Pablo Neira Ayuso says:

====================
Netfilter fixes for net

The following patchset contains Netfilter fixes for net:

1) Flush the cleanup xtables worker to make sure destructors
   have completed, from Florian Westphal.

2) iifgroup is matching erroneously, also from Florian.

3) Add selftest for meta interface matching, from Florian Westphal.

4) Move nf_ct_offload_timeout() to header, from Roi Dayan.

5) Call nf_ct_offload_timeout() from flow_offload_add() to
   make sure garbage collection does not evict offloaded flow,
   from Roi Dayan.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

A dead lock was triggered on thunderx driver:

        CPU0                    CPU1
        ----                    ----
   [01] lock(&(&nic->rx_mode_wq_lock)->rlock);
                           [11] lock(&(&mc->mca_lock)->rlock);
                           [12] lock(&(&nic->rx_mode_wq_lock)->rlock);
   [02] <Interrupt> lock(&(&mc->mca_lock)->rlock);

The path for each is:

  [01] worker_thread() -> process_one_work() -> nicvf_set_rx_mode_task()
  [02] mld_ifc_timer_expire()
  [11] ipv6_add_dev() -> ipv6_dev_mc_inc() -> igmp6_group_added() ->
  [12] dev_mc_add() -> __dev_set_rx_mode() -> nicvf_set_rx_mode()

To fix it, it needs to disable bh on [1], so that the timer on [2]
wouldn't be triggered until rx_mode_wq_lock is released. So change
to use spin_lock_bh() instead of spin_lock().

Thanks to Paolo for helping with this.

v1->v2:
  - post to netdev.
Reported-by: Rafael P. <rparrazo@redhat.com>
Tested-by: Dean Nelson <dnelson@redhat.com>
Fixes: 469998c8 ("net: thunderx: prevent concurrent data re-writing by nicvf_set_rx_mode")
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Stefano Brivio says:

====================
Support PMTU discovery with bridged UDP tunnels

Currently, PMTU discovery for UDP tunnels only works if packets are
routed to the encapsulating interfaces, not bridged.

This results from the fact that we generally don't have valid routes
to the senders we can use to relay ICMP and ICMPv6 errors, and makes
PMTU discovery completely non-functional for VXLAN and GENEVE ports of
both regular bridges and Open vSwitch instances.

If the sender is local, and packets are forwarded to the port by a
regular bridge, all it takes is to generate a corresponding route
exception on the encapsulating device. The bridge then finds the route
exception carrying the PMTU value estimate as it forwards frames, and
relays ICMP messages back to the socket of the local sender. Patch 1/6
fixes this case.

If the sender resides on another node, we actually need to reply to
IP and IPv6 packets ourselves and send these ICMP or ICMPv6 errors
back, using the same encapsulating device. Patch 2/6, based on an
original idea by Florian Westphal, adds the needed functionality,
while patches 3/6 and 4/6 add matching support for VXLAN and GENEVE.

Finally, 5/6 and 6/6 introduce selftests for all combinations of
inner and outer IP versions, covering both VXLAN and GENEVE, with
both regular bridges and Open vSwitch instances.

v2: Add helper to check for any bridge port, skip oif check for PMTU
    routes for bridge ports only, split IPv4 and IPv6 helpers and
    functions (all suggested by David Ahern)
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

The new tests check that IP and IPv6 packets exceeding the local PMTU
estimate, forwarded by an Open vSwitch instance from another node,
result in the correct route exceptions being created, and that
communication with end-to-end fragmentation, over GENEVE and VXLAN
Open vSwitch ports, is now possible as a result of PMTU discovery.
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The new tests check that IP and IPv6 packets exceeding the local PMTU
estimate, both locally generated and forwarded by a bridge from
another node, result in the correct route exceptions being created,
and that communication with end-to-end fragmentation over VXLAN and
GENEVE tunnels is now possible as a result of PMTU discovery.

Part of the existing setup functions aren't generic enough to simply
add a namespace and a bridge to the existing routing setup. This
rework is in progress and we can easily shrink this once more generic
topology functions are available.
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

If the interface is a bridge or Open vSwitch port, and we can't
forward a packet because it exceeds the local PMTU estimate,
trigger an ICMP or ICMPv6 reply to the sender, using the same
interface to forward it back.

