Quoting Ben Hutchings: we presumably won't be defining features that
can only be enabled on 64-bit architectures.
Occurences found by `grep -r` on net/, drivers/net, include/
[ Move features and vlan_features next to each other in
struct netdev, as per Eric Dumazet's suggestion -DaveM ]
Signed-off-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace skb->csum_start - skb_headroom(skb) with skb_checksum_start_offset().
Note for usb/smsc95xx: skb->data - skb->head == skb_headroom(skb).
Signed-off-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
Special care is taken inside sk_port_alloc to avoid overwriting
skc_node/skc_nulls_node. We should also avoid overwriting
skc_bind_node/skc_portaddr_node.
The patch fixes the following crash:
BUG: unable to handle kernel paging request at fffffffffffffff0
IP: [<ffffffff812ec6dd>] udp4_lib_lookup2+0xad/0x370
[<ffffffff812ecc22>] __udp4_lib_lookup+0x282/0x360
[<ffffffff812ed63e>] __udp4_lib_rcv+0x31e/0x700
[<ffffffff812bba45>] ? ip_local_deliver_finish+0x65/0x190
[<ffffffff812bbbf8>] ? ip_local_deliver+0x88/0xa0
[<ffffffff812eda35>] udp_rcv+0x15/0x20
[<ffffffff812bba45>] ip_local_deliver_finish+0x65/0x190
[<ffffffff812bbbf8>] ip_local_deliver+0x88/0xa0
[<ffffffff812bb2cd>] ip_rcv_finish+0x32d/0x6f0
[<ffffffff8128c14c>] ? netif_receive_skb+0x99c/0x11c0
[<ffffffff812bb94b>] ip_rcv+0x2bb/0x350
[<ffffffff8128c14c>] netif_receive_skb+0x99c/0x11c0
Signed-off-by: Leonard Crestez <lcrestez@ixiacom.com>
Signed-off-by: Octavian Purdila <opurdila@ixiacom.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use the macros defined for the members of flowi to clean the code up.
Signed-off-by: Changli Gao <xiaosuo@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
UDP sockets refcount is usually 2, unless an incoming frame is going to
be queued in receive or backlog queue.
Using atomic_inc_not_zero_hint() permits to reduce latency, because
processor issues less memory transactions.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Robin Holt tried to boot a 16TB machine and found some limits were
reached : sysctl_tcp_mem[2], sysctl_udp_mem[2]
We can switch infrastructure to use long "instead" of "int", now
atomic_long_t primitives are available for free.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Reported-by: Robin Holt <holt@sgi.com>
Reviewed-by: Robin Holt <holt@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add __rcu annotation to :
(struct sock)->sk_filter
And use appropriate rcu primitives to reduce sparse warnings if
CONFIG_SPARSE_RCU_POINTER=y
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 30fff923 introduced in linux-2.6.33 (udp: bind() optimisation)
added a secondary hash on UDP, hashed on (local addr, local port).
Problem is that following sequence :
fd = socket(...)
connect(fd, &remote, ...)
not only selects remote end point (address and port), but also sets
local address, while UDP stack stored in secondary hash table the socket
while its local address was INADDR_ANY (or ipv6 equivalent)
Sequence is :
- autobind() : choose a random local port, insert socket in hash tables
[while local address is INADDR_ANY]
- connect() : set remote address and port, change local address to IP
given by a route lookup.
When an incoming UDP frame comes, if more than 10 sockets are found in
primary hash table, we switch to secondary table, and fail to find
socket because its local address changed.
One solution to this problem is to rehash datagram socket if needed.
We add a new rehash(struct socket *) method in "struct proto", and
implement this method for UDP v4 & v6, using a common helper.
This rehashing only takes care of secondary hash table, since primary
hash (based on local port only) is not changed.
Reported-by: Krzysztof Piotr Oledzki <ole@ans.pl>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Tested-by: Krzysztof Piotr Oledzki <ole@ans.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch removes the abstraction introduced by the union skb_shared_tx in
the shared skb data.
The access of the different union elements at several places led to some
confusion about accessing the shared tx_flags e.g. in skb_orphan_try().
http://marc.info/?l=linux-netdev&m=128084897415886&w=2
Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
remove useless union keyword in rtable, rt6_info and dn_route.
Since there is only one member in a union, the union keyword isn't useful.
Signed-off-by: Changli Gao <xiaosuo@gmail.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Correct sk_forward_alloc handling for error_queue would need to use a
backlog of frames that softirq handler could not deliver because socket
is owned by user thread. Or extend backlog processing to be able to
process normal and error packets.
