rxrpc calls shouldn't hold refs on the sock struct.  This was done so that
the socket wouldn't go away whilst the call was in progress, such that the
call could reach the socket's queues.

However, we can mark the socket as requiring an RCU release and rely on the
RCU read lock.

To make this work, we do:

 (1) rxrpc_release_call() removes the call's call user ID.  This is now
     only called from socket operations and not from the call processor:

	rxrpc_accept_call() / rxrpc_kernel_accept_call()
	rxrpc_reject_call() / rxrpc_kernel_reject_call()
	rxrpc_kernel_end_call()
	rxrpc_release_calls_on_socket()
	rxrpc_recvmsg()

     Though it is also called in the cleanup path of
     rxrpc_accept_incoming_call() before we assign a user ID.

 (2) Pass the socket pointer into rxrpc_release_call() rather than getting
     it from the call so that we can get rid of uninitialised calls.

 (3) Fix call processor queueing to pass a ref to the work queue and to
     release that ref at the end of the processor function (or to pass it
     back to the work queue if we have to requeue).

 (4) Skip out of the call processor function asap if the call is complete
     and don't requeue it if the call is complete.

 (5) Clean up the call immediately that the refcount reaches 0 rather than
     trying to defer it.  Actual deallocation is deferred to RCU, however.

 (6) Don't hold socket refs for allocated calls.

 (7) Use the RCU read lock when queueing a message on a socket and treat
     the call's socket pointer according to RCU rules and check it for
     NULL.

     We also need to use the RCU read lock when viewing a call through
     procfs.

 (8) Transmit the final ACK/ABORT to a client call in rxrpc_release_call()
     if this hasn't been done yet so that we can then disconnect the call.
     Once the call is disconnected, it won't have any access to the
     connection struct and the UDP socket for the call work processor to be
     able to send the ACK.  Terminal retransmission will be handled by the
     connection processor.

 (9) Release all calls immediately on the closing of a socket rather than
     trying to defer this.  Incomplete calls will be aborted.

The call refcount model is much simplified.  Refs are held on the call by:

 (1) A socket's user ID tree.

 (2) A socket's incoming call secureq and acceptq.

 (3) A kernel service that has a call in progress.

 (4) A queued call work processor.  We have to take care to put any call
     that we failed to queue.

 (5) sk_buffs on a socket's receive queue.  A future patch will get rid of
     this.

Whilst we're at it, we can do:

 (1) Get rid of the RXRPC_CALL_EV_RELEASE event.  Release is now done
     entirely from the socket routines and never from the call's processor.

 (2) Get rid of the RXRPC_CALL_DEAD state.  Calls now end in the
     RXRPC_CALL_COMPLETE state.

 (3) Get rid of the rxrpc_call::destroyer work item.  Calls are now torn
     down when their refcount reaches 0 and then handed over to RCU for
     final cleanup.

 (4) Get rid of the rxrpc_call::deadspan timer.  Calls are cleaned up
     immediately they're finished with and don't hang around.
     Post-completion retransmission is handled by the connection processor
     once the call is disconnected.

 (5) Get rid of the dead call expiry setting as there's no longer a timer
     to set.

 (6) rxrpc_destroy_all_calls() can just check that the call list is empty.
Signed-off-by: David Howells <dhowells@redhat.com>

Use rxrpc_is_service_call() rather than rxrpc_conn_is_service() if the call
is available just in case call->conn is NULL.
Signed-off-by: David Howells <dhowells@redhat.com>

Pass the connection pointer to rxrpc_post_packet_to_call() as the call
might get disconnected whilst we're looking at it, but the connection
pointer determined by rxrpc_data_read() is guaranteed by RCU for the
duration of the call.
Signed-off-by: David Howells <dhowells@redhat.com>

Cache the security index in the rxrpc_call struct so that we can get at it
even when the call has been disconnected and the connection pointer
cleared.
Signed-off-by: David Howells <dhowells@redhat.com>

Use call->peer rather than call->conn->params.peer to avoid the possibility
of call->conn being NULL and, whilst we're at it, check it for NULL before we
access it.
Signed-off-by: David Howells <dhowells@redhat.com>

Improve the call tracking tracepoint by showing more differentiation
between some of the put and get events, including:

  (1) Getting and putting refs for the socket call user ID tree.

