don't clone skb when skb isn't shared

When the tcf_action is TC_ACT_STOLEN, and the skb isn't shared, we don't need
to clone a new skb. As the skb will be freed after this function returns, we
can use it freely once we get a reference to it.
Signed-off-by: NChangli Gao <xiaosuo@gmail.com>
----
 include/net/sch_generic.h |   11 +++++++++--
 net/sched/act_mirred.c    |    6 +++---
 2 files changed, 12 insertions(+), 5 deletions(-)
Signed-off-by: NJamal Hadi Salim <hadi@cyberus.ca>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We can pass a gfp argument to tso_fragment() and avoid GFP_ATOMIC
allocations sometimes.
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use u64_stats_sync infrastructure to implement 64bit rx stats.

(tx stats are addressed later)
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Acked-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use u64_stats_sync infrastructure to implement 64bit stats.
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Acked-by: NPatrick McHardy <kaber@trash.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

- Must disable preemption in case of 32bit UP in u64_stats_fetch_begin()
and u64_stats_fetch_retry()

- Add new u64_stats_fetch_begin_bh() and u64_stats_fetch_retry_bh() for
network usage, disabling BH on 32bit UP only.
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

use this_cpu_ptr(p) instead of per_cpu_ptr(p, smp_processor_id())
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

- Add a comment about interrupts:

6) If counter might be written by an interrupt, readers should block
interrupts.

- Fix a typo in sample of use.
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Currently much of the code assumes that a specific window has been
selected, while a few functions save and restore the window.  This
makes it impossible to introduce fine-grained locking.

Make those assumptions explicit by introducing wrapper functions
to set the window and read/write a register.  Use these everywhere
except vortex_interrupt(), vortex_start_xmit() and vortex_rx().
These set the window just once, or not at all in the case of
vortex_rx() as it should always be called from vortex_interrupt().

Cache the current window in struct vortex_private to avoid
unnecessary hardware writes.
Signed-off-by: NBen Hutchings <ben@decadent.org.uk>
Tested-by: Arne Nordmark <nordmark@mech.kth.se> [against 2.6.32]
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch adds dynamic LRO diable support for mlx4 net driver.
It also fixes a bug of mlx4, which checks NETIF_F_LRO flag in rx
path without rtnl lock.

(I don't have mlx4 card, so only did compiling test. Anyone who wants
to test this is more than welcome.)

This is based on Neil's initial work too, and heavily modified based
on Stanislaw's suggestions.
Signed-off-by: NWANG Cong <amwang@redhat.com>
Signed-off-by: NNeil Horman <nhorman@redhat.com>
Acked-by: NNeil Horman <nhorman@redhat.com>
Reviewed-by: NStanislaw Gruszka <sgruszka@redhat.com>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch adds dynamic LRO disable support for s2io net driver,
enables LRO by default, increases the driver version number, and
corrects the name of the LRO modparm.

This is mostly Wang's patch based on Neil's initial work, heavily
modified based on Ramkrishna's suggestions.  This has been tested on
a Neterion Xframe adapter and verified via adapter LRO statistics.
Signed-off-by: NJon Mason <jon.mason@exar.com>
Signed-off-by: NWANG Cong <amwang@redhat.com>
Signed-off-by: NNeil Horman <nhorman@redhat.com>
Acked-by: NNeil Horman <nhorman@redhat.com>
Reviewed-by: NStanislaw Gruszka <sgruszka@redhat.com>
Cc: Ramkrishna Vepa <Ramkrishna.Vepa@exar.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Allows use of ECN when syncookies are in effect by encoding ecn_ok
into the syn-ack tcp timestamp.

While at it, remove a uneeded #ifdef CONFIG_SYN_COOKIES.
With CONFIG_SYN_COOKIES=nm want_cookie is ifdef'd to 0 and gcc
removes the "if (0)".
Signed-off-by: NFlorian Westphal <fw@strlen.de>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

As pointed out by Fernando Gont there is no need to encode rcv_wscale
into the cookie.

We did not use the restored rcv_wscale anyway; it is recomputed
via tcp_select_initial_window().

Thus we can save 4 bits in the ts option space by removing rcv_wscale.
In case window scaling was not supported, we set the (invalid) wscale
value 0xf.
Signed-off-by: NFlorian Westphal <fw@strlen.de>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

commit 9261e537 (ipv6: making ip and icmp statistics per/namespace)
forgot to remove ipv6_statistics variable.

commit bc417d99 (ipv6: remove stale MIB definitions) took care of
icmpv6_statistics & icmpv6msg_statistics
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
CC: Denis V. Lunev <den@openvz.org>
CC: Alexey Dobriyan <adobriyan@gmail.com>
CC: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In preparation for 64bit snmp counters for some mibs,
add an 'align' parameter to snmp_mib_init(), instead
of assuming mibs only contain 'unsigned long' fields.