Another possibility is to not use mem charge for error queue, this is
what I implemented in this patch.
Note: this reverts commit 29030374
(net: fix sk_forward_alloc corruptions), since we dont need to lock
socket anymore.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As David found out, sock_queue_err_skb() should be called with socket
lock hold, or we risk sk_forward_alloc corruption, since we use non
atomic operations to update this field.
This patch adds bh_lock_sock()/bh_unlock_sock() pair to three spots.
(BH already disabled)
1) skb_tstamp_tx()
2) Before calling ip_icmp_error(), in __udp4_lib_err()
3) Before calling ipv6_icmp_error(), in __udp6_lib_err()
Reported-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (22 commits)
netlink: bug fix: wrong size was calculated for vfinfo list blob
netlink: bug fix: don't overrun skbs on vf_port dump
xt_tee: use skb_dst_drop()
netdev/fec: fix ifconfig eth0 down hang issue
cnic: Fix context memory init. on 5709.
drivers/net: Eliminate a NULL pointer dereference
drivers/net/hamradio: Eliminate a NULL pointer dereference
be2net: Patch removes redundant while statement in loop.
ipv6: Add GSO support on forwarding path
net: fix __neigh_event_send()
vhost: fix the memory leak which will happen when memory_access_ok fails
vhost-net: fix to check the return value of copy_to/from_user() correctly
vhost: fix to check the return value of copy_to/from_user() correctly
vhost: Fix host panic if ioctl called with wrong index
net: fix lock_sock_bh/unlock_sock_bh
net/iucv: Add missing spin_unlock
net: ll_temac: fix checksum offload logic
net: ll_temac: fix interrupt bug when interrupt 0 is used
sctp: dubious bitfields in sctp_transport
ipmr: off by one in __ipmr_fill_mroute()
...
This new sock lock primitive was introduced to speedup some user context
socket manipulation. But it is unsafe to protect two threads, one using
regular lock_sock/release_sock, one using lock_sock_bh/unlock_sock_bh
This patch changes lock_sock_bh to be careful against 'owned' state.
If owned is found to be set, we must take the slow path.
lock_sock_bh() now returns a boolean to say if the slow path was taken,
and this boolean is used at unlock_sock_bh time to call the appropriate
unlock function.
After this change, BH are either disabled or enabled during the
lock_sock_bh/unlock_sock_bh protected section. This might be misleading,
so we rename these functions to lock_sock_fast()/unlock_sock_fast().
Reported-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Tested-by: Anton Blanchard <anton@samba.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
- C99 knows about USHRT_MAX/SHRT_MAX/SHRT_MIN, not
USHORT_MAX/SHORT_MAX/SHORT_MIN.
- Make SHRT_MIN of type s16, not int, for consistency.
[akpm@linux-foundation.org: fix drivers/dma/timb_dma.c]
[akpm@linux-foundation.org: fix security/keys/keyring.c]
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Acked-by: WANG Cong <xiyou.wangcong@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(Dropped the infiniband part, because Tetsuo modified the related code,
I will send a separate patch for it once this is accepted.)
This patch introduces /proc/sys/net/ipv4/ip_local_reserved_ports which
allows users to reserve ports for third-party applications.
The reserved ports will not be used by automatic port assignments
(e.g. when calling connect() or bind() with port number 0). Explicit
port allocation behavior is unchanged.
Signed-off-by: Octavian Purdila <opurdila@ixiacom.com>
Signed-off-by: WANG Cong <amwang@redhat.com>
Cc: Neil Horman <nhorman@tuxdriver.com>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 2783ef23 moved the initialisation of saddr and daddr after
pskb_may_pull() to avoid a potential data corruption. Unfortunately
also placing it after the short packet and bad checksum error paths,
where these variables are used for logging. The result is bogus
output like
[92238.389505] UDP: short packet: From 2.0.0.0:65535 23715/178 to 0.0.0.0:65535
Moving the saddr and daddr initialisation above the error paths, while still
keeping it after the pskb_may_pull() to keep the fix from commit 2783ef23.
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Cc: stable@kernel.org
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When queueing a skb to socket, we can immediately release its dst if
target socket do not use IP_CMSG_PKTINFO.
tcp_data_queue() can drop dst too.
This to benefit from a hot cache line and avoid the receiver, possibly
on another cpu, to dirty this cache line himself.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit 95766fff ([UDP]: Add memory accounting.),
each received packet needs one extra sock_lock()/sock_release() pair.
This added latency because of possible backlog handling. Then later,
ticket spinlocks added yet another latency source in case of DDOS.