If metadata collection is enabled, set destination and source
addresses for the flow as if we were receiving the packet, so that
Open vSwitch can match the ICMP error against the existing
association.

v2: Use netif_is_any_bridge_port() (David Ahern)
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

If the interface is a bridge or Open vSwitch port, and we can't
forward a packet because it exceeds the local PMTU estimate,
trigger an ICMP or ICMPv6 reply to the sender, using the same
interface to forward it back.

If metadata collection is enabled, reverse destination and source
addresses, so that Open vSwitch is able to match this packet against
the existing, reverse flow.

v2: Use netif_is_any_bridge_port() (David Ahern)
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

It's currently possible to bridge Ethernet tunnels carrying IP
packets directly to external interfaces without assigning them
addresses and routes on the bridged network itself: this is the case
for UDP tunnels bridged with a standard bridge or by Open vSwitch.

PMTU discovery is currently broken with those configurations, because
the encapsulation effectively decreases the MTU of the link, and
while we are able to account for this using PMTU discovery on the
lower layer, we don't have a way to relay ICMP or ICMPv6 messages
needed by the sender, because we don't have valid routes to it.

On the other hand, as a tunnel endpoint, we can't fragment packets
as a general approach: this is for instance clearly forbidden for
VXLAN by RFC 7348, section 4.3:

   VTEPs MUST NOT fragment VXLAN packets.  Intermediate routers may
   fragment encapsulated VXLAN packets due to the larger frame size.
   The destination VTEP MAY silently discard such VXLAN fragments.

The same paragraph recommends that the MTU over the physical network
accomodates for encapsulations, but this isn't a practical option for
complex topologies, especially for typical Open vSwitch use cases.

Further, it states that:

   Other techniques like Path MTU discovery (see [RFC1191] and
   [RFC1981]) MAY be used to address this requirement as well.

Now, PMTU discovery already works for routed interfaces, we get
route exceptions created by the encapsulation device as they receive
ICMP Fragmentation Needed and ICMPv6 Packet Too Big messages, and
we already rebuild those messages with the appropriate MTU and route
them back to the sender.

Add the missing bits for bridged cases:

- checks in skb_tunnel_check_pmtu() to understand if it's appropriate
  to trigger a reply according to RFC 1122 section 3.2.2 for ICMP and
  RFC 4443 section 2.4 for ICMPv6. This function is already called by
  UDP tunnels

- a new function generating those ICMP or ICMPv6 replies. We can't
  reuse icmp_send() and icmp6_send() as we don't see the sender as a
  valid destination. This doesn't need to be generic, as we don't
  cover any other type of ICMP errors given that we only provide an
  encapsulation function to the sender

While at it, make the MTU check in skb_tunnel_check_pmtu() accurate:
we might receive GSO buffers here, and the passed headroom already
includes the inner MAC length, so we don't have to account for it
a second time (that would imply three MAC headers on the wire, but
there are just two).

This issue became visible while bridging IPv6 packets with 4500 bytes
of payload over GENEVE using IPv4 with a PMTU of 4000. Given the 50
bytes of encapsulation headroom, we would advertise MTU as 3950, and
we would reject fragmented IPv6 datagrams of 3958 bytes size on the
wire. We're exclusively dealing with network MTU here, though, so we
could get Ethernet frames up to 3964 octets in that case.

v2:
- moved skb_tunnel_check_pmtu() to ip_tunnel_core.c (David Ahern)
- split IPv4/IPv6 functions (David Ahern)
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently, processes sending traffic to a local bridge with an
encapsulation device as a port don't get ICMP errors if they exceed
the PMTU of the encapsulated link.

David Ahern suggested this as a hack, but it actually looks like
the correct solution: when we update the PMTU for a given destination
by means of updating or creating a route exception, the encapsulation
might trigger this because of PMTU discovery happening either on the
encapsulation device itself, or its lower layer. This happens on
bridged encapsulations only.

The output interface shouldn't matter, because we already have a
valid destination. Drop the output interface restriction from the
associated route lookup.

For UDP tunnels, we will now have a route exception created for the
encapsulation itself, with a MTU value reflecting its headroom, which
allows a bridge forwarding IP packets originated locally to deliver
errors back to the sending socket.

The behaviour is now consistent with IPv6 and verified with selftests
pmtu_ipv{4,6}_br_{geneve,vxlan}{4,6}_exception introduced later in
this series.

v2:
- reset output interface only for bridge ports (David Ahern)
- add and use netif_is_any_bridge_port() helper (David Ahern)
Suggested-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge tag 'wireless-drivers-next-2020-08-04' of git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers-next

Kalle Valo says:

====================
wireless-drivers-next patches for v5.9

Second set of patches for v5.9. mt76 has most of patches this time.
Otherwise it's just smaller fixes and cleanups to other drivers.

