linux-toradex.git/include/net, branch v2.6.34.15

net: fix a race in sock_queue_err_skb()

2014-02-10T21:10:55+00:00

commit 110c43304db6f06490961529536c362d9ac5732f upstream.

As soon as an skb is queued into socket error queue, another thread
can consume it, so we are not allowed to reference skb anymore, or risk
use after free.

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller 
[PG: net/core/skbuff.c  -->  include/net/sock.h on 2.6.34 baseline]
Signed-off-by: Paul Gortmaker

ipv6: call udp_push_pending_frames when uncorking a socket with AF_INET pending data

2014-02-10T21:10:45+00:00

commit 8822b64a0fa64a5dd1dfcf837c5b0be83f8c05d1 upstream.

We accidentally call down to ip6_push_pending_frames when uncorking
pending AF_INET data on a ipv6 socket. This results in the following
splat (from Dave Jones):

skbuff: skb_under_panic: text:ffffffff816765f6 len:48 put:40 head:ffff88013deb6df0 data:ffff88013deb6dec tail:0x2c end:0xc0 dev:
------------[ cut here ]------------
kernel BUG at net/core/skbuff.c:126!
invalid opcode: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC
Modules linked in: dccp_ipv4 dccp 8021q garp bridge stp dlci mpoa snd_seq_dummy sctp fuse hidp tun bnep nfnetlink scsi_transport_iscsi rfcomm can_raw can_bcm af_802154 appletalk caif_socket can caif ipt_ULOG x25 rose af_key pppoe pppox ipx phonet irda llc2 ppp_generic slhc p8023 psnap p8022 llc crc_ccitt atm bluetooth
+netrom ax25 nfc rfkill rds af_rxrpc coretemp hwmon kvm_intel kvm crc32c_intel snd_hda_codec_realtek ghash_clmulni_intel microcode pcspkr snd_hda_codec_hdmi snd_hda_intel snd_hda_codec snd_hwdep usb_debug snd_seq snd_seq_device snd_pcm e1000e snd_page_alloc snd_timer ptp snd pps_core soundcore xfs libcrc32c
CPU: 2 PID: 8095 Comm: trinity-child2 Not tainted 3.10.0-rc7+ #37
task: ffff8801f52c2520 ti: ffff8801e6430000 task.ti: ffff8801e6430000
RIP: 0010:[]  [] skb_panic+0x63/0x65
RSP: 0018:ffff8801e6431de8  EFLAGS: 00010282
RAX: 0000000000000086 RBX: ffff8802353d3cc0 RCX: 0000000000000006
RDX: 0000000000003b90 RSI: ffff8801f52c2ca0 RDI: ffff8801f52c2520
RBP: ffff8801e6431e08 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000001 R11: 0000000000000001 R12: ffff88022ea0c800
R13: ffff88022ea0cdf8 R14: ffff8802353ecb40 R15: ffffffff81cc7800
FS:  00007f5720a10740(0000) GS:ffff880244c00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000005862000 CR3: 000000022843c000 CR4: 00000000001407e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000600
Stack:
 ffff88013deb6dec 000000000000002c 00000000000000c0 ffffffff81a3f6e4
 ffff8801e6431e18 ffffffff8159a9aa ffff8801e6431e90 ffffffff816765f6
 ffffffff810b756b 0000000700000002 ffff8801e6431e40 0000fea9292aa8c0
Call Trace:
 [] skb_push+0x3a/0x40
 [] ip6_push_pending_frames+0x1f6/0x4d0
 [] ? mark_held_locks+0xbb/0x140
 [] udp_v6_push_pending_frames+0x2b9/0x3d0
 [] ? udplite_getfrag+0x20/0x20
 [] udp_lib_setsockopt+0x1aa/0x1f0
 [] ? fget_light+0x387/0x4f0
 [] udpv6_setsockopt+0x34/0x40
 [] sock_common_setsockopt+0x14/0x20
 [] SyS_setsockopt+0x71/0xd0
 [] tracesys+0xdd/0xe2
Code: 00 00 48 89 44 24 10 8b 87 d8 00 00 00 48 89 44 24 08 48 8b 87 e8 00 00 00 48 c7 c7 c0 04 aa 81 48 89 04 24 31 c0 e8 e1 7e ff ff <0f> 0b 55 48 89 e5 0f 0b 55 48 89 e5 0f 0b 55 48 89 e5 0f 0b 55
RIP  [] skb_panic+0x63/0x65
 RSP 

This patch adds a check if the pending data is of address family AF_INET
and directly calls udp_push_ending_frames from udp_v6_push_pending_frames
if that is the case.