This patch introduces lock_sock_bh() and unlock_sock_bh()
synchronization primitives, avoiding one atomic operation and backlog
processing.
skb_free_datagram_locked() uses them instead of full blown
lock_sock()/release_sock(). skb is orphaned inside locked section for
proper socket memory reclaim, and finally freed outside of it.
UDP receive path now take the socket spinlock only once.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
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 <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Idea from Eric Dumazet.
As for placement inside of struct sock, I tried to choose a place
that otherwise has a 32-bit hole on 64-bit systems.
Signed-off-by: David S. Miller <davem@davemloft.net>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Sparse can help us find endianness bugs, but we need to make some
cleanups to be able to more easily spot real bugs.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch implements receive flow steering (RFS). RFS steers
received packets for layer 3 and 4 processing to the CPU where
the application for the corresponding flow is running. RFS is an
extension of Receive Packet Steering (RPS).
The basic idea of RFS is that when an application calls recvmsg
(or sendmsg) the application's running CPU is stored in a hash
table that is indexed by the connection's rxhash which is stored in
the socket structure. The rxhash is passed in skb's received on
the connection from netif_receive_skb. For each received packet,
the associated rxhash is used to look up the CPU in the hash table,
if a valid CPU is set then the packet is steered to that CPU using
the RPS mechanisms.
The convolution of the simple approach is that it would potentially
allow OOO packets. If threads are thrashing around CPUs or multiple
threads are trying to read from the same sockets, a quickly changing
CPU value in the hash table could cause rampant OOO packets--
we consider this a non-starter.
To avoid OOO packets, this solution implements two types of hash
tables: rps_sock_flow_table and rps_dev_flow_table.
rps_sock_table is a global hash table. Each entry is just a CPU
number and it is populated in recvmsg and sendmsg as described above.
This table contains the "desired" CPUs for flows.
rps_dev_flow_table is specific to each device queue. Each entry
contains a CPU and a tail queue counter. The CPU is the "current"
CPU for a matching flow. The tail queue counter holds the value
of a tail queue counter for the associated CPU's backlog queue at
the time of last enqueue for a flow matching the entry.
Each backlog queue has a queue head counter which is incremented
on dequeue, and so a queue tail counter is computed as queue head
count + queue length. When a packet is enqueued on a backlog queue,
the current value of the queue tail counter is saved in the hash
entry of the rps_dev_flow_table.
And now the trick: when selecting the CPU for RPS (get_rps_cpu)
the rps_sock_flow table and the rps_dev_flow table for the RX queue
are consulted. When the desired CPU for the flow (found in the
rps_sock_flow table) does not match the current CPU (found in the
rps_dev_flow table), the current CPU is changed to the desired CPU
if one of the following is true:
- The current CPU is unset (equal to RPS_NO_CPU)
- Current CPU is offline
- The current CPU's queue head counter >= queue tail counter in the
rps_dev_flow table. This checks if the queue tail has advanced
beyond the last packet that was enqueued using this table entry.
This guarantees that all packets queued using this entry have been
dequeued, thus preserving in order delivery.
Making each queue have its own rps_dev_flow table has two advantages:
1) the tail queue counters will be written on each receive, so
keeping the table local to interrupting CPU s good for locality. 2)
this allows lockless access to the table-- the CPU number and queue
tail counter need to be accessed together under mutual exclusion
from netif_receive_skb, we assume that this is only called from
device napi_poll which is non-reentrant.
This patch implements RFS for TCP and connected UDP sockets.
It should be usable for other flow oriented protocols.
There are two configuration parameters for RFS. The
"rps_flow_entries" kernel init parameter sets the number of
entries in the rps_sock_flow_table, the per rxqueue sysfs entry
"rps_flow_cnt" contains the number of entries in the rps_dev_flow
table for the rxqueue. Both are rounded to power of two.
The obvious benefit of RFS (over just RPS) is that it achieves
CPU locality between the receive processing for a flow and the
applications processing; this can result in increased performance
(higher pps, lower latency).
The benefits of RFS are dependent on cache hierarchy, application
load, and other factors. On simple benchmarks, we don't necessarily
see improvement and sometimes see degradation. However, for more
complex benchmarks and for applications where cache pressure is
much higher this technique seems to perform very well.