  (2) Getting and putting refs for queueing and failing to queue the call
      processor work item.

Note that these aren't necessarily used in this patch, but will be taken
advantage of in future patches.

An enum is added for the event subtype numbers rather than coding them
directly as decimal numbers and a table of 3-letter strings is provided
rather than a sequence of ?: operators.
Signed-off-by: David Howells <dhowells@redhat.com>

Delete rxrpc_kernel_free_skb() as it's unused.
Signed-off-by: David Howells <dhowells@redhat.com>

Remove some whitespace.
Signed-off-by: David Howells <dhowells@redhat.com>

These messages are unnecessary as OOM allocation failures already do
a dump_stack() giving more or less the same information.

$ size drivers/net/ethernet/qlogic/qed/built-in.o* (defconfig x86-64)
   text	   data	    bss	    dec	    hex	filename
 127817	  27969	  32800	 188586	  2e0aa	drivers/net/ethernet/qlogic/qed/built-in.o.new
 132474	  27969	  32800	 193243	  2f2db	drivers/net/ethernet/qlogic/qed/built-in.o.old

Miscellanea:

o Change allocs to the generally preferred forms where possible.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge tag 'rxrpc-rewrite-20160904-2' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs

David Howells says:

====================
rxrpc: Split output code from sendmsg code

Here's a set of small patches that split the packet transmission code from
the sendmsg code and simply rearrange the new file to make it more
logically laid out ready for being rewritten.  An enum is also moved out of
the header file to there as it's only used there.  This needs to be applied
on top of the just-posted fixes patch set.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge tag 'rxrpc-rewrite-20160904-1' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs

David Howells says:

====================
rxrpc: Small fixes

Here's a set of small fix patches:

 (1) Fix some uninitialised variables.

 (2) Set the client call state before making it live by attaching it to the
     conn struct.

 (3) Randomise the epoch and starting client conn ID values, and don't
     change the epoch when the client conn ID rolls round.

 (4) Replace deprecated create_singlethread_workqueue() calls.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

If vxlan_build_skb return err < 0, tx_errors should be also increased.
Signed-off-by: Haishuang Yan <yanhaishuang@cmss.chinamobile.com>
Acked-by: Jiri Benc <jbenc@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Depending on the preempt mode, the bpf_prog stored in xdp_prog may be
freed despite the use of call_rcu inside bpf_prog_put. The situation is
possible when running in PREEMPT_RCU=y mode, for instance, since the rcu
callback for destroying the bpf prog can run even during the bh handling
in the mlx4 rx path.

Several options were considered before this patch was settled on:

Add a napi_synchronize loop in mlx4_xdp_set, which would occur after all
of the rings are updated with the new program.
This approach has the disadvantage that as the number of rings
increases, the speed of update will slow down significantly due to
napi_synchronize's msleep(1).

Add a new rcu_head in bpf_prog_aux, to be used by a new bpf_prog_put_bh.
The action of the bpf_prog_put_bh would be to then call bpf_prog_put
later. Those drivers that consume a bpf prog in a bh context (like mlx4)
would then use the bpf_prog_put_bh instead when the ring is up. This has
the problem of complexity, in maintaining proper refcnts and rcu lists,
and would likely be harder to review. In addition, this approach to
freeing must be exclusive with other frees of the bpf prog, for instance
a _bh prog must not be referenced from a prog array that is consumed by
a non-_bh prog.

The placement of rcu_read_lock in this patch is functionally the same as
putting an rcu_read_lock in napi_poll. Actually doing so could be a
potentially controversial change, but would bring the implementation in
line with sk_busy_loop (though of course the nature of those two paths
is substantially different), and would also avoid future copy/paste
problems with future supporters of XDP. Still, this patch does not take
that opinionated option.

