More and more projects use libbpf and one day it'll likely be packaged
and distributed as DSO and that requires ABI versioning so that both
compatible and incompatible changes to ABI can be introduced in a safe
way in the future without breaking executables dynamically linked with a
previous version of the library.

Usual way to do ABI versioning is version script for the linker. Add
such a script for libbpf. All global symbols currently exported via
LIBBPF_API macro are added to the version script libbpf.map.

The version name LIBBPF_0.0.1 is constructed from the name of the
library + version specified by $(LIBBPF_VERSION) in Makefile.

Version script does not duplicate the work done by LIBBPF_API macro, it
rather complements it. The macro is used at compile time and can be used
by compiler to do optimization that can't be done at link time, it is
purely about global symbol visibility. The version script, in turn, is
used at link time and takes care of ABI versioning. Both techniques are
described in details in [1].

Whenever ABI is changed in the future, version script should be changed
appropriately.

[1] https://www.akkadia.org/drepper/dsohowto.pdfSigned-off-by: Andrey Ignatov <rdna@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

s/btf_get_from_id/btf__get_from_id/ to restore the API naming convention.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Andrey Ignatov <rdna@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Yonghong Song says:

====================
Commit 838e9690 ("bpf: Introduce bpf_func_info")
added bpf func info support. The userspace is able
to get better ksym's for bpf programs with jit, and
is able to print out func prototypes.

For a program containing func-to-func calls, the existing
implementation returns user specified number of function
calls and BTF types if jit is enabled. If the jit is not
enabled, it only returns the type for the main function.

This is undesirable. Interpreter may still be used
and we should keep feature identical regardless of
whether jit is enabled or not.
This patch fixed this discrepancy.

The following example shows bpftool output for
the bpf program in selftests test_btf_haskv.o when jit
is disabled:
  $ bpftool prog dump xlated id 1490
  int _dummy_tracepoint(struct dummy_tracepoint_args * arg):
     0: (85) call pc+2#__bpf_prog_run_args32
     1: (b7) r0 = 0
     2: (95) exit
  int test_long_fname_1(struct dummy_tracepoint_args * arg):
     3: (85) call pc+1#__bpf_prog_run_args32
     4: (95) exit
  int test_long_fname_2(struct dummy_tracepoint_args * arg):
     5: (b7) r2 = 0
     6: (63) *(u32 *)(r10 -4) = r2
     7: (79) r1 = *(u64 *)(r1 +8)
     8: (15) if r1 == 0x0 goto pc+9
     9: (bf) r2 = r10
    10: (07) r2 += -4
    11: (18) r1 = map[id:1173]
    13: (85) call bpf_map_lookup_elem#77088
    14: (15) if r0 == 0x0 goto pc+3
    15: (61) r1 = *(u32 *)(r0 +4)
    16: (07) r1 += 1
    17: (63) *(u32 *)(r0 +4) = r1
    18: (95) exit
  $ bpftool prog dump jited id 1490
    no instructions returned
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

The selftest test_btf is changed to test both jit and non-jit.
The test result should be the same regardless of whether jit
is enabled or not.
Signed-off-by: Yonghong Song <yhs@fb.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Commit 838e9690 ("bpf: Introduce bpf_func_info")
added bpf func info support. The userspace is able
to get better ksym's for bpf programs with jit, and
is able to print out func prototypes.

For a program containing func-to-func calls, the existing
implementation returns user specified number of function
calls and BTF types if jit is enabled. If the jit is not
enabled, it only returns the type for the main function.

This is undesirable. Interpreter may still be used
and we should keep feature identical regardless of
whether jit is enabled or not.
This patch fixed this discrepancy.

Fixes: 838e9690 ("bpf: Introduce bpf_func_info")
Signed-off-by: Yonghong Song <yhs@fb.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

'offset' is constant and if it is zero, no need to subtract it
from BPF_REG_TMP.
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Daniel Borkmann says:

====================
pull-request: bpf-next 2018-11-26

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

The main changes are:

1) Extend BTF to support function call types and improve the BPF
   symbol handling with this info for kallsyms and bpftool program
   dump to make debugging easier, from Martin and Yonghong.