Below are some benchmark results which show the potential benfit of
this patch. The netperf test has 500 instances of netperf TCP_RR
test with 1 byte req. and resp. The RPC test is an request/response
test similar in structure to netperf RR test ith 100 threads on
each host, but does more work in userspace that netperf.
e1000e on 8 core Intel
No RFS or RPS 104K tps at 30% CPU
No RFS (best RPS config): 290K tps at 63% CPU
RFS 303K tps at 61% CPU
RPC test tps CPU% 50/90/99% usec latency Latency StdDev
No RFS/RPS 103K 48% 757/900/3185 4472.35
RPS only: 174K 73% 415/993/2468 491.66
RFS 223K 73% 379/651/1382 315.61
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commits 5051ebd275 and
5051ebd275 ("ipv[46]: udp: optimize unicast RX
path") broke some programs.
After upgrading a L2TP server to 2.6.33 it started to fail, tunnels going up an
down, after the 10th tunnel came up. My modified rp-l2tp uses a global
unconnected socket bound to (INADDR_ANY, 1701) and one connected socket per
tunnel after parameter negotiation.
After ten sockets were open and due to mixed parameters to
udp[46]_lib_lookup2() kernel started to drop packets.
Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
sk_add_backlog -> __sk_add_backlog
sk_add_backlog_limited -> sk_add_backlog
Signed-off-by: Zhu Yi <yi.zhu@intel.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Make udp adapt to the limited socket backlog change.
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Cc: "Pekka Savola (ipv6)" <pekkas@netcore.fi>
Cc: Patrick McHardy <kaber@trash.net>
Signed-off-by: Zhu Yi <yi.zhu@intel.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The variable 'copied' is used in udp_recvmsg() to emphasize that the passed
'len' is adjusted to fit the actual datagram length. But the same can be
done by adjusting 'len' directly. This patch thus removes the indirection.
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
__net_init/__net_exit are apparently not going away, so use them
to full extent.
In some cases __net_init was removed, because it was called from
__net_exit code.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Now we can have a large udp hash table, udp_lib_get_port() loop
should be converted to a do {} while (cond) form,
or we dont enter it at all if hash table size is exactly 65536.
Reported-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
On Sun, 2009-11-22 at 16:31 -0800, David Miller wrote:
> It should be of the form:
> if (x &&
> y)
>
> or:
> if (x && y)
>
> Fix patches, rather than complaints, for existing cases where things
> do not follow this pattern are certainly welcome.
Also collapsed some multiple tabs to single space.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
UDP bind() can be O(N^2) in some pathological cases.
Thanks to secondary hash tables, we can make it O(N)
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When skb_clone() fails, we should increment sk_drops and SNMP counters.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
UDP multicast rx path is a bit complex and can hold a spinlock
for a long time.
Using a small (32 or 64 entries) stack of socket pointers can help
to perform expensive operations (skb_clone(), udp_queue_rcv_skb())
outside of the lock, in most cases.
It's also a base for a future RCU conversion of multicast recption.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Lucian Adrian Grijincu <lgrijincu@ixiacom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We first locate the (local port) hash chain head
If few sockets are in this chain, we proceed with previous lookup algo.
If too many sockets are listed, we take a look at the secondary
(port, address) hash chain we added in previous patch.
We choose the shortest chain and proceed with a RCU lookup on the elected chain.
But, if we chose (port, address) chain, and fail to find a socket on given address,
we must try another lookup on (port, INADDR_ANY) chain to find socket not bound
to a particular IP.
-> No extra cost for typical setups, where the first lookup will probabbly
be performed.
RCU lookups everywhere, we dont acquire spinlock.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Extends udp_table to contain a secondary hash table.
socket anchor for this second hash is free, because UDP
doesnt use skc_bind_node : We define an union to hold
both skc_bind_node & a new hlist_nulls_node udp_portaddr_node
udp_lib_get_port() inserts sockets into second hash chain
(additional cost of one atomic op)
udp_lib_unhash() deletes socket from second hash chain
(additional cost of one atomic op)
Note : No spinlock lockdep annotation is needed, because
lock for the secondary hash chain is always get after
lock for primary hash chain.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Union sk_hash with two u16 hashes for udp (no extra memory taken)
One 16 bits hash on (local port) value (the previous udp 'hash')
One 16 bits hash on (local address, local port) values, initialized
but not yet used. This second hash is using jenkin hash for better
distribution.
Because the 'port' is xored later, a partial hash is performed
on local address + net_hash_mix(net)
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Adds a counter in udp_hslot to keep an accurate count
of sockets present in chain.
This will permit to upcoming UDP lookup algo to chose
the shortest chain when secondary hash is added.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/usb/cdc_ether.c
All CDC ethernet devices of type USB_CLASS_COMM need to use
'&mbm_info'.
