From: Doug Ledford <dledford@redhat.com>
Date: Tue, 14 Apr 2009 15:23:31 -0400
Subject: [openib] core: update core code to OFED 1.4.1-rc3
Message-id: 1239737023-31222-5-git-send-email-dledford@redhat.com
O-Subject: [Patch RHEL5.4 04/16] [infiniband/core] Update core code to OFED 1.4.1-rc3
Bugzilla: 476301
Signed-off-by: Doug Ledford <dledford@redhat.com>
diff --git a/drivers/infiniband/core/Makefile b/drivers/infiniband/core/Makefile
index f646040..b2e522d 100644
--- a/drivers/infiniband/core/Makefile
+++ b/drivers/infiniband/core/Makefile
@@ -30,3 +30,5 @@ ib_umad-y := user_mad.o
ib_ucm-y := ucm.o
ib_uverbs-y := uverbs_main.o uverbs_cmd.o uverbs_marshall.o
+
+ib_core-y += namespace.o writeback.o
diff --git a/drivers/infiniband/core/addr.c b/drivers/infiniband/core/addr.c
index 485f81e..09a2bec 100644
--- a/drivers/infiniband/core/addr.c
+++ b/drivers/infiniband/core/addr.c
@@ -4,28 +4,33 @@
* Copyright (c) 1999-2005, Mellanox Technologies, Inc. All rights reserved.
* Copyright (c) 2005 Intel Corporation. All rights reserved.
*
- * This Software is licensed under one of the following licenses:
+ * 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:
*
- * 1) under the terms of the "Common Public License 1.0" a copy of which is
- * available from the Open Source Initiative, see
- * http://www.opensource.org/licenses/cpl.php.
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
*
- * 2) under the terms of the "The BSD License" a copy of which is
- * available from the Open Source Initiative, see
- * http://www.opensource.org/licenses/bsd-license.php.
+ * - Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer.
*
- * 3) under the terms of the "GNU General Public License (GPL) Version 2" a
- * copy of which is available from the Open Source Initiative, see
- * http://www.opensource.org/licenses/gpl-license.php.
+ * - 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.
*
- * Licensee has the right to choose one of the above licenses.
- *
- * Redistributions of source code must retain the above copyright
- * notice and one of the license notices.
- *
- * Redistributions in binary form must reproduce both the above copyright
- * notice, one of the license notices 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/mutex.h>
@@ -55,7 +60,7 @@ struct addr_req {
int status;
};
-static void process_req(void *work);
+static void process_req(struct work_struct *work);
static DEFINE_MUTEX(lock);
static LIST_HEAD(req_list);
@@ -100,6 +105,7 @@ int rdma_copy_addr(struct rdma_dev_addr *dev_addr, struct net_device *dev,
memcpy(dev_addr->broadcast, dev->broadcast, MAX_ADDR_LEN);
if (dst_dev_addr)
memcpy(dev_addr->dst_dev_addr, dst_dev_addr, MAX_ADDR_LEN);
+ dev_addr->src_dev = dev;
return 0;
}
EXPORT_SYMBOL(rdma_copy_addr);
@@ -110,7 +116,7 @@ int rdma_translate_ip(struct sockaddr *addr, struct rdma_dev_addr *dev_addr)
__be32 ip = ((struct sockaddr_in *) addr)->sin_addr.s_addr;
int ret;
- dev = ip_dev_find(ip);
+ dev = ip_dev_find(&init_net, ip);
if (!dev)
return -EADDRNOTAVAIL;
@@ -154,11 +160,11 @@ static void addr_send_arp(struct sockaddr_in *dst_in)
{
struct rtable *rt;
struct flowi fl;
- u32 dst_ip = dst_in->sin_addr.s_addr;
+ __be32 dst_ip = dst_in->sin_addr.s_addr;
memset(&fl, 0, sizeof fl);
fl.nl_u.ip4_u.daddr = dst_ip;
- if (ip_route_output_key(&rt, &fl))
+ if (ip_route_output_key(&init_net, &rt, &fl))
return;
neigh_event_send(rt->u.dst.neighbour, NULL);
@@ -169,8 +175,8 @@ static int addr_resolve_remote(struct sockaddr_in *src_in,
struct sockaddr_in *dst_in,
struct rdma_dev_addr *addr)
{
- u32 src_ip = src_in->sin_addr.s_addr;
- u32 dst_ip = dst_in->sin_addr.s_addr;
+ __be32 src_ip = src_in->sin_addr.s_addr;
+ __be32 dst_ip = dst_in->sin_addr.s_addr;
struct flowi fl;
struct rtable *rt;
struct neighbour *neigh;
@@ -179,7 +185,7 @@ static int addr_resolve_remote(struct sockaddr_in *src_in,
memset(&fl, 0, sizeof fl);
fl.nl_u.ip4_u.daddr = dst_ip;
fl.nl_u.ip4_u.saddr = src_ip;
- ret = ip_route_output_key(&rt, &fl);
+ ret = ip_route_output_key(&init_net, &rt, &fl);
if (ret)
goto out;
@@ -214,7 +220,7 @@ out:
return ret;
}
-static void process_req(void *work)
+static void process_req(struct work_struct *work)
{
struct addr_req *req, *temp_req;
struct sockaddr_in *src_in, *dst_in;
@@ -257,19 +263,19 @@ static int addr_resolve_local(struct sockaddr_in *src_in,
struct rdma_dev_addr *addr)
{
struct net_device *dev;
- u32 src_ip = src_in->sin_addr.s_addr;
+ __be32 src_ip = src_in->sin_addr.s_addr;
__be32 dst_ip = dst_in->sin_addr.s_addr;
int ret;
- dev = ip_dev_find(dst_ip);
+ dev = ip_dev_find(&init_net, dst_ip);
if (!dev)
return -EADDRNOTAVAIL;
- if (ZERONET(src_ip)) {
+ if (ipv4_is_zeronet(src_ip)) {
src_in->sin_family = dst_in->sin_family;
src_in->sin_addr.s_addr = dst_ip;
ret = rdma_copy_addr(addr, dev, dev->dev_addr);
- } else if (LOOPBACK(src_ip)) {
+ } else if (ipv4_is_loopback(src_ip)) {
ret = rdma_translate_ip((struct sockaddr *)dst_in, addr);
if (!ret)
memcpy(addr->dst_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
diff --git a/drivers/infiniband/core/agent.h b/drivers/infiniband/core/agent.h
index fb9ed14..6669287 100644
--- a/drivers/infiniband/core/agent.h
+++ b/drivers/infiniband/core/agent.h
@@ -32,8 +32,6 @@
* 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.
- *
- * $Id: agent.h 1389 2004-12-27 22:56:47Z roland $
*/
#ifndef __AGENT_H_
diff --git a/drivers/infiniband/core/cache.c b/drivers/infiniband/core/cache.c
index e85f701..6888356 100644
--- a/drivers/infiniband/core/cache.c
+++ b/drivers/infiniband/core/cache.c
@@ -31,8 +31,6 @@
* 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.
- *
- * $Id: cache.c 1349 2004-12-16 21:09:43Z roland $
*/
#include <linux/module.h>
diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c
index dc3bc17..73774b4 100644
--- a/drivers/infiniband/core/cm.c
+++ b/drivers/infiniband/core/cm.c
@@ -1,5 +1,5 @@
/*
- * Copyright (c) 2004-2006 Intel Corporation. All rights reserved.
+ * Copyright (c) 2004-2007 Intel Corporation. All rights reserved.
* Copyright (c) 2004 Topspin Corporation. All rights reserved.
* Copyright (c) 2004, 2005 Voltaire Corporation. All rights reserved.
* Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
@@ -31,19 +31,20 @@
* 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.
- *
- * $Id: cm.c 4311 2005-12-05 18:42:01Z sean.hefty $
*/
#include <linux/completion.h>
#include <linux/dma-mapping.h>
+#include <linux/device.h>
#include <linux/err.h>
#include <linux/idr.h>
#include <linux/interrupt.h>
#include <linux/random.h>
#include <linux/rbtree.h>
#include <linux/spinlock.h>
+#include <linux/sysfs.h>
#include <linux/workqueue.h>
+#include <linux/kdev_t.h>
#include <rdma/ib_cache.h>
#include <rdma/ib_cm.h>
@@ -89,17 +90,94 @@ static struct ib_cm {
struct workqueue_struct *wq;
} cm;
+/* Counter indexes ordered by attribute ID */
+enum {
+ CM_REQ_COUNTER,
+ CM_MRA_COUNTER,
+ CM_REJ_COUNTER,
+ CM_REP_COUNTER,
+ CM_RTU_COUNTER,
+ CM_DREQ_COUNTER,
+ CM_DREP_COUNTER,
+ CM_SIDR_REQ_COUNTER,
+ CM_SIDR_REP_COUNTER,
+ CM_LAP_COUNTER,
+ CM_APR_COUNTER,
+ CM_ATTR_COUNT,
+ CM_ATTR_ID_OFFSET = 0x0010,
+};
+
+enum {
+ CM_XMIT,
+ CM_XMIT_RETRIES,
+ CM_RECV,
+ CM_RECV_DUPLICATES,
+ CM_COUNTER_GROUPS
+};
+
+static char const counter_group_names[CM_COUNTER_GROUPS]
+ [sizeof("cm_rx_duplicates")] = {
+ "cm_tx_msgs", "cm_tx_retries",
+ "cm_rx_msgs", "cm_rx_duplicates"
+};
+
+struct cm_counter_group {
+ struct kobject obj;
+ atomic_long_t counter[CM_ATTR_COUNT];
+};
+
+struct cm_counter_attribute {
+ struct attribute attr;
+ int index;
+};
+
+#define CM_COUNTER_ATTR(_name, _index) \
+struct cm_counter_attribute cm_##_name##_counter_attr = { \
+ .attr = { .name = __stringify(_name), .mode = 0444, .owner = THIS_MODULE }, \
+ .index = _index \
+}
+
+static CM_COUNTER_ATTR(req, CM_REQ_COUNTER);
+static CM_COUNTER_ATTR(mra, CM_MRA_COUNTER);
+static CM_COUNTER_ATTR(rej, CM_REJ_COUNTER);
+static CM_COUNTER_ATTR(rep, CM_REP_COUNTER);
+static CM_COUNTER_ATTR(rtu, CM_RTU_COUNTER);
+static CM_COUNTER_ATTR(dreq, CM_DREQ_COUNTER);
+static CM_COUNTER_ATTR(drep, CM_DREP_COUNTER);
+static CM_COUNTER_ATTR(sidr_req, CM_SIDR_REQ_COUNTER);
+static CM_COUNTER_ATTR(sidr_rep, CM_SIDR_REP_COUNTER);
+static CM_COUNTER_ATTR(lap, CM_LAP_COUNTER);
+static CM_COUNTER_ATTR(apr, CM_APR_COUNTER);
+
+static struct attribute *cm_counter_default_attrs[] = {
+ &cm_req_counter_attr.attr,
+ &cm_mra_counter_attr.attr,
+ &cm_rej_counter_attr.attr,
+ &cm_rep_counter_attr.attr,
+ &cm_rtu_counter_attr.attr,
+ &cm_dreq_counter_attr.attr,
+ &cm_drep_counter_attr.attr,
+ &cm_sidr_req_counter_attr.attr,
+ &cm_sidr_rep_counter_attr.attr,
+ &cm_lap_counter_attr.attr,
+ &cm_apr_counter_attr.attr,
+ NULL
+};
+
struct cm_port {
struct cm_device *cm_dev;
struct ib_mad_agent *mad_agent;
+ struct kobject port_obj;
u8 port_num;
+ struct cm_counter_group counter_group[CM_COUNTER_GROUPS];
};
struct cm_device {
struct list_head list;
- struct ib_device *device;
+ struct ib_device *ib_device;
+ struct device *device;
u8 ack_delay;
- struct cm_port port[0];
+ struct cm_port *port[0];
};
struct cm_av {
@@ -273,7 +351,7 @@ static void cm_init_av_for_response(struct cm_port *port, struct ib_wc *wc,
{
av->port = port;
av->pkey_index = wc->pkey_index;
- ib_init_ah_from_wc(port->cm_dev->device, port->port_num, wc,
+ ib_init_ah_from_wc(port->cm_dev->ib_device, port->port_num, wc,
grh, &av->ah_attr);
}
@@ -287,9 +365,9 @@ static int cm_init_av_by_path(struct ib_sa_path_rec *path, struct cm_av *av)
read_lock_irqsave(&cm.device_lock, flags);
list_for_each_entry(cm_dev, &cm.device_list, list) {
- if (!ib_find_cached_gid(cm_dev->device, &path->sgid,
+ if (!ib_find_cached_gid(cm_dev->ib_device, &path->sgid,
&p, NULL)) {
- port = &cm_dev->port[p-1];
+ port = cm_dev->port[p-1];
break;
}
}
@@ -298,13 +376,13 @@ static int cm_init_av_by_path(struct ib_sa_path_rec *path, struct cm_av *av)
if (!port)
return -EINVAL;
- ret = ib_find_cached_pkey(cm_dev->device, port->port_num,
+ ret = ib_find_cached_pkey(cm_dev->ib_device, port->port_num,
be16_to_cpu(path->pkey), &av->pkey_index);
if (ret)
return ret;
av->port = port;
- ib_init_ah_from_path(cm_dev->device, port->port_num, path,
+ ib_init_ah_from_path(cm_dev->ib_device, port->port_num, path,
&av->ah_attr);
av->timeout = path->packet_life_time + 1;
return 0;
@@ -325,7 +403,7 @@ static int cm_alloc_id(struct cm_id_private *cm_id_priv)
spin_unlock_irqrestore(&cm.lock, flags);
} while( (ret == -EAGAIN) && idr_pre_get(&cm.local_id_table, GFP_KERNEL) );
- cm_id_priv->id.local_id = (__force __be32) (id ^ cm.random_id_operand);
+ cm_id_priv->id.local_id = (__force __be32)id ^ cm.random_id_operand;
return ret;
}
@@ -399,6 +477,31 @@ static int cm_compare_private_data(u8 *private_data,
return memcmp(src, dst_data->data, IB_CM_COMPARE_SIZE);
}
+/*
+ * Trivial helpers to strip endian annotation and compare; the
+ * endianness doesn't actually matter since we just need a stable
+ * order for the RB tree.
