This patch adds a new operation to dump the content of the dying and
unconfirmed lists.

Under some situations, the global conntrack counter can be inconsistent
with the number of entries that we can dump from the conntrack table.
The way to resolve this is to allow dumping the content of the unconfirmed
and dying lists, so far it was not possible to look at its content.

This provides some extra instrumentation to resolve problematic situations
in which anyone suspects memory leaks.
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

This patch modifies the conntrack subsystem so that all existing
allocated conntrack objects can be found in any of the following
places:

* the hash table, this is the typical place for alive conntrack objects.
* the unconfirmed list, this is the place for newly created conntrack objects
  that are still traversing the stack.
* the dying list, this is where you can find conntrack objects that are dying
  or that should die anytime soon (eg. once the destroy event is delivered to
  the conntrackd daemon).

Thus, we make sure that we follow the track for all existing conntrack
objects. This patch, together with some extension of the ctnetlink interface
to dump the content of the dying and unconfirmed lists, will help in case
to debug suspected nf_conn object leaks.
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

The max number of sets was hardcoded at kernel cofiguration time and
could only be modified via a module parameter. The patch adds the support
of increasing the max number of sets automatically, as needed.

The array of sets is incremented by 64 new slots if we run out of
empty slots. The absolute limit for the maximal number of sets
is limited by 65534.
Signed-off-by: NJozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

These calls are followed by calls to memcpy() on the same memory area, so they
can safely be removed.
Signed-off-by: NCyril Roelandt <tipecaml@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

TCP coalescing added a regression in splice(socket->pipe) performance,
for some workloads because of the way tcp_read_sock() is implemented.

The reason for this is the break when (offset + 1 != skb->len).

As we released the socket lock, this condition is possible if TCP stack
added a fragment to the skb, which can happen with TCP coalescing.

So let's go back to the beginning of the loop when this happens,
to give a chance to splice more frags per system call.

Doing so fixes the issue and makes GRO 10% faster than LRO
on CPU-bound splice() workloads instead of the opposite.
Signed-off-by: NWilly Tarreau <w@1wt.eu>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

# make C=2 CF=-D__CHECK_ENDIAN__ net/ipv4/inet_hashtables.o
...
net/ipv4/inet_hashtables.c:242:7: warning: restricted __portpair degrades to integer
net/ipv4/inet_hashtables.c:242:7: warning: restricted __addrpair degrades to integer
...

Move __portpair/__addrpair from include/net/inet_hashtables.h
to include/net/sock.h where we need them in struct sock_common
Reported-by: NFengguang Wu <fengguang.wu@intel.com>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Ling Ma <ling.ma.program@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

bnx2x driver could only have enabled pfc via usage of dcbnl; now, it can
also correctly disable it.
Signed-off-by: NBarak Witkowski <barak@broadcom.com>
Signed-off-by: NYuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When configuring pauses using 'ethtool -A', the requested values have
effect when used together with autoneg (up to this point, when configured
for autoneg, driver ignored requested pause configuration)
Signed-off-by: NYuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NBarak Witkowski <barak@broadcom.com>
Signed-off-by: NYuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

A rare case of no link due to a missed interrupt may occur due to a
race condition between acknowledging the IGU via the BAR and restoring the NIG
interrupt mask via the GRC.
To solve it, we wait for the IGU ack command to finish prior to restoring the
NIG interrupt mask.
Signed-off-by: NYaniv Rosner <yaniv.rosner@broadcom.com>
Signed-off-by: NYuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Unmasked interrupt caused "FATAL HW block attention set2 0x20" messages
to erroneously appear, as the associated interrupt is fully recoverable.
Signed-off-by: NYuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If IGU parse error is encountered during the probing process, the error
propagates and the probe gracefully fails (until now, such errors were ignored,
later causing mischief).
Signed-off-by: NBarak Witkowski <barak@broadcom.com>
Signed-off-by: NYuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NYuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

It is now possible to enable dropless flow control via nvram.
Signed-off-by: NYuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If link is down due to management (and not due to actual phy link being lost),
driver should still behave as if the link is down; Querying via ethtool about
speed/duplex state should result in 'UNKNOWN' (same behaviour as when link is
actually down).
Signed-off-by: NYuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Whenever bnx2x fails to transmit a packet due to a full Tx ring, if the
ring size is zero (indicating an FCoE ring) driver filters the packet out
and gracefully continues.
Driver also gathers statistics on such filtered packets.
Signed-off-by: NDmitry Kravkov <dmitry@broadcom.com>
Signed-off-by: NYuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If configured for PFC/ETS by management, configure chip regardless of the
presence of a remote peer which supports DCBX.
Signed-off-by: NBarak Witkowski <barak@broadcom.com>
Signed-off-by: NYuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Parity recovery was enhanced in order to handle a few more corner cases.
Signed-off-by: NBarak Witkowski <barak@broadcom.com>
Signed-off-by: NYuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Changed naming convention of SPIO macros, and prevented access to invalid SPIOs.
Signed-off-by: NYuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: NEilon Greenstein <eilong@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add support to changing number of rx/tx channels using
ethtool ('ethtool -[lL]'). Where the number of tx channels specified in ethtool
is the number of rings per user priority - not total number of tx rings.
Signed-off-by: NAmir Vadai <amirv@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We need to re-set tx moderation information after calling set_ringparam
else default tx moderation will be used.
Also avoid related code duplication, by putting it in a utility function.
Signed-off-by: NAmir Vadai <amirv@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Set mlx4_en maintainer to Amir Vadai instead of Yevgeny Petrilin.
Signed-off-by: NAmir Vadai <amirv@mellanox.com>
Cc: Yevgeny Petrilin <yevgenyp@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

David Woodhouse says:

====================
This is the result of pulling on the thread started by Krzysztof Mazur's
original patch 'pppoatm: don't send frames to destroyed vcc'.

