linux-toradex.git/net/rds/tcp_send.c, branch v3.0.63

rds: make local functions/variables static

2010-10-21T11:26:39+00:00

The RDS protocol has lots of functions that should be
declared static. rds_message_get/add_version_extension is
removed since it defined but never used.

Signed-off-by: Stephen Hemminger 
Signed-off-by: David S. Miller

Merge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6

2010-09-27T08:03:03+00:00

Conflicts:
	drivers/net/qlcnic/qlcnic_init.c
	net/ipv4/ip_output.c

net: fix a lockdep splat

2010-09-25T05:26:10+00:00

We have for each socket :

One spinlock (sk_slock.slock)
One rwlock (sk_callback_lock)

Possible scenarios are :

(A) (this is used in net/sunrpc/xprtsock.c)
read_lock(&sk->sk_callback_lock) (without blocking BH)

spin_lock(&sk->sk_slock.slock);
...
read_lock(&sk->sk_callback_lock);
...

(B)
write_lock_bh(&sk->sk_callback_lock)
stuff
write_unlock_bh(&sk->sk_callback_lock)

(C)
spin_lock_bh(&sk->sk_slock)
...
write_lock_bh(&sk->sk_callback_lock)
stuff
write_unlock_bh(&sk->sk_callback_lock)
spin_unlock_bh(&sk->sk_slock)

This (C) case conflicts with (A) :

CPU1 [A]                         CPU2 [C]
read_lock(callback_lock)
                             spin_lock_bh(slock)

                                 

We have one problematic (C) use case in inet_csk_listen_stop() :

local_bh_disable();
bh_lock_sock(child); // spin_lock_bh(&sk->sk_slock)
WARN_ON(sock_owned_by_user(child));
...
sock_orphan(child); // write_lock_bh(&sk->sk_callback_lock)

lockdep is not happy with this, as reported by Tetsuo Handa

It seems only way to deal with this is to use read_lock_bh(callbacklock)
everywhere.

Thanks to Jarek for pointing a bug in my first attempt and suggesting
this solution.

Reported-by: Tetsuo Handa 
Tested-by: Tetsuo Handa 
Signed-off-by: Eric Dumazet 
CC: Jarek Poplawski 
Tested-by: Eric Dumazet 
Signed-off-by: David S. Miller

RDS: Stop supporting old cong map sending method

2010-09-09T01:12:10+00:00

We now ask the transport to give us a rm for the congestion
map, and then we handle it normally. Previously, the
transport defined a function that we would call to send
a congestion map.

Convert TCP and loop transports to new cong map method.

Signed-off-by: Andy Grover

RDS: Rename data op members prefix from m_ to op_

2010-09-09T01:11:59+00:00

For consistency.

Signed-off-by: Andy Grover

RDS: break out rdma and data ops into nested structs in rds_message

2010-09-09T01:11:33+00:00

Clearly separate rdma-related variables in rm from data-related ones.
This is in anticipation of adding atomic support.

Signed-off-by: Andy Grover

RDS: cleanup: remove "== NULL"s and "!= NULL"s in ptr comparisons

2010-09-09T01:11:32+00:00

Favor "if (foo)" style over "if (foo != NULL)".

Signed-off-by: Andy Grover

RDS/TCP: Wait to wake thread when write space available

2010-03-17T04:16:55+00:00

Instead of waking the send thread whenever any send space is available,
wait until it is at least half empty. This is modeled on how
sock_def_write_space() does it, and may help to minimize context
switches.

Signed-off-by: Andy Grover 
Signed-off-by: David S. Miller

net/rds: remove uses of NIPQUAD, use %pI4

2010-02-04T04:16:48+00:00

Signed-off-by: Joe Perches 
Cc: Andy Grover 
Signed-off-by: Andrew Morton 
Signed-off-by: David S. Miller

RDS: Add TCP transport to RDS

2009-08-24T02:13:02+00:00

This code allows RDS to be tunneled over a TCP connection.

RDMA operations are disabled when using TCP transport,
but this frees RDS from the IB/RDMA stack dependency, and allows
it to be used with standard Ethernet adapters, or in a VM.

Signed-off-by: Andy Grover 
Signed-off-by: David S. Miller