+ */
+static int be32_lt(__be32 a, __be32 b)
+{
+ return (__force u32) a < (__force u32) b;
+}
+
+static int be32_gt(__be32 a, __be32 b)
+{
+ return (__force u32) a > (__force u32) b;
+}
+
+static int be64_lt(__be64 a, __be64 b)
+{
+ return (__force u64) a < (__force u64) b;
+}
+
+static int be64_gt(__be64 a, __be64 b)
+{
+ return (__force u64) a > (__force u64) b;
+}
+
static struct cm_id_private * cm_insert_listen(struct cm_id_private *cm_id_priv)
{
struct rb_node **link = &cm.listen_service_table.rb_node;
@@ -424,9 +527,9 @@ static struct cm_id_private * cm_insert_listen(struct cm_id_private *cm_id_priv)
link = &(*link)->rb_left;
else if (cm_id_priv->id.device > cur_cm_id_priv->id.device)
link = &(*link)->rb_right;
- else if (service_id < cur_cm_id_priv->id.service_id)
+ else if (be64_lt(service_id, cur_cm_id_priv->id.service_id))
link = &(*link)->rb_left;
- else if (service_id > cur_cm_id_priv->id.service_id)
+ else if (be64_gt(service_id, cur_cm_id_priv->id.service_id))
link = &(*link)->rb_right;
else if (data_cmp < 0)
link = &(*link)->rb_left;
@@ -459,9 +562,9 @@ static struct cm_id_private * cm_find_listen(struct ib_device *device,
node = node->rb_left;
else if (device > cm_id_priv->id.device)
node = node->rb_right;
- else if (service_id < cm_id_priv->id.service_id)
+ else if (be64_lt(service_id, cm_id_priv->id.service_id))
node = node->rb_left;
- else if (service_id > cm_id_priv->id.service_id)
+ else if (be64_gt(service_id, cm_id_priv->id.service_id))
node = node->rb_right;
else if (data_cmp < 0)
node = node->rb_left;
@@ -484,13 +587,13 @@ static struct cm_timewait_info * cm_insert_remote_id(struct cm_timewait_info
parent = *link;
cur_timewait_info = rb_entry(parent, struct cm_timewait_info,
remote_id_node);
- if (remote_id < cur_timewait_info->work.remote_id)
+ if (be32_lt(remote_id, cur_timewait_info->work.remote_id))
link = &(*link)->rb_left;
- else if (remote_id > cur_timewait_info->work.remote_id)
+ else if (be32_gt(remote_id, cur_timewait_info->work.remote_id))
link = &(*link)->rb_right;
- else if (remote_ca_guid < cur_timewait_info->remote_ca_guid)
+ else if (be64_lt(remote_ca_guid, cur_timewait_info->remote_ca_guid))
link = &(*link)->rb_left;
- else if (remote_ca_guid > cur_timewait_info->remote_ca_guid)
+ else if (be64_gt(remote_ca_guid, cur_timewait_info->remote_ca_guid))
link = &(*link)->rb_right;
else
return cur_timewait_info;
@@ -510,13 +613,13 @@ static struct cm_timewait_info * cm_find_remote_id(__be64 remote_ca_guid,
while (node) {
timewait_info = rb_entry(node, struct cm_timewait_info,
remote_id_node);
- if (remote_id < timewait_info->work.remote_id)
+ if (be32_lt(remote_id, timewait_info->work.remote_id))
node = node->rb_left;
- else if (remote_id > timewait_info->work.remote_id)
+ else if (be32_gt(remote_id, timewait_info->work.remote_id))
node = node->rb_right;
- else if (remote_ca_guid < timewait_info->remote_ca_guid)
+ else if (be64_lt(remote_ca_guid, timewait_info->remote_ca_guid))
node = node->rb_left;
- else if (remote_ca_guid > timewait_info->remote_ca_guid)
+ else if (be64_gt(remote_ca_guid, timewait_info->remote_ca_guid))
node = node->rb_right;
else
return timewait_info;
@@ -537,13 +640,13 @@ static struct cm_timewait_info * cm_insert_remote_qpn(struct cm_timewait_info
parent = *link;
cur_timewait_info = rb_entry(parent, struct cm_timewait_info,
remote_qp_node);
- if (remote_qpn < cur_timewait_info->remote_qpn)
+ if (be32_lt(remote_qpn, cur_timewait_info->remote_qpn))
link = &(*link)->rb_left;
- else if (remote_qpn > cur_timewait_info->remote_qpn)
+ else if (be32_gt(remote_qpn, cur_timewait_info->remote_qpn))
link = &(*link)->rb_right;
- else if (remote_ca_guid < cur_timewait_info->remote_ca_guid)
+ else if (be64_lt(remote_ca_guid, cur_timewait_info->remote_ca_guid))
link = &(*link)->rb_left;
- else if (remote_ca_guid > cur_timewait_info->remote_ca_guid)
+ else if (be64_gt(remote_ca_guid, cur_timewait_info->remote_ca_guid))
link = &(*link)->rb_right;
else
return cur_timewait_info;
@@ -567,9 +670,9 @@ static struct cm_id_private * cm_insert_remote_sidr(struct cm_id_private
parent = *link;
cur_cm_id_priv = rb_entry(parent, struct cm_id_private,
sidr_id_node);
- if (remote_id < cur_cm_id_priv->id.remote_id)
+ if (be32_lt(remote_id, cur_cm_id_priv->id.remote_id))
link = &(*link)->rb_left;
- else if (remote_id > cur_cm_id_priv->id.remote_id)
+ else if (be32_gt(remote_id, cur_cm_id_priv->id.remote_id))
link = &(*link)->rb_right;
else {
int cmp;
@@ -906,6 +1009,9 @@ static void cm_format_req(struct cm_req_msg *req_msg,
struct cm_id_private *cm_id_priv,
struct ib_cm_req_param *param)
{
+ struct ib_sa_path_rec *pri_path = param->primary_path;
+ struct ib_sa_path_rec *alt_path = param->alternate_path;
+
cm_format_mad_hdr(&req_msg->hdr, CM_REQ_ATTR_ID,
cm_form_tid(cm_id_priv, CM_MSG_SEQUENCE_REQ));
@@ -941,35 +1047,46 @@ static void cm_format_req(struct cm_req_msg *req_msg,
cm_req_set_max_cm_retries(req_msg, param->max_cm_retries);
cm_req_set_srq(req_msg, param->srq);
- req_msg->primary_local_lid = param->primary_path->slid;
- req_msg->primary_remote_lid = param->primary_path->dlid;
- req_msg->primary_local_gid = param->primary_path->sgid;
- req_msg->primary_remote_gid = param->primary_path->dgid;
- cm_req_set_primary_flow_label(req_msg, param->primary_path->flow_label);
- cm_req_set_primary_packet_rate(req_msg, param->primary_path->rate);
- req_msg->primary_traffic_class = param->primary_path->traffic_class;
- req_msg->primary_hop_limit = param->primary_path->hop_limit;
- cm_req_set_primary_sl(req_msg, param->primary_path->sl);
- cm_req_set_primary_subnet_local(req_msg, 1); /* local only... */
+ if (pri_path->hop_limit <= 1) {
+ req_msg->primary_local_lid = pri_path->slid;
+ req_msg->primary_remote_lid = pri_path->dlid;
+ } else {
+ /* Work-around until there's a way to obtain remote LID info */
+ req_msg->primary_local_lid = IB_LID_PERMISSIVE;
+ req_msg->primary_remote_lid = IB_LID_PERMISSIVE;
+ }
+ req_msg->primary_local_gid = pri_path->sgid;
+ req_msg->primary_remote_gid = pri_path->dgid;
+ cm_req_set_primary_flow_label(req_msg, pri_path->flow_label);
+ cm_req_set_primary_packet_rate(req_msg, pri_path->rate);
+ req_msg->primary_traffic_class = pri_path->traffic_class;
+ req_msg->primary_hop_limit = pri_path->hop_limit;
+ cm_req_set_primary_sl(req_msg, pri_path->sl);
+ cm_req_set_primary_subnet_local(req_msg, (pri_path->hop_limit <= 1));
cm_req_set_primary_local_ack_timeout(req_msg,
cm_ack_timeout(cm_id_priv->av.port->cm_dev->ack_delay,
- param->primary_path->packet_life_time));
+ pri_path->packet_life_time));
- if (param->alternate_path) {
- req_msg->alt_local_lid = param->alternate_path->slid;
- req_msg->alt_remote_lid = param->alternate_path->dlid;
- req_msg->alt_local_gid = param->alternate_path->sgid;
- req_msg->alt_remote_gid = param->alternate_path->dgid;
+ if (alt_path) {
+ if (alt_path->hop_limit <= 1) {
+ req_msg->alt_local_lid = alt_path->slid;
+ req_msg->alt_remote_lid = alt_path->dlid;
+ } else {
+ req_msg->alt_local_lid = IB_LID_PERMISSIVE;
+ req_msg->alt_remote_lid = IB_LID_PERMISSIVE;
+ }
+ req_msg->alt_local_gid = alt_path->sgid;
+ req_msg->alt_remote_gid = alt_path->dgid;
cm_req_set_alt_flow_label(req_msg,
- param->alternate_path->flow_label);
- cm_req_set_alt_packet_rate(req_msg, param->alternate_path->rate);
- req_msg->alt_traffic_class = param->alternate_path->traffic_class;
- req_msg->alt_hop_limit = param->alternate_path->hop_limit;
- cm_req_set_alt_sl(req_msg, param->alternate_path->sl);
- cm_req_set_alt_subnet_local(req_msg, 1); /* local only... */
+ alt_path->flow_label);
+ cm_req_set_alt_packet_rate(req_msg, alt_path->rate);
+ req_msg->alt_traffic_class = alt_path->traffic_class;
+ req_msg->alt_hop_limit = alt_path->hop_limit;
+ cm_req_set_alt_sl(req_msg, alt_path->sl);
+ cm_req_set_alt_subnet_local(req_msg, (alt_path->hop_limit <= 1));
cm_req_set_alt_local_ack_timeout(req_msg,
cm_ack_timeout(cm_id_priv->av.port->cm_dev->ack_delay,
- param->alternate_path->packet_life_time));
+ alt_path->packet_life_time));
}
if (param->private_data && param->private_data_len)
@@ -1298,6 +1415,9 @@ static void cm_dup_req_handler(struct cm_work *work,
struct ib_mad_send_buf *msg = NULL;
int ret;
+ atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES].