Signed-off-by: David S. Miller <davem@davemloft.net>
On UDP sockets, we must call skb_free_datagram() with socket locked,
or risk sk_forward_alloc corruption. This requirement is not respected
in SUNRPC.
Add a convenient helper, skb_free_datagram_locked() and use it in SUNRPC
Reported-by: Francis Moreau <francis.moro@gmail.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
- skb_kill_datagram() can increment sk->sk_drops itself, not callers.
- UDP on IPV4 & IPV6 dropped frames (because of bad checksum or policy checks) increment sk_drops
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In order to have better cache layouts of struct sock (separate zones
for rx/tx paths), we need this preliminary patch.
Goal is to transfert fields used at lookup time in the first
read-mostly cache line (inside struct sock_common) and move sk_refcnt
to a separate cache line (only written by rx path)
This patch adds inet_ prefix to daddr, rcv_saddr, dport, num, saddr,
sport and id fields. This allows a future patch to define these
fields as macros, like sk_refcnt, without name clashes.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sock_queue_rcv_skb() can update sk_drops itself, removing need for
callers to take care of it. This is more consistent since
sock_queue_rcv_skb() also reads sk_drops when queueing a skb.
This adds sk_drops managment to many protocols that not cared yet.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
udp_poll() can in some circumstances drop frames with incorrect checksums.
Problem is we now have to lock the socket while dropping frames, or risk
sk_forward corruption.
This bug is present since commit 95766fff6b
([UDP]: Add memory accounting.)
While we are at it, we can correct ioctl(SIOCINQ) to also drop bad frames.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Create a new socket level option to report number of queue overflows
Recently I augmented the AF_PACKET protocol to report the number of frames lost
on the socket receive queue between any two enqueued frames. This value was
exported via a SOL_PACKET level cmsg. AFter I completed that work it was
requested that this feature be generalized so that any datagram oriented socket
could make use of this option. As such I've created this patch, It creates a
new SOL_SOCKET level option called SO_RXQ_OVFL, which when enabled exports a
SOL_SOCKET level cmsg that reports the nubmer of times the sk_receive_queue
overflowed between any two given frames. It also augments the AF_PACKET
protocol to take advantage of this new feature (as it previously did not touch
sk->sk_drops, which this patch uses to record the overflow count). Tested
successfully by me.
Notes:
1) Unlike my previous patch, this patch simply records the sk_drops value, which
is not a number of drops between packets, but rather a total number of drops.
Deltas must be computed in user space.
2) While this patch currently works with datagram oriented protocols, it will
also be accepted by non-datagram oriented protocols. I'm not sure if thats
agreeable to everyone, but my argument in favor of doing so is that, for those
protocols which aren't applicable to this option, sk_drops will always be zero,
and reporting no drops on a receive queue that isn't used for those
non-participating protocols seems reasonable to me. This also saves us having
to code in a per-protocol opt in mechanism.
3) This applies cleanly to net-next assuming that commit
977750076d (my af packet cmsg patch) is reverted
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
UDP_HTABLE_SIZE was initialy defined to 128, which is a bit small for
several setups.
4000 active UDP sockets -> 32 sockets per chain in average. An
incoming frame has to lookup all sockets to find best match, so long
chains hurt latency.
Instead of a fixed size hash table that cant be perfect for every
needs, let UDP stack choose its table size at boot time like tcp/ip
route, using alloc_large_system_hash() helper
Add an optional boot parameter, uhash_entries=x so that an admin can
force a size between 256 and 65536 if needed, like thash_entries and
rhash_entries.
dmesg logs two new lines :
[ 0.647039] UDP hash table entries: 512 (order: 0, 4096 bytes)
[ 0.647099] UDP Lite hash table entries: 512 (order: 0, 4096 bytes)
Maximal size on 64bit arches would be 65536 slots, ie 1 MBytes for non
debugging spinlocks.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch against v2.6.31 adds support for route lookup using sk_mark in some
more places. The benefits from this patch are the following.
First, SO_MARK option now has effect on UDP sockets too.
Second, ip_queue_xmit() and inet_sk_rebuild_header() could fail to do routing
lookup correctly if TCP sockets with SO_MARK were used.
Signed-off-by: Atis Elsts <atis@mikrotik.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
This provides safety against negative optlen at the type
level instead of depending upon (sometimes non-trivial)
checks against this sprinkled all over the the place, in
each and every implementation.
Based upon work done by Arjan van de Ven and feedback
from Linus Torvalds.
Signed-off-by: David S. Miller <davem@davemloft.net>
Christoph Lameter pointed out that packet drops at qdisc level where not
accounted in SNMP counters. Only if application sets IP_RECVERR, drops
are reported to user (-ENOBUFS errors) and SNMP counters updated.