Various problems in the pppoatm and br2684 code are solved, some of which
were easily triggered and would panic the kernel.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

No idea why we've gone so long dumping a list of VCCs with vci==0 on
every ->open() call...
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

 - Flush pending TX skbs from the queue rather than waiting for them all to
   complete (suggested by Krzysztof Mazur <krzysiek@podlesie.net>).
 - Clear ATM_VF_ADDR only when the PKT_PCLOSE packet has been submitted.
 - Don't clear ATM_VF_READY at all — vcc_destroy_socket() does that for us.
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

We don't need to schedule the wakeup tasklet on *every* unlock; only if we
actually blocked the channel in the first place.
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>
Acked-by: NKrzysztof Mazur <krzysiek@podlesie.net>

The br2684 does not check if used vcc is in connected state,
causing potential Oops in pppoatm_send() when vcc->send() is called
on not fully connected socket.

Now br2684 can be assigned only on connected sockets; otherwise
-EINVAL error is returned.
Signed-off-by: NKrzysztof Mazur <krzysiek@podlesie.net>
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

... and ensure that the next skb is set up for RX in the DMA case.
Signed-off-by: NNathan Williams <nathan@traverse.com.au>
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

The br2684 code used module_put() during unassignment from vcc with
hope that we have BKL. This assumption is no longer true.

Now owner field in atmvcc is used to move this module_put()
to vcc_destroy_socket().
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>
Acked-by: NKrzysztof Mazur <krzysiek@podlesie.net>

Now that we can return zero from pppoatm_send() for reasons *other* than
the queue being full, that means we can't depend on a subsequent call to
pppoatm_pop() waking the queue, and we might leave it stalled
indefinitely.

Use the ->release_cb() callback to wake the queue after the sock is
unlocked.
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>
Acked-by: NKrzysztof Mazur <krzysiek@podlesie.net>

Avoid submitting packets to a vcc which is being closed. Things go badly
wrong when the ->pop method gets later called after everything's been
torn down.

Use the ATM socket lock for synchronisation with vcc_destroy_socket(),
which clears the ATM_VF_READY bit under the same lock. Otherwise, we
could end up submitting a packet to the device driver even after its
->ops->close method has been called. And it could call the vcc's ->pop
method after the protocol has been shut down. Which leads to a panic.
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>
Acked-by: NKrzysztof Mazur <krzysiek@podlesie.net>

The immediate use case for this is that it will allow us to ensure that a
pppoatm queue is woken after it has to drop a packet due to the sock being
locked.

Note that 'release_cb' is called when the socket is *unlocked*. This is
not to be confused with vcc_release() — which probably ought to be called
vcc_close().
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>
Acked-by: NKrzysztof Mazur <krzysiek@podlesie.net>

We should no longer be calling the old pop routine for the vcc, after
vcc_release() has completed. Make sure we wait for any pending TX skbs
to complete, by waiting for our own PKT_PCLOSE control skb to be sent.
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

Remove duplicated include.
Signed-off-by: NWei Yongjun <yongjun_wei@trendmicro.com.cn>
Acked-by: NSony Chacko <sony.chacko@qlogic.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

1b4c44e6 incorrectly used
ntohs() rather than htons() in myri10ge_vlan_rx().

Thanks to Fengguang Wu, Yuanhan Liu's kernel-build tester
for pointing out this bug.
Signed-off-by: NAndrew Gallatin <gallatin@myri.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

As of 026359bc [ipv6: Send ICMPv6 RSes only when RAs are accepted], the
logic determining whether to send Router Solicitations is identical
to the logic determining whether kernel accepts Router Advertisements.

However the condition itself is repeated in several code locations.

Unify it by introducing 'ipv6_accept_ra()' accessor.

Also, simplify the condition expression, making it more readable.
No semantic change.
Signed-off-by: NShmulik Ladkani <shmulik.ladkani@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

As time passed, available memory increased faster than number of
concurrent tcp sockets.

As a result, a machine with 4GB of ram gets a hash table
with 524288 slots, using 8388608 bytes of memory.

Lets change that by a 16x factor (one slot for 128 KB of ram)

Even if a small machine needs a _lot_ of sockets, tcp lookups are now
very efficient, using one cache line per socket.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Ben Hutchings says:

====================
1. More workarounds for TX queue flush failures that can occur during
   interface reconfiguration.
2. Fix spurious failure of a firmware request running during a system
   clock change, e.g. ntpd started at the same time as driver load.
3. Fix inconsistent statistics after a firmware upgrade.
4. Fix a variable (non-)initialisation in offline self-test that can
   make it more disruptive than intended.
5. Fix a race that can (at least) cause an assertion failure.
6. Miscellaneous cleanup.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Jeff Kirsher says:

====================
This series contains updates to ixgbe, igb and e1000e.  Majority of the
changes are against igb.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>