+ counter[CM_REQ_COUNTER]);
+
/* Quick state check to discard duplicate REQs. */
if (cm_id_priv->id.state == IB_CM_REQ_RCVD)
return;
@@ -1387,6 +1507,34 @@ out:
return listen_cm_id_priv;
}
+/*
+ * Work-around for inter-subnet connections. If the LIDs are permissive,
+ * we need to override the LID/SL data in the REQ with the LID information
+ * in the work completion.
+ */
+static void cm_process_routed_req(struct cm_req_msg *req_msg, struct ib_wc *wc)
+{
+ if (!cm_req_get_primary_subnet_local(req_msg)) {
+ if (req_msg->primary_local_lid == IB_LID_PERMISSIVE) {
+ req_msg->primary_local_lid = cpu_to_be16(wc->slid);
+ cm_req_set_primary_sl(req_msg, wc->sl);
+ }
+
+ if (req_msg->primary_remote_lid == IB_LID_PERMISSIVE)
+ req_msg->primary_remote_lid = cpu_to_be16(wc->dlid_path_bits);
+ }
+
+ if (!cm_req_get_alt_subnet_local(req_msg)) {
+ if (req_msg->alt_local_lid == IB_LID_PERMISSIVE) {
+ req_msg->alt_local_lid = cpu_to_be16(wc->slid);
+ cm_req_set_alt_sl(req_msg, wc->sl);
+ }
+
+ if (req_msg->alt_remote_lid == IB_LID_PERMISSIVE)
+ req_msg->alt_remote_lid = cpu_to_be16(wc->dlid_path_bits);
+ }
+}
+
static int cm_req_handler(struct cm_work *work)
{
struct ib_cm_id *cm_id;
@@ -1396,7 +1544,7 @@ static int cm_req_handler(struct cm_work *work)
req_msg = (struct cm_req_msg *)work->mad_recv_wc->recv_buf.mad;
- cm_id = ib_create_cm_id(work->port->cm_dev->device, NULL, NULL);
+ cm_id = ib_create_cm_id(work->port->cm_dev->ib_device, NULL, NULL);
if (IS_ERR(cm_id))
return PTR_ERR(cm_id);
@@ -1427,10 +1575,11 @@ static int cm_req_handler(struct cm_work *work)
cm_id_priv->id.service_id = req_msg->service_id;
cm_id_priv->id.service_mask = __constant_cpu_to_be64(~0ULL);
+ cm_process_routed_req(req_msg, work->mad_recv_wc->wc);
cm_format_paths_from_req(req_msg, &work->path[0], &work->path[1]);
ret = cm_init_av_by_path(&work->path[0], &cm_id_priv->av);
if (ret) {
- ib_get_cached_gid(work->port->cm_dev->device,
+ ib_get_cached_gid(work->port->cm_dev->ib_device,
work->port->port_num, 0, &work->path[0].sgid);
ib_send_cm_rej(cm_id, IB_CM_REJ_INVALID_GID,
&work->path[0].sgid, sizeof work->path[0].sgid,
@@ -1651,6 +1800,8 @@ static void cm_dup_rep_handler(struct cm_work *work)
if (!cm_id_priv)
return;
+ atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES].
+ counter[CM_REP_COUNTER]);
ret = cm_alloc_response_msg(work->port, work->mad_recv_wc, &msg);
if (ret)
goto deref;
@@ -1816,6 +1967,8 @@ static int cm_rtu_handler(struct cm_work *work)
if (cm_id_priv->id.state != IB_CM_REP_SENT &&
cm_id_priv->id.state != IB_CM_MRA_REP_RCVD) {
spin_unlock_irq(&cm_id_priv->lock);
+ atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES].
+ counter[CM_RTU_COUNTER]);
goto out;
}
cm_id_priv->id.state = IB_CM_ESTABLISHED;
@@ -1993,6 +2146,8 @@ static int cm_dreq_handler(struct cm_work *work)
cm_id_priv = cm_acquire_id(dreq_msg->remote_comm_id,
dreq_msg->local_comm_id);
if (!cm_id_priv) {
+ atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES].
+ counter[CM_DREQ_COUNTER]);
cm_issue_drep(work->port, work->mad_recv_wc);
return -EINVAL;
}
@@ -2012,6 +2167,8 @@ static int cm_dreq_handler(struct cm_work *work)
case IB_CM_MRA_REP_RCVD:
break;
case IB_CM_TIMEWAIT:
+ atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES].
+ counter[CM_DREQ_COUNTER]);
if (cm_alloc_response_msg(work->port, work->mad_recv_wc, &msg))
goto unlock;
@@ -2023,6 +2180,10 @@ static int cm_dreq_handler(struct cm_work *work)
if (ib_post_send_mad(msg, NULL))
cm_free_msg(msg);
goto deref;
+ case IB_CM_DREQ_RCVD:
+ atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES].
+ counter[CM_DREQ_COUNTER]);
+ goto unlock;
default:
goto unlock;
}
@@ -2380,10 +2541,20 @@ static int cm_mra_handler(struct cm_work *work)
if (cm_mra_get_msg_mraed(mra_msg) != CM_MSG_RESPONSE_OTHER ||
cm_id_priv->id.lap_state != IB_CM_LAP_SENT ||
ib_modify_mad(cm_id_priv->av.port->mad_agent,
- cm_id_priv->msg, timeout))
+ cm_id_priv->msg, timeout)) {
+ if (cm_id_priv->id.lap_state == IB_CM_MRA_LAP_RCVD)
+ atomic_long_inc(&work->port->
+ counter_group[CM_RECV_DUPLICATES].
+ counter[CM_MRA_COUNTER]);
goto out;
+ }
cm_id_priv->id.lap_state = IB_CM_MRA_LAP_RCVD;
break;
+ case IB_CM_MRA_REQ_RCVD:
+ case IB_CM_MRA_REP_RCVD:
+ atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES].
+ counter[CM_MRA_COUNTER]);
+ /* fall through */
default:
goto out;
}
@@ -2543,6 +2714,8 @@ static int cm_lap_handler(struct cm_work *work)
case IB_CM_LAP_IDLE:
break;
case IB_CM_MRA_LAP_SENT:
+ atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES].
+ counter[CM_LAP_COUNTER]);
if (cm_alloc_response_msg(work->port, work->mad_recv_wc, &msg))
goto unlock;
@@ -2556,6 +2729,10 @@ static int cm_lap_handler(struct cm_work *work)
if (ib_post_send_mad(msg, NULL))
cm_free_msg(msg);
goto deref;
+ case IB_CM_LAP_RCVD:
+ atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES].
+ counter[CM_LAP_COUNTER]);
+ goto unlock;
default:
goto unlock;
}
@@ -2736,7 +2913,7 @@ static void cm_format_sidr_req(struct cm_sidr_req_msg *sidr_req_msg,
cm_format_mad_hdr(&sidr_req_msg->hdr, CM_SIDR_REQ_ATTR_ID,
cm_form_tid(cm_id_priv, CM_MSG_SEQUENCE_SIDR));
sidr_req_msg->request_id = cm_id_priv->id.local_id;
- sidr_req_msg->pkey = cpu_to_be16(param->path->pkey);
+ sidr_req_msg->pkey = param->path->pkey;
sidr_req_msg->service_id = param->service_id;
if (param->private_data && param->private_data_len)
@@ -2821,7 +2998,7 @@ static int cm_sidr_req_handler(struct cm_work *work)
struct cm_sidr_req_msg *sidr_req_msg;
struct ib_wc *wc;
- cm_id = ib_create_cm_id(work->port->cm_dev->device, NULL, NULL);
+ cm_id = ib_create_cm_id(work->port->cm_dev->ib_device, NULL, NULL);
if (IS_ERR(cm_id))
return PTR_ERR(cm_id);
cm_id_priv = container_of(cm_id, struct cm_id_private, id);
@@ -2843,6 +3020,8 @@ static int cm_sidr_req_handler(struct cm_work *work)
cur_cm_id_priv = cm_insert_remote_sidr(cm_id_priv);
if (cur_cm_id_priv) {
spin_unlock_irq(&cm.lock);
+ atomic_long_inc(&work->port->counter_group[CM_RECV_DUPLICATES].
+ counter[CM_SIDR_REQ_COUNTER]);
goto out; /* Duplicate message. */
}
cm_id_priv->id.state = IB_CM_SIDR_REQ_RCVD;
@@ -3037,6 +3216,27 @@ static void cm_send_handler(struct ib_mad_agent *mad_agent,
struct ib_mad_send_wc *mad_send_wc)
{
struct ib_mad_send_buf *msg = mad_send_wc->send_buf;
+ struct cm_port *port;
+ u16 attr_index;
+
+ port = mad_agent->context;
+ attr_index = be16_to_cpu(((struct ib_mad_hdr *)
+ msg->mad)->attr_id) - CM_ATTR_ID_OFFSET;
+
+ /*
+ * If the send was in response to a received message (context[0] is not
+ * set to a cm_id), and is not a REJ, then it is a send that was
+ * manually retried.
+ */
+ if (!msg->context[0] && (attr_index != CM_REJ_COUNTER))
+ msg->retries = 1;
+
+ atomic_long_add(1 + msg->retries,
+ &port->counter_group[CM_XMIT].counter[attr_index]);
+ if (msg->retries)
+ atomic_long_add(msg->retries,
+ &port->counter_group[CM_XMIT_RETRIES].