IP_RECVERR is used to enable extended reliable error message passing,
but these are not needed to update system wide SNMP stats.
This patch changes things a bit to allow SNMP counters to be updated,
regardless of IP_RECVERR being set or not on the socket.
Example after an UDP tx flood
# netstat -s
...
IP:
1487048 outgoing packets dropped
...
Udp:
...
SndbufErrors: 1487048
send() syscalls, do however still return an OK status, to not
break applications.
Note : send() manual page explicitly says for -ENOBUFS error :
"The output queue for a network interface was full.
This generally indicates that the interface has stopped sending,
but may be caused by transient congestion.
(Normally, this does not occur in Linux. Packets are just silently
dropped when a device queue overflows.) "
This is not true for IP_RECVERR enabled sockets : a send() syscall
that hit a qdisc drop returns an ENOBUFS error.
Many thanks to Christoph, David, and last but not least, Alexey !
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
- validate and forward GSO UDP/IPv4 packets from untrusted sources.
- do software UFO if the outgoing device doesn't support UFO.
Signed-off-by: Sridhar Samudrala <sri@us.ibm.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 2b85a34e91
(net: No more expensive sock_hold()/sock_put() on each tx)
changed initial sk_wmem_alloc value.
We need to take into account this offset when reporting
sk_wmem_alloc to user, in PROC_FS files or various
ioctls (SIOCOUTQ/TIOCOUTQ)
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Define three accessors to get/set dst attached to a skb
struct dst_entry *skb_dst(const struct sk_buff *skb)
void skb_dst_set(struct sk_buff *skb, struct dst_entry *dst)
void skb_dst_drop(struct sk_buff *skb)
This one should replace occurrences of :
dst_release(skb->dst)
skb->dst = NULL;
Delete skb->dst field
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The ipv6 version of bind_conflict code calls ipv6_rcv_saddr_equal()
which at times wrongly identified intersections between addresses.
It particularly broke down under a few instances and caused erroneous
bind conflicts.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Instructions for time stamping outgoing packets are take from the
socket layer and later copied into the new skb.
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Like the UDP header fix, pskb_may_pull() can potentially
alter the SKB buffer. Thus the saddr and daddr, pointers
may point to the old skb->data buffer.
I haven't seen corruptions, as its only seen if the old
skb->data buffer were reallocated by another user and
written into very quickly (or poison'd by SLAB debugging).
Signed-off-by: Jesper Dangaard Brouer <hawk@comx.dk>
Signed-off-by: David S. Miller <davem@davemloft.net>
The UDP header pointer assignment must happen after calling
pskb_may_pull(). As pskb_may_pull() can potentially alter the SKB
buffer.
This was exposted by running multicast traffic through the NIU driver,
as it won't prepull the protocol headers into the linear area on
receive.
Signed-off-by: Jesper Dangaard Brouer <hawk@comx.dk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 93821778de (udp: Fix rcv socket
locking) accidentally removed sk_drops increments for UDP IPV4
sockets.
This field can be used to detect incorrect sizing of socket receive
buffers.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 9088c56095
(udp: Improve port randomization) introduced a regression for UDP bind() syscall
to null port (getting a random port) in case lot of ports are already in use.
This is because we do about 28000 scans of very long chains (220 sockets per chain),
with many spin_lock_bh()/spin_unlock_bh() calls.
Fix this using a bitmap (64 bytes for current value of UDP_HTABLE_SIZE)
so that we scan chains at most once.
Instead of 250 ms per bind() call, we get after patch a time of 2.9 ms
Based on a report from Vitaly Mayatskikh
Reported-by: Vitaly Mayatskikh <v.mayatskih@gmail.com>
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Tested-by: Vitaly Mayatskikh <v.mayatskih@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Impact: Optimization
Like done in inet_unhash(), we can avoid taking a chain lock if
socket is not hashed in udp_unhash()
Triggered by close(socket(AF_INET, SOCK_DGRAM, 0));
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We can reduce pressure on dst entry refcount that slowdown UDP transmit
path on SMP machines. This pressure is visible on RTP servers when
delivering content to mediagateways, especially big ones, handling
thousand of streams. Several cpus send UDP frames to the same
destination, hence use the same dst entry.
This patch makes ip_append_data() eventually steal the refcount its
callers had to take on the dst entry.
This doesnt avoid all refcounting, but still gives speedups on SMP,
on UDP/RAW transmit path
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
udp_sendmsg() didn't fill struct flowi->flags, which means that
the route lookup would fail for non-local IPs even if the
IP_TRANSPARENT sockopt was set.