There was a major conflict in mt76 driver between wireless-drivers and
wireless-drivers-next. I solved that by merging the former to the
latter.

Major changes:

rtw88

* add support for ieee80211_ops::change_interface

* add support for enabling and disabling beacon

* add debugfs file for testing h2c

mt76

* ARP filter offload for 7663

* runtime power management for 7663

* testmode support for mfg calibration

* support for more channels
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Use netdev_<level> in place of VELOCITY_PRT.
Use pr_<level> in place of printk(KERN_<LEVEL>.

Miscellanea:

o Add pr_fmt to prefix pr_<level> output with "via-velocity: "
o Remove now unused functions and macros
o Realign some logging lines
o Remove devname where pr_<level> is also used
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Luo bin says:

====================
hinic: mailbox channel enhancement

add support to generate mailbox random id for VF to ensure that
the mailbox message from VF is valid and PF should check whether
the cmd from VF is supported before passing it to hw.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

PF should check whether the cmd from VF is supported and its content
is right before passing it to hw.
Signed-off-by: Luo bin <luobin9@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

add support to generate mailbox random id of VF to ensure that
mailbox messages PF received are from the correct VF.
Signed-off-by: Luo bin <luobin9@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

mt76 driver had major conflicts within mt7615 directory. To make it easier for
every merge wireless-drivers to wireless-drivers-next and solve those
conflicts.

   drivers/net/ethernet/sfc/ef100_nic.c:835:3: error: 'const struct efx_nic_type' has no member named 'filter_rfs_expire_one'
     835 |  .filter_rfs_expire_one = efx_mcdi_filter_rfs_expire_one,
         |   ^~~~~~~~~~~~~~~~~~~~~
>> drivers/net/ethernet/sfc/ef100_nic.c:835:27: error: initialization of 'void (*)(struct efx_nic *, u32)' {aka 'void (*)(struct efx_nic *, unsigned int)'} from incompatible pointer type 'bool (*)(struct efx_nic *, u32,  unsigned int)' {aka '_Bool (*)(struct efx_nic *, unsigned int,  unsigned int)'} [-Werror=incompatible-pointer-types]
     835 |  .filter_rfs_expire_one = efx_mcdi_filter_rfs_expire_one,
         |                           ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Daniel Borkmann says:

====================
pull-request: bpf-next 2020-08-04

The following pull-request contains BPF updates for your *net-next* tree.

We've added 73 non-merge commits during the last 9 day(s) which contain
a total of 135 files changed, 4603 insertions(+), 1013 deletions(-).

The main changes are:

1) Implement bpf_link support for XDP. Also add LINK_DETACH operation for the BPF
   syscall allowing processes with BPF link FD to force-detach, from Andrii Nakryiko.

2) Add BPF iterator for map elements and to iterate all BPF programs for efficient
   in-kernel inspection, from Yonghong Song and Alexei Starovoitov.

3) Separate bpf_get_{stack,stackid}() helpers for perf events in BPF to avoid
   unwinder errors, from Song Liu.

4) Allow cgroup local storage map to be shared between programs on the same
   cgroup. Also extend BPF selftests with coverage, from YiFei Zhu.

5) Add BPF exception tables to ARM64 JIT in order to be able to JIT BPF_PROBE_MEM
   load instructions, from Jean-Philippe Brucker.

6) Follow-up fixes on BPF socket lookup in combination with reuseport group
   handling. Also add related BPF selftests, from Jakub Sitnicki.

7) Allow to use socket storage in BPF_PROG_TYPE_CGROUP_SOCK-typed programs for
   socket create/release as well as bind functions, from Stanislav Fomichev.

8) Fix an info leak in xsk_getsockopt() when retrieving XDP stats via old struct
   xdp_statistics, from Peilin Ye.

9) Fix PT_REGS_RC{,_CORE}() macros in libbpf for MIPS arch, from Jerry Crunchtime.

10) Extend BPF kernel test infra with skb->family and skb->{local,remote}_ip{4,6}
    fields and allow user space to specify skb->dev via ifindex, from Dmitry Yakunin.