+ counter[attr_index]);
switch (mad_send_wc->status) {
case IB_WC_SUCCESS:
@@ -3195,8 +3395,10 @@ EXPORT_SYMBOL(ib_cm_notify);
static void cm_recv_handler(struct ib_mad_agent *mad_agent,
struct ib_mad_recv_wc *mad_recv_wc)
{
+ struct cm_port *port = mad_agent->context;
struct cm_work *work;
enum ib_cm_event_type event;
+ u16 attr_id;
int paths = 0;
switch (mad_recv_wc->recv_buf.mad->mad_hdr.attr_id) {
@@ -3241,6 +3443,10 @@ static void cm_recv_handler(struct ib_mad_agent *mad_agent,
return;
}
+ attr_id = be16_to_cpu(mad_recv_wc->recv_buf.mad->mad_hdr.attr_id);
+ atomic_long_inc(&port->counter_group[CM_RECV].
+ counter[attr_id - CM_ATTR_ID_OFFSET]);
+
work = kmalloc(sizeof *work + sizeof(struct ib_sa_path_rec) * paths,
GFP_KERNEL);
if (!work) {
@@ -3251,7 +3457,7 @@ static void cm_recv_handler(struct ib_mad_agent *mad_agent,
INIT_DELAYED_WORK(&work->work, cm_work_handler);
work->cm_event.event = event;
work->mad_recv_wc = mad_recv_wc;
- work->port = (struct cm_port *)mad_agent->context;
+ work->port = port;
queue_delayed_work(cm.wq, &work->work, 0);
}
@@ -3420,13 +3626,93 @@ static void cm_get_ack_delay(struct cm_device *cm_dev)
{
struct ib_device_attr attr;
- if (ib_query_device(cm_dev->device, &attr))
+ if (ib_query_device(cm_dev->ib_device, &attr))
cm_dev->ack_delay = 0; /* acks will rely on packet life time */
else
cm_dev->ack_delay = attr.local_ca_ack_delay;
}
-static void cm_add_one(struct ib_device *device)
+static ssize_t cm_show_counter(struct kobject *obj, struct attribute *attr,
+ char *buf)
+{
+ struct cm_counter_group *group;
+ struct cm_counter_attribute *cm_attr;
+
+ group = container_of(obj, struct cm_counter_group, obj);
+ cm_attr = container_of(attr, struct cm_counter_attribute, attr);
+
+ return sprintf(buf, "%ld\n",
+ atomic_long_read(&group->counter[cm_attr->index]));
+}
+
+static struct sysfs_ops cm_counter_ops = {
+ .show = cm_show_counter
+};
+
+static struct kobj_type cm_counter_obj_type = {
+ .sysfs_ops = &cm_counter_ops,
+ .default_attrs = cm_counter_default_attrs
+};
+
+static void cm_release_port_obj(struct kobject *obj)
+{
+ struct cm_port *cm_port;
+
+ cm_port = container_of(obj, struct cm_port, port_obj);
+ kfree(cm_port);
+}
+
+static struct kobj_type cm_port_obj_type = {
+ .release = cm_release_port_obj
+};
+
+struct class cm_class = {
+ .name = "infiniband_cm",
+};
+EXPORT_SYMBOL(cm_class);
+
+static int cm_create_port_fs(struct cm_port *port)
+{
+ int i, ret;
+
+ ret = kobject_init_and_add(&port->port_obj, &cm_port_obj_type,
+ &port->cm_dev->device->kobj,
+ "%d", port->port_num);
+ if (ret) {
+ kfree(port);
+ return ret;
+ }
+
+ for (i = 0; i < CM_COUNTER_GROUPS; i++) {
+ ret = kobject_init_and_add(&port->counter_group[i].obj,
+ &cm_counter_obj_type,
+ &port->port_obj,
+ "%s", counter_group_names[i]);
+ if (ret)
+ goto error;
+ }
+
+ return 0;
+
+error:
+ while (i--)
+ kobject_unregister(&port->counter_group[i].obj);
+ kobject_unregister(&port->port_obj);
+ return ret;
+
+}
+
+static void cm_remove_port_fs(struct cm_port *port)
+{
+ int i;
+
+ for (i = 0; i < CM_COUNTER_GROUPS; i++)
+ kobject_unregister(&port->counter_group[i].obj);
+
+ kobject_unregister(&port->port_obj);
+}
+
+static void cm_add_one(struct ib_device *ib_device)
{
struct cm_device *cm_dev;
struct cm_port *port;
@@ -3441,23 +3727,40 @@ static void cm_add_one(struct ib_device *device)
int ret;
u8 i;
- if (rdma_node_get_transport(device->node_type) != RDMA_TRANSPORT_IB)
+ if (rdma_node_get_transport(ib_device->node_type) != RDMA_TRANSPORT_IB)
return;
- cm_dev = kmalloc(sizeof(*cm_dev) + sizeof(*port) *
- device->phys_port_cnt, GFP_KERNEL);
+ cm_dev = kzalloc(sizeof(*cm_dev) + sizeof(*port) *
+ ib_device->phys_port_cnt, GFP_KERNEL);
if (!cm_dev)
return;
- cm_dev->device = device;
+ cm_dev->ib_device = ib_device;
cm_get_ack_delay(cm_dev);
+ cm_dev->device = device_create(&cm_class, ib_device->class_dev.dev,
+ MKDEV(0, 0),
+ "%s", ib_device->name);
+ if (!cm_dev->device) {
+ kfree(cm_dev);
+ return;
+ }
+
set_bit(IB_MGMT_METHOD_SEND, reg_req.method_mask);
- for (i = 1; i <= device->phys_port_cnt; i++) {
- port = &cm_dev->port[i-1];
+ for (i = 1; i <= ib_device->phys_port_cnt; i++) {
+ port = kzalloc(sizeof *port, GFP_KERNEL);
+ if (!port)
+ goto error1;
+
+ cm_dev->port[i-1] = port;
port->cm_dev = cm_dev;
port->port_num = i;
- port->mad_agent = ib_register_mad_agent(device, i,
+
+ ret = cm_create_port_fs(port);
+ if (ret)
+ goto error1;
+
+ port->mad_agent = ib_register_mad_agent(ib_device, i,
IB_QPT_GSI,
®_req,
0,
@@ -3465,33 +3768,37 @@ static void cm_add_one(struct ib_device *device)
cm_recv_handler,
port);
if (IS_ERR(port->mad_agent))
- goto error1;
+ goto error2;
- ret = ib_modify_port(device, i, 0, &port_modify);
+ ret = ib_modify_port(ib_device, i, 0, &port_modify);
if (ret)
- goto error2;
+ goto error3;
}
- ib_set_client_data(device, &cm_client, cm_dev);
+ ib_set_client_data(ib_device, &cm_client, cm_dev);
write_lock_irqsave(&cm.device_lock, flags);
list_add_tail(&cm_dev->list, &cm.device_list);
write_unlock_irqrestore(&cm.device_lock, flags);
return;
-error2:
+error3:
ib_unregister_mad_agent(port->mad_agent);
+error2:
+ cm_remove_port_fs(port);
error1:
port_modify.set_port_cap_mask = 0;
port_modify.clr_port_cap_mask = IB_PORT_CM_SUP;
while (--i) {
- port = &cm_dev->port[i-1];
- ib_modify_port(device, port->port_num, 0, &port_modify);
+ port = cm_dev->port[i-1];
+ ib_modify_port(ib_device, port->port_num, 0, &port_modify);
ib_unregister_mad_agent(port->mad_agent);
+ cm_remove_port_fs(port);
}
+ device_unregister(cm_dev->device);
kfree(cm_dev);
}
-static void cm_remove_one(struct ib_device *device)
+static void cm_remove_one(struct ib_device *ib_device)
{
struct cm_device *cm_dev;
struct cm_port *port;
@@ -3501,7 +3808,7 @@ static void cm_remove_one(struct ib_device *device)
unsigned long flags;
int i;
- cm_dev = ib_get_client_data(device, &cm_client);
+ cm_dev = ib_get_client_data(ib_device, &cm_client);
if (!cm_dev)
return;
@@ -3509,12 +3816,14 @@ static void cm_remove_one(struct ib_device *device)
list_del(&cm_dev->list);
write_unlock_irqrestore(&cm.device_lock, flags);
- for (i = 1; i <= device->phys_port_cnt; i++) {
- port = &cm_dev->port[i-1];
- ib_modify_port(device, port->port_num, 0, &port_modify);
+ for (i = 1; i <= ib_device->phys_port_cnt; i++) {
+ port = cm_dev->port[i-1];
+ ib_modify_port(ib_device, port->port_num, 0, &port_modify);
ib_unregister_mad_agent(port->mad_agent);
flush_workqueue(cm.wq);
+ cm_remove_port_fs(port);
}
+ device_unregister(cm_dev->device);
kfree(cm_dev);
}
@@ -3536,17 +3845,25 @@ static int __init ib_cm_init(void)
idr_pre_get(&cm.local_id_table, GFP_KERNEL);
INIT_LIST_HEAD(&cm.timewait_list);
- cm.wq = create_workqueue("ib_cm");
- if (!cm.wq)
+ ret = class_register(&cm_class);
+ if (ret)
return -ENOMEM;
+ cm.wq = create_workqueue("ib_cm");
+ if (!cm.wq) {
+ ret = -ENOMEM;
+ goto error1;
+ }
+
ret = ib_register_client(&cm_client);
if (ret)
- goto error;
+ goto error2;
return 0;
-error:
+error2:
destroy_workqueue(cm.wq);
+error1:
+ class_unregister(&cm_class);
return ret;
}
@@ -3567,6 +3884,7 @@ static void __exit ib_cm_cleanup(void)
kfree(timewait_info);
}
+ class_unregister(&cm_class);
idr_destroy(&cm.local_id_table);
}
diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c
index 19f75d2..faef87b 100644
--- a/drivers/infiniband/core/cma.c
+++ b/drivers/infiniband/core/cma.c
@@ -4,29 +4,33 @@
* Copyright (c) 1999-2005, Mellanox Technologies, Inc. All rights reserved.
* Copyright (c) 2005-2006 Intel Corporation. All rights reserved.
*
- * This Software is licensed under one of the following licenses:
+ * 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:
*
- * 1) under the terms of the "Common Public License 1.0" a copy of which is
- * available from the Open Source Initiative, see
- * http://www.opensource.org/licenses/cpl.php.
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
*
- * 2) under the terms of the "The BSD License" a copy of which is
- * available from the Open Source Initiative, see
- * http://www.opensource.org/licenses/bsd-license.php.
+ * - Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer.
*
- * 3) under the terms of the "GNU General Public License (GPL) Version 2" a
- * copy of which is available from the Open Source Initiative, see
- * http://www.opensource.org/licenses/gpl-license.php.
- *
- * Licensee has the right to choose one of the above licenses.
- *
- * Redistributions of source code must retain the above copyright
- * notice and one of the license notices.
- *
- * Redistributions in binary form must reproduce both the above copyright
- * notice, one of the license notices in the documentation
- * and/or other materials provided with the distribution.
