Delay update of hw tail descriptor if we know that another skb is going
to be sent.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

Ying Xue says:

====================
Involve rhashtable_lookup_insert routine

The series aims to involve rhashtable_lookup_insert() to guarantee
that the process of lookup and insertion of an object from/into hash
table is finished atomically, allowing rhashtable's users not to
introduce an extra lock during search and insertion. For example,
tipc socket is the first user benefiting from this enhancement.

v2 changes:
 - fix the issue of waking up worker thread under a wrong condition in
   patch #2, which is pointed by Thomas.
 - move a comment from rhashtable_inser() to rhashtable_wakeup_worker()
   according to Thomas's suggestion in patch #2.
 - indent the third line of condition statement in
   rhashtable_wakeup_worker() to inner bracket in patch #2.
 - drop patch #3 of v1 series
 - fix an issue of being unable to remove an object from hash table in
   certain special case in patch #4.
 - involve a new patch #5 to avoid unnecessary wakeup for worker queue
   thread
 - involve a new patch #6 to initialize atomic "nelems" variable
 - adjust "nelem_hint" value from 256 to 192 avoiding to unnecessarily
   to shrink hash table from the beginning phase in patch #7.

v1 changes:
 But before rhashtable_lookup_insert() is involved, the following
 optimizations need to be first done:
- simplify rhashtable_lookup by reusing rhashtable_lookup_compare()
- introduce rhashtable_wakeup_worker() to further reduce duplicated
  code in patch #2
- fix an issue in patch #3
- involve rhashtable_lookup_insert(). But in this version, we firstly
  use rhashtable_lookup() to search duplicate key in both old and new
  bucket table; secondly introduce another __rhashtable_insert() helper
  function to reduce the duplicated code between rhashtable_insert()
  and rhashtable_lookup_insert().
- add patch #5 into the series as it depends on above patches. But in
  this version, no change is made comparing with its previous version.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

As tipc reference table is statically allocated, its memory size
requested on stack initialization stage is quite big even if the
maximum port number is just restricted to 8191 currently, however,
the number already becomes insufficient in practice. But if the
maximum ports is allowed to its theory value - 2^32, its consumed
memory size will reach a ridiculously unacceptable value. Apart from
this, heavy tipc users spend a considerable amount of time in
tipc_sk_get() due to the read-lock on ref_table_lock.

If tipc reference table is converted with generic rhashtable, above
mentioned both disadvantages would be resolved respectively: making
use of the new resizable hash table can avoid locking on the lookup;
smaller memory size is required at initial stage, for example, 256
hash bucket slots are requested at the beginning phase instead of
allocating the entire 8191 slots in old mode. The hash table will
grow if entries exceeds 75% of table size up to a total table size
of 1M, and it will automatically shrink if usage falls below 30%,
but the minimum table size is allowed down to 256.

Also converts ref_table_lock to a separate mutex to protect hash table
mutations on write side. Lastly defers the release of the socket
reference using call_rcu() to allow using an RCU read-side protected
call to rhashtable_lookup().
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Acked-by: Erik Hugne <erik.hugne@ericsson.com>
Cc: Thomas Graf <tgraf@suug.ch>
Acked-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Ying Xue <ying.xue@windriver.com>
Cc: Thomas Graf <tgraf@suug.ch>
Acked-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Move condition statements of verifying whether hash table size exceeds
its maximum threshold or reaches its minimum threshold from resizing
functions to resizing decision functions, avoiding unnecessary wakeup
for worker queue thread.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Cc: Thomas Graf <tgraf@suug.ch>
Acked-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

When remove an object from hash table, we currently only traverse old
bucket table to check whether the object exists. If the object is not
found in it, we will try again. But in the second search loop, we still
search the object from the old table instead of future table. As a
result, the object may be not removed from hash table especially when
resizing is currently in progress and the object is just saved in the
future table.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Cc: Thomas Graf <tgraf@suug.ch>
Acked-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Involve a new function called rhashtable_lookup_insert() which makes
lookup and insertion atomic under bucket lock protection, helping us
avoid to introduce an extra lock when we search and insert an object
into hash table.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Acked-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Introduce rhashtable_wakeup_worker() helper function to reduce
duplicated code where to wake up worker.