Testing was done with kernels in either PREEMPT_RCU=y or
CONFIG_PREEMPT_VOLUNTARY=y+PREEMPT_RCU=n modes, with neither exhibiting
any drawback. With PREEMPT_RCU=n, the extra call to rcu_read_lock did
not show up in the perf report whatsoever, and with PREEMPT_RCU=y the
overhead of rcu_read_lock (according to perf) was the same before/after.
In the rx path, rcu_read_lock is eventually called for every packet
from netif_receive_skb_internal, so the napi poll call's rcu_read_lock
is easily amortized.

v2:
Remove extra rcu_read_lock in mlx4_en_process_rx_cq body
Annotate xdp_prog with __rcu, and convert all usages to rcu_assign or
rcu_dereference[_protected] as appropriate.
Add explicit mutex lock around rcu_assign instead of xchg loop.

Fixes: d576acf0 ("net/mlx4_en: add page recycle to prepare rx ring for tx support")
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <alexei.starovoitov@gmail.com>
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Sean Wang says:

====================
net: ethernet: mediatek: add enhancements to RX path

Changes since v1:
- fix message typos and add coverletter

Changes since v2:
- split from the previous series for submitting add enhancements as
a series targeting 'net-next' and add indents before comments.

Changes since v3:
- merge the patch using PDMA RX path
- fixed the input of mtk_poll_rx is with the remaining budget

Changes since v4:
- save one wmb and register update when no packet is being handled
inside mtk_poll_rx call
- fixed incorrect return packet count from mtk_napi_rx
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

The patch adds support for aggregating more SKBs feed into NAPI in
order to get more benefits from generic receive offload (GRO) by
peeking at the RX ring status and moving more packets right before
returning from NAPI RX polling handler if NAPI budgets are still
available and some packets already present in hardware.
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The patch makes move wmb() to outside the loop that could help
RX path handling more faster although that RX descriptors aren't
freed for DMA to use as soon as possible, but based on my experiment
and the result shows it still can reach about 943Mbpis without
performance drop that is tested based on the setup with one port
using Giga PHY and 256 RX descriptors for DMA to move.
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Joe Perches says:

====================
hso: neatening

This seems to be the only code in the kernel that uses
macro defines with a trailing underscore.  Fix that.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Use a more common logging style

Miscellanea:

o Add pr_fmt to prefix each output message
o Realign arguments
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Macros that end in an underscore are just odd.
Add hso_dbg(level, fmt, ...) and use it everwhere instead.

Several uses had additional unnecessary newlines as the
macro added a newline.  Remove the newline from the macro
and add newlines to each use as appropriate.

Remove now unused D<digit> macros.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add mdix control through ethtool.
Signed-off-by: Woojung Huh <Woojung.huh@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add STRAP_STATUS defines.
Signed-off-by: Woojung Huh <Woojung.huh@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add Microchip Linux Driver Support as maintainer
because this driver is maintaining by Microchip.
Signed-off-by: Woojung Huh <Woojung.huh@gmail.com>
Acked-by: Steve Glendinning <steve.glendinning@shawell.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Vivien Didelot says:

====================
net: dsa: mv88e6xxx: isolate Global2 support

Registers of Marvell chips are organized in internal SMI devices.

One of them at address 0x1C is called Global2. It provides an extended
set of registers, used for interrupt control, EEPROM access, indirect
PHY access (to bypass the PHY Polling Unit) and cross-chip setup.

Most chips have it, but some others don't (older ones such as 6060).

Now that its related code is isolated in mv88e6xxx_g2_* functions, move
it to its own global2.c file, making most of its setup code static.

Then make its compilation optional, which allows to reduce the size of
the mv88e6xxx driver for devices such as home routers embedding Ethernet
chips without Global2 support.

It is present on most recent chips, thus enable its support by default.

Changes in v2: fail probe if GLOBAL2 is required but not enabled.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Since not every chip has a Global2 set of registers, make its support
optional, in which case the related functions will return -EOPNOTSUPP.

This also allows to reduce the size of the mv88e6xxx driver for devices
such as home routers embedding Ethernet chips without Global2 support.