11) Fix a bpftool segfault due to missing program type name and make it more robust
    to prevent them in future gaps, from Quentin Monnet.

12) Consolidate cgroup helper functions across selftests and fix a v6 localhost
    resolver issue, from John Fastabend.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Saeed Mahameed says:

====================
mlx5-updates-2020-08-03

This patchset introduces some updates to mlx5 driver.

1) Jakub converts mlx5 to use the new udp tunnel infrastructure.
   Starting with a hack to allow drivers to request a static configuration
   of the default vxlan port, and then a patch that converts mlx5.

2) Parav implements change_carrier ndo for VF eswitch representors,
   to speedup link state control of representors netdevices.

3) Alex Vesker, makes a simple update to software steering to fix an issue
   with push vlan action sequence

4) Leon removes a redundant dump stack on error flow.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Edward Cree says:

====================
sfc: driver for EF100 family NICs, part 2

This series implements the data path and various other functionality
 for Xilinx/Solarflare EF100 NICs.

Changed from v2:
 * Improved error handling of design params (patch #3)
 * Removed 'inline' from .c file in patch #4
 * Don't report common stats to ethtool -S (patch #8)

Changed from v1:
 * Fixed build errors on CONFIG_RFS_ACCEL=n (patch #5) and 32-bit
   (patch #8)
 * Dropped patch #10 (ethtool ops) as it's buggy and will need a
   bigger rework to fix.
====================
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

We don't yet have a .sriov_configure() to create them, though.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We'll need it later, for VF representors.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Self-tests for event and interrupt reception and NVRAM.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

MAC stats work much the same as on EF10, with a periodic DMA to a region
 specified via an MCDI.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Bring down the TX and RX queues at ifdown, so that we can then fini the
 EVQs (otherwise the MC would return EBUSY because they're still in use).
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Includes RSS spreading.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Includes checksum offload and TSO, so declare those in our netdev features.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Several parts of the EF100 architecture are parameterised (to allow
 varying capabilities on FPGAs according to resource constraints), and
 these parameters are exposed to the driver through a TLV-encoded
 region of the BAR.
For the most part we either don't care about these values at all or
 just need to sanity-check them against the driver's assumptions, but
 there are a number of TSO limits which we record so that we will be
 able to check against them in the TX path when handling GSO skbs.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

In the future, EF100 is planned to have a credit-based scheme for
 handling unsolicited events, which drivers will need to use in order
 to function correctly.  However, current EF100 hardware does not yet
 generate unsolicited events and the credit scheme has not yet been
 implemented in firmware.  To prevent compatibility problems later if
 the current driver is used with future firmware which does implement
 it, we check for the corresponding capability flag (which that
 future firmware will set), and if found, we refuse to probe.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Early in EF100 development there was a different format of event
 descriptor; if the NIC is somehow running the very old firmware
 which will use that format, fail the probe.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The driver calls napi_schedule_irqoff() from a context where, in RT,
hardirqs are not disabled, since the IRQ handler is force-threaded.

In the call path of this function, __raise_softirq_irqoff() is modifying
its per-CPU mask of pending softirqs that must be processed, using
or_softirq_pending(). The or_softirq_pending() function is not atomic,
but since interrupts are supposed to be disabled, nobody should be
preempting it, and the operation should be safe.

Nonetheless, when running with hardirqs on, as in the PREEMPT_RT case,
it isn't safe, and the pending softirqs mask can get corrupted,
resulting in softirqs being lost and never processed.

To have common code that works with PREEMPT_RT and with mainline Linux,
we can use plain napi_schedule() instead. The difference is that
napi_schedule() (via __napi_schedule) also calls local_irq_save, which
disables hardirqs if they aren't already. But, since they already are
disabled in non-RT, this means that in practice we don't see any
measurable difference in throughput or latency with this patch.
Signed-off-by: Jiafei Pan <Jiafei.Pan@nxp.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The driver calls napi_schedule_irqoff() from a context where, in RT,
hardirqs are not disabled, since the IRQ handler is force-threaded.

In the call path of this function, __raise_softirq_irqoff() is modifying
its per-CPU mask of pending softirqs that must be processed, using
or_softirq_pending(). The or_softirq_pending() function is not atomic,
but since interrupts are supposed to be disabled, nobody should be
preempting it, and the operation should be safe.

Nonetheless, when running with hardirqs on, as in the PREEMPT_RT case,
it isn't safe, and the pending softirqs mask can get corrupted,
resulting in softirqs being lost and never processed.