By the way, as long as the both "future_tbl" and "tbl" bucket table
pointers point to the same bucket array, we should try to wake up
the resizing worker thread, otherwise, it indicates the work of
resizing hash table is not finished yet. However, currently we will
wake up the worker thread only when the two pointers point to
different bucket array. Obviously this is wrong. So, the issue is
also fixed as well in the patch.
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Cc: Thomas Graf <tgraf@suug.ch>
Acked-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Define an internal compare function and relevant compare argument,
and then make use of rhashtable_lookup_compare() to lookup key in
hash table, reducing duplicated code between rhashtable_lookup()
and rhashtable_lookup_compare().
Signed-off-by: Ying Xue <ying.xue@windriver.com>
Cc: Thomas Graf <tgraf@suug.ch>
Acked-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Hariprasad Shenai says:

====================
Add support for few debugfs entries

This patch series adds support for devlog, cim_la, cim_qcfg and mps_tcam
debugfs entries.

The patches series is created against 'net-next' tree.
And includes patches on cxgb4 driver.

We have included all the maintainers of respective drivers. Kindly review the
change and let us know in case of any review comments.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Debug log to get the MPS TCAM table
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Adds debug log to get cim queue config
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The CIM LA captures the embedded processor’s internal state. Optionally, it can
also trace the flow of data in and out of the embedded processor. Therefore, the
CIM LA output contains detailed information of what code the embedded processor
executed prior to the CIM LA capture.
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add support for device log entry in debugfs
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Vinson reported:

  HOSTCC  Documentation/networking/timestamping/txtimestamp
Documentation/networking/timestamping/txtimestamp.c:64:8: error:
redefinition of ‘struct in6_pktinfo’
 struct in6_pktinfo {
        ^
In file included from /usr/include/arpa/inet.h:23:0,
                 from Documentation/networking/timestamping/txtimestamp.c:33:
/usr/include/netinet/in.h:456:8: note: originally defined here
 struct in6_pktinfo
        ^

After we sync with libc header, we don't need this ugly hack any more.
Reported-by: Vinson Lee <vlee@twopensource.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Both netinet/in.h and linux/ipv6.h define these two structs,
if we include both of them, we got:

	/usr/include/linux/ipv6.h:19:8: error: redefinition of ‘struct in6_pktinfo’
	 struct in6_pktinfo {
		^
	In file included from /usr/include/arpa/inet.h:22:0,
			 from txtimestamp.c:33:
	/usr/include/netinet/in.h:524:8: note: originally defined here
	 struct in6_pktinfo
		^
	In file included from txtimestamp.c:40:0:
	/usr/include/linux/ipv6.h:24:8: error: redefinition of ‘struct ip6_mtuinfo’
	 struct ip6_mtuinfo {
		^
	In file included from /usr/include/arpa/inet.h:22:0,
			 from txtimestamp.c:33:
	/usr/include/netinet/in.h:531:8: note: originally defined here
	 struct ip6_mtuinfo
		^
So similarly to what we did for in6_addr, we need to sync with
libc header on their definitions.
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Pull networking fixes from David Miller:
 "Just a pile of random fixes, including:

   1) Do not apply TSO limits to non-TSO packets, fix from Herbert Xu.

   2) MDI{,X} eeprom check in e100 driver is reversed, from John W.
      Linville.

   3) Missing error return assignments in several ethernet drivers, from
      Julia Lawall.

   4) Altera TSE device doesn't come back up after ifconfig down/up
      sequence, fix from Kostya Belezko.

   5) Add more cases to the check for whether the qmi_wwan device has a
      bogus MAC address and needs to be assigned a random one.  From
      Kristian Evensen.

   6) Fix interrupt hangs in CPSW, from Felipe Balbi.