2) Optimize LPM lookups by making longest_prefix_match() handle
   multiple bytes at a time, from Eric.

3) Adds support for loading and attaching flow dissector BPF progs
   from bpftool, from Stanislav.

4) Extend the sk_lookup() helper to be supported from XDP, from Nitin.

5) Enable verifier to support narrow context loads with offset > 0
   to adapt to LLVM code generation (currently only offset of 0 was
   supported). Add test cases as well, from Andrey.

6) Simplify passing device functions for offloaded BPF progs by
   adding callbacks to bpf_prog_offload_ops instead of ndo_bpf.
   Also convert nfp and netdevsim to make use of them, from Quentin.

7) Add support for sock_ops based BPF programs to send events to
   the perf ring-buffer through perf_event_output helper, from
   Sowmini and Daniel.

8) Add read / write support for skb->tstamp from tc BPF and cg BPF
   programs to allow for supporting rate-limiting in EDT qdiscs
   like fq from BPF side, from Vlad.

9) Extend libbpf API to support map in map types and add test cases
   for it as well to BPF kselftests, from Nikita.

10) Account the maximum packet offset accessed by a BPF program in
    the verifier and use it for optimizing nfp JIT, from Jiong.

11) Fix error handling regarding kprobe_events in BPF sample loader,
    from Daniel T.

12) Add support for queue and stack map type in bpftool, from David.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Make the formatting for map_type_name array consistent.
Signed-off-by: David Calavera <david.calavera@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

There is a spelling mistake in a btf_verifier_log_member message,
fix it.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Building tags produces warning:

  ctags: Warning: kernel/bpf/local_storage.c:10: null expansion of name pattern "\1"

Let's use the same fix as in commit 25528213 ("tags: Fix DEFINE_PER_CPU
expansions"), even though it violates the usual code style.
Signed-off-by: Rustam Kovhaev <rkovhaev@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

I do not see how one can effectively use skb_insert() without holding
some kind of lock. Otherwise other cpus could have changed the list
right before we have a chance of acquiring list->lock.

Only existing user is in drivers/infiniband/hw/nes/nes_mgt.c and this
one probably meant to use __skb_insert() since it appears nesqp->pau_list
is protected by nesqp->pau_lock. This looks like nesqp->pau_lock
could be removed, since nesqp->pau_list.lock could be used instead.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Faisal Latif <faisal.latif@intel.com>
Cc: Doug Ledford <dledford@redhat.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: linux-rdma <linux-rdma@vger.kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

A recent change added a null check on p->dev after p->dev was being
dereferenced by the ns_capable check on p->dev. It turns out that
neither the p->dev and p->br null checks are necessary, and can be
removed, which cleans up a static analyis warning.

As Nikolay Aleksandrov noted, these checks can be removed because:

"My reasoning of why it shouldn't be possible:
- On port add new_nbp() sets both p->dev and p->br before creating
  kobj/sysfs

- On port del (trickier) del_nbp() calls kobject_del() before call_rcu()
  to destroy the port which in turn calls sysfs_remove_dir() which uses
  kernfs_remove() which deactivates (shouldn't be able to open new
  files) and calls kernfs_drain() to drain current open/mmaped files in
  the respective dir before continuing, thus making it impossible to
  open a bridge port sysfs file with p->dev and p->br equal to NULL.

So I think it's safe to remove those checks altogether. It'd be nice to
get a second look over my reasoning as I might be missing something in
sysfs/kernfs call path."

Thanks to Nikolay Aleksandrov's suggestion to remove the check and
David Miller for sanity checking this.

Detected by CoverityScan, CID#751490 ("Dereference before null check")

Fixes: a5f3ea54 ("net: bridge: add support for raw sysfs port options")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Heiner Kallweit says:

====================
r8169: make use of xmit_more and __netdev_sent_queue

This series adds helper __netdev_sent_queue to the core and makes use
of it in the r8169 driver.

