linux/net/rds/tcp_recv.c

337 lines
8.9 KiB
C
Raw Normal View History

/*
* Copyright (c) 2006 Oracle. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*
*/
#include <linux/kernel.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 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>
2010-03-24 09:04:11 +01:00
#include <linux/slab.h>
#include <net/tcp.h>
#include "rds.h"
#include "tcp.h"
static struct kmem_cache *rds_tcp_incoming_slab;
static void rds_tcp_inc_purge(struct rds_incoming *inc)
{
struct rds_tcp_incoming *tinc;
tinc = container_of(inc, struct rds_tcp_incoming, ti_inc);
rdsdebug("purging tinc %p inc %p\n", tinc, inc);
skb_queue_purge(&tinc->ti_skb_list);
}
void rds_tcp_inc_free(struct rds_incoming *inc)
{
struct rds_tcp_incoming *tinc;
tinc = container_of(inc, struct rds_tcp_incoming, ti_inc);
rds_tcp_inc_purge(inc);
rdsdebug("freeing tinc %p inc %p\n", tinc, inc);
kmem_cache_free(rds_tcp_incoming_slab, tinc);
}
/*
* this is pretty lame, but, whatever.
*/
int rds_tcp_inc_copy_to_user(struct rds_incoming *inc, struct iov_iter *to)
{
struct rds_tcp_incoming *tinc;
struct sk_buff *skb;
int ret = 0;
if (!iov_iter_count(to))
goto out;
tinc = container_of(inc, struct rds_tcp_incoming, ti_inc);
skb_queue_walk(&tinc->ti_skb_list, skb) {
unsigned long to_copy, skb_off;
for (skb_off = 0; skb_off < skb->len; skb_off += to_copy) {
to_copy = iov_iter_count(to);
to_copy = min(to_copy, skb->len - skb_off);
if (skb_copy_datagram_iter(skb, skb_off, to, to_copy))
return -EFAULT;
rds_stats_add(s_copy_to_user, to_copy);
ret += to_copy;
if (!iov_iter_count(to))
goto out;
}
}
out:
return ret;
}
/*
* We have a series of skbs that have fragmented pieces of the congestion
* bitmap. They must add up to the exact size of the congestion bitmap. We
* use the skb helpers to copy those into the pages that make up the in-memory
* congestion bitmap for the remote address of this connection. We then tell
* the congestion core that the bitmap has been changed so that it can wake up
* sleepers.
*
* This is racing with sending paths which are using test_bit to see if the
* bitmap indicates that their recipient is congested.
*/
static void rds_tcp_cong_recv(struct rds_connection *conn,
struct rds_tcp_incoming *tinc)
{
struct sk_buff *skb;
unsigned int to_copy, skb_off;
unsigned int map_off;
unsigned int map_page;
struct rds_cong_map *map;
int ret;
/* catch completely corrupt packets */
if (be32_to_cpu(tinc->ti_inc.i_hdr.h_len) != RDS_CONG_MAP_BYTES)
return;
map_page = 0;
map_off = 0;
map = conn->c_fcong;
skb_queue_walk(&tinc->ti_skb_list, skb) {
skb_off = 0;
while (skb_off < skb->len) {
to_copy = min_t(unsigned int, PAGE_SIZE - map_off,
skb->len - skb_off);
BUG_ON(map_page >= RDS_CONG_MAP_PAGES);
/* only returns 0 or -error */
ret = skb_copy_bits(skb, skb_off,
(void *)map->m_page_addrs[map_page] + map_off,
to_copy);
BUG_ON(ret != 0);
skb_off += to_copy;
map_off += to_copy;
if (map_off == PAGE_SIZE) {
map_off = 0;
map_page++;
}
}
}
rds_cong_map_updated(map, ~(u64) 0);
}
struct rds_tcp_desc_arg {
struct rds_connection *conn;
gfp_t gfp;
};
static int rds_tcp_data_recv(read_descriptor_t *desc, struct sk_buff *skb,
unsigned int offset, size_t len)
{
struct rds_tcp_desc_arg *arg = desc->arg.data;
struct rds_connection *conn = arg->conn;
struct rds_tcp_connection *tc = conn->c_transport_data;
struct rds_tcp_incoming *tinc = tc->t_tinc;
struct sk_buff *clone;
size_t left = len, to_copy;
rdsdebug("tcp data tc %p skb %p offset %u len %zu\n", tc, skb, offset,
len);
/*
* tcp_read_sock() interprets partial progress as an indication to stop
* processing.
*/
while (left) {
if (!tinc) {
tinc = kmem_cache_alloc(rds_tcp_incoming_slab,
arg->gfp);
if (!tinc) {
desc->error = -ENOMEM;
goto out;
}
tc->t_tinc = tinc;
rdsdebug("alloced tinc %p\n", tinc);
rds_inc_init(&tinc->ti_inc, conn, conn->c_faddr);
/*
* XXX * we might be able to use the __ variants when
* we've already serialized at a higher level.