+ * - 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/completion.h>
@@ -54,6 +58,11 @@ static int tavor_quirk = 0;
module_param_named(tavor_quirk, tavor_quirk, int, 0644);
MODULE_PARM_DESC(tavor_quirk, "Tavor performance quirk: limit MTU to 1K if > 0");
+int unify_tcp_port_space = 0;
+module_param(unify_tcp_port_space, int, 0644);
+MODULE_PARM_DESC(unify_tcp_port_space, "Unify the host TCP and RDMA port "
+ "space allocation (default=0)");
+
#define CMA_CM_RESPONSE_TIMEOUT 20
#define CMA_MAX_CM_RETRIES 15
#define CMA_CM_MRA_SETTING (IB_CM_MRA_FLAG_DELAY | 24)
@@ -121,6 +130,7 @@ struct rdma_id_private {
struct rdma_cm_id id;
struct rdma_bind_list *bind_list;
+ struct socket *sock;
struct hlist_node node;
struct list_head list; /* listen_any_list or cma_device.list */
struct list_head listen_list; /* per device listens */
@@ -134,8 +144,7 @@ struct rdma_id_private {
struct completion comp;
atomic_t refcount;
- wait_queue_head_t wait_remove;
- atomic_t dev_remove;
+ struct mutex handler_mutex;
int backlog;
int timeout_ms;
@@ -160,9 +169,7 @@ struct cma_multicast {
} multicast;
struct list_head list;
void *context;
- struct sockaddr addr;
- u8 pad[sizeof(struct sockaddr_in6) -
- sizeof(struct sockaddr)];
+ struct sockaddr_storage addr;
};
struct cma_work {
@@ -173,18 +180,24 @@ struct cma_work {
struct rdma_cm_event event;
};
+struct cma_ndev_work {
+ struct work_struct work;
+ struct rdma_id_private *id;
+ struct rdma_cm_event event;
+};
+
union cma_ip_addr {
struct in6_addr ip6;
struct {
- __u32 pad[3];
- __u32 addr;
+ __be32 pad[3];
+ __be32 addr;
} ip4;
};
struct cma_hdr {
u8 cma_version;
u8 ip_version; /* IP version: 7:4 */
- __u16 port;
+ __be16 port;
union cma_ip_addr src_addr;
union cma_ip_addr dst_addr;
};
@@ -194,8 +207,8 @@ struct sdp_hh {
u8 sdp_version; /* Major version: 7:4 */
u8 ip_version; /* IP version: 7:4 */
u8 sdp_specific1[10];
- __u16 port;
- __u16 sdp_specific2;
+ __be16 port;
+ __be16 sdp_specific2;
union cma_ip_addr src_addr;
union cma_ip_addr dst_addr;
};
@@ -297,21 +310,25 @@ static void cma_detach_from_dev(struct rdma_id_private *id_priv)
id_priv->cma_dev = NULL;
}
-static int cma_set_qkey(struct ib_device *device, u8 port_num,
- enum rdma_port_space ps,
- struct rdma_dev_addr *dev_addr, u32 *qkey)
+static int cma_set_qkey(struct rdma_id_private *id_priv)
{
struct ib_sa_mcmember_rec rec;
int ret = 0;
- switch (ps) {
+ if (id_priv->qkey)
+ return 0;
+
+ switch (id_priv->id.ps) {
case RDMA_PS_UDP:
- *qkey = RDMA_UDP_QKEY;
+ id_priv->qkey = RDMA_UDP_QKEY;
break;
case RDMA_PS_IPOIB:
- ib_addr_get_mgid(dev_addr, &rec.mgid);
- ret = ib_sa_get_mcmember_rec(device, port_num, &rec.mgid, &rec);
- *qkey = be32_to_cpu(rec.qkey);
+ ib_addr_get_mgid(&id_priv->id.route.addr.dev_addr, &rec.mgid);
+ ret = ib_sa_get_mcmember_rec(id_priv->id.device,
+ id_priv->id.port_num, &rec.mgid,
+ &rec);
+ if (!ret)
+ id_priv->qkey = be32_to_cpu(rec.qkey);
break;
default:
break;
@@ -341,12 +358,7 @@ static int cma_acquire_dev(struct rdma_id_private *id_priv)
ret = ib_find_cached_gid(cma_dev->device, &gid,
&id_priv->id.port_num, NULL);
if (!ret) {
- ret = cma_set_qkey(cma_dev->device,
- id_priv->id.port_num,
- id_priv->id.ps, dev_addr,
- &id_priv->qkey);
- if (!ret)
- cma_attach_to_dev(id_priv, cma_dev);
+ cma_attach_to_dev(id_priv, cma_dev);
break;
}
}
@@ -359,26 +371,15 @@ static void cma_deref_id(struct rdma_id_private *id_priv)
complete(&id_priv->comp);
}
-static int cma_disable_remove(struct rdma_id_private *id_priv,
+static int cma_disable_callback(struct rdma_id_private *id_priv,
enum cma_state state)
{
- unsigned long flags;
- int ret;
-
- spin_lock_irqsave(&id_priv->lock, flags);
- if (id_priv->state == state) {
- atomic_inc(&id_priv->dev_remove);
- ret = 0;
- } else
- ret = -EINVAL;
- spin_unlock_irqrestore(&id_priv->lock, flags);
- return ret;
-}
-
-static void cma_enable_remove(struct rdma_id_private *id_priv)
-{
- if (atomic_dec_and_test(&id_priv->dev_remove))
- wake_up(&id_priv->wait_remove);
+ mutex_lock(&id_priv->handler_mutex);
+ if (id_priv->state != state) {
+ mutex_unlock(&id_priv->handler_mutex);
+ return -EINVAL;
+ }
+ return 0;
}
static int cma_has_cm_dev(struct rdma_id_private *id_priv)
@@ -403,8 +404,7 @@ struct rdma_cm_id *rdma_create_id(rdma_cm_event_handler event_handler,
mutex_init(&id_priv->qp_mutex);
init_completion(&id_priv->comp);
atomic_set(&id_priv->refcount, 1);
- init_waitqueue_head(&id_priv->wait_remove);
- atomic_set(&id_priv->dev_remove, 0);
+ mutex_init(&id_priv->handler_mutex);
INIT_LIST_HEAD(&id_priv->listen_list);
INIT_LIST_HEAD(&id_priv->mc_list);
get_random_bytes(&id_priv->seq_num, sizeof id_priv->seq_num);
@@ -496,7 +496,8 @@ void rdma_destroy_qp(struct rdma_cm_id *id)
}
EXPORT_SYMBOL(rdma_destroy_qp);
-static int cma_modify_qp_rtr(struct rdma_id_private *id_priv)
+static int cma_modify_qp_rtr(struct rdma_id_private *id_priv,
+ struct rdma_conn_param *conn_param)
{
struct ib_qp_attr qp_attr;
int qp_attr_mask, ret;
@@ -522,13 +523,16 @@ static int cma_modify_qp_rtr(struct rdma_id_private *id_priv)
if (ret)
goto out;
+ if (conn_param)
+ qp_attr.max_dest_rd_atomic = conn_param->responder_resources;
ret = ib_modify_qp(id_priv->id.qp, &qp_attr, qp_attr_mask);
out:
mutex_unlock(&id_priv->qp_mutex);
return ret;
}
-static int cma_modify_qp_rts(struct rdma_id_private *id_priv)
+static int cma_modify_qp_rts(struct rdma_id_private *id_priv,
+ struct rdma_conn_param *conn_param)
{
struct ib_qp_attr qp_attr;
int qp_attr_mask, ret;
@@ -544,6 +548,8 @@ static int cma_modify_qp_rts(struct rdma_id_private *id_priv)
if (ret)
goto out;
+ if (conn_param)
+ qp_attr.max_rd_atomic = conn_param->initiator_depth;
ret = ib_modify_qp(id_priv->id.qp, &qp_attr, qp_attr_mask);
out:
mutex_unlock(&id_priv->qp_mutex);
@@ -584,6 +590,10 @@ static int cma_ib_init_qp_attr(struct rdma_id_private *id_priv,
*qp_attr_mask = IB_QP_STATE | IB_QP_PKEY_INDEX | IB_QP_PORT;
if (cma_is_ud_ps(id_priv->id.ps)) {
+ ret = cma_set_qkey(id_priv);
+ if (ret)
+ return ret;
+
qp_attr->qkey = id_priv->qkey;
*qp_attr_mask |= IB_QP_QKEY;
} else {
@@ -632,7 +642,8 @@ static inline int cma_zero_addr(struct sockaddr *addr)
struct in6_addr *ip6;
if (addr->sa_family == AF_INET)
- return ZERONET(((struct sockaddr_in *) addr)->sin_addr.s_addr);
+ return ipv4_is_zeronet(
+ ((struct sockaddr_in *)addr)->sin_addr.s_addr);
else {
ip6 = &((struct sockaddr_in6 *) addr)->sin6_addr;
return (ip6->s6_addr32[0] | ip6->s6_addr32[1] |
@@ -642,7 +653,7 @@ static inline int cma_zero_addr(struct sockaddr *addr)
static inline int cma_loopback_addr(struct sockaddr *addr)
{
- return LOOPBACK(((struct sockaddr_in *) addr)->sin_addr.s_addr);
+ return ipv4_is_loopback(((struct sockaddr_in *) addr)->sin_addr.s_addr);
}
static inline int cma_any_addr(struct sockaddr *addr)
@@ -664,7 +675,7 @@ static inline int cma_any_port(struct sockaddr *addr)
}
static int cma_get_net_info(void *hdr, enum rdma_port_space ps,
- u8 *ip_ver, __u16 *port,
+ u8 *ip_ver, __be16 *port,
union cma_ip_addr **src, union cma_ip_addr **dst)
{
switch (ps) {
@@ -696,7 +707,7 @@ static int cma_get_net_info(void *hdr, enum rdma_port_space ps,
static void cma_save_net_info(struct rdma_addr *addr,
struct rdma_addr *listen_addr,
- u8 ip_ver, __u16 port,
+ u8 ip_ver, __be16 port,
union cma_ip_addr *src, union cma_ip_addr *dst)
{
struct sockaddr_in *listen4, *ip4;
@@ -790,8 +801,8 @@ static void cma_cancel_operation(struct rdma_id_private *id_priv,
cma_cancel_route(id_priv);
break;
case CMA_LISTEN:
- if (cma_any_addr(&id_priv->id.route.addr.src_addr) &&
- !id_priv->cma_dev)
+ if (cma_any_addr((struct sockaddr *) &id_priv->id.route.addr.src_addr)
+ && !id_priv->cma_dev)
cma_cancel_listens(id_priv);
break;
default:
@@ -813,6 +824,8 @@ static void cma_release_port(struct rdma_id_private *id_priv)
kfree(bind_list);
}
mutex_unlock(&lock);
+ if (id_priv->sock)
+ sock_release(id_priv->sock);
}
static void cma_leave_mc_groups(struct rdma_id_private *id_priv)
@@ -874,11 +887,11 @@ static int cma_rep_recv(struct rdma_id_private *id_priv)
{
int ret;
- ret = cma_modify_qp_rtr(id_priv);
+ ret = cma_modify_qp_rtr(id_priv, NULL);
if (ret)
goto reject;
- ret = cma_modify_qp_rts(id_priv);
+ ret = cma_modify_qp_rts(id_priv, NULL);
if (ret)
goto reject;
@@ -924,7 +937,10 @@ static int cma_ib_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
struct rdma_cm_event event;
int ret = 0;
- if (cma_disable_remove(id_priv, CMA_CONNECT))