Heiner Kallweit (2):
  net: core: add __netdev_sent_queue as variant of __netdev_tx_sent_queue
  r8169: make use of xmit_more and __netdev_sent_queue

v2:
- fix minor style issue
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Make use of xmit_more and add the functionality introduced with
3e59020a ("net: bql: add __netdev_tx_sent_queue()").
I used the mlx4 driver as template.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Similar to netdev_sent_queue add helper __netdev_sent_queue as variant
of __netdev_tx_sent_queue.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Packet sockets with PACKET_TX_RING send skbs with user data in frags.

Before commit 5cd8d46e ("packet: copy user buffers before orphan
or clone") ring slots could be released prematurely, possibly allowing
a process to overwrite data still in flight.

This test opens two packet sockets, one to send and one to read.
The sender has a tx ring of one slot. It sends two packets with
different payload, then reads both and verifies their payload.

Before the above commit, both receive calls return the same data as
the send calls use the same buffer. From the commit, the clone
needed for looping onto a packet socket triggers an skb_copy_ubufs
to create a private copy. The separate sends each arrive correctly.
Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently dev is dereferenced by the call dev_net(dev) before dev is null
checked.  Fix this by null checking dev before the potential null
pointer dereference.

Detected by CoverityScan, CID#1462955 ("Dereference before null check")

Fixes: 23790ef1 ("net: qualcomm: rmnet: Allow to configure flags for existing devices")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fixes gcc '-Wunused-but-set-variable' warning:

drivers/net/ethernet/chelsio/cxgb4/t4_hw.c:5883:6:
 warning: variable 'multitrc' set but not used [-Wunused-but-set-variable]

drivers/net/ethernet/chelsio/cxgb4/t4_hw.c:8585:32:
 warning: variable 'speed' set but not used [-Wunused-but-set-variable]

'multitrc' never used since introduction in
commit 8e3d04fd ("cxgb4: Add MPS tracing support")

'speed' never used since introduction in
commit c3168cab ("cxgb4/cxgbvf: Handle 32-bit fw port capabilities")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Return code should be formally "netdev_tx_t".
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Pull arm64 fixes from Catalin Marinas::

 - Fix wrong conflict resolution around CONFIG_ARM64_SSBD

 - Fix sparse warning on unsigned long constant

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: cpufeature: Fix mismerge of CONFIG_ARM64_SSBD block
  arm64: sysreg: fix sparse warnings

Pull networking fixes from David Miller:

 1) Need to take mutex in ath9k_add_interface(), from Dan Carpenter.

 2) Fix mt76 build without CONFIG_LEDS_CLASS, from Arnd Bergmann.

 3) Fix socket wmem accounting in SCTP, from Xin Long.

 4) Fix failed resume crash in ena driver, from Arthur Kiyanovski.

 5) qed driver passes bytes instead of bits into second arg of
    bitmap_weight(). From Denis Bolotin.

 6) Fix reset deadlock in ibmvnic, from Juliet Kim.

 7) skb_scrube_packet() needs to scrub the fwd marks too, from Petr
    Machata.

 8) Make sure older TCP stacks see enough dup ACKs, and avoid doing SACK
    compression during this period, from Eric Dumazet.

 9) Add atomicity to SMC protocol cursor handling, from Ursula Braun.

10) Don't leave dangling error pointer if bpf_prog_add() fails in
    thunderx driver, from Lorenzo Bianconi. Also, when we unmap TSO
    headers, set sq->tso_hdrs to NULL.

11) Fix race condition over state variables in act_police, from Davide
    Caratti.

