linux/net/core, branch v3.8

net: fix infinite loop in __skb_recv_datagram()

2013-02-12T21:07:19Z

Tommi was fuzzing with trinity and reported the following problem : commit 3f518bf745 (datagram: Add offset argument to __skb_recv_datagram) missed that a raw socket receive queue can contain skbs with no payload. We can loop in __skb_recv_datagram() with MSG_PEEK mode, because wait_for_packet() is not prepared to skip these skbs. [ 83.541011] INFO: rcu_sched detected stalls on CPUs/tasks: {} (detected by 0, t=26002 jiffies, g=27673, c=27672, q=75) [ 83.541011] INFO: Stall ended before state dump start [ 108.067010] BUG: soft lockup - CPU#0 stuck for 22s! [trinity-child31:2847] ... [ 108.067010] Call Trace: [ 108.067010] [] __skb_recv_datagram+0x1a3/0x3b0 [ 108.067010] [] skb_recv_datagram+0x2d/0x30 [ 108.067010] [] rawv6_recvmsg+0xad/0x240 [ 108.067010] [] sock_common_recvmsg+0x34/0x50 [ 108.067010] [] sock_recvmsg+0xbc/0xf0 [ 108.067010] [] sys_recvfrom+0xde/0x150 [ 108.067010] [] system_call_fastpath+0x16/0x1b Reported-by: Tommi Rantala Tested-by: Tommi Rantala Signed-off-by: Eric Dumazet Cc: Pavel Emelyanov Acked-by: Pavel Emelyanov Signed-off-by: David S. Miller

net: Fix inner_network_header assignment in skb-copy.

2013-02-03T21:10:36Z

Use correct inner offset to set inner_network_offset. Found by inspection. Signed-off-by: Pravin B Shelar Signed-off-by: David S. Miller

pktgen: correctly handle failures when adding a device

2013-01-29T20:43:03Z

The return value of pktgen_add_device() is not checked, so even if we fail to add some device, for example, non-exist one, we still see "OK:...". This patch fixes it. After this patch, I got: # echo "add_device non-exist" > /proc/net/pktgen/kpktgend_0 -bash: echo: write error: No such device # cat /proc/net/pktgen/kpktgend_0 Running: Stopped: Result: ERROR: can not add device non-exist # echo "add_device eth0" > /proc/net/pktgen/kpktgend_0 # cat /proc/net/pktgen/kpktgend_0 Running: Stopped: eth0 Result: OK: add_device=eth0 (Candidate for -stable) Cc: David S. Miller Signed-off-by: Cong Wang Signed-off-by: David S. Miller

net: net_cls: fd passed in SCM_RIGHTS datagram not set correctly

2013-01-22T19:17:38Z

Commit 6a328d8c6f03501657ad580f6f98bf9a42583ff7 changed the update logic for the socket but it does not update the SCM_RIGHTS update as well. This patch is based on the net_prio fix commit 48a87cc26c13b68f6cce4e9d769fcb17a6b3e4b8 net: netprio: fd passed in SCM_RIGHTS datagram not set correctly A socket fd passed in a SCM_RIGHTS datagram was not getting updated with the new tasks cgrp prioidx. This leaves IO on the socket tagged with the old tasks priority. To fix this add a check in the scm recvmsg path to update the sock cgrp prioidx with the new tasks value. Let's apply the same fix for net_cls. Signed-off-by: Daniel Wagner Reported-by: Li Zefan Cc: "David S. Miller" Cc: "Eric W. Biederman" Cc: Al Viro Cc: John Fastabend Cc: Neil Horman Cc: netdev@vger.kernel.org Cc: cgroups@vger.kernel.org Signed-off-by: David S. Miller

net: splice: fix __splice_segment()

2013-01-21T04:19:36Z

commit 9ca1b22d6d2 (net: splice: avoid high order page splitting) forgot that skb->head could need a copy into several page frags. This could be the case for loopback traffic mostly. Also remove now useless skb argument from linear_to_page() and __splice_segment() prototypes. Signed-off-by: Eric Dumazet Cc: Willy Tarreau Signed-off-by: David S. Miller

net: splice: avoid high order page splitting

2013-01-21T04:19:35Z

splice() can handle pages of any order, but network code tries hard to split them in PAGE_SIZE units. Not quite successfully anyway, as __splice_segment() assumed poff < PAGE_SIZE. This is true for the skb->data part, not necessarily for the fragments. This patch removes this logic to give the pages as they are in the skb. Signed-off-by: Eric Dumazet Cc: Willy Tarreau Signed-off-by: David S. Miller

tcp: fix a panic on UP machines in reqsk_fastopen_remove

2013-01-14T23:10:05Z

spin_is_locked() on a non !SMP build is kind of useless. BUG_ON(!spin_is_locked(xx)) is guaranteed to crash. Just remove this check in reqsk_fastopen_remove() as the callers do hold the socket lock. Reported-by: Ketan Kulkarni Signed-off-by: Eric Dumazet Cc: Jerry Chu Cc: Yuchung Cheng Cc: Dave Taht Acked-by: H.K. Jerry Chu Signed-off-by: David S. Miller

net, wireless: overwrite default_ethtool_ops

2013-01-11T23:55:48Z

Since: commit 2c60db037034d27f8c636403355d52872da92f81 Author: Eric Dumazet Date: Sun Sep 16 09:17:26 2012 +0000 net: provide a default dev->ethtool_ops wireless core does not correctly assign ethtool_ops. After alloc_netdev*() call, some cfg80211 drivers provide they own ethtool_ops, but some do not. For them, wireless core provide generic cfg80211_ethtool_ops, which is assigned in NETDEV_REGISTER notify call: if (!dev->ethtool_ops) dev->ethtool_ops = &cfg80211_ethtool_ops; But after Eric's commit, dev->ethtool_ops is no longer NULL (on cfg80211 drivers without custom ethtool_ops), but points to &default_ethtool_ops. In order to fix the problem, provide function which will overwrite default_ethtool_ops and use it by wireless core. Signed-off-by: Stanislaw Gruszka Acked-by: Johannes Berg Acked-by: Ben Hutchings Signed-off-by: David S. Miller

CONFIG_HOTPLUG removal from networking core

2012-12-22T08:03:00Z

CONFIG_HOTPLUG is always enabled now, so remove the unused code that was trying to be compiled out when this option was disabled, in the networking core. Cc: Bill Pemberton Signed-off-by: Greg Kroah-Hartman Signed-off-by: David S. Miller

net: devnet_rename_seq should be a seqcount

2012-12-21T21:14:01Z

Using a seqlock for devnet_rename_seq is not a good idea, as device_rename() can sleep. As we hold RTNL, we dont need a protection for writers, and only need a seqcount so that readers can catch a change done by a writer. Bug added in commit c91f6df2db4972d3 (sockopt: Change getsockopt() of SO_BINDTODEVICE to return an interface name) Reported-by: Dave Jones Signed-off-by: Eric Dumazet Cc: Brian Haley Signed-off-by: David S. Miller