This bug was found by Dave Jones with trinity.

(Also move the initialization of fl6 below the AF_INET check, even if
not strictly necessary.)

Cc: Dave Jones 
Cc: YOSHIFUJI Hideaki 
Signed-off-by: Hannes Frederic Sowa 
Signed-off-by: David S. Miller 
[PG: The line "flowi6 *fl6 = &inet->cork.fl.u.ip6"  was
              "flowi *fl = &inet->cork.fl" in 2.6.34 kernel.]
Signed-off-by: Paul Gortmaker

inet: add RCU protection to inet->opt

2014-02-10T21:10:42+00:00

commit f6d8bd051c391c1c0458a30b2a7abcd939329259 upstream.

We lack proper synchronization to manipulate inet->opt ip_options

Problem is ip_make_skb() calls ip_setup_cork() and
ip_setup_cork() possibly makes a copy of ipc->opt (struct ip_options),
without any protection against another thread manipulating inet->opt.

Another thread can change inet->opt pointer and free old one under us.

Use RCU to protect inet->opt (changed to inet->inet_opt).

Instead of handling atomic refcounts, just copy ip_options when
necessary, to avoid cache line dirtying.

We cant insert an rcu_head in struct ip_options since its included in
skb->cb[], so this patch is large because I had to introduce a new
ip_options_rcu structure.

Signed-off-by: Eric Dumazet 
Cc: Herbert Xu 
Signed-off-by: David S. Miller 
[dannf/bwh: backported to Debian's 2.6.32]
Signed-off-by: Ben Hutchings 
Signed-off-by: Willy Tarreau 
[PG: use 2.6.32 patch, since it is closer to 2.6.34 than original
 baseline; drop net/l2tp/l2tp_ip.c chunk as we don't have that file]
Signed-off-by: Paul Gortmaker

sctp: Enforce retransmission limit during shutdown

2013-01-16T21:45:06+00:00

commit f8d9605243280f1870dd2c6c37a735b925c15f3c upstream.

When initiating a graceful shutdown while having data chunks
on the retransmission queue with a peer which is in zero
window mode the shutdown is never completed because the
retransmission error count is reset periodically by the
following two rules:

 - Do not timeout association while doing zero window probe.
 - Reset overall error count when a heartbeat request has
   been acknowledged.

The graceful shutdown will wait for all outstanding TSN to
be acknowledged before sending the SHUTDOWN request. This
never happens due to the peer's zero window not acknowledging
the continuously retransmitted data chunks. Although the
error counter is incremented for each failed retransmission,
the receiving of the SACK announcing the zero window clears
the error count again immediately. Also heartbeat requests
continue to be sent periodically. The peer acknowledges these
requests causing the error counter to be reset as well.

This patch changes behaviour to only reset the overall error
counter for the above rules while not in shutdown. After
reaching the maximum number of retransmission attempts, the
T5 shutdown guard timer is scheduled to give the receiver
some additional time to recover. The timer is stopped as soon
as the receiver acknowledges any data.

The issue can be easily reproduced by establishing a sctp
association over the loopback device, constantly queueing
data at the sender while not reading any at the receiver.
Wait for the window to reach zero, then initiate a shutdown
by killing both processes simultaneously. The association
will never be freed and the chunks on the retransmission
queue will be retransmitted indefinitely.

