linux-stable.git/net/core, branch linux-3.17.y

rtnetlink: release net refcnt on error in do_setlink()

2014-12-16T17:37:10+00:00

[ Upstream commit e0ebde0e131b529fd721b24f62872def5ec3718c ]

rtnl_link_get_net() holds a reference on the 'struct net', we need to release
it in case of error.

CC: Eric W. Biederman 
Fixes: b51642f6d77b ("net: Enable a userns root rtnl calls that are safe for unprivilged users")
Signed-off-by: Nicolas Dichtel 
Reviewed-by: "Eric W. Biederman" 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

net: tso: fix unaligned access to crafted TCP header in helper API

2014-11-14T18:10:21+00:00

[ Upstream commit a63ba13eec092b70d4e5522d692eaeb2f9747387 ]

The crafted header start address is from a driver supplied buffer, which
one can reasonably expect to be aligned on a 4-bytes boundary.
However ATM the TSO helper API is only used by ethernet drivers and
the tcp header will then be aligned to a 2-bytes only boundary from the
header start address.

Signed-off-by: Karl Beldan 
Cc: Ezequiel Garcia 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

gro: fix aggregation for skb using frag_list

2014-09-29T19:17:59+00:00

In commit 8a29111c7ca6 ("net: gro: allow to build full sized skb")
I added a regression for linear skb that traditionally force GRO
to use the frag_list fallback.

Erez Shitrit found that at most two segments were aggregated and
the "if (skb_gro_len(p) != pinfo->gso_size)" test was failing.

This is because pinfo at this spot still points to the last skb in the
chain, instead of the first one, where we find the correct gso_size
information.

Signed-off-by: Eric Dumazet 
Fixes: 8a29111c7ca6 ("net: gro: allow to build full sized skb")
Reported-by: Erez Shitrit 
Signed-off-by: David S. Miller

net: fix creation adjacent device symlinks

2014-09-15T18:24:53+00:00

__netdev_adjacent_dev_insert may add adjust device of different net
namespace, without proper check it leads to emergence of broken
sysfs links from/to devices in another namespace.
Fix: rewrite netdev_adjacent_is_neigh_list macro as a function,
     move net_eq check into netdev_adjacent_is_neigh_list.
     (thanks David)
     related to: 4c75431ac3520631f1d9e74aa88407e6374dbbc4

Signed-off-by: Alexander Fomichev 
Signed-off-by: David S. Miller

net: fix skb_page_frag_refill() kerneldoc

2014-09-08T21:12:14+00:00

In commit d9b2938aabf7 ("net: attempt a single high order allocation)
I forgot to update kerneldoc, as @prio parameter was renamed to @gfp

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

net: treewide: Fix typo found in DocBook/networking.xml

2014-09-06T00:35:28+00:00

This patch fix spelling typo found in DocBook/networking.xml.
It is because the neworking.xml is generated from comments
in the source, I have to fix typo in comments within the source.

Signed-off-by: Masanari Iida 
Acked-by: Randy Dunlap 
Signed-off-by: David S. Miller

net: attempt a single high order allocation

2014-08-30T03:28:23+00:00

In commit ed98df3361f0 ("net: use __GFP_NORETRY for high order
allocations") we tried to address one issue caused by order-3
allocations.

We still observe high latencies and system overhead in situations where
compaction is not successful.

Instead of trying order-3, order-2, and order-1, do a single order-3
best effort and immediately fallback to plain order-0.

This mimics slub strategy to fallback to slab min order if the high
order allocation used for performance failed.

Order-3 allocations give a performance boost only if they can be done
without recurring and expensive memory scan.

Quoting David :

The page allocator relies on synchronous (sync light) memory compaction
after direct reclaim for allocations that don't retry and deferred
compaction doesn't work with this strategy because the allocation order
is always decreasing from the previous failed attempt.

This means sync light compaction will always be encountered if memory
cannot be defragmented or reclaimed several times during the
skb_page_frag_refill() iteration.

