAlexander Duyck <alexander.h.duyck@redhat.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch cleans up the page reuse code getting it into a state where all
the workarounds needed are in place as well as cleaning up a few minor
oversights such as using __free_pages instead of put_page to drop a locally
allocated page.

It also cleans up how we clear the descriptor status bits.  Previously they
were zeroed as a part of clearing the hdr_addr.  However the hdr_addr is a
64 bit field and 64 bit writes can be a bit more expensive on on 32 bit
systems.  Since we are no longer using the header split feature the upper
32 bits of the address no longer need to be cleared.  As a result we can
just clear the status bits and leave the length and VLAN fields as-is which
should provide more information in debugging.

Cc: Don Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: NAlexander Duyck <alexander.h.duyck@redhat.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

The elements must be u32 sized for the used hash function.
Signed-off-by: NJozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

Signed-off-by: NJozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

Signed-off-by: NJozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

Signed-off-by: NJozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

Sven-Haegar Koch reported the issue:

sims:~# iptables -A OUTPUT -m set --match-set testset src -j ACCEPT
iptables: Invalid argument. Run `dmesg' for more information.

In syslog:
x_tables: ip_tables: set.3 match: invalid size 48 (kernel) != (user) 32

which was introduced by the counter extension in ipset.

The patch fixes the alignment issue with introducing a new set match
revision with the fixed underlying 'struct ip_set_counter_match'
structure.
Signed-off-by: NJozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

When the set was full (hash type and maxelem reached), it was not
possible to update the extension part of already existing elements.
The patch removes this limitation.

Fixes: https://bugzilla.netfilter.org/show_bug.cgi?id=880Signed-off-by: NJozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

When we write the GSO meta-data in tun_get_user we end up advancing
the IO vector twice, thus exhausting the user buffer before we can
finish writing the packet.

Fixes: f5ff53b4 ("{macvtap,tun}_get_user(): switch to iov_iter")
Reported-by: NMarcelo Ricardo Leitner <mleitner@redhat.com>
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>
Acked-by: NJason Wang <jasowang@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Jiri Pirko says:

====================
introduce rocker switch driver with hardware accelerated datapath api - phase 1: bridge fdb offload

This patchset is just the first phase of switch and switch-ish device
support api in kernel. Note that the api will extend.

So what this patchset includes:
- introduce switchdev api skeleton for implementing switch drivers
- introduce rocker switch driver which implements switchdev api fdb and
  bridge set/get link ndos

As to the discussion if there is need to have specific class of device
representing the switch itself, so far we found no need to introduce that.
But we are generally ok with the idea and when the time comes and it will
be needed, it can be easily introduced without any disturbance.

This patchset introduces switch id export through rtnetlink and sysfs,
which is similar to what we have for port id in SR-IOV. I will send iproute2
patchset for showing the switch id for port netdevs once this is applied.
This applies also for the PF_BRIDGE and fdb iproute2 patches.

iproute2 patches are now available here:
https://github.com/jpirko/iproute2-rocker

For detailed description and version history, please see individual patches.

In v4 I reordered the patches leaving rocker patches on the end of the patchset.

In v5 I only fixed whitespace issues of patch #13

We have a TODO for related items we want to work on in near future:
https://etherpad.wikimedia.org/p/netdev-swdev-todo
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Silences various sparse warnings
Signed-off-by: NThomas Graf <tgraf@suug.ch>
Signed-off-by: NJiri Pirko <jiri@resnulli.us>
Signed-off-by: NScott Feldman <sfeldma@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NThomas Graf <tgraf@suug.ch>
Signed-off-by: NJiri Pirko <jiri@resnulli.us>
Signed-off-by: NScott Feldman <sfeldma@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Rocker ports will use new "swdev" hwmode for bridge port offload policy.
Current supported policy settings are BR_LEARNING and BR_LEARNING_SYNC.
User can turn on/off device port FDB learning and syncing to bridge.
Signed-off-by: NScott Feldman <sfeldma@gmail.com>
Signed-off-by: NJiri Pirko <jiri@resnulli.us>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NJiri Pirko <jiri@resnulli.us>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add L2 bridge offloading support to rocker driver.  Here, the Linux bridge
driver is used to collect swdev ports into a tagged (or untagged) VLAN
bridge.  The switchdev will offload from the bridge driver the following L2
bridging functions:

 - Learning of neighbor MAC addresses on VLAN X  Learned mac/vlan is
installed in bridge FDB.  (And removed when device unlearns mac/vlan).
Learning must be turned off on each bridge port to disable the feature in
the bridge driver.