12) Disable guest csum in the presence of XDP in virtio_net, from Jason
    Wang.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (64 commits)
  net: gemini: Fix copy/paste error
  net: phy: mscc: fix deadlock in vsc85xx_default_config
  dt-bindings: dsa: Fix typo in "probed"
  net: thunderx: set tso_hdrs pointer to NULL in nicvf_free_snd_queue
  net: amd: add missing of_node_put()
  team: no need to do team_notify_peers or team_mcast_rejoin when disabling port
  virtio-net: fail XDP set if guest csum is negotiated
  virtio-net: disable guest csum during XDP set
  net/sched: act_police: add missing spinlock initialization
  net: don't keep lonely packets forever in the gro hash
  net/ipv6: re-do dad when interface has IFF_NOARP flag change
  packet: copy user buffers before orphan or clone
  ibmvnic: Update driver queues after change in ring size support
  ibmvnic: Fix RX queue buffer cleanup
  net: thunderx: set xdp_prog to NULL if bpf_prog_add fails
  net/dim: Update DIM start sample after each DIM iteration
  net: faraday: ftmac100: remove netif_running(netdev) check before disabling interrupts
  net/smc: use after free fix in smc_wr_tx_put_slot()
  net/smc: atomic SMCD cursor handling
  net/smc: add SMC-D shutdown signal
  ...

Pull xfs fixes from Darrick Wong:
 "Dave and I have continued our work fixing corruption problems that can
  be found when running long-term burn-in exercisers on xfs. Here are
  some patches fixing most of the problems, but there will likely be
  more. :/

   - Numerous corruption fixes for copy on write

   - Numerous corruption fixes for blocksize < pagesize writes

   - Don't miscalculate AG reservations for small final AGs

   - Fix page cache truncation to work properly for reflink and extent
     shifting

   - Fix use-after-free when retrying failed inode/dquot buffer logging

   - Fix corruptions seen when using copy_file_range in directio mode"

* tag 'xfs-4.20-fixes-2' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
  iomap: readpages doesn't zero page tail beyond EOF
  vfs: vfs_dedupe_file_range() doesn't return EOPNOTSUPP
  iomap: dio data corruption and spurious errors when pipes fill
  iomap: sub-block dio needs to zeroout beyond EOF
  iomap: FUA is wrong for DIO O_DSYNC writes into unwritten extents
  xfs: delalloc -> unwritten COW fork allocation can go wrong
  xfs: flush removing page cache in xfs_reflink_remap_prep
  xfs: extent shifting doesn't fully invalidate page cache
  xfs: finobt AG reserves don't consider last AG can be a runt
  xfs: fix transient reference count error in xfs_buf_resubmit_failed_buffers
  xfs: uncached buffer tracing needs to print bno
  xfs: make xfs_file_remap_range() static
  xfs: fix shared extent data corruption due to missing cow reservation

The TX stats should be started with the tx_stats_syncp,
there seems to be a copy/paste error in the driver.
Signed-off-by: Andreas Fiedler <andreas.fiedler@gmx.net>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

The vsc85xx_default_config function called in the vsc85xx_config_init
function which is used by VSC8530, VSC8531, VSC8540 and VSC8541 PHYs
mistakenly calls phy_read and phy_write in-between phy_select_page and
phy_restore_page.

phy_select_page and phy_restore_page actually take and release the MDIO
bus lock and phy_write and phy_read take and release the lock to write
or read to a PHY register.

Let's fix this deadlock by using phy_modify_paged which handles
correctly a read followed by a write in a non-standard page.

Fixes: 6a0bfbbe ("net: phy: mscc: migrate to phy_select/restore_page functions")
Signed-off-by: Quentin Schulz <quentin.schulz@bootlin.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

The correct form is "can be probed", so fix the typo.
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Reset snd_queue tso_hdrs pointer to NULL in nicvf_free_snd_queue routine
since it is used to check if tso dma descriptor queue has been previously
allocated. The issue can be triggered with the following reproducer:

$ip link set dev enP2p1s0v0 xdpdrv obj xdp_dummy.o
$ip link set dev enP2p1s0v0 xdpdrv off