Signed-off-by: Thomas Graf 
Acked-by: Vlad Yasevich 
Signed-off-by: David S. Miller 
Signed-off-by: Paul Gortmaker

sctp: ABORT if receive, reassmbly, or reodering queue is not empty while closing socket

2013-01-16T21:45:06+00:00

commit cd4fcc704f30f2064ab30b5300d44d431e46db50 upstream.

Trigger user ABORT if application closes a socket which has data
queued on the socket receive queue or chunks waiting on the
reassembly or ordering queue as this would imply data being lost
which defeats the point of a graceful shutdown.

This behavior is already practiced in TCP.

We do not check the input queue because that would mean to parse
all chunks on it to look for unacknowledged data which seems too
much of an effort. Control chunks or duplicated chunks may also
be in the input queue and should not be stopping a graceful
shutdown.

Signed-off-by: Thomas Graf 
Acked-by: Vlad Yasevich 
Signed-off-by: David S. Miller 
Signed-off-by: Paul Gortmaker

scm: lower SCM_MAX_FD

2012-08-17T19:35:40+00:00

commit bba14de98753cb6599a2dae0e520714b2153522d upstream.

Lower SCM_MAX_FD from 255 to 253 so that allocations for scm_fp_list are
halved. (commit f8d570a4 added two pointers in this structure)

scm_fp_dup() should not copy whole structure (and trigger kmemcheck
warnings), but only the used part. While we are at it, only allocate
needed size.

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

rose: Add length checks to CALL_REQUEST parsing

2012-08-17T19:34:59+00:00

commit e0bccd315db0c2f919e7fcf9cb60db21d9986f52 upstream.

Define some constant offsets for CALL_REQUEST based on the description
at  and the
definition of ROSE as using 10-digit (5-byte) addresses.  Use them
consistently.  Validate all implicit and explicit facilities lengths.
Validate the address length byte rather than either trusting or
assuming its value.

Signed-off-by: Ben Hutchings 
Signed-off-by: David S. Miller 
Signed-off-by: Paul Gortmaker

net: Compute protocol sequence numbers and fragment IDs using MD5.

2012-05-17T15:21:27+00:00

commit 6e5714eaf77d79ae1c8b47e3e040ff5411b717ec upstream.

Computers have become a lot faster since we compromised on the
partial MD4 hash which we use currently for performance reasons.

MD5 is a much safer choice, and is inline with both RFC1948 and
other ISS generators (OpenBSD, Solaris, etc.)

Furthermore, only having 24-bits of the sequence number be truly
unpredictable is a very serious limitation.  So the periodic
regeneration and 8-bit counter have been removed.  We compute and
use a full 32-bit sequence number.

For ipv6, DCCP was found to use a 32-bit truncated initial sequence
number (it needs 43-bits) and that is fixed here as well.

Reported-by: Dan Kaminsky 
Tested-by: Willy Tarreau 
Signed-off-by: David S. Miller 
[PG: diffstat vs. 6e5714 differs, since no secure_ipv6_id to delete in 34]
Signed-off-by: Paul Gortmaker

net: sk_add_backlog() take rmem_alloc into account

2012-03-14T14:58:05+00:00

commit c377411f2494a931ff7facdbb3a6839b1266bcf6 upstream.

Current socket backlog limit is not enough to really stop DDOS attacks,
because user thread spend many time to process a full backlog each
round, and user might crazy spin on socket lock.

We should add backlog size and receive_queue size (aka rmem_alloc) to
pace writers, and let user run without being slow down too much.

Introduce a sk_rcvqueues_full() helper, to avoid taking socket lock in
stress situations.

Under huge stress from a multiqueue/RPS enabled NIC, a single flow udp
receiver can now process ~200.000 pps (instead of ~100 pps before the
patch) on a 8 core machine.

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

netlink: Make nlmsg_find_attr take a const nlmsghdr*.

2011-06-26T16:47:18+00:00

commit 6b8c92ba07287578718335ce409de8e8d7217e40 upstream.

This will let us use it on a nlmsghdr stored inside a netlink_callback.

Signed-off-by: Nelson Elhage 
Signed-off-by: David S. Miller 
Signed-off-by: Paul Gortmaker