This prevents sendto() to work properly for UDP sockets, whereas
bind(foreign-ip) + connect() + send() worked fine.
Signed-off-by: Balazs Scheidler <bazsi@balabit.hu>
Signed-off-by: David S. Miller <davem@davemloft.net>
This is a straightforward patch, using hlist_nulls infrastructure.
RCUification already done on UDP two weeks ago.
Using hlist_nulls permits us to avoid some memory barriers, both
at lookup time and delete time.
Patch is large because it adds new macros to include/net/sock.h.
These macros will be used by TCP & DCCP in next patch.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Current UDP multicast delivery is not namespace aware.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Acked-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Corey Minyard spotted a missing memory barrier in udp_lib_get_port()
We need to make sure a reader cannot read the new 'sk->sk_next' value
and previous value of 'sk->sk_hash'. Or else, an item could be deleted
from a chain, and inserted into another chain. If new chain was empty
before the move, 'next' pointer is NULL, and lockless reader can
not detect it missed following items in original chain.
This patch is temporary, since we expect an upcoming patch
to introduce another way of handling the problem.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Using NIPQUAD() with NIPQUAD_FMT, %d.%d.%d.%d or %u.%u.%u.%u
can be replaced with %pI4
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Corey Minyard found a race added in commit 271b72c7fa
(udp: RCU handling for Unicast packets.)
"If the socket is moved from one list to another list in-between the
time the hash is calculated and the next field is accessed, and the
socket has moved to the end of the new list, the traversal will not
complete properly on the list it should have, since the socket will
be on the end of the new list and there's not a way to tell it's on a
new list and restart the list traversal. I think that this can be
solved by pre-fetching the "next" field (with proper barriers) before
checking the hash."
This patch corrects this problem, introducing a new
sk_for_each_rcu_safenext() macro.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch mimics commit 57413ebc4e
(tcp: calculate tcp_mem based on low memory instead of all memory)
The udp_mem array which contains limits on the total amount of memory
used by UDP sockets is calculated based on nr_all_pages. On a 32 bits
x86 system, we should base this on the number of lowmem pages.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Goals are :
1) Optimizing handling of incoming Unicast UDP frames, so that no memory
writes should happen in the fast path.
Note: Multicasts and broadcasts still will need to take a lock,
because doing a full lockless lookup in this case is difficult.
2) No expensive operations in the socket bind/unhash phases :
- No expensive synchronize_rcu() calls.
- No added rcu_head in socket structure, increasing memory needs,
but more important, forcing us to use call_rcu() calls,
that have the bad property of making sockets structure cold.
(rcu grace period between socket freeing and its potential reuse
make this socket being cold in CPU cache).
David did a previous patch using call_rcu() and noticed a 20%
impact on TCP connection rates.
Quoting Cristopher Lameter :
"Right. That results in cacheline cooldown. You'd want to recycle
the object as they are cache hot on a per cpu basis. That is screwed
up by the delayed regular rcu processing. We have seen multiple
regressions due to cacheline cooldown.
The only choice in cacheline hot sensitive areas is to deal with the
complexity that comes with SLAB_DESTROY_BY_RCU or give up on RCU."
- Because udp sockets are allocated from dedicated kmem_cache,
use of SLAB_DESTROY_BY_RCU can help here.
Theory of operation :
---------------------
As the lookup is lockfree (using rcu_read_lock()/rcu_read_unlock()),
special attention must be taken by readers and writers.
Use of SLAB_DESTROY_BY_RCU is tricky too, because a socket can be freed,
reused, inserted in a different chain or in worst case in the same chain
while readers could do lookups in the same time.
In order to avoid loops, a reader must check each socket found in a chain
really belongs to the chain the reader was traversing. If it finds a
mismatch, lookup must start again at the begining. This *restart* loop
is the reason we had to use rdlock for the multicast case, because
we dont want to send same message several times to the same socket.
We use RCU only for fast path.
Thus, /proc/net/udp still takes spinlocks.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
UDP sockets are hashed in a 128 slots hash table.
This hash table is protected by *one* rwlock.
This rwlock is readlocked each time an incoming UDP message is handled.
This rwlock is writelocked each time a socket must be inserted in
hash table (bind time), or deleted from this table (close time)
This is not scalable on SMP machines :
1) Even in read mode, lock() and unlock() are atomic operations and
must dirty a contended cache line, shared by all cpus.