[  341.467649] WARNING: CPU: 74 PID: 2158 at mm/vmalloc.c:1511 __vunmap+0x98/0xe0
[  341.515010] Hardware name: GIGABYTE H270-T70/MT70-HD0, BIOS T49 02/02/2018
[  341.521874] pstate: 60400005 (nZCv daif +PAN -UAO)
[  341.526654] pc : __vunmap+0x98/0xe0
[  341.530132] lr : __vunmap+0x98/0xe0
[  341.533609] sp : ffff00001c5db860
[  341.536913] x29: ffff00001c5db860 x28: 0000000000020000
[  341.542214] x27: ffff810feb5090b0 x26: ffff000017e57000
[  341.547515] x25: 0000000000000000 x24: 00000000fbd00000
[  341.552816] x23: 0000000000000000 x22: ffff810feb5090b0
[  341.558117] x21: 0000000000000000 x20: 0000000000000000
[  341.563418] x19: ffff000017e57000 x18: 0000000000000000
[  341.568719] x17: 0000000000000000 x16: 0000000000000000
[  341.574020] x15: 0000000000000010 x14: ffffffffffffffff
[  341.579321] x13: ffff00008985eb27 x12: ffff00000985eb2f
[  341.584622] x11: ffff0000096b3000 x10: ffff00001c5db510
[  341.589923] x9 : 00000000ffffffd0 x8 : ffff0000086868e8
[  341.595224] x7 : 3430303030303030 x6 : 00000000000006ef
[  341.600525] x5 : 00000000003fffff x4 : 0000000000000000
[  341.605825] x3 : 0000000000000000 x2 : ffffffffffffffff
[  341.611126] x1 : ffff0000096b3728 x0 : 0000000000000038
[  341.616428] Call trace:
[  341.618866]  __vunmap+0x98/0xe0
[  341.621997]  vunmap+0x3c/0x50
[  341.624961]  arch_dma_free+0x68/0xa0
[  341.628534]  dma_direct_free+0x50/0x80
[  341.632285]  nicvf_free_resources+0x160/0x2d8 [nicvf]
[  341.637327]  nicvf_config_data_transfer+0x174/0x5e8 [nicvf]
[  341.642890]  nicvf_stop+0x298/0x340 [nicvf]
[  341.647066]  __dev_close_many+0x9c/0x108
[  341.650977]  dev_close_many+0xa4/0x158
[  341.654720]  rollback_registered_many+0x140/0x530
[  341.659414]  rollback_registered+0x54/0x80
[  341.663499]  unregister_netdevice_queue+0x9c/0xe8
[  341.668192]  unregister_netdev+0x28/0x38
[  341.672106]  nicvf_remove+0xa4/0xa8 [nicvf]
[  341.676280]  nicvf_shutdown+0x20/0x30 [nicvf]
[  341.680630]  pci_device_shutdown+0x44/0x88
[  341.684720]  device_shutdown+0x144/0x250
[  341.688640]  kernel_restart_prepare+0x44/0x50
[  341.692986]  kernel_restart+0x20/0x68
[  341.696638]  __se_sys_reboot+0x210/0x238
[  341.700550]  __arm64_sys_reboot+0x24/0x30
[  341.704555]  el0_svc_handler+0x94/0x110
[  341.708382]  el0_svc+0x8/0xc
[  341.711252] ---[ end trace 3f4019c8439959c9 ]---
[  341.715874] page:ffff7e0003ef4000 count:0 mapcount:0 mapping:0000000000000000 index:0x4
[  341.723872] flags: 0x1fffe000000000()
[  341.727527] raw: 001fffe000000000 ffff7e0003f1a008 ffff7e0003ef4048 0000000000000000
[  341.735263] raw: 0000000000000004 0000000000000000 00000000ffffffff 0000000000000000
[  341.742994] page dumped because: VM_BUG_ON_PAGE(page_ref_count(page) == 0)

where xdp_dummy.c is a simple bpf program that forwards the incoming
frames to the network stack (available here:
https://github.com/altoor/xdp_walkthrough_examples/blob/master/sample_1/xdp_dummy.c)

Fixes: 05c773f5 ("net: thunderx: Add basic XDP support")
Fixes: 4863dea3 ("net: Adding support for Cavium ThunderX network controller")
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fix smatch warning:

drivers/ptp/ptp_clock.c:298 ptp_clock_register() warn:
 passing zero to 'ERR_PTR'