   7) Implement ndo_features_check in r8152 so that the stack doesn't
      feed GSO packets which are outside of the chip's capabilities.
      From Hayes Wang"

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (26 commits)
  qla3xxx: don't allow never end busy loop
  xen-netback: fixing the propagation of the transmit shaper timeout
  r8152: support ndo_features_check
  batman-adv: fix potential TT client + orig-node memory leak
  batman-adv: fix multicast counter when purging originators
  batman-adv: fix counter for multicast supporting nodes
  batman-adv: fix lock class for decoding hash in network-coding.c
  batman-adv: fix delayed foreign originator recognition
  batman-adv: fix and simplify condition when bonding should be used
  Revert "mac80211: Fix accounting of the tailroom-needed counter"
  net: ethernet: cpsw: fix hangs with interrupts
  enic: free all rq buffs when allocation fails
  qmi_wwan: Set random MAC on devices with buggy fw
  openvswitch: Consistently include VLAN header in flow and port stats.
  tcp: Do not apply TSO segment limit to non-TSO packets
  Altera TSE: Add missing phydev
  net/mlx4_core: Fix error flow in mlx4_init_hca()
  net/mlx4_core: Correcly update the mtt's offset in the MR re-reg flow
  qlcnic: Fix return value in qlcnic_probe()
  net: axienet: fix error return code
  ...

Pull IPMI fixlet from Corey Minyard:
 "Fix a compile warning"

* tag 'for-linus-3' of git://git.code.sf.net/p/openipmi/linux-ipmi:
  ipmi: Fix compile warning with tv_usec

This adds support for APM X-Gene ethernet driver to use ACPI table to derive
ethernet driver parameter.
Signed-off-by: Feng Kan <fkan@apm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The counter variable wasn't increased at all which may stuck under
certain circumstances.
Signed-off-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The Freescale mEMAC supports operating at 10/100/1000/10G, and
its associated MDIO controller is likewise capable of operating
both Clause 22 and Clause 45 MDIO buses. It is nearly identical
to the MDIO controller on the XGMAC, so we just modify that
driver.

Portions of this driver developed by:

Sandeep Singh <sandeep@freescale.com>
Roy Zang <tie-fei.zang@freescale.com>
Signed-off-by: Andy Fleming <afleming@gmail.com>
Signed-off-by: Shaohui Xie <Shaohui.Xie@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch extends the ethtool plugin module eeprom API to support cards
whose phy support is delegated to a separate driver.

The handlers for ETHTOOL_GMODULEINFO and ETHTOOL_GMODULEEEPROM call the
module_info and module_eeprom functions if the phy driver provides them;
otherwise the handlers call the equivalent ethtool_ops functions provided
by network drivers with built-in phy support.
Signed-off-by: Ed Swierk <eswierk@skyportsystems.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Pull ext4 bugfixes from Ted Ts'o:
 "Revert a potential seek_data/hole regression which shows up when using
  ext4 to handle ext3 file systems, plus two minor bug fixes"

* tag 'ext4_for_linus_stable' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  ext4: remove spurious KERN_INFO from ext4_warning call
  Revert "ext4: fix suboptimal seek_{data,hole} extents traversial"
  ext4: prevent online resize with backup superblock

Jay Foad reports that the address sanitizer test (asan) sometimes gets
confused by a stack pointer that ends up being outside the stack vma
that is reported by /proc/maps.

This happens due to an interaction between RLIMIT_STACK and the guard
page: when we do the guard page check, we ignore the potential error
from the stack expansion, which effectively results in a missing guard
page, since the expected stack expansion won't have been done.

And since /proc/maps explicitly ignores the guard page (commit
d7824370: "mm: fix up some user-visible effects of the stack guard
page"), the stack pointer ends up being outside the reported stack area.

This is the minimal patch: it just propagates the error.  It also
effectively makes the guard page part of the stack limit, which in turn
measn that the actual real stack is one page less than the stack limit.