- Flooding of multicast/broadcast and unknown unicast pkts to (STP)
active ports in bridge.  The bridge driver is unaware of the flooding happening
at the device level.  Flooding must be turned off on each bridge port to
disable the feature on the bridge driver.

- STP port state is pushed down to driver/device.  The bridge still processes
STP BDPUs and maintains port STP state (for all VLANs in bridge), but
the driver/device must be notified of port STP state change to program
the device.

Multiple (VLAN) bridges are supported.  The device (implemented per
the OF-DPA spec) must use a portion of the VLAN namespace for
internal VLANs.  Right now, the upper 255 VLANs (0xf00 to 0xffe) are
used as internal VLAN IDs for untagged traffic and are not available
as port VLANs.

The driver uses the following interfaces:

1. To track VLAN add/del on ports in bridge:

.ndo_vlan_rx_add_vid
.ndo_vlan_rx_kill_vid

2. To track port add/del membership in bridge:

NETDEV_CHANGEUPPER netdevice notifier

3. To catch static FDB entries installed on bridge/vlan by user using netlink:

.ndo_fdb_add
.ndo_fdb_del

4. To be notified on port STP state change:

.ndo_switch_port_stp_update

5. To notify bridge driver on learned/forgotten mac/vlans on bridge port:

br_fdb_external_learn_add
br_fdb_external_learn_del
Signed-off-by: NScott Feldman <sfeldma@gmail.com>
Signed-off-by: NJiri Pirko <jiri@resnulli.us>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The rocker driver maintains 4 hash tables: flows, groups, FDB, and VLANs.

Flow and group tables track the entries installed to OF-DPA tables,
per the OF-DPA spec.  See OF-DPA spec for full description of fields
in each flow and group table.  New table entries are pushed to the
device with ADD cmd.  Updated entries are pushed to the device with
MOD cmd.  For flow table entries, a crc32 key is made from fields of
the particular field.  For group table entries, the group_id is used
as the key.

The FDB table tracks fdb entries learned by the device or manually
pushed to the bridge by the user.  A crc32 key is made from the
port/mac/vlan tuple for the fdb entry.

The VLAN table tracks the ifindex-to-internal-vlan mapping for
untagged pkts.  On ingress, an untagged pkt is inserted with an
internal VLAN ID based on the input port's current internal VLAN ID.
The input port's internal VLAN will either be referenced by the port's
ifindex, if not bridged, or the containing bridge's ifindex, if
bridged.  Since the ifindex space isn't within a fixed range, uses a
hash table (with ifindex as key) to track internal VLAN ID for a given
ifindex.  The internal VLAN ID range is fixed and currently uses the
upper 255 VLAN IDs, starting at 0xf00.
Signed-off-by: NScott Feldman <sfeldma@gmail.com>
Signed-off-by: NJiri Pirko <jiri@resnulli.us>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch introduces the first driver to benefit from the switchdev
infrastructure and to implement newly introduced switch ndos. This is a
driver for emulated switch chip implemented in qemu:
https://github.com/sfeldma/qemu-rocker/

This patch is a result of joint work with Scott Feldman.
Signed-off-by: NScott Feldman <sfeldma@gmail.com>
Signed-off-by: NJiri Pirko <jiri@resnulli.us>
Reviewed-by: NThomas Graf <tgraf@suug.ch>
Reviewed-by: NJohn Fastabend <john.r.fastabend@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>