Refactor how we add new filters to firmware to avoid a race condition
that can occur due to removing filters from the hash temporarily.

To understand the race condition, suppose that you have a number of MAC
filters, but have not yet added any VLANs. Now, add two VLANs in rapid
succession. A possible resulting flow would look something like the
following:

(1) lock hash for add VLAN
(2) add the new MAC/VLAN combos for each current MAC filter
(3) unlock hash
(4) lock hash for filter sync
(5) notice that we have a VLAN, so prepare to update all MAC filters
    with VLAN=-1 to be VLAN=0.
(6) move NEW and REMOVE filters to temporary list
(7) unlock hash
(8) lock hash for add VLAN
(9) add new MAC/VLAN combos. Notice that no MAC filters are currently in
    the hash list, so we don't add any VLANs <--- BUG!
(10) unlock hash
(11) sync the temporary lists to firmware
(12) lock hash for post-sync
(13) move the temporary elements back to the main list
....

Because we take filters out of the main hash into temporary lists, we
introduce a narrow window where it is possible that other callers to the
list will not see some of the filters which were previously added but
have not yet been finalized. This results in sometimes dropping VLAN
additions, and could also result in failing to add a MAC address on the
newly added VLAN.

One obvious way to avoid this race condition would be to lock the entire
firmware process. Unfortunately this does not work because adminq
firmware commands take a mutex which results in a sleep while atomic
BUG(). So, we can't use the simplest approach.

An alternative approach is to simply not remove the filters from the
hash list while adding. Instead, add an i40e_new_mac_filter structure
which we will use to track added filters. This avoids the need to remove
the filter from the hash list. We'll store a pointer to the original
i40e_mac_filter, along with our own copy of the state.

We won't update the state directly, so as to avoid race with other code
that may modify the state while under the lock. We are safe to read
f->macaddr and f->vlan since these only change in two locations. The
first is on filter creation, which must have already occurred. The
second is inside i40e_correct_vlan_filters which was previously run
after creation of this object and can't be run again until after. Thus,
we should be safe to read the MAC address and VLAN while outside the
lock.

We also aren't going to run into a use-after-free issue because the only
place where we free filters is when they are marked FAILED or when we
remove them inside the sync subtask. Since the subtask has its own
critical flag to prevent duplicate runs, we know this won't happen. We
also know that the only location to transition a filter from NEW to
FAILED is inside the subtask also, so we aren't worried about that
either.

Use the wrapper i40e_new_mac_filter for additions, and once we've
finalized the addition to firmware, we will update the filter state
inside a lock, and then free the wrapper structure.

In order to avoid a possible race condition with filter deletion, we
won't update the original filter state unless it is still
I40E_FILTER_NEW when we finish the firmware sync.

This approach is more complex, but avoids race conditions related to
filters being temporarily removed from the list. We do not need the same
behavior for deletion because we always unconditionally removed the
filters from the list regardless of the firmware status.

Change-Id: I14b74bc2301f8e69433fbe77ebca532db20c5317
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Fix a bug where we modified the mac_filter_hash while outside a lock,
when handling addition of broadcast filters.

Normally, we add filters to firmware by batching the additions into
lists and issuing 1 update for every few filters. Broadcast filters are
handled differently, by instead setting the broadcast promiscuous mode
flags. In order to make sure the 1<->1 mapping of filters in our
addition array lined up with filters in the hlist tmp_add_list, we had
to remove the filter and move it back to the main hash. However, we
didn't do this under lock, which could cause consistency problems for
the list.

Fix this by updating i40e_update_filter_state logic so that it knows to
avoid broadcast filters. This ensures that we don't have to remove the
filter separately, and can put it back using the normal flow.

Change-ID: Id288fade80b3e3a9a54b68cc249188cb95147518
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

The intent of this message was to indicate to a user that we might have
missed a timestamp event for a valid packet. The original method of
detecting the missed events relied on waiting until all 4 registers were
filled.

A recent commit d55458c0cd7a5 ("i40e: replace PTP Rx timestamp hang
logic") replaced this logic with much better detection
scheme that could detect a stalled Rx timestamp register even when other
registers were still functional.

The new logic means that a message will be displayed almost as soon as
a timestamp for a dropped frame occurs. This new logic highlights that
the hardware will attempt timestamp for frames which it later decides to
drop. The most prominent example is when a multicast PTP frame is
received on a multicast address that we are not subscribed to.

Because the hardware initiates the Rx timestamp as soon as possible, it
will latch an RXTIME register, but then drop the packet.