+ if ((ib_event->event != IB_CM_TIMEWAIT_EXIT &&
+ cma_disable_callback(id_priv, CMA_CONNECT)) ||
+ (ib_event->event == IB_CM_TIMEWAIT_EXIT &&
+ cma_disable_callback(id_priv, CMA_DISCONNECT)))
return 0;
memset(&event, 0, sizeof event);
@@ -960,6 +976,8 @@ static int cma_ib_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
event.event = RDMA_CM_EVENT_DISCONNECTED;
break;
case IB_CM_TIMEWAIT_EXIT:
+ event.event = RDMA_CM_EVENT_TIMEWAIT_EXIT;
+ break;
case IB_CM_MRA_RECEIVED:
/* ignore event */
goto out;
@@ -971,7 +989,7 @@ static int cma_ib_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
event.param.conn.private_data_len = IB_CM_REJ_PRIVATE_DATA_SIZE;
break;
default:
- printk(KERN_ERR "RDMA CMA: unexpected IB CM event: %d",
+ printk(KERN_ERR "RDMA CMA: unexpected IB CM event: %d\n",
ib_event->event);
goto out;
}
@@ -981,12 +999,12 @@ static int cma_ib_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
/* Destroy the CM ID by returning a non-zero value. */
id_priv->cm_id.ib = NULL;
cma_exch(id_priv, CMA_DESTROYING);
- cma_enable_remove(id_priv);
+ mutex_unlock(&id_priv->handler_mutex);
rdma_destroy_id(&id_priv->id);
return ret;
}
out:
- cma_enable_remove(id_priv);
+ mutex_unlock(&id_priv->handler_mutex);
return ret;
}
@@ -997,8 +1015,9 @@ static struct rdma_id_private *cma_new_conn_id(struct rdma_cm_id *listen_id,
struct rdma_cm_id *id;
struct rdma_route *rt;
union cma_ip_addr *src, *dst;
- __u16 port;
+ __be16 port;
u8 ip_ver;
+ int ret;
if (cma_get_net_info(ib_event->private_data, listen_id->ps,
&ip_ver, &port, &src, &dst))
@@ -1023,10 +1042,11 @@ static struct rdma_id_private *cma_new_conn_id(struct rdma_cm_id *listen_id,
if (rt->num_paths == 2)
rt->path_rec[1] = *ib_event->param.req_rcvd.alternate_path;
- ib_addr_set_sgid(&rt->addr.dev_addr, &rt->path_rec[0].sgid);
ib_addr_set_dgid(&rt->addr.dev_addr, &rt->path_rec[0].dgid);
- ib_addr_set_pkey(&rt->addr.dev_addr, be16_to_cpu(rt->path_rec[0].pkey));
- rt->addr.dev_addr.dev_type = RDMA_NODE_IB_CA;
+ ret = rdma_translate_ip((struct sockaddr *) &id->route.addr.src_addr,
+ &id->route.addr.dev_addr);
+ if (ret)
+ goto destroy_id;
id_priv = container_of(id, struct rdma_id_private, id);
id_priv->state = CMA_CONNECT;
@@ -1044,7 +1064,7 @@ static struct rdma_id_private *cma_new_udp_id(struct rdma_cm_id *listen_id,
struct rdma_id_private *id_priv;
struct rdma_cm_id *id;
union cma_ip_addr *src, *dst;
- __u16 port;
+ __be16 port;
u8 ip_ver;
int ret;
@@ -1061,7 +1081,7 @@ static struct rdma_id_private *cma_new_udp_id(struct rdma_cm_id *listen_id,
cma_save_net_info(&id->route.addr, &listen_id->route.addr,
ip_ver, port, src, dst);
- ret = rdma_translate_ip(&id->route.addr.src_addr,
+ ret = rdma_translate_ip((struct sockaddr *) &id->route.addr.src_addr,
&id->route.addr.dev_addr);
if (ret)
goto err;
@@ -1096,7 +1116,7 @@ static int cma_req_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
int offset, ret;
listen_id = cm_id->context;
- if (cma_disable_remove(listen_id, CMA_LISTEN))
+ if (cma_disable_callback(listen_id, CMA_LISTEN))
return -ECONNABORTED;
memset(&event, 0, sizeof event);
@@ -1117,7 +1137,7 @@ static int cma_req_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
goto out;
}
- atomic_inc(&conn_id->dev_remove);
+ mutex_lock_nested(&conn_id->handler_mutex, SINGLE_DEPTH_NESTING);
mutex_lock(&lock);
ret = cma_acquire_dev(conn_id);
mutex_unlock(&lock);
@@ -1139,7 +1159,7 @@ static int cma_req_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
!cma_is_ud_ps(conn_id->id.ps))
ib_send_cm_mra(cm_id, CMA_CM_MRA_SETTING, NULL, 0);
mutex_unlock(&lock);
- cma_enable_remove(conn_id);
+ mutex_unlock(&conn_id->handler_mutex);
goto out;
}
@@ -1148,11 +1168,11 @@ static int cma_req_handler(struct ib_cm_id *cm_id, struct ib_cm_event *ib_event)
release_conn_id:
cma_exch(conn_id, CMA_DESTROYING);
- cma_enable_remove(conn_id);
+ mutex_unlock(&conn_id->handler_mutex);
rdma_destroy_id(&conn_id->id);
out:
- cma_enable_remove(listen_id);
+ mutex_unlock(&listen_id->handler_mutex);
return ret;
}
@@ -1166,7 +1186,7 @@ static void cma_set_compare_data(enum rdma_port_space ps, struct sockaddr *addr,
{
struct cma_hdr *cma_data, *cma_mask;
struct sdp_hh *sdp_data, *sdp_mask;
- __u32 ip4_addr;
+ __be32 ip4_addr;
struct in6_addr ip6_addr;
memset(compare, 0, sizeof *compare);
@@ -1182,12 +1202,12 @@ static void cma_set_compare_data(enum rdma_port_space ps, struct sockaddr *addr,
sdp_set_ip_ver(sdp_data, 4);
sdp_set_ip_ver(sdp_mask, 0xF);
sdp_data->dst_addr.ip4.addr = ip4_addr;
- sdp_mask->dst_addr.ip4.addr = ~0;
+ sdp_mask->dst_addr.ip4.addr = htonl(~0);
} else {
cma_set_ip_ver(cma_data, 4);
cma_set_ip_ver(cma_mask, 0xF);
cma_data->dst_addr.ip4.addr = ip4_addr;
- cma_mask->dst_addr.ip4.addr = ~0;
+ cma_mask->dst_addr.ip4.addr = htonl(~0);
}
break;
case AF_INET6:
@@ -1218,7 +1238,7 @@ static int cma_iw_handler(struct iw_cm_id *iw_id, struct iw_cm_event *iw_event)
struct sockaddr_in *sin;
int ret = 0;
- if (cma_disable_remove(id_priv, CMA_CONNECT))
+ if (cma_disable_callback(id_priv, CMA_CONNECT))
return 0;
memset(&event, 0, sizeof event);
@@ -1262,12 +1282,12 @@ static int cma_iw_handler(struct iw_cm_id *iw_id, struct iw_cm_event *iw_event)
/* Destroy the CM ID by returning a non-zero value. */
id_priv->cm_id.iw = NULL;
cma_exch(id_priv, CMA_DESTROYING);
- cma_enable_remove(id_priv);
+ mutex_unlock(&id_priv->handler_mutex);
rdma_destroy_id(&id_priv->id);
return ret;
}
- cma_enable_remove(id_priv);
+ mutex_unlock(&id_priv->handler_mutex);
return ret;
}
@@ -1283,31 +1303,31 @@ static int iw_conn_req_handler(struct iw_cm_id *cm_id,
struct ib_device_attr attr;
listen_id = cm_id->context;
- if (cma_disable_remove(listen_id, CMA_LISTEN))
+ if (cma_disable_callback(listen_id, CMA_LISTEN))
return -ECONNABORTED;
/* Create a new RDMA id for the new IW CM ID */
new_cm_id = rdma_create_id(listen_id->id.event_handler,
listen_id->id.context,
RDMA_PS_TCP);
- if (!new_cm_id) {
+ if (IS_ERR(new_cm_id)) {
ret = -ENOMEM;
goto out;
}
conn_id = container_of(new_cm_id, struct rdma_id_private, id);
- atomic_inc(&conn_id->dev_remove);
+ mutex_lock_nested(&conn_id->handler_mutex, SINGLE_DEPTH_NESTING);
conn_id->state = CMA_CONNECT;
- dev = ip_dev_find(iw_event->local_addr.sin_addr.s_addr);
+ dev = ip_dev_find(&init_net, iw_event->local_addr.sin_addr.s_addr);
if (!dev) {
ret = -EADDRNOTAVAIL;
- cma_enable_remove(conn_id);
+ mutex_unlock(&conn_id->handler_mutex);
rdma_destroy_id(new_cm_id);
goto out;
}
ret = rdma_copy_addr(&conn_id->id.route.addr.dev_addr, dev, NULL);
if (ret) {
- cma_enable_remove(conn_id);
+ mutex_unlock(&conn_id->handler_mutex);
rdma_destroy_id(new_cm_id);
goto out;
}
@@ -1316,7 +1336,7 @@ static int iw_conn_req_handler(struct iw_cm_id *cm_id,
ret = cma_acquire_dev(conn_id);
mutex_unlock(&lock);
if (ret) {
- cma_enable_remove(conn_id);
+ mutex_unlock(&conn_id->handler_mutex);
rdma_destroy_id(new_cm_id);
goto out;
}
@@ -1332,30 +1352,33 @@ static int iw_conn_req_handler(struct iw_cm_id *cm_id,
ret = ib_query_device(conn_id->id.device, &attr);
if (ret) {
- cma_enable_remove(conn_id);
+ mutex_unlock(&conn_id->handler_mutex);
rdma_destroy_id(new_cm_id);
goto out;
}
-
+
memset(&event, 0, sizeof event);
event.event = RDMA_CM_EVENT_CONNECT_REQUEST;
event.param.conn.private_data = iw_event->private_data;
event.param.conn.private_data_len = iw_event->private_data_len;
- event.param.conn.initiator_depth = attr.max_qp_init_rd_atom;
+ event.param.conn.initiator_depth = attr.max_qp_init_rd_atom;
event.param.conn.responder_resources = attr.max_qp_rd_atom;
ret = conn_id->id.event_handler(&conn_id->id, &event);
if (ret) {
/* User wants to destroy the CM ID */
conn_id->cm_id.iw = NULL;
cma_exch(conn_id, CMA_DESTROYING);
- cma_enable_remove(conn_id);
+ mutex_unlock(&conn_id->handler_mutex);
rdma_destroy_id(&conn_id->id);
+ goto out;
}
+ mutex_unlock(&conn_id->handler_mutex);
+
out:
if (dev)
dev_put(dev);
- cma_enable_remove(listen_id);
+ mutex_unlock(&listen_id->handler_mutex);
return ret;
}
@@ -1371,7 +1394,7 @@ static int cma_ib_listen(struct rdma_id_private *id_priv)
if (IS_ERR(id_priv->cm_id.ib))
return PTR_ERR(id_priv->cm_id.ib);
- addr = &id_priv->id.route.addr.src_addr;
+ addr = (struct sockaddr *) &id_priv->id.route.addr.src_addr;
svc_id = cma_get_service_id(id_priv->id.ps, addr);
if (cma_any_addr(addr))
ret = ib_cm_listen(id_priv->cm_id.ib, svc_id, 0, NULL);