'err' should be set while device_create_with_groups and
pps_register_source fails

Fixes: 85a66e55 ("ptp: create "pins" together with the rest of attributes")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Petr Machata says:

====================
switchdev: Convert switchdev_port_obj_{add,del}() to notifiers

An offloading driver may need to have access to switchdev events on
ports that aren't directly under its control. An example is a VXLAN port
attached to a bridge offloaded by a driver. The driver needs to know
about VLANs configured on the VXLAN device. However the VXLAN device
isn't stashed between the bridge and a front-panel-port device (such as
is the case e.g. for LAG devices), so the usual switchdev ops don't
reach the driver.

VXLAN is likely not the only device type like this: in theory any L2
tunnel device that needs offloading will prompt requirement of this
sort.

A way to fix this is to give up the notion of port object addition /
deletion as a switchdev operation, which assumes somewhat tight coupling
between the message producer and consumer. And instead send the message
over a notifier chain.

The series starts with a clean-up patch #1, where
SWITCHDEV_OBJ_PORT_{VLAN, MDB}() are fixed up to lift the constraint
that the passed-in argument be a simple variable named "obj".

switchdev_port_obj_add and _del are invoked in a context that permits
blocking. Not only that, at least for the VLAN notification, being able
to signal failure is actually important. Therefore introduce a new
blocking notifier chain that the new events will be sent on. That's done
in patch #2. Retain the current (atomic) notifier chain for the
preexisting notifications.

In patch #3, introduce two new switchdev notifier types,
SWITCHDEV_PORT_OBJ_ADD and SWITCHDEV_PORT_OBJ_DEL. These notifier types
communicate the same event as the corresponding switchdev op, except in
a form of a notification. struct switchdev_notifier_port_obj_info was
added to carry the fields that correspond to the switchdev op arguments.
An additional field, handled, will be used to communicate back to
switchdev that the event has reached an interested party, which will be
important for the two-phase commit.

In patches #4, #5, and #7, rocker, DSA resp. ethsw are updated to
subscribe to the switchdev blocking notifier chain, and handle the new
notifier types. #6 introduces a helper to determine whether a
netdevice corresponds to a front panel port.

What these three drivers have in common is that their ports don't
support any uppers besides bridge. That makes it possible to ignore any
notifiers that don't reference a front-panel port device, because they
are certainly out of scope.

Unlike the previous three, mlxsw and ocelot drivers admit stacked
devices as uppers. While the current switchdev code recursively descends
through layers of lower devices, eventually calling the op on a
front-panel port device, the notifier would reference a stacking device
that's one of front-panel ports uppers. The filtering is thus more
complex.

For ocelot, such iteration is currently pretty much required, because
there's no bookkeeping of LAG devices. mlxsw does keep the list of LAGs,
however it iterates the lower devices anyway when deciding whether an
event on a tunnel device pertains to the driver or not.

Therefore this patch set instead introduces, in patch #8, a helper to
iterate through lowers, much like the current switchdev code does,
looking for devices that match a given predicate.

Then in patches #9 and #10, first mlxsw and then ocelot are updated to
dispatch the newly-added notifier types to the preexisting
port_obj_add/_del handlers. The dispatch is done via the new helper, to
recursively descend through lower devices.

Finally in patch #11, the actual switch is made, retiring the current
SDO-based code in favor of a notifier.

Now that the event is distributed through a notifier, the explicit
netdevice check in rocker, DSA and ethsw doesn't let through any events
except those done on a front-panel port itself. It is therefore
unnecessary to check in VLAN-handling code whether a VLAN was added to
the bridge itself: such events will simply be ignored much sooner.
Therefore remove it in patch #12.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Due to an explicit check in rocker_world_port_obj_vlan_add(),
dsa_slave_switchdev_event() resp. port_switchdev_event(), VLAN objects
that are added to a device that is not a front-panel port device are
ignored. Therefore this check is immaterial.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Drop switchdev_ops.switchdev_port_obj_add and _del. Drop the uses of
this field from all clients, which were migrated to use switchdev
notification in the previous patches.