This results in users being confused by the message as they are not
expecting to see dropped timestamp messages unless their application
also indicates that timestamps were missing.

Resolve this by reducing the severity and frequency of the displayed
message. We now only print the message if 3 or 4 of the RXTIME registers
are stalled and get cleared within the same watchdog event. This ensures
that the common case does not constantly display the message.
Additionally, since the message is likely not as meaningful to most
users, reduce the message to a dev_dbg instead of a dev_warn.

Users can still get a count of the number of timestamps dropped by
reading the ethtool statistics value, if necessary.

Change-ID: I35494442226a444c418dfb4f91a3070d06c8435c
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Store the FEC status bits from the link up event into the
hw_link_info structure.

Change-ID: I9a7b256f6dfb0dce89c2f503075d0d383526832e
Signed-off-by: Henry Tieman <henry.w.tieman@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Currently i40e_bus_info has PCI device and function info only and log
messages print device number as bus number. Added field to provide bus
number info and modified log statements to print bus, device and
function information.

Change-ID: I811617cee2714cc0d6bade8d369f57040990756f
Signed-off-by: Sudheer Mogilappagari <sudheer.mogilappagari@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

The function i40e_client_prepare() can never return an error. So make it
void and quit checking its return value.

Change-ID: I9ff311e2324dde329eb68648efb2c94aaff856db
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Signed-off-by: Bimmy Pujari <bimmy.pujari@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

The original comment implies that the only location where the raw_packet
buffer will be freed is in i40e_clean_tx_ring() which is incorrect. In
fact this isn't even the normal case. Update the comment explaining
where the memory is freed.

Change-ID: Ie0defc35ed1c3af183f81fdc60b6d783707a5595
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Reorganize the i40e_pull_tail() logic, doing it in i40e_add_rx_frag()
where it's cheaper.  The igb driver does this the same way.

Also renames i40e_page_is_reserved() to reflect what it actually
tests.

Change-ID: Icd9cc507aae1fcdc02308b3a09034111b4c24071
Signed-off-by: Scott Peterson <scott.d.peterson@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch reduces the size of struct i40e_rx_buffer by one pointer,
and makes the i40e driver a little more consistent with the igb driver
in terms of packets that span buffers.

We do this by moving the skb field from struct i40e_rx_buffer to
struct i40e_ring. We pass the skb we already have (or NULL if we
don't) to i40e_fetch_rx_buffer(), which skips the skb allocation if we
already have one for this packet.

Change-ID: I4ad48a531844494ba0c5d8e1a62209a057f661b0
Signed-off-by: Scott Peterson <scott.d.peterson@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

On packet RX, we perform a DMA sync for CPU before passing the
packet up.  Here we limit that sync to the actual length of the
incoming packet, rather than always syncing the entire buffer.

Change-ID: I626aaf6c37275a8ce9e81efcaa773f327b331487
Signed-off-by: Scott Peterson <scott.d.peterson@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

The iWarp client cannot continue until this operation has been completed
by the PF driver. Sleep (with timeout) until the reply from the PF
driver has been received.

Change-ID: I5dc41b857bba32d0218b7ce167b5da122dadf349
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

We can avoid the minor bit of work by calling check params after we
check for the client instance, since we're about to return early in
cases where we do not have a client.

Change-ID: I56f8ea2ba48d4f571fa331c9ace50819a022fa1c
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

If skb_padto failed the skb has been dropped already, so it was
consumed, but it doesn't mean it was sent, thus no need to update
queue tx time, etc. So, return NET_XMIT_DROP as more appropriate.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronzhuk@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch drops the vni zero check for COLLECT_METADATA mode.
It is not really needed, vni zero is a valid vni.

Fixes: 3ad7a4b1 ("vxlan: support fdb and learning in COLLECT_METADATA mode"
Reported-by: Joe Stringer <joe@ovn.org>
Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

After the dst->pending_confirm flag was removed, we do not
need anymore to provide dst arg to dst_neigh_output.
So, rename it to neigh_output as before commit 5110effe
("net: Do delayed neigh confirmation.").
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fixes: a0ee3541 ("sfc: process RX event inner checksum flags")
Reported-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Sainath Grandhi says:

====================
Refactor macvtap to re-use tap functionality by other virtual intefaces

Tap character devices can be implemented on other virtual interfaces like
ipvlan, similar to macvtap. Source code for tap functionality in macvtap
can be re-used for this purpose.