It is present on most recent chips, thus enable its support by default.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Marvell chips are composed of multiple SMI devices. One of them at
address 0x1C is called Global2. It provides an extended set of
registers, used for interrupt control, EEPROM access, indirect PHY
access (to bypass the PHY Polling Unit) and cross-chip related setup.

Most chips have it, but some others don't (older ones such as 6060).

Now that its related code is isolated in mv88e6xxx_g2_* functions, move
it to its own global2.c file, making most of its setup code static.
Document each registers in the meantime.

Its compilation can be later avoided for chips without such registers.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Since the mv88e6xxx.c file has been renamed, the driver compiled as a
module is called chip.ko instead of mv88e6xxx.ko. Fix this.

Fixes: fad09c73 ("net: dsa: mv88e6xxx: rename single-chip support")
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Pablo Neira Ayuso says:

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

The following patchset contains Netfilter updates for your net-next
tree.  Most relevant updates are the removal of per-conntrack timers to
use a workqueue/garbage collection approach instead from Florian
Westphal, the hash and numgen expression for nf_tables from Laura
Garcia, updates on nf_tables hash set to honor the NLM_F_EXCL flag,
removal of ip_conntrack sysctl and many other incremental updates on our
Netfilter codebase.

More specifically, they are:

1) Retrieve only 4 bytes to fetch ports in case of non-linear skb
   transport area in dccp, sctp, tcp, udp and udplite protocol
   conntrackers, from Gao Feng.

2) Missing whitespace on error message in physdev match, from Hangbin Liu.

3) Skip redundant IPv4 checksum calculation in nf_dup_ipv4, from Liping Zhang.

4) Add nf_ct_expires() helper function and use it, from Florian Westphal.

5) Replace opencoded nf_ct_kill() call in IPVS conntrack support, also
   from Florian.

6) Rename nf_tables set implementation to nft_set_{name}.c

7) Introduce the hash expression to allow arbitrary hashing of selector
   concatenations, from Laura Garcia Liebana.

8) Remove ip_conntrack sysctl backward compatibility code, this code has
   been around for long time already, and we have two interfaces to do
   this already: nf_conntrack sysctl and ctnetlink.

9) Use nf_conntrack_get_ht() helper function whenever possible, instead
   of opencoding fetch of hashtable pointer and size, patch from Liping Zhang.

10) Add quota expression for nf_tables.

11) Add number generator expression for nf_tables, this supports
    incremental and random generators that can be combined with maps,
    very useful for load balancing purpose, again from Laura Garcia Liebana.

12) Fix a typo in a debug message in FTP conntrack helper, from Colin Ian King.

13) Introduce a nft_chain_parse_hook() helper function to parse chain hook
    configuration, this is used by a follow up patch to perform better chain
    update validation.

14) Add rhashtable_lookup_get_insert_key() to rhashtable and use it from the
    nft_set_hash implementation to honor the NLM_F_EXCL flag.

15) Missing nulls check in nf_conntrack from nf_conntrack_tuple_taken(),
    patch from Florian Westphal.

16) Don't use the DYING bit to know if the conntrack event has been already
    delivered, instead a state variable to track event re-delivery
    states, also from Florian.

17) Remove the per-conntrack timer, use the workqueue approach that was
    discussed during the NFWS, from Florian Westphal.

18) Use the netlink conntrack table dump path to kill stale entries,
    again from Florian.

19) Add a garbage collector to get rid of stale conntracks, from
    Florian.

20) Reschedule garbage collector if eviction rate is high.

21) Get rid of the __nf_ct_kill_acct() helper.

22) Use ARPHRD_ETHER instead of hardcoded 1 from ARP logger.

23) Make nf_log_set() interface assertive on unsupported families.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Move enum rxrpc_command to sendmsg.c as it's now only used in that file.
Signed-off-by: David Howells <dhowells@redhat.com>

The workqueue "afs_lock_manager" queues work item &vnode->lock_work,
per vnode. Since there can be multiple vnodes and since their work items
can be executed concurrently, alloc_workqueue has been used to replace
the deprecated create_singlethread_workqueue instance.