@@ -1437,7 +1460,7 @@ static void cma_listen_on_dev(struct rdma_id_private *id_priv,
dev_id_priv->state = CMA_ADDR_BOUND;
memcpy(&id->route.addr.src_addr, &id_priv->id.route.addr.src_addr,
- ip_addr_size(&id_priv->id.route.addr.src_addr));
+ ip_addr_size((struct sockaddr *) &id_priv->id.route.addr.src_addr));
cma_attach_to_dev(dev_id_priv, cma_dev);
list_add_tail(&dev_id_priv->listen_list, &id_priv->listen_list);
@@ -1447,7 +1470,7 @@ static void cma_listen_on_dev(struct rdma_id_private *id_priv,
ret = rdma_listen(id, id_priv->backlog);
if (ret)
printk(KERN_WARNING "RDMA CMA: cma_listen_on_dev, error %d, "
- "listening on device %s", ret, cma_dev->device->name);
+ "listening on device %s\n", ret, cma_dev->device->name);
}
static void cma_listen_on_all(struct rdma_id_private *id_priv)
@@ -1557,13 +1580,14 @@ static int cma_query_ib_route(struct rdma_id_private *id_priv, int timeout_ms,
path_rec.pkey = cpu_to_be16(ib_addr_get_pkey(&addr->dev_addr));
path_rec.numb_path = 1;
path_rec.reversible = 1;
- path_rec.service_id = cma_get_service_id(id_priv->id.ps, &addr->dst_addr);
+ path_rec.service_id = cma_get_service_id(id_priv->id.ps,
+ (struct sockaddr *) &addr->dst_addr);
comp_mask = IB_SA_PATH_REC_DGID | IB_SA_PATH_REC_SGID |
IB_SA_PATH_REC_PKEY | IB_SA_PATH_REC_NUMB_PATH |
IB_SA_PATH_REC_REVERSIBLE | IB_SA_PATH_REC_SERVICE_ID;
- if (addr->src_addr.sa_family == AF_INET) {
+ if (addr->src_addr.ss_family == AF_INET) {
path_rec.qos_class = cpu_to_be16((u16) id_priv->tos);
comp_mask |= IB_SA_PATH_REC_QOS_CLASS;
} else {
@@ -1592,7 +1616,7 @@ static void cma_work_handler(struct work_struct *_work)
struct rdma_id_private *id_priv = work->id;
int destroy = 0;
- atomic_inc(&id_priv->dev_remove);
+ mutex_lock(&id_priv->handler_mutex);
if (!cma_comp_exch(id_priv, work->old_state, work->new_state))
goto out;
@@ -1601,7 +1625,31 @@ static void cma_work_handler(struct work_struct *_work)
destroy = 1;
}
out:
- cma_enable_remove(id_priv);
+ mutex_unlock(&id_priv->handler_mutex);
+ cma_deref_id(id_priv);
+ if (destroy)
+ rdma_destroy_id(&id_priv->id);
+ kfree(work);
+}
+
+static void cma_ndev_work_handler(struct work_struct *_work)
+{
+ struct cma_ndev_work *work = container_of(_work, struct cma_ndev_work, work);
+ struct rdma_id_private *id_priv = work->id;
+ int destroy = 0;
+
+ mutex_lock(&id_priv->handler_mutex);
+ if (id_priv->state == CMA_DESTROYING ||
+ id_priv->state == CMA_DEVICE_REMOVAL)
+ goto out;
+
+ if (id_priv->id.event_handler(&id_priv->id, &work->event)) {
+ cma_exch(id_priv, CMA_DESTROYING);
+ destroy = 1;
+ }
+
+out:
+ mutex_unlock(&id_priv->handler_mutex);
cma_deref_id(id_priv);
if (destroy)
rdma_destroy_id(&id_priv->id);
@@ -1764,7 +1812,7 @@ static void addr_handler(int status, struct sockaddr *src_addr,
struct rdma_cm_event event;
memset(&event, 0, sizeof event);
- atomic_inc(&id_priv->dev_remove);
+ mutex_lock(&id_priv->handler_mutex);
/*
* Grab mutex to block rdma_destroy_id() from removing the device while
@@ -1793,13 +1841,13 @@ static void addr_handler(int status, struct sockaddr *src_addr,
if (id_priv->id.event_handler(&id_priv->id, &event)) {
cma_exch(id_priv, CMA_DESTROYING);
- cma_enable_remove(id_priv);
+ mutex_unlock(&id_priv->handler_mutex);
cma_deref_id(id_priv);
rdma_destroy_id(&id_priv->id);
return;
}
out:
- cma_enable_remove(id_priv);
+ mutex_unlock(&id_priv->handler_mutex);
cma_deref_id(id_priv);
}
@@ -1823,7 +1871,7 @@ static int cma_resolve_loopback(struct rdma_id_private *id_priv)
ib_addr_get_sgid(&id_priv->id.route.addr.dev_addr, &gid);
ib_addr_set_dgid(&id_priv->id.route.addr.dev_addr, &gid);
- if (cma_zero_addr(&id_priv->id.route.addr.src_addr)) {
+ if (cma_zero_addr((struct sockaddr *) &id_priv->id.route.addr.src_addr)) {
src_in = (struct sockaddr_in *)&id_priv->id.route.addr.src_addr;
dst_in = (struct sockaddr_in *)&id_priv->id.route.addr.dst_addr;
src_in->sin_family = dst_in->sin_family;
@@ -1872,7 +1920,7 @@ int rdma_resolve_addr(struct rdma_cm_id *id, struct sockaddr *src_addr,
if (cma_any_addr(dst_addr))
ret = cma_resolve_loopback(id_priv);
else
- ret = rdma_resolve_ip(&addr_client, &id->route.addr.src_addr,
+ ret = rdma_resolve_ip(&addr_client, (struct sockaddr *) &id->route.addr.src_addr,
dst_addr, &id->route.addr.dev_addr,
timeout_ms, addr_handler, id_priv);
if (ret)
@@ -1996,11 +2044,11 @@ static int cma_use_port(struct idr *ps, struct rdma_id_private *id_priv)
* We don't support binding to any address if anyone is bound to
* a specific address on the same port.
*/
- if (cma_any_addr(&id_priv->id.route.addr.src_addr))
+ if (cma_any_addr((struct sockaddr *) &id_priv->id.route.addr.src_addr))
return -EADDRNOTAVAIL;
hlist_for_each_entry(cur_id, node, &bind_list->owners, node) {
- if (cma_any_addr(&cur_id->id.route.addr.src_addr))
+ if (cma_any_addr((struct sockaddr *) &cur_id->id.route.addr.src_addr))
return -EADDRNOTAVAIL;
cur_sin = (struct sockaddr_in *) &cur_id->id.route.addr.src_addr;
@@ -2012,6 +2060,34 @@ static int cma_use_port(struct idr *ps, struct rdma_id_private *id_priv)
return 0;
}
+static int cma_get_tcp_port(struct rdma_id_private *id_priv)
+{
+ int ret;
+ int size;
+ struct socket *sock;
+
+ ret = sock_create_kern(AF_INET, SOCK_STREAM, IPPROTO_TCP, &sock);
+ if (ret)
+ return ret;
+ ret = sock->ops->bind(sock,
+ (struct sockaddr *) &id_priv->id.route.addr.src_addr,
+ ip_addr_size((struct sockaddr *) &id_priv->id.route.addr.src_addr));
+ if (ret) {
+ sock_release(sock);
+ return ret;
+ }
+ size = ip_addr_size((struct sockaddr *) &id_priv->id.route.addr.src_addr);
+ ret = sock->ops->getname(sock,
+ (struct sockaddr *) &id_priv->id.route.addr.src_addr,
+ &size, 0);
+ if (ret) {
+ sock_release(sock);
+ return ret;
+ }
+ id_priv->sock = sock;
+ return 0;
+}
+
static int cma_get_port(struct rdma_id_private *id_priv)
{
struct idr *ps;
@@ -2023,6 +2099,11 @@ static int cma_get_port(struct rdma_id_private *id_priv)
break;
case RDMA_PS_TCP:
ps = &tcp_ps;
+ if (unify_tcp_port_space) {
+ ret = cma_get_tcp_port(id_priv);
+ if (ret)
+ goto out;
+ }
break;
case RDMA_PS_UDP:
ps = &udp_ps;
@@ -2035,12 +2116,12 @@ static int cma_get_port(struct rdma_id_private *id_priv)
}
mutex_lock(&lock);
- if (cma_any_port(&id_priv->id.route.addr.src_addr))
+ if (cma_any_port((struct sockaddr *) &id_priv->id.route.addr.src_addr))
ret = cma_alloc_any_port(ps, id_priv);
else
ret = cma_use_port(ps, id_priv);
mutex_unlock(&lock);
-
+out:
return ret;
}
@@ -2126,7 +2207,7 @@ static int cma_sidr_rep_handler(struct ib_cm_id *cm_id,
struct ib_cm_sidr_rep_event_param *rep = &ib_event->param.sidr_rep_rcvd;
int ret = 0;
- if (cma_disable_remove(id_priv, CMA_CONNECT))
+ if (cma_disable_callback(id_priv, CMA_CONNECT))
return 0;
memset(&event, 0, sizeof event);
@@ -2143,6 +2224,12 @@ static int cma_sidr_rep_handler(struct ib_cm_id *cm_id,
event.status = ib_event->param.sidr_rep_rcvd.status;
break;
}
+ ret = cma_set_qkey(id_priv);
+ if (ret) {
+ event.event = RDMA_CM_EVENT_ADDR_ERROR;
+ event.status = -EINVAL;
+ break;
+ }
if (id_priv->qkey != rep->qkey) {
event.event = RDMA_CM_EVENT_UNREACHABLE;
event.status = -EINVAL;
@@ -2157,7 +2244,7 @@ static int cma_sidr_rep_handler(struct ib_cm_id *cm_id,
event.status = 0;
break;
default:
- printk(KERN_ERR "RDMA CMA: unexpected IB CM event: %d",
+ printk(KERN_ERR "RDMA CMA: unexpected IB CM event: %d\n",
ib_event->event);
goto out;
}
@@ -2167,12 +2254,12 @@ static int cma_sidr_rep_handler(struct ib_cm_id *cm_id,
/* Destroy the CM ID by returning a non-zero value. */
id_priv->cm_id.ib = NULL;
cma_exch(id_priv, CMA_DESTROYING);
- cma_enable_remove(id_priv);
+ mutex_unlock(&id_priv->handler_mutex);
rdma_destroy_id(&id_priv->id);
return ret;
}
out:
- cma_enable_remove(id_priv);
+ mutex_unlock(&id_priv->handler_mutex);
return ret;
}
@@ -2207,8 +2294,8 @@ static int cma_resolve_ib_udp(struct rdma_id_private *id_priv,
req.path = route->path_rec;
req.service_id = cma_get_service_id(id_priv->id.ps,
- &route->addr.dst_addr);
- req.timeout_ms = 1 << (cma_response_timeout - 8);
+ (struct sockaddr *) &route->addr.dst_addr);
+ req.timeout_ms = 1 << (cma_response_timeout - 8);
req.max_cm_retries = CMA_MAX_CM_RETRIES;
ret = ib_send_cm_sidr_req(id_priv->cm_id.ib, &req);
@@ -2258,7 +2345,7 @@ static int cma_connect_ib(struct rdma_id_private *id_priv,
req.alternate_path = &route->path_rec[1];
req.service_id = cma_get_service_id(id_priv->id.ps,
- &route->addr.dst_addr);
+ (struct sockaddr *) &route->addr.dst_addr);