This patch series splits macvtap source into two modules, macvtap and tap.
This patch series also includes a patch for implementing tap character
device driver based on the IP-VLAN network interface, called ipvtap.

These patches are tested on x86 platform.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds a tap character device driver that is based on the
IP-VLAN network interface, called ipvtap. An ipvtap device can be created
in the same way as an ipvlan device, using 'type ipvtap', and then accessed
using the tap user space interface.
Signed-off-by: Sainath Grandhi <sainath.grandhi@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch makes tap a separate module for other types of virtual interfaces, for example,
ipvlan to use.
Signed-off-by: Sainath Grandhi <sainath.grandhi@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Extending tap APIs get/free_minor and create/destroy_cdev to handle more than one
type of virtual interface.
Signed-off-by: Sainath Grandhi <sainath.grandhi@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

macvlan object is re-structured to hold tap related elements in a separate
entity, tap_dev. Upon NETDEV_REGISTER device_event, tap_dev is registered with
idr and fetched again on tap_open. Few of the tap functions are modified to
accepted tap_dev as argument. tap_dev object includes callbacks to be used by
underlying virtual interface to take care of tx and rx accounting.
Signed-off-by: Sainath Grandhi <sainath.grandhi@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch provides tap device create/destroy APIs in tap.c.
Signed-off-by: Sainath Grandhi <sainath.grandhi@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Renaming tap related APIs, data structures and macros in tap.c from macvtap_.* to tap_.*
Signed-off-by: Sainath Grandhi <sainath.grandhi@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

macvtap module has code for tap/queue management and link management. This patch splits
the code into macvtap_main.c for link management and tap.c for tap/queue management.
Functionality in tap.c can be re-used for implementing tap on other virtual interfaces.
Signed-off-by: Sainath Grandhi <sainath.grandhi@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Pull networking fixes from David Miller:

 1) If the timing is wrong we can indefinitely stop generating new ipv6
    temporary addresses, from Marcus Huewe.

 2) Don't double free per-cpu stats in ipv6 SIT tunnel driver, from Cong
    Wang.

 3) Put protections in place so that AF_PACKET is not able to submit
    packets which don't even have a link level header to drivers. From
    Willem de Bruijn.

 4) Fix memory leaks in ipv4 and ipv6 multicast code, from Hangbin Liu.

 5) Don't use udp_ioctl() in l2tp code, UDP version expects a UDP socket
    and that doesn't go over very well when it is passed an L2TP one.
    Fix from Eric Dumazet.

 6) Don't crash on NULL pointer in phy_attach_direct(), from Florian
    Fainelli.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
  l2tp: do not use udp_ioctl()
  xen-netfront: Delete rx_refill_timer in xennet_disconnect_backend()
  NET: mkiss: Fix panic
  net: hns: Fix the device being used for dma mapping during TX
  net: phy: Initialize mdio clock at probe function
  igmp, mld: Fix memory leak in igmpv3/mld_del_delrec()
  xen-netfront: Improve error handling during initialization
  sierra_net: Skip validating irrelevant fields for IDLE LSIs
  sierra_net: Add support for IPv6 and Dual-Stack Link Sense Indications
  kcm: fix 0-length case for kcm_sendmsg()
  xen-netfront: Rework the fix for Rx stall during OOM and network stress
  net: phy: Fix PHY module checks and NULL deref in phy_attach_direct()
  net: thunderx: Fix PHY autoneg for SGMII QLM mode
  net: dsa: Do not destroy invalid network devices
  ping: fix a null pointer dereference
  packet: round up linear to header len
  net: introduce device min_header_len
  sit: fix a double free on error path
  lwtunnel: valid encap attr check should return 0 when lwtunnel is disabled
  ipv6: addrconf: fix generation of new temporary addresses

Pull rdma fixes from Doug Ledford:
 "Third round of -rc fixes for 4.10 kernel:

   - two security related issues in the rxe driver

   - one compile issue in the RDMA uapi header"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dledford/rdma:
  RDMA: Don't reference kernel private header from UAPI header
  IB/rxe: Fix mem_check_range integer overflow
  IB/rxe: Fix resid update

Pull i2c bugfixes from Wolfram Sang:
 "Two bugfixes (proper IO mapping and use of mutex) for a driver feature
  we introduced in this cycle"

* 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  i2c: piix4: Request the SMBUS semaphore inside the mutex
  i2c: piix4: Fix request_region size

Pull MMC host fix from Ulf Hansson:
 "mmci: Fix hang while waiting for busy-end interrupt"