Add a new function switchdev_port_obj_notify() that sends the switchdev
notifications SWITCHDEV_PORT_OBJ_ADD and _DEL.

Update switchdev_port_obj_del_now() to dispatch to this new function.
Drop __switchdev_port_obj_add() and update switchdev_port_obj_add()
likewise.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Following patches will change the way of distributing port object
changes from a switchdev operation to a switchdev notifier. The
switchdev code currently recursively descends through layers of lower
devices, eventually calling the op on a front-panel port device. The
notifier will instead be sent referencing the bridge port device, which
may be a stacking device that's one of front-panel ports uppers, or a
completely unrelated device.

Dispatch the new events to ocelot_port_obj_add() resp. _del() to
maintain the same behavior that the switchdev operation based code
currently has. Pass through switchdev_handle_port_obj_add() / _del() to
handle the recursive descend, because Ocelot supports LAG uppers.

Register to the new switchdev blocking notifier chain to get the new
events when they start getting distributed.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Following patches will change the way of distributing port object
changes from a switchdev operation to a switchdev notifier. The
switchdev code currently recursively descends through layers of lower
devices, eventually calling the op on a front-panel port device. The
notifier will instead be sent referencing the bridge port device, which
may be a stacking device that's one of front-panel ports uppers, or a
completely unrelated device.

To handle SWITCHDEV_PORT_OBJ_ADD and _DEL, subscribe to the blocking
notifier chain. Dispatch to mlxsw_sp_port_obj_add() resp. _del() to
maintain the behavior that the switchdev operation based code currently
has. Defer to switchdev_handle_port_obj_add() / _del() to handle the
recursive descend, because mlxsw supports a number of upper types.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

After the transition from switchdev operations to notifier chain (which
will take place in following patches), the onus is on the driver to find
its own devices below possible layer of LAG or other uppers.

The logic to do so is fairly repetitive: each driver is looking for its
own devices among the lowers of the notified device. For those that it
finds, it calls a handler. To indicate that the event was handled,
struct switchdev_notifier_port_obj_info.handled is set. The differences
lie only in what constitutes an "own" device and what handler to call.

Therefore abstract this logic into two helpers,
switchdev_handle_port_obj_add() and switchdev_handle_port_obj_del(). If
a driver only supports physical ports under a bridge device, it will
simply avoid this layer of indirection.

One area where this helper diverges from the current switchdev behavior
is the case of mixed lowers, some of which are switchdev ports and some
of which are not. Previously, such scenario would fail with -EOPNOTSUPP.
The helper could do that for lowers for which the passed-in predicate
doesn't hold. That would however break the case that switchdev ports
from several different drivers are stashed under one master, a scenario
that switchdev currently happily supports. Therefore tolerate any and
all unknown netdevices, whether they are backed by a switchdev driver
or not.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Following patches will change the way of distributing port object
changes from a switchdev operation to a switchdev notifier. The
switchdev code currently recursively descends through layers of lower
devices, eventually calling the op on a front-panel port device. The
notifier will instead be sent referencing the bridge port device, which
may be a stacking device that's one of front-panel ports uppers, or a
completely unrelated device.

ethsw currently doesn't support any uppers other than bridge.
SWITCHDEV_OBJ_ID_HOST_MDB and _PORT_MDB objects are always notified on
the bridge port device. Thus the only case that a stacked device could
be validly referenced by port object notifications are bridge
notifications for VLAN objects added to the bridge itself. But the
driver explicitly rejects such notifications in port_vlans_add(). It is
therefore safe to assume that the only interesting case is that the
notification is on a front-panel port netdevice.