To have common code that works with PREEMPT_RT and with mainline Linux,
we can use plain napi_schedule() instead. The difference is that
napi_schedule() (via __napi_schedule) also calls local_irq_save, which
disables hardirqs if they aren't already. But, since they already are
disabled in non-RT, this means that in practice we don't see any
measurable difference in throughput or latency with this patch.
Signed-off-by: Jiafei Pan <Jiafei.Pan@nxp.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: dsa: loop: Preparatory changes for 802.1Q data path
Florian Fainelli says:

====================
These patches are all meant to help pave the way for a 802.1Q data path
added to the mockup driver, making it more useful than just testing for
configuration. Sending those out now since there is no real need to
wait.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

We only support DSA_LOOP_NUM_PORTS in the switch, do not tell the DSA
core to allocate up to DSA_MAX_PORTS which is nearly the double (6 vs.
11).
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

For now we simply store the port MTU into a per-port member.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

In preparation for adding support for a mockup data path, move the
driver data structures to include/linux/dsa/loop.h such that we can
share them between net/dsa/ and drivers/net/dsa/ later on.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Allocate a 4K array of VLANs instead of limiting ourselves to just 5
which is arbitrary.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

The PVID should be per-port, this is a preliminary change to support a
802.1Q data path in the driver.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Matching IPv6 traffic require allocating their own individual slots
in TCAM. So, fetch additional slots to insert IPv6 rules. Also, fetch
the cumulative stats of all the slots occupied by the Matchall rule.
Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The current poll interval is enough to ensure that rising and falling
edge events are not lost for a 1 PPS signal with 50% duty cycle.

But when we deliver the events to user space, it will try to infer if
they were corresponding to a rising or to a falling edge (the kernel
driver doesn't know that either). User space will try to make that
inference based on the time at which the PPS master had emitted the
pulse (i.e. if it's a .0 time, it's rising edge, if it's .5 time, it's
falling edge).

But there is no in-kernel API for retrieving the precise timestamp
corresponding to a PPS master (aka perout) pulse. So user space has to
guess even that. It will read the PTP time on the PPS master right after
we've delivered the extts event, and declare that the PPS master time
was just the closest integer second, based on 2 thresholds (lower than
.25, or higher than .75, and ignore anything else).

Except that, if we poll for extts events (and our hardware doesn't
really help us, by not providing an interrupt), then there is a risk
that the poll period (and therefore the time at which the event is
delivered) might confuse user space.

Because we are always scheduling the next extts poll at
SJA1105_EXTTS_INTERVAL "from now" (that's the only thing that the
schedule_delayed_work() API gives us), it means that the start time of
the next delayed workqueue will always be shifted to the right a little
bit (shifted with the SPI access duration of this workqueue run).
In turn, because user space sees extts events that are non-periodic
compared to the PPS master's time, this means that it might start making
wrong guesses about rising/falling edge.

To understand the effect, here is the output of ts2phc currently. Notice
the 'src' timestamps of the 'SKIP extts' events, and how they have a
large wander. They keep increasing until the upper limit for the ignore
threshold (.75 seconds), after which the application starts ignoring the
_other_ edge.