The WQ_MEM_RECLAIM flag has been set to ensure forward progress under
memory pressure because the workqueue is being used on a memory reclaim
path.

Since there are fixed number of work items, explicit concurrency
limit is unnecessary here.
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Signed-off-by: David Howells <dhowells@redhat.com>

Rearrange net/rxrpc/sendmsg.c to be in a more logical order.  This makes it
easier to follow and eliminates forward declarations.
Signed-off-by: David Howells <dhowells@redhat.com>

The workqueue "afs_callback_update_worker" queues multiple work items
viz  &vnode->cb_broken_work, &server->cb_break_work which require strict
execution ordering. Hence, an ordered dedicated workqueue has been used.

Since the workqueue is being used on a memory reclaim path, WQ_MEM_RECLAIM
has been set to ensure forward progress under memory pressure.
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Signed-off-by: David Howells <dhowells@redhat.com>

Split the sendmsg code from the packet transmission code (mostly to be
found in output.c).
Signed-off-by: David Howells <dhowells@redhat.com>

The workqueue "afs_async_calls" queues work item
&call->async_work per afs_call. Since there could be multiple calls and since
these calls can be run concurrently, alloc_workqueue has been used to replace
the deprecated create_singlethread_workqueue instance.

The WQ_MEM_RECLAIM flag has been set to ensure forward progress under
memory pressure because the workqueue is being used on a memory reclaim
path.

Since there are fixed number of work items, explicit concurrency
limit is unnecessary here.
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Signed-off-by: David Howells <dhowells@redhat.com>

The workqueue "afs_vlocation_update_worker" queues a single work item
&afs_vlocation_update and hence it doesn't require execution ordering.
Hence, alloc_workqueue has been used to replace the deprecated
create_singlethread_workqueue instance.

Since the workqueue is being used on a memory reclaim path, WQ_MEM_RECLAIM
flag has been set to ensure forward progress under memory pressure.

Since there are fixed number of work items, explicit concurrency
limit is unnecessary here.
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Signed-off-by: David Howells <dhowells@redhat.com>

It seems the local epoch should only be changed on boot, so remove the code
that changes it for client connections.
Signed-off-by: David Howells <dhowells@redhat.com>

Create a random epoch value rather than a time-based one on startup and set
the top bit to indicate that this is the case.

Also create a random starting client connection ID value.  This will be
incremented from here as new client connections are created.
Signed-off-by: David Howells <dhowells@redhat.com>

cpmac_start_xmit() used the max() macro on skb->len (an unsigned int)
and ETH_ZLEN (a signed int literal). This led to the following compiler
warning:

  In file included from include/linux/list.h:8:0,
                   from include/linux/module.h:9,
                   from drivers/net/ethernet/ti/cpmac.c:19:
  drivers/net/ethernet/ti/cpmac.c: In function 'cpmac_start_xmit':
  include/linux/kernel.h:748:17: warning: comparison of distinct pointer
  types lacks a cast
    (void) (&_max1 == &_max2);  \
                   ^
  drivers/net/ethernet/ti/cpmac.c:560:8: note: in expansion of macro 'max'
    len = max(skb->len, ETH_ZLEN);
          ^

On top of this, it assigned the result of the max() macro to a signed
integer whilst all further uses of it result in it being cast to varying
widths of unsigned integer.

Fix this up by using max_t to ensure the comparison is performed as
unsigned integers, and for consistency change the type of the len
variable to unsigned int.
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Adds support for ndo_get_vf_config, also fill the default mac address
that will be provided to the VF by firmware, in case user doesn't
provide one. So user can get the default MAC address address also
through ndo_get_vf_config.
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Cong Wang says:

====================
net: some minor optimization for netns id

Cong Wang (2):
  vxlan: call peernet2id() in fdb notification
  netns: avoid disabling irq for netns id
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

We never read or change netns id in hardirq context,
the only place we read netns id in softirq context
is in vxlan_xmit(). So, it should be enough to just
disable BH.

Cc: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>