* tag 'mmc-v4.10-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc:
  mmc: mmci: avoid clearing ST Micro busy end interrupt mistakenly

Pull sound fixes from Takashi Iwai:
 "Here are some last-minute fixes: two fixes for races in ALSA sequencer
  queue spotted by syzkaller, a revert for a regression of LINE6 driver
  (since 4.9), and a trivial new codec ID addition for Nvidia HDMI"

* tag 'sound-4.10' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: hda - adding a new NV HDMI/DP codec ID in the driver
  ALSA: seq: Fix race at creating a queue
  Revert "ALSA: line6: Only determine control port properties if needed"
  ALSA: seq: Don't handle loop timeout at snd_seq_pool_done()

Pull nfsd revert from Bruce Fields:
 "This patch turned out to have a couple problems. The problems are
  fixable, but at least one of the fixes is a little ugly. The original
  bug has always been there, so we can wait another week or two to get
  this right"

* tag 'nfsd-4.10-3' of git://linux-nfs.org/~bfields/linux:
  nfsd: Revert "nfsd: special case truncates some more"

Pull powerpc fixes friom Michael Ellerman:
 "Apologies for the late pull request, but Ben has been busy finding bugs.

   - Userspace was semi-randomly segfaulting on radix due to us
     incorrectly handling a fault triggered by autonuma, caused by a
     patch we merged earlier in v4.10 to prevent the kernel executing
     userspace.

   - We weren't marking host IPIs properly for KVM in the OPAL ICP
     backend.

   - The ERAT flushing on radix was missing an isync and was incorrectly
     marked as DD1 only.

   - The powernv CPU hotplug code was missing a wakeup type and failing
     to flush the interrupt correctly when using OPAL ICP

  Thanks to Benjamin Herrenschmidt"

* tag 'powerpc-4.10-4' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
  powerpc/powernv: Properly set "host-ipi" on IPIs
  powerpc/powernv: Fix CPU hotplug to handle waking on HVI
  powerpc/mm/radix: Update ERAT flushes when invalidating TLB
  powerpc/mm: Fix spurrious segfaults on radix with autonuma

Add maintainers for this tunnel driver. Include main osmocom.org mailist
list too.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

udp_ioctl(), as its name suggests, is used by UDP protocols,
but is also used by L2TP :(

L2TP should use its own handler, because it really does not
look the same.

SIOCINQ for instance should not assume UDP checksum or headers.

Thanks to Andrey and syzkaller team for providing the report
and a nice reproducer.

While crashes only happen on recent kernels (after commit
7c13f97f ("udp: do fwd memory scheduling on dequeue")), this
probably needs to be backported to older kernels.

Fixes: 7c13f97f ("udp: do fwd memory scheduling on dequeue")
Fixes: 85584672 ("udp: Fix udp_poll() and ioctl()")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Andrey Konovalov <andreyknvl@google.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Mickaël Salaün says:

====================
Improve BPF selftests and use the library (net-next tree)

This series brings some fixes to selftests, add the ability to test
unprivileged BPF programs as root and replace bpf_sys.h with calls to the BPF
library.

This is intended for the net-next tree and apply on c0e4dadb ("net: dsa:
mv88e6xxx: Move forward declaration to where it is needed").

Changes since v4:
* align text for function calls as requested by Daniel Borkmann
  (bpf_load_program and bpf_map_update_elem)
* rebase

Changes since v3:
* keep the bzero() calls

Changes since v2:
* use the patches from two previous series (unprivileged tests and bpf_sys.h
  replacement)
* include one more stdint.h
* rebase on net-next
* add this cover letter

Changes since v1:
* exclude patches not intended for the net-next tree
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Mickaël Salaün <mic@digikod.net>
Cc: Alexei Starovoitov <ast@fb.com>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Shuah Khan <shuah@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add require dependency headers.
Signed-off-by: Mickaël Salaün <mic@digikod.net>
Cc: Alexei Starovoitov <ast@fb.com>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Shuah Khan <shuah@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Replace bpf_map_create() with bpf_create_map() calls.
Signed-off-by: Mickaël Salaün <mic@digikod.net>
Cc: Alexei Starovoitov <ast@fb.com>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Shuah Khan <shuah@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Replace bpf_map_next_key() with bpf_map_get_next_key() calls.
Signed-off-by: Mickaël Salaün <mic@digikod.net>
Cc: Alexei Starovoitov <ast@fb.com>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Shuah Khan <shuah@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>