ts2phc[26.624]: /dev/ptp3 SKIP extts index 0 at 21.449898912 src 21.657784518
ts2phc[27.133]: adding tstamp 21.949894240 to clock /dev/ptp3
ts2phc[27.133]: adding tstamp 22.000000000 to clock /dev/ptp1
ts2phc[27.133]: /dev/ptp3 offset        640 s2 freq   +5112
ts2phc[27.636]: /dev/ptp3 SKIP extts index 0 at 22.449889360 src 22.669398022
ts2phc[28.140]: adding tstamp 22.949884376 to clock /dev/ptp3
ts2phc[28.140]: adding tstamp 23.000000000 to clock /dev/ptp1
ts2phc[28.140]: /dev/ptp3 offset         96 s2 freq   +4760
ts2phc[28.644]: /dev/ptp3 SKIP extts index 0 at 23.449879504 src 23.677420422
ts2phc[29.153]: adding tstamp 23.949874704 to clock /dev/ptp3
ts2phc[29.153]: adding tstamp 24.000000000 to clock /dev/ptp1
ts2phc[29.153]: /dev/ptp3 offset       -264 s2 freq   +4429
ts2phc[29.656]: /dev/ptp3 SKIP extts index 0 at 24.449870008 src 24.689407238
ts2phc[30.160]: adding tstamp 24.949865376 to clock /dev/ptp3
ts2phc[30.160]: adding tstamp 25.000000000 to clock /dev/ptp1
ts2phc[30.160]: /dev/ptp3 offset       -280 s2 freq   +4334
ts2phc[30.664]: /dev/ptp3 SKIP extts index 0 at 25.449860760 src 25.697449926
ts2phc[31.168]: adding tstamp 25.949856176 to clock /dev/ptp3
ts2phc[31.168]: adding tstamp 26.000000000 to clock /dev/ptp1
ts2phc[31.168]: /dev/ptp3 offset       -176 s2 freq   +4354
ts2phc[31.672]: /dev/ptp3 SKIP extts index 0 at 26.449851584 src 26.705433606
ts2phc[32.180]: adding tstamp 26.949846992 to clock /dev/ptp3
ts2phc[32.180]: adding tstamp 27.000000000 to clock /dev/ptp1
ts2phc[32.180]: /dev/ptp3 offset        -80 s2 freq   +4397
ts2phc[32.684]: /dev/ptp3 SKIP extts index 0 at 27.449842384 src 27.717415110
ts2phc[33.192]: adding tstamp 27.949837768 to clock /dev/ptp3
ts2phc[33.192]: adding tstamp 28.000000000 to clock /dev/ptp1
ts2phc[33.192]: /dev/ptp3 offset          0 s2 freq   +4453
ts2phc[33.696]: /dev/ptp3 SKIP extts index 0 at 28.449833128 src 28.729412902
ts2phc[34.200]: adding tstamp 28.949828472 to clock /dev/ptp3
ts2phc[34.200]: adding tstamp 29.000000000 to clock /dev/ptp1
ts2phc[34.200]: /dev/ptp3 offset          8 s2 freq   +4461
ts2phc[34.704]: /dev/ptp3 SKIP extts index 0 at 29.449823816 src 29.737416038
ts2phc[35.208]: adding tstamp 29.949819152 to clock /dev/ptp3
ts2phc[35.208]: adding tstamp 30.000000000 to clock /dev/ptp1
ts2phc[35.208]: /dev/ptp3 offset         -8 s2 freq   +4447
ts2phc[35.712]: /dev/ptp3 SKIP extts index 0 at 30.449814496 src 30.745554982
ts2phc[36.216]: adding tstamp 30.949809840 to clock /dev/ptp3
ts2phc[36.216]: adding tstamp 31.000000000 to clock /dev/ptp1
ts2phc[36.216]: /dev/ptp3 offset         -8 s2 freq   +4445
ts2phc[36.468]: /dev/ptp3 SKIP extts index 0 at 31.449805184 src 31.501109446
ts2phc[36.972]: adding tstamp 31.949800536 to clock /dev/ptp3
ts2phc[36.972]: adding tstamp 32.000000000 to clock /dev/ptp1
ts2phc[36.972]: /dev/ptp3 offset         -8 s2 freq   +4442
ts2phc[37.480]: /dev/ptp3 SKIP extts index 0 at 32.449795896 src 32.513320070
ts2phc[37.984]: adding tstamp 32.949791248 to clock /dev/ptp3
ts2phc[37.984]: adding tstamp 33.000000000 to clock /dev/ptp1
ts2phc[37.984]: /dev/ptp3 offset          0 s2 freq   +4448

Fix that by taking the following measures:
- Schedule the poll from a timer. Because we are really scheduling the
  timer periodically, the extts events delivered to user space are
  periodic too, and don't suffer from the "shift-to-the-right" effect.
- Increase the poll period to 6 times a second. This imposes a smaller
  upper bound to the shift that can occur to the delivery time of extts
  events, and makes user space (ts2phc) to always interpret correctly
  which events should be skipped and which shouldn't.
- Move the SPI readout itself to the main PTP kernel thread, instead of
  the generic workqueue. This is because the timer runs in atomic
  context, but is also better than before, because if needed, we can
  chrt & taskset this kernel thread, to ensure it gets enough priority
  under load.