Callers can use __alignof__(type) to provide correct
alignment.
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
CC: Herbert Xu <herbert@gondor.apana.org.au>
CC: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
CC: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
CC: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit 6b10de38 (loopback: Implement 64bit stats on 32bit arches)
introduced 64bit stats in loopback driver, using a private seqcount and
private helpers.

David suggested to introduce a generic infrastructure, added in (net:
Introduce u64_stats_sync infrastructure)

This patch reimplements loopback 64bit stats using the u64_stats_sync
infrastructure.
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Avoid two atomic ops in arp_solicit()
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch is thanks to Andre Noll who reported the issue and helped testing.

The Syn-RTT sampled during the initial handshake currently only works for
the client sending the DCCP-Request. TFRC penalizes the absence of an RTT
sample with a very slow initial speed (1 packet per second), which delays
slow-start significantly, resulting in sluggish performance.

This patch mirrors the "Syn RTT" principle by adding a timestamp also onto
the DCCP-Response, producing an RTT sample  when the (Data)Ack completing
the handshake arrives.

Also changed the documentation to 'TFRC' since Syn RTTs are also used by CCID-4.
Signed-off-by: NGerrit Renker <gerrit@erg.abdn.ac.uk>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This removes an unused 'sk' argument from several option-inserting functions.
Signed-off-by: NGerrit Renker <gerrit@erg.abdn.ac.uk>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit 59d4289b converted fec to dev_pm_ops but
didn't update the suspend/resume functions thus leading to the following warning :
"initialization from incompatible pointer type" when CONFIG_PM is set.

This patch also fixe a few indentation and style around CONFIG_PM area.
Signed-off-by: NEric Bénard <eric@eukrea.com>
Cc: netdev@vger.kernel.org
Cc: davem@davemloft.net
Cc: amit.kucheria@canonical.com
Cc: s.hauer@pengutronix.de
Cc: linux-arm-kernel@lists.infradead.org
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The driver requests FW 7.10
Bump up driver version.
Signed-off-by: NDivy Le Ray <divy@chelsio.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Update FW to 7.10
Signed-off-by: NDivy Le Ray <divy@chelsio.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add pr_fmt(fmt) KBUILD_MODNAME ": " fmt
Remove "pktgen: " from formats
Convert printks to pr_<level>
Added func_enter() for debugging
Moved version to end of string at module_init
Coalesced long formats
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Gcc is currenlty not in the ability to optimize the switch statement in
sk_run_filter() because of dense case labels. This patch replace the
OR'd labels with ordered sequenced case labels. The sk_chk_filter()
function is modified to patch/replace the original OPCODES in a
ordered but equivalent form. gcc is now in the ability to transform the
switch statement in sk_run_filter into a jump table of complexity O(1).

Until this patch gcc generates a sequence of conditional branches (O(n) of 567
byte .text segment size (arch x86_64):

7ff: 8b 06                 mov    (%rsi),%eax
801: 66 83 f8 35           cmp    $0x35,%ax
805: 0f 84 d0 02 00 00     je     adb <sk_run_filter+0x31d>
80b: 0f 87 07 01 00 00     ja     918 <sk_run_filter+0x15a>
811: 66 83 f8 15           cmp    $0x15,%ax
815: 0f 84 c5 02 00 00     je     ae0 <sk_run_filter+0x322>
81b: 77 73                 ja     890 <sk_run_filter+0xd2>
81d: 66 83 f8 04           cmp    $0x4,%ax
821: 0f 84 17 02 00 00     je     a3e <sk_run_filter+0x280>
827: 77 29                 ja     852 <sk_run_filter+0x94>
829: 66 83 f8 01           cmp    $0x1,%ax
[...]

With the modification the compiler translate the switch statement into
the following jump table fragment:

7ff: 66 83 3e 2c           cmpw   $0x2c,(%rsi)
803: 0f 87 1f 02 00 00     ja     a28 <sk_run_filter+0x26a>
809: 0f b7 06              movzwl (%rsi),%eax
80c: ff 24 c5 00 00 00 00  jmpq   *0x0(,%rax,8)
813: 44 89 e3              mov    %r12d,%ebx
816: e9 43 03 00 00        jmpq   b5e <sk_run_filter+0x3a0>
81b: 41 89 dc              mov    %ebx,%r12d
81e: e9 3b 03 00 00        jmpq   b5e <sk_run_filter+0x3a0>