Signed-off-by: Eric Dumazet 
Acked-by: David Rientjes 
Signed-off-by: David S. Miller

net: fix checksum features handling in netif_skb_features()

2014-08-26T00:23:03+00:00

This is follow-up to

  da08143b8520 ("vlan: more careful checksum features handling")

which introduced more careful feature intersection in vlan code,
taking into account that HW_CSUM should be considered superset
of IP_CSUM/IPV6_CSUM. The same is needed in netif_skb_features()
in order to avoid offloading mismatch warning when vlan is
created on top of a bond consisting of slaves supporting IP/IPv6
checksumming but not vlan Tx offloading.

Signed-off-by: Michal Kubecek 
Signed-off-by: David S. Miller

net: prevent of emerging cross-namespace symlinks

2014-08-25T22:17:43+00:00

Code manipulating sysfs symlinks on adjacent net_devices(s)
currently doesn't take into account that devices potentially
belong to different namespaces.

This patch trying to fix an issue as follows:
- check for net_ns before creating / deleting symlink.
  for now only netdev_adjacent_rename_links and
  __netdev_adjacent_dev_remove are affected, afaics
  __netdev_adjacent_dev_insert implies both net_devs
  belong to the same namespace.
- Drop all existing symlinks to / from all adj_devs before
  switching namespace and recreate them just after.

Signed-off-by: Alexander Y. Fomichev 
Signed-off-by: David S. Miller

Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net

2014-08-14T00:27:40+00:00

Pull networking fixes from David Miller:
"Several networking final fixes and tidies for the merge window:

1) Changes during the merge window unintentionally took away the
ability to build bluetooth modular, fix from Geert Uytterhoeven.

2) Several phy_node reference count bug fixes from Uwe Kleine-König.

3) Fix ucc_geth build failures, also from Uwe Kleine-König.

4) Fix klog false positivies when netlink messages go to network
taps, by properly resetting the network header. Fix from Daniel
Borkmann.

5) Sizing estimate of VF netlink messages is too small, from Jiri
Benc.

6) New APM X-Gene SoC ethernet driver, from Iyappan Subramanian.

7) VLAN untagging is erroneously dependent upon whether the VLAN
module is loaded or not, but there are generic dependencies that
matter wrt what can be expected as the SKB enters the stack.
Make the basic untagging generic code, and do it unconditionally.
From Vlad Yasevich.

8) xen-netfront only has so many slots in it's transmit queue so
linearize packets that have too many frags. From Zoltan Kiss.

9) Fix suspend/resume PHY handling in bcmgenet driver, from Florian
Fainelli"

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (55 commits)
net: bcmgenet: correctly resume adapter from Wake-on-LAN
net: bcmgenet: update UMAC_CMD only when link is detected
net: bcmgenet: correctly suspend and resume PHY device
net: bcmgenet: request and enable main clock earlier
net: ethernet: myricom: myri10ge: myri10ge.c: Cleaning up missing null-terminate after strncpy call
xen-netfront: Fix handling packets on compound pages with skb_linearize
net: fec: Support phys probed from devicetree and fixed-link
smsc: replace WARN_ON() with WARN_ON_SMP()
xen-netback: Don't deschedule NAPI when carrier off
net: ethernet: qlogic: qlcnic: Remove duplicate object file from Makefile
wan: wanxl: Remove typedefs from struct names
m68k/atari: EtherNEC - ethernet support (ne)
net: ethernet: ti: cpmac.c: Cleaning up missing null-terminate after strncpy call
hdlc: Remove typedefs from struct names
airo_cs: Remove typedef local_info_t
atmel: Remove typedef atmel_priv_ioctl
com20020_cs: Remove typedef com20020_dev_t
ethernet: amd: Remove typedef local_info_t
net: Always untag vlan-tagged traffic on input.
drivers: net: Add APM X-Gene SoC ethernet driver support.
...