After this patch, one can notice that the wander is greatly reduced, and
that the latencies of one extts poll are not propagated to the next. The
'src' timestamp that is skipped is never larger than .65 seconds (which
means .15 seconds larger than the time at which the real event occurred
at, and .10 seconds smaller than the .75 upper threshold for ignoring
the falling edge):

ts2phc[40.076]: adding tstamp 34.949261296 to clock /dev/ptp3
ts2phc[40.076]: adding tstamp 35.000000000 to clock /dev/ptp1
ts2phc[40.076]: /dev/ptp3 offset         48 s2 freq   +4631
ts2phc[40.568]: /dev/ptp3 SKIP extts index 0 at 35.449256496 src 35.595791078
ts2phc[41.064]: adding tstamp 35.949251744 to clock /dev/ptp3
ts2phc[41.064]: adding tstamp 36.000000000 to clock /dev/ptp1
ts2phc[41.064]: /dev/ptp3 offset       -224 s2 freq   +4374
ts2phc[41.552]: /dev/ptp3 SKIP extts index 0 at 36.449247088 src 36.579825574
ts2phc[42.044]: adding tstamp 36.949242456 to clock /dev/ptp3
ts2phc[42.044]: adding tstamp 37.000000000 to clock /dev/ptp1
ts2phc[42.044]: /dev/ptp3 offset       -240 s2 freq   +4290
ts2phc[42.536]: /dev/ptp3 SKIP extts index 0 at 37.449237848 src 37.563828774
ts2phc[43.028]: adding tstamp 37.949233264 to clock /dev/ptp3
ts2phc[43.028]: adding tstamp 38.000000000 to clock /dev/ptp1
ts2phc[43.028]: /dev/ptp3 offset       -144 s2 freq   +4314
ts2phc[43.520]: /dev/ptp3 SKIP extts index 0 at 38.449228656 src 38.547823238
ts2phc[44.012]: adding tstamp 38.949224048 to clock /dev/ptp3
ts2phc[44.012]: adding tstamp 39.000000000 to clock /dev/ptp1
ts2phc[44.012]: /dev/ptp3 offset        -80 s2 freq   +4335
ts2phc[44.508]: /dev/ptp3 SKIP extts index 0 at 39.449219432 src 39.535846118
ts2phc[44.996]: adding tstamp 39.949214816 to clock /dev/ptp3
ts2phc[44.996]: adding tstamp 40.000000000 to clock /dev/ptp1
ts2phc[44.996]: /dev/ptp3 offset        -32 s2 freq   +4359
ts2phc[45.488]: /dev/ptp3 SKIP extts index 0 at 40.449210192 src 40.515824678
ts2phc[45.980]: adding tstamp 40.949205568 to clock /dev/ptp3
ts2phc[45.980]: adding tstamp 41.000000000 to clock /dev/ptp1
ts2phc[45.980]: /dev/ptp3 offset          8 s2 freq   +4390
ts2phc[46.636]: /dev/ptp3 SKIP extts index 0 at 41.449200928 src 41.664176902
ts2phc[47.132]: adding tstamp 41.949196288 to clock /dev/ptp3
ts2phc[47.132]: adding tstamp 42.000000000 to clock /dev/ptp1
ts2phc[47.132]: /dev/ptp3 offset          0 s2 freq   +4384
ts2phc[47.620]: /dev/ptp3 SKIP extts index 0 at 42.449191656 src 42.648117190
ts2phc[48.112]: adding tstamp 42.949187016 to clock /dev/ptp3
ts2phc[48.112]: adding tstamp 43.000000000 to clock /dev/ptp1
ts2phc[48.112]: /dev/ptp3 offset          0 s2 freq   +4384
ts2phc[48.604]: /dev/ptp3 SKIP extts index 0 at 43.449182384 src 43.632112582
ts2phc[49.100]: adding tstamp 43.949177736 to clock /dev/ptp3
ts2phc[49.100]: adding tstamp 44.000000000 to clock /dev/ptp1
ts2phc[49.100]: /dev/ptp3 offset         -8 s2 freq   +4376
ts2phc[49.588]: /dev/ptp3 SKIP extts index 0 at 44.449173096 src 44.616136774
ts2phc[50.080]: adding tstamp 44.949168464 to clock /dev/ptp3
ts2phc[50.080]: adding tstamp 45.000000000 to clock /dev/ptp1
ts2phc[50.080]: /dev/ptp3 offset          8 s2 freq   +4390