Furthermore, I reordered the instructions to reduce cache line misses by
order the most common instruction to the start.
Signed-off-by: NHagen Paul Pfeifer <hagen@jauu.net>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NBen Hutchings <bhutchings@solarflare.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Insertion of the Falcon hash is unreliable.
Signed-off-by: NBen Hutchings <bhutchings@solarflare.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We will use this hash key for Toeplitz IPv4 hashing too.
Signed-off-by: NBen Hutchings <bhutchings@solarflare.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The hash appears immediately before the packet data, not at the
beginning of the buffer. This means we can easily use negative offsets
from the start of packet data, so adjust the data and length at the
top of __efx_rx_packet() instead of wherever we consume the hash.
Signed-off-by: NBen Hutchings <bhutchings@solarflare.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

1) Update copyright
2) Fix hardware queue descriptor field size CQ_ENET_RQ_DESC_FCOE_SOF_BITS
3) Include rtnetlink.h instead of if_link.h
4) Selectively flush writes to interrupt mask register
5) Use pci_enable_device_mem
6) Remove unused variables and header files
7) Fix size mismatch between memory alloc and free operations of a variable
8) Check for non null arguments to vic_provinfo_alloc
Signed-off-by: NScott Feldman <scofeldm@cisco.com>
Signed-off-by: NVasanthy Kolluri <vkolluri@cisco.com>
Signed-off-by: NRoopa Prabhu <roprabhu@cisco.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Handle surprise hardware removals gracefully during devcmd issue and init,
cleanup of queues.
Signed-off-by: NScott Feldman <scofeldm@cisco.com>
Signed-off-by: NVasanthy Kolluri <vkolluri@cisco.com>
Signed-off-by: NRoopa Prabhu <roprabhu@cisco.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Hardware has the loopback capability to queue the packets transmitted from
a device to the receive queue of the same device. enic now supports the
loopback capability.
Signed-off-by: NScott Feldman <scofeldm@cisco.com>
Signed-off-by: NVasanthy Kolluri <vkolluri@cisco.com>
Signed-off-by: NRoopa Prabhu <roprabhu@cisco.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Change the receive queue buffer allocations into blocks of 32 entries when
ring size is less than 64, otherwise use 64 entries per block.
Signed-off-by: NScott Feldman <scofeldm@cisco.com>
Signed-off-by: NVasanthy Kolluri <vkolluri@cisco.com>
Signed-off-by: NRoopa Prabhu <roprabhu@cisco.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add new firmware devcmds - CMD_PROXY_BY_BDF, CMD_PACKET_FILTER_ALL,
CMD_ENABLE_WAIT.
Signed-off-by: NScott Feldman <scofeldm@cisco.com>
Signed-off-by: NVasanthy Kolluri <vkolluri@cisco.com>
Signed-off-by: NRoopa Prabhu <roprabhu@cisco.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Replace all printk routines with the (netdev|dev|pr)_<level> macros that
provide verbose logs.
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NScott Feldman <scofeldm@cisco.com>
Signed-off-by: NVasanthy Kolluri <vkolluri@cisco.com>
Signed-off-by: NRoopa Prabhu <roprabhu@cisco.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add wrapper routines that issue devcmds to firmware and ensure that a
devcmd lock is held for each devcmd call.
Signed-off-by: NScott Feldman <scofeldm@cisco.com>
Signed-off-by: NVasanthy Kolluri <vkolluri@cisco.com>
Signed-off-by: NRoopa Prabhu <roprabhu@cisco.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The port profile information for a dynamic enic device is set by the upper
layers, that are oblivious to the device reset operation. We do not want a
reset operation erase the network state of a dynamic enic device as there
is no way to set up the port profile information again. Hence a lighter
reset operation called hang reset is used. Hang reset, unlike soft reset
does not reset the network state and resets the host side state only.
Signed-off-by: NScott Feldman <scofeldm@cisco.com>
Signed-off-by: NVasanthy Kolluri <vkolluri@cisco.com>
Signed-off-by: NRoopa Prabhu <roprabhu@cisco.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The current ingress vlan rewrite mode setting lets the hardware strip off
the tag control information of a packet received on native vlan. As a
result, the priority bits are also lost. The fix is to change the ingress
vlan rewrite mode setting such that the complete tag control information is
retained for packets that belong to native vlan.
Signed-off-by: NScott Feldman <scofeldm@cisco.com>
Signed-off-by: NVasanthy Kolluri <vkolluri@cisco.com>
Signed-off-by: NRoopa Prabhu <roprabhu@cisco.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

enic now uses the GRO mechanism instead of LRO to pass skbs to upper
layers.
Signed-off-by: NScott Feldman <scofeldm@cisco.com>
Signed-off-by: NVasanthy Kolluri <vkolluri@cisco.com>
Signed-off-by: NRoopa Prabhu <roprabhu@cisco.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This eliminates some of the duplicate code for the various devices
that require the same basic kcq handling.
Signed-off-by: NMichael Chan <mchan@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

By doing more work in the common function cnic_get_kcqes(), and
making full use of the kcq_info structure.
Signed-off-by: NMichael Chan <mchan@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>