Let's see if anybody notices.  We could teach acct_stack_growth() to
allow an extra page for a grow-up/grow-down stack in the rlimit test,
but I don't want to add more complexity if it isn't needed.
Reported-and-tested-by: Jay Foad <jay.foad@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Included changes:
- ensure bonding is used (if enabled) for packets coming in the soft
  interface
- fix race condition to avoid orig_nodes to be deleted right after
  being added
- avoid false positive lockdep splats by assigning lockclass to
  the proper hashtable lock objects
- avoid miscounting of multicast 'disabled' nodes in the network
- fix memory leak in the Global Translation Table in case of
  originator interval change
Signed-off-by: David S. Miller <davem@davemloft.net>

Since e9ce7cb6 ("xen-netback: Factor queue-specific data into queue struct"),
the transimt shaper timeout is always set to 0.  The value the user sets via
xenbus is never propagated to the transmit shaper.

This patch fixes the issue.

Cc: Anthony Liguori <aliguori@amazon.com>
Signed-off-by: Imre Palik <imrep@amazon.de>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Rx ring 2 size can be configured by adjusting rx-jumbo parameter
of ethtool -G.
Signed-off-by: Ramya Bolla <bollar@vmware.com>
Signed-off-by: Shreyas Bhatewara <sbhatewara@vmware.com>
Signed-off-by: Shrikrishna Khare <skhare@vmware.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge tag 'mac80211-for-davem-2015-01-06' of git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211

Here's just a single fix - a revert of a patch that broke the
p54 and cw2100 drivers (arguably due to bad assumptions there.)
Since this affects kernels since 3.17, I decided to revert for
now and we'll revisit this optimisation properly for -next.
Signed-off-by: David S. Miller <davem@davemloft.net>

Support ndo_features_check to avoid:
 - the transport offset is more than the hw limitation when using hw checksum.
 - the skb->len of a GSO packet is more than the limitation.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This driver makes use of the clocksource code. Previously it had only
included the proper header indirectly, but that chain was inadvertently
broken by 74d23cc7 "time: move the timecounter/cyclecounter code into its
own file."

This patch fixes the issue by including clocksource.h directly.
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch fixes a potential memory leak which can occur once an
originator times out. On timeout the according global translation table
entry might not get purged correctly. Furthermore, the non purged TT
entry will cause its orig-node to leak, too. Which additionally can lead
to the new multicast optimization feature not kicking in because of a
therefore bogus counter.

In detail: The batadv_tt_global_entry->orig_list holds the reference to
the orig-node. Usually this reference is released after
BATADV_PURGE_TIMEOUT through: _batadv_purge_orig()->
batadv_purge_orig_node()->batadv_update_route()->_batadv_update_route()->
batadv_tt_global_del_orig() which purges this global tt entry and
releases the reference to the orig-node.

However, if between two batadv_purge_orig_node() calls the orig-node
timeout grew to 2*BATADV_PURGE_TIMEOUT then this call path isn't
reached. Instead the according orig-node is removed from the
originator hash in _batadv_purge_orig(), the batadv_update_route()
part is skipped and won't be reached anymore.

Fixing the issue by moving batadv_tt_global_del_orig() out of the rcu
callback.
Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue>
Acked-by: Antonio Quartulli <antonio@meshcoding.com>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>

When purging an orig_node we should only decrease counter tracking the
number of nodes without multicast optimizations support if it was
increased through this orig_node before.

A not yet quite initialized orig_node (meaning it did not have its turn
in the mcast-tvlv handler so far) which gets purged would not adhere to
this and will lead to a counter imbalance.

Fixing this by adding a check whether the orig_node is mcast-initalized
before decreasing the counter in the mcast-orig_node-purging routine.

Introduced by 60432d75
("batman-adv: Announce new capability via multicast TVLV")
Reported-by: Tobias Hachmer <tobias@hachmer.de>
Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>

A miscounting of nodes having multicast optimizations enabled can lead
to multicast packet loss in the following scenario:

If the first OGM a node receives from another one has no multicast
optimizations support (no multicast tvlv) then we are missing to
increase the counter. This potentially leads to the wrong assumption
that we could safely use multicast optimizations.

Fixings this by increasing the counter if the initial OGM has the
multicast TVLV unset, too.