ts2phc[50.572]: /dev/ptp3 SKIP extts index 0 at 45.449163816 src 45.600134662
ts2phc[51.064]: adding tstamp 45.949159160 to clock /dev/ptp3
ts2phc[51.064]: adding tstamp 46.000000000 to clock /dev/ptp1
ts2phc[51.064]: /dev/ptp3 offset         -8 s2 freq   +4376
ts2phc[51.556]: /dev/ptp3 SKIP extts index 0 at 46.449154528 src 46.584588550
ts2phc[52.048]: adding tstamp 46.949149896 to clock /dev/ptp3
ts2phc[52.048]: adding tstamp 47.000000000 to clock /dev/ptp1
ts2phc[52.048]: /dev/ptp3 offset          0 s2 freq   +4382
ts2phc[52.540]: /dev/ptp3 SKIP extts index 0 at 47.449145256 src 47.568132198
ts2phc[53.032]: adding tstamp 47.949140616 to clock /dev/ptp3
ts2phc[53.032]: adding tstamp 48.000000000 to clock /dev/ptp1
ts2phc[53.032]: /dev/ptp3 offset          0 s2 freq   +4382
ts2phc[53.524]: /dev/ptp3 SKIP extts index 0 at 48.449135968 src 48.552121446
ts2phc[54.016]: adding tstamp 48.949131320 to clock /dev/ptp3
ts2phc[54.016]: adding tstamp 49.000000000 to clock /dev/ptp1
ts2phc[54.016]: /dev/ptp3 offset          0 s2 freq   +4382
ts2phc[54.512]: /dev/ptp3 SKIP extts index 0 at 49.449126680 src 49.540147014
ts2phc[55.000]: adding tstamp 49.949122040 to clock /dev/ptp3
ts2phc[55.000]: adding tstamp 50.000000000 to clock /dev/ptp1
ts2phc[55.000]: /dev/ptp3 offset          0 s2 freq   +4382
ts2phc[55.492]: /dev/ptp3 SKIP extts index 0 at 50.449117400 src 50.520119078
ts2phc[55.988]: adding tstamp 50.949112768 to clock /dev/ptp3
ts2phc[55.988]: adding tstamp 51.000000000 to clock /dev/ptp1
ts2phc[55.988]: /dev/ptp3 offset          8 s2 freq   +4390
ts2phc[56.476]: /dev/ptp3 SKIP extts index 0 at 51.449108120 src 51.504175910
ts2phc[57.132]: adding tstamp 51.949103480 to clock /dev/ptp3
ts2phc[57.132]: adding tstamp 52.000000000 to clock /dev/ptp1
ts2phc[57.132]: /dev/ptp3 offset          0 s2 freq   +4384
ts2phc[57.624]: /dev/ptp3 SKIP extts index 0 at 52.449098840 src 52.651833574
ts2phc[58.116]: adding tstamp 52.949094200 to clock /dev/ptp3
ts2phc[58.116]: adding tstamp 53.000000000 to clock /dev/ptp1
ts2phc[58.116]: /dev/ptp3 offset          8 s2 freq   +4392
ts2phc[58.612]: /dev/ptp3 SKIP extts index 0 at 53.449089560 src 53.639826918
ts2phc[59.100]: adding tstamp 53.949084920 to clock /dev/ptp3
ts2phc[59.100]: adding tstamp 54.000000000 to clock /dev/ptp1
ts2phc[59.100]: /dev/ptp3 offset          8 s2 freq   +4394
ts2phc[59.592]: /dev/ptp3 SKIP extts index 0 at 54.449080272 src 54.619842278
ts2phc[60.084]: adding tstamp 54.949075624 to clock /dev/ptp3
ts2phc[60.084]: adding tstamp 55.000000000 to clock /dev/ptp1
ts2phc[60.084]: /dev/ptp3 offset          8 s2 freq   +4397
ts2phc[60.576]: /dev/ptp3 SKIP extts index 0 at 55.449070968 src 55.603885542
ts2phc[61.068]: adding tstamp 55.949066312 to clock /dev/ptp3
ts2phc[61.068]: adding tstamp 56.000000000 to clock /dev/ptp1
ts2phc[61.068]: /dev/ptp3 offset          0 s2 freq   +4391
ts2phc[61.560]: /dev/ptp3 SKIP extts index 0 at 56.449061680 src 56.587885798
ts2phc[62.052]: adding tstamp 56.949057032 to clock /dev/ptp3
ts2phc[62.052]: adding tstamp 57.000000000 to clock /dev/ptp1
ts2phc[62.052]: /dev/ptp3 offset         -8 s2 freq   +4383
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>