2) A writer might be starved if many readers are 'in flight'. This can
happen on a machine with some NIC receiving many UDP messages. User
process can be delayed a long time at socket creation/dismantle time.
This patch prepares RCU migration, by introducing 'struct udp_table
and struct udp_hslot', and using one spinlock per chain, to reduce
contention on central rwlock.
Introducing one spinlock per chain reduces latencies, for port
randomization on heavily loaded UDP servers. This also speedup
bindings to specific ports.
udp_lib_unhash() was uninlined, becoming to big.
Some cleanups were done to ease review of following patch
(RCUification of UDP Unicast lookups)
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Clean up the various different email addresses of mine listed in the code
to a single current and valid address. As Dave says his network merges
for 2.6.28 are now done this seems a good point to send them in where
they won't risk disrupting real changes.
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While looking at UDP port randomization, I noticed it
was litle bit pessimistic, not looking at type of sockets
(IPV6/IPV4) and not looking at bound addresses if any.
We should perform same tests than when binding to a
specific port.
This permits a cleanup of udp_lib_get_port()
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Current UDP port allocation is suboptimal.
We select the shortest chain to chose a port (out of 512)
that will hash in this shortest chain.
First, it can lead to give not so ramdom ports and ease
give attackers more opportunities to break the system.
Second, it can consume a lot of CPU to scan all table
in order to find the shortest chain.
Third, in some pathological cases we can fail to find
a free port even if they are plenty of them.
This patch zap the search for a short chain and only
use one random seed. Problem of getting long chains
should be addressed in another way, since we can
obtain long chains with non random ports.
Based on a report and patch from Vitaly Mayatskikh
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use the socket cached in the skb if it's present.
Signed-off-by: KOVACS Krisztian <hidden@sch.bme.hu>
Acked-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
To be able to use the cached socket reference in the skb during input
processing we add a new set of lookup functions that receive the skb on
their argument list.
Signed-off-by: KOVACS Krisztian <hidden@sch.bme.hu>
Acked-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The iptables tproxy code has to be able to do UDP socket hash lookups,
so we have to provide an exported lookup function for this purpose.
Signed-off-by: KOVACS Krisztian <hidden@sch.bme.hu>
Signed-off-by: David S. Miller <davem@davemloft.net>
The previous patch in response to the recursive locking on IPsec
reception is broken as it tries to drop the BH socket lock while in
user context.
This patch fixes it by shrinking the section protected by the
socket lock to sock_queue_rcv_skb only. The only reason we added
the lock is for the accounting which happens in that function.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
The socket lock is there to protect the normal UDP receive path.
Encapsulation UDP sockets don't need that protection. In fact
the locking is deadly for them as they may contain another UDP
packet within, possibly with the same addresses.
Also the nested bit was copied from TCP. TCP needs it because
of accept(2) spawning sockets. This simply doesn't apply to UDP
so I've removed it.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch clamps the cscov setsockopt values to a maximum of 0xFFFF.
Setsockopt values greater than 0xffff can cause an unwanted
wrap-around. Further, IPv6 jumbograms are not supported (RFC 3838,
3.5), so that values greater than 0xffff are not even useful.
Further changes: fixed a typo in the documentation.
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Some places, that deal with IP statistics already have where to
get a struct net from, but use it directly, without declaring
a separate variable on the stack.
So, save this net on the stack for future IP_XXX_STATS macros.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Some places, that deal with ICMP statistics already have where
to get a struct net from, but use it directly, without declaring
a separate variable on the stack.
Since I will need this net soon, I declare a struct net on the
stack and use it in the existing places in a separate patch not
to spoil the future ones.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Two special cases here - one is rxrpc - I put init_net there
explicitly, since we haven't touched this part yet. The second
place is in __udp4_lib_rcv - we already have a struct net there,
but I have to move its initialization above to make it ready
at the "drop" label.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Acked-by: Denis V. Lunev <den@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Michał Mirosław
David S. Miller
Octavian Purdila
Changli Gao
Eric Dumazet
Oliver Hartkopp
Linus Torvalds
Alexey Dobriyan
Amerigo Wang
Bjørn Mork
Tom Herbert
Jorge Boncompte [DTI2]
Tejun Heo
Zhu Yi
Gerrit Renker
Joe Perches
Neil Horman
Atis Elsts
Sridhar Samudrala
Vlad Yasevich
Vitaly Mayatskikh
Patrick Ohly
Jesper Dangaard Brouer
Eric Dumazet
Balazs Scheidler
Harvey Harrison
Alan Cox
Denis V. Lunev
KOVACS Krisztian
Herbert Xu
Pavel Emelyanov