To handle SWITCHDEV_PORT_OBJ_ADD and _DEL, subscribe to the blocking
notifier chain. Dispatch to swdev_port_obj_add() resp. _del() to
maintain the behavior that the switchdev operation based code currently
has.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ethsw currently uses an open-coded comparison of netdev_ops to determine
whether whether a device represents a front panel port. Wrap this into a
named function to simplify reuse.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Following patches will change the way of distributing port object
changes from a switchdev operation to a switchdev notifier. The
switchdev code currently recursively descends through layers of lower
devices, eventually calling the op on a front-panel port device. The
notifier will instead be sent referencing the bridge port device, which
may be a stacking device that's one of front-panel ports uppers, or a
completely unrelated device.

DSA currently doesn't support any other uppers than bridge.
SWITCHDEV_OBJ_ID_HOST_MDB and _PORT_MDB objects are always notified on
the bridge port device. Thus the only case that a stacked device could
be validly referenced by port object notifications are bridge
notifications for VLAN objects added to the bridge itself. But the
driver explicitly rejects such notifications in dsa_port_vlan_add(). It
is therefore safe to assume that the only interesting case is that the
notification is on a front-panel port netdevice. Therefore keep the
filtering by dsa_slave_dev_check() in place.

To handle SWITCHDEV_PORT_OBJ_ADD and _DEL, subscribe to the blocking
notifier chain. Dispatch to rocker_port_obj_add() resp. _del() to
maintain the behavior that the switchdev operation based code currently
has.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Following patches will change the way of distributing port object
changes from a switchdev operation to a switchdev notifier. The
switchdev code currently recursively descends through layers of lower
devices, eventually calling the op on a front-panel port device. The
notifier will instead be sent referencing the bridge port device, which
may be a stacking device that's one of front-panel ports uppers, or a
completely unrelated device.

rocker currently doesn't support any uppers other than bridge. Thus the
only case that a stacked device could be validly referenced by port
object notifications are bridge notifications for VLAN objects added to
the bridge itself. But the driver explicitly rejects such notifications
in rocker_world_port_obj_vlan_add(). It is therefore safe to assume that
the only interesting case is that the notification is on a front-panel
port netdevice.

Subscribe to the blocking notifier chain. In the handler, filter out
notifications on any foreign netdevices. Dispatch the new notifiers to
rocker_port_obj_add() resp. _del() to maintain the behavior that the
switchdev operation based code currently has.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

An offloading driver may need to have access to switchdev events on
ports that aren't directly under its control. An example is a VXLAN port
attached to a bridge offloaded by a driver. The driver needs to know
about VLANs configured on the VXLAN device. However the VXLAN device
isn't stashed between the bridge and a front-panel-port device (such as
is the case e.g. for LAG devices), so the usual switchdev ops don't
reach the driver.

VXLAN is likely not the only device type like this: in theory any L2
tunnel device that needs offloading will prompt requirement of this
sort. This falsifies the assumption that only the lower devices of a
front panel port need to be notified to achieve flawless offloading.

A way to fix this is to give up the notion of port object addition /
deletion as a switchdev operation, which assumes somewhat tight coupling
between the message producer and consumer. And instead send the message
over a notifier chain.

To that end, introduce two new switchdev notifier types,
SWITCHDEV_PORT_OBJ_ADD and SWITCHDEV_PORT_OBJ_DEL. These notifier types
communicate the same event as the corresponding switchdev op, except in
a form of a notification. struct switchdev_notifier_port_obj_info was
added to carry the fields that the switchdev op carries. An additional
field, handled, will be used to communicate back to switchdev that the
event has reached an interested party, which will be important for the
two-phase commit.

The two switchdev operations themselves are kept in place. Following
patches first convert individual clients to the notifier protocol, and
only then are the operations removed.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

In general one can't assume that a switchdev notifier is called in a
non-atomic context, and correspondingly, the switchdev notifier chain is
an atomic one.

However, port object addition and deletion messages are delivered from a
process context. Even the MDB addition messages, whose delivery is
scheduled from atomic context, are queued and the delivery itself takes
place in blocking context. For VLAN messages in particular, keeping the
blocking nature is important for error reporting.

Therefore introduce a blocking notifier chain and related service
functions to distribute the notifications for which a blocking context
can be assumed.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>