Introduced by 60432d75
("batman-adv: Announce new capability via multicast TVLV")
Reported-by: Tobias Hachmer <tobias@hachmer.de>
Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>

batadv_has_set_lock_class() is called with the wrong hash table as first
argument (probably due to a copy-paste error), which leads to false
positives when running with lockdep.

Introduced-by: 612d2b4f
("batman-adv: network coding - save overheard and tx packets for decoding")
Signed-off-by: Martin Hundebøll <martin@hundeboll.net>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>

Currently it can happen that the reception of an OGM from a new
originator is not being accepted. More precisely it can happen that
an originator struct gets allocated and initialized
(batadv_orig_node_new()), even the TQ gets calculated and set correctly
(batadv_iv_ogm_calc_tq()) but still the periodic orig_node purging
thread will decide to delete it if it has a chance to jump between
these two function calls.

This is because batadv_orig_node_new() initializes the last_seen value
to zero and its caller (batadv_iv_ogm_orig_get()) makes it visible to
other threads by adding it to the hash table already.
batadv_iv_ogm_calc_tq() will set the last_seen variable to the correct,
current time a few lines later but if the purging thread jumps in between
that it will think that the orig_node timed out and will wrongly
schedule it for deletion already.

If the purging interval is the same as the originator interval (which is
the default: 1 second), then this game can continue for several rounds
until the random OGM jitter added enough difference between these
two (in tests, two to about four rounds seemed common).

Fixing this by initializing the last_seen variable of an orig_node
to the current time before adding it to the hash table.
Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>

The current condition actually does NOT consider bonding when the
interface the packet came in from is the soft interface, which is the
opposite of what it should do (and the comment describes). Fix that and
slightly simplify the condition.
Reported-by: Ray Gibson <booray@gmail.com>
Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>

Daniel Borkmann says:

====================
net: allow setting congctl via routing table

This is the second part of our work and allows for setting the congestion
control algorithm via routing table. For details, please see individual
patches.

Since patch 1 is a bug fix, we suggest applying patch 1 to net, and then
merging net into net-next, for example, and following up with the remaining
feature patches wrt dependencies.

Joint work with Florian Westphal, suggested by Hannes Frederic Sowa.

Patch for iproute2 is available under [1], but will be reposted with along
with the man-page update when this set hits net-next.

  [1] http://patchwork.ozlabs.org/patch/418149/

Thanks!

v2 -> v3:
 - Added module auto-loading as suggested by David Miller, thanks!
  - Added patch 2 for handling possible sleeps in fib6
  - While working on this, we discovered a bug, hence fix in patch 1
  - Added auto-loading to patch 4
 - Rebased, retested, rest the same.
v1 -> v2:
 - Very sorry, I noticed I had decnet disabled during testing.
   Added missing header include in decnet, rest as is.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

This work adds the possibility to define a per route/destination
congestion control algorithm. Generally, this opens up the possibility
for a machine with different links to enforce specific congestion
control algorithms with optimal strategies for each of them based
on their network characteristics, even transparently for a single
application listening on all links.

For our specific use case, this additionally facilitates deployment
of DCTCP, for example, applications can easily serve internal
traffic/dsts in DCTCP and external one with CUBIC. Other scenarios
would also allow for utilizing e.g. long living, low priority
background flows for certain destinations/routes while still being
able for normal traffic to utilize the default congestion control
algorithm. We also thought about a per netns setting (where different
defaults are possible), but given its actually a link specific
property, we argue that a per route/destination setting is the most
natural and flexible.

The administrator can utilize this through ip-route(8) by appending
"congctl [lock] <name>", where <name> denotes the name of a
congestion control algorithm and the optional lock parameter allows
to enforce the given algorithm so that applications in user space
would not be allowed to overwrite that algorithm for that destination.

The dst metric lookups are being done when a dst entry is already
available in order to avoid a costly lookup and still before the
algorithms are being initialized, thus overhead is very low when the
feature is not being used. While the client side would need to drop
the current reference on the module, on server side this can actually
even be avoided as we just got a flat-copied socket clone.

Joint work with Florian Westphal.
Suggested-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>