*/
skb_queue_head_init(&tinc->ti_skb_list);
}
if (left && tc->t_tinc_hdr_rem) {
to_copy = min(tc->t_tinc_hdr_rem, left);
rdsdebug("copying %zu header from skb %p\n", to_copy,
skb);
skb_copy_bits(skb, offset,
(char *)&tinc->ti_inc.i_hdr +
sizeof(struct rds_header) -
tc->t_tinc_hdr_rem,
to_copy);
tc->t_tinc_hdr_rem -= to_copy;
left -= to_copy;
offset += to_copy;
if (tc->t_tinc_hdr_rem == 0) {
/* could be 0 for a 0 len message */
tc->t_tinc_data_rem =
be32_to_cpu(tinc->ti_inc.i_hdr.h_len);
}
}
if (left && tc->t_tinc_data_rem) {
clone = skb_clone(skb, arg->gfp);
if (!clone) {
desc->error = -ENOMEM;
goto out;
}
to_copy = min(tc->t_tinc_data_rem, left);
pskb_pull(clone, offset);
pskb_trim(clone, to_copy);
skb_queue_tail(&tinc->ti_skb_list, clone);
rdsdebug("skb %p data %p len %d off %u to_copy %zu -> "
"clone %p data %p len %d\n",
skb, skb->data, skb->len, offset, to_copy,
clone, clone->data, clone->len);
tc->t_tinc_data_rem -= to_copy;
left -= to_copy;
offset += to_copy;
}
if (tc->t_tinc_hdr_rem == 0 && tc->t_tinc_data_rem == 0) {
if (tinc->ti_inc.i_hdr.h_flags == RDS_FLAG_CONG_BITMAP)
rds_tcp_cong_recv(conn, tinc);
else
rds_recv_incoming(conn, conn->c_faddr,
conn->c_laddr, &tinc->ti_inc,
arg->gfp);
tc->t_tinc_hdr_rem = sizeof(struct rds_header);
tc->t_tinc_data_rem = 0;
tc->t_tinc = NULL;
rds_inc_put(&tinc->ti_inc);
tinc = NULL;
}
}
out:
rdsdebug("returning len %zu left %zu skb len %d rx queue depth %d\n",
len, left, skb->len,
skb_queue_len(&tc->t_sock->sk->sk_receive_queue));
return len - left;
}
/* the caller has to hold the sock lock */
static int rds_tcp_read_sock(struct rds_connection *conn, gfp_t gfp)
{
struct rds_tcp_connection *tc = conn->c_transport_data;
struct socket *sock = tc->t_sock;
read_descriptor_t desc;
struct rds_tcp_desc_arg arg;
/* It's like glib in the kernel! */
arg.conn = conn;
arg.gfp = gfp;
desc.arg.data = &arg;
desc.error = 0;
desc.count = 1; /* give more than one skb per call */
tcp_read_sock(sock->sk, &desc, rds_tcp_data_recv);
rdsdebug("tcp_read_sock for tc %p gfp 0x%x returned %d\n", tc, gfp,
desc.error);
return desc.error;
}
/*
* We hold the sock lock to serialize our rds_tcp_recv->tcp_read_sock from
* data_ready.
*
* if we fail to allocate we're in trouble.. blindly wait some time before
* trying again to see if the VM can free up something for us.
*/
int rds_tcp_recv(struct rds_connection *conn)
{
struct rds_tcp_connection *tc = conn->c_transport_data;
struct socket *sock = tc->t_sock;
int ret = 0;
rdsdebug("recv worker conn %p tc %p sock %p\n", conn, tc, sock);
lock_sock(sock->sk);
ret = rds_tcp_read_sock(conn, GFP_KERNEL);
release_sock(sock->sk);
return ret;
}
void rds_tcp_data_ready(struct sock *sk)
{
void (*ready)(struct sock *sk);
struct rds_connection *conn;
struct rds_tcp_connection *tc;
rdsdebug("data ready sk %p\n", sk);
read_lock(&sk->sk_callback_lock);
conn = sk->sk_user_data;
if (!conn) { /* check for teardown race */
ready = sk->sk_data_ready;
goto out;
}
tc = conn->c_transport_data;
ready = tc->t_orig_data_ready;
rds_tcp_stats_inc(s_tcp_data_ready_calls);
if (rds_tcp_read_sock(conn, GFP_ATOMIC) == -ENOMEM)
queue_delayed_work(rds_wq, &conn->c_recv_w, 0);
out:
read_unlock(&sk->sk_callback_lock);
ready(sk);
}
int rds_tcp_recv_init(void)
{
rds_tcp_incoming_slab = kmem_cache_create("rds_tcp_incoming",
sizeof(struct rds_tcp_incoming),
0, 0, NULL);
if (!rds_tcp_incoming_slab)
return -ENOMEM;
return 0;
}
void rds_tcp_recv_exit(void)
{
kmem_cache_destroy(rds_tcp_incoming_slab);
}