req.qp_num = id_priv->qp_num;
req.qp_type = IB_QPT_RC;
req.starting_psn = id_priv->seq_num;
@@ -2305,7 +2392,7 @@ static int cma_connect_iw(struct rdma_id_private *id_priv,
sin = (struct sockaddr_in*) &id_priv->id.route.addr.dst_addr;
cm_id->remote_addr = *sin;
- ret = cma_modify_qp_rtr(id_priv);
+ ret = cma_modify_qp_rtr(id_priv, conn_param);
if (ret)
goto out;
@@ -2368,25 +2455,15 @@ static int cma_accept_ib(struct rdma_id_private *id_priv,
struct rdma_conn_param *conn_param)
{
struct ib_cm_rep_param rep;
- struct ib_qp_attr qp_attr;
- int qp_attr_mask, ret;
-
- if (id_priv->id.qp) {
- ret = cma_modify_qp_rtr(id_priv);
- if (ret)
- goto out;
+ int ret;
- qp_attr.qp_state = IB_QPS_RTS;
- ret = ib_cm_init_qp_attr(id_priv->cm_id.ib, &qp_attr,
- &qp_attr_mask);
- if (ret)
- goto out;
+ ret = cma_modify_qp_rtr(id_priv, conn_param);
+ if (ret)
+ goto out;
- qp_attr.max_rd_atomic = conn_param->initiator_depth;
- ret = ib_modify_qp(id_priv->id.qp, &qp_attr, qp_attr_mask);
- if (ret)
- goto out;
- }
+ ret = cma_modify_qp_rts(id_priv, conn_param);
+ if (ret)
+ goto out;
memset(&rep, 0, sizeof rep);
rep.qp_num = id_priv->qp_num;
@@ -2411,7 +2488,7 @@ static int cma_accept_iw(struct rdma_id_private *id_priv,
struct iw_cm_conn_param iw_param;
int ret;
- ret = cma_modify_qp_rtr(id_priv);
+ ret = cma_modify_qp_rtr(id_priv, conn_param);
if (ret)
return ret;
@@ -2432,10 +2509,14 @@ static int cma_send_sidr_rep(struct rdma_id_private *id_priv,
const void *private_data, int private_data_len)
{
struct ib_cm_sidr_rep_param rep;
+ int ret;
memset(&rep, 0, sizeof rep);
rep.status = status;
if (status == IB_SIDR_SUCCESS) {
+ ret = cma_set_qkey(id_priv);
+ if (ret)
+ return ret;
rep.qp_num = id_priv->qp_num;
rep.qkey = id_priv->qkey;
}
@@ -2580,8 +2661,8 @@ static int cma_ib_mc_handler(int status, struct ib_sa_multicast *multicast)
int ret;
id_priv = mc->id_priv;
- if (cma_disable_remove(id_priv, CMA_ADDR_BOUND) &&
- cma_disable_remove(id_priv, CMA_ADDR_RESOLVED))
+ if (cma_disable_callback(id_priv, CMA_ADDR_BOUND) &&
+ cma_disable_callback(id_priv, CMA_ADDR_RESOLVED))
return 0;
mutex_lock(&id_priv->qp_mutex);
@@ -2606,12 +2687,12 @@ static int cma_ib_mc_handler(int status, struct ib_sa_multicast *multicast)
ret = id_priv->id.event_handler(&id_priv->id, &event);
if (ret) {
cma_exch(id_priv, CMA_DESTROYING);
- cma_enable_remove(id_priv);
+ mutex_unlock(&id_priv->handler_mutex);
rdma_destroy_id(&id_priv->id);
return 0;
}
- cma_enable_remove(id_priv);
+ mutex_unlock(&id_priv->handler_mutex);
return 0;
}
@@ -2631,11 +2712,9 @@ static void cma_set_mgid(struct rdma_id_private *id_priv,
/* IPv6 address is an SA assigned MGID. */
memcpy(mgid, &sin6->sin6_addr, sizeof *mgid);
} else {
- ip_ib_mc_map(sin->sin_addr.s_addr, mc_map);
+ ip_ib_mc_map(sin->sin_addr.s_addr, dev_addr->broadcast, mc_map);
if (id_priv->id.ps == RDMA_PS_UDP)
mc_map[7] = 0x01; /* Use RDMA CM signature */
- mc_map[8] = ib_addr_get_pkey(dev_addr) >> 8;
- mc_map[9] = (unsigned char) ib_addr_get_pkey(dev_addr);
*mgid = *(union ib_gid *) (mc_map + 4);
}
}
@@ -2654,7 +2733,7 @@ static int cma_join_ib_multicast(struct rdma_id_private *id_priv,
if (ret)
return ret;
- cma_set_mgid(id_priv, &mc->addr, &rec.mgid);
+ cma_set_mgid(id_priv, (struct sockaddr *) &mc->addr, &rec.mgid);
if (id_priv->id.ps == RDMA_PS_UDP)
rec.qkey = cpu_to_be32(RDMA_UDP_QKEY);
ib_addr_get_sgid(dev_addr, &rec.port_gid);
@@ -2667,6 +2746,10 @@ static int cma_join_ib_multicast(struct rdma_id_private *id_priv,
IB_SA_MCMEMBER_REC_FLOW_LABEL |
IB_SA_MCMEMBER_REC_TRAFFIC_CLASS;
+ if (id_priv->id.ps == RDMA_PS_IPOIB)
+ comp_mask |= IB_SA_MCMEMBER_REC_RATE |
+ IB_SA_MCMEMBER_REC_RATE_SELECTOR;
+
mc->multicast.ib = ib_sa_join_multicast(&sa_client, id_priv->id.device,
id_priv->id.port_num, &rec,
comp_mask, GFP_KERNEL,
@@ -2745,6 +2828,62 @@ void rdma_leave_multicast(struct rdma_cm_id *id, struct sockaddr *addr)
}
EXPORT_SYMBOL(rdma_leave_multicast);
+static int cma_netdev_change(struct net_device *ndev, struct rdma_id_private *id_priv)
+{
+ struct rdma_dev_addr *dev_addr;
+ struct cma_ndev_work *work;
+
+ dev_addr = &id_priv->id.route.addr.dev_addr;
+
+ if ((dev_addr->src_dev == ndev) &&
+ memcmp(dev_addr->src_dev_addr, ndev->dev_addr, ndev->addr_len)) {
+ printk(KERN_INFO "RDMA CM addr change for ndev %s used by id %p\n",
+ ndev->name, &id_priv->id);
+ work = kzalloc(sizeof *work, GFP_KERNEL);
+ if (!work)
+ return -ENOMEM;
+
+ INIT_WORK(&work->work, cma_ndev_work_handler);
+ work->id = id_priv;
+ work->event.event = RDMA_CM_EVENT_ADDR_CHANGE;
+ atomic_inc(&id_priv->refcount);
+ queue_work(cma_wq, &work->work);
+ }
+
+ return 0;
+}
+
+static int cma_netdev_callback(struct notifier_block *self, unsigned long event,
+ void *ctx)
+{
+ struct net_device *ndev = (struct net_device *)ctx;
+ struct cma_device *cma_dev;
+ struct rdma_id_private *id_priv;
+ int ret = NOTIFY_DONE;
+
+ if (event != NETDEV_BONDING_FAILOVER)
+ return NOTIFY_DONE;
+
+ if (!(ndev->flags & IFF_MASTER))
+ return NOTIFY_DONE;
+
+ mutex_lock(&lock);
+ list_for_each_entry(cma_dev, &dev_list, list)
+ list_for_each_entry(id_priv, &cma_dev->id_list, list) {
+ ret = cma_netdev_change(ndev, id_priv);
+ if (ret)
+ goto out;
+ }
+
+out:
+ mutex_unlock(&lock);
+ return ret;
+}
+
+static struct notifier_block cma_nb = {
+ .notifier_call = cma_netdev_callback
+};
+
static void cma_add_one(struct ib_device *device)
{
struct cma_device *cma_dev;
@@ -2772,6 +2911,7 @@ static int cma_remove_id_dev(struct rdma_id_private *id_priv)
{
struct rdma_cm_event event;
enum cma_state state;
+ int ret = 0;
/* Record that we want to remove the device */
state = cma_exch(id_priv, CMA_DEVICE_REMOVAL);
@@ -2779,15 +2919,18 @@ static int cma_remove_id_dev(struct rdma_id_private *id_priv)
return 0;
cma_cancel_operation(id_priv, state);
- wait_event(id_priv->wait_remove, !atomic_read(&id_priv->dev_remove));
+ mutex_lock(&id_priv->handler_mutex);
/* Check for destruction from another callback. */
if (!cma_comp(id_priv, CMA_DEVICE_REMOVAL))
- return 0;
+ goto out;
memset(&event, 0, sizeof event);
event.event = RDMA_CM_EVENT_DEVICE_REMOVAL;
- return id_priv->id.event_handler(&id_priv->id, &event);
+ ret = id_priv->id.event_handler(&id_priv->id, &event);
+out:
+ mutex_unlock(&id_priv->handler_mutex);
+ return ret;
}
static void cma_process_remove(struct cma_device *cma_dev)
@@ -2849,6 +2992,7 @@ static int cma_init(void)
ib_sa_register_client(&sa_client);
rdma_addr_register_client(&addr_client);
+ register_netdevice_notifier(&cma_nb);
ret = ib_register_client(&cma_client);
if (ret)
@@ -2856,6 +3000,7 @@ static int cma_init(void)
return 0;
err:
+ unregister_netdevice_notifier(&cma_nb);
rdma_addr_unregister_client(&addr_client);
ib_sa_unregister_client(&sa_client);
destroy_workqueue(cma_wq);
@@ -2865,6 +3010,7 @@ err:
static void cma_cleanup(void)
{
ib_unregister_client(&cma_client);
+ unregister_netdevice_notifier(&cma_nb);
rdma_addr_unregister_client(&addr_client);
ib_sa_unregister_client(&sa_client);
destroy_workqueue(cma_wq);
diff --git a/drivers/infiniband/core/core_priv.h b/drivers/infiniband/core/core_priv.h
index 7ad47a4..05ac36e 100644
--- a/drivers/infiniband/core/core_priv.h
+++ b/drivers/infiniband/core/core_priv.h
@@ -28,8 +28,6 @@
* 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.
- *
- * $Id: core_priv.h 1349 2004-12-16 21:09:43Z roland $
*/
#ifndef _CORE_PRIV_H
diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index 858b6ea..179f753 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -29,8 +29,6 @@
* 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.
- *
- * $Id: device.c 1349 2004-12-16 21:09:43Z roland $
*/
#include <linux/module.h>
@@ -180,9 +178,14 @@ static int end_port(struct ib_device *device)
*/
struct ib_device *ib_alloc_device(size_t size)
{
+ struct ib_device *ibdev;
+
BUG_ON(size < sizeof (struct ib_device));
- return kzalloc(size, GFP_KERNEL);
+ ibdev = kzalloc(size, GFP_KERNEL);
+ if (ibdev)
+ mutex_init(&ibdev->sysfs_mutex);
+ return ibdev;
}
EXPORT_SYMBOL(ib_alloc_device);
@@ -313,9 +316,10 @@ int ib_register_device(struct ib_device *device)
goto out;
}
+ mutex_lock(&device->sysfs_mutex);
list_add_tail(&device->core_list, &device_list);
-
device->reg_state = IB_DEV_REGISTERED;
+ mutex_unlock(&device->sysfs_mutex);
{
struct ib_client *client;
@@ -361,7 +365,9 @@ void ib_unregister_device(struct ib_device *device)
kfree(context);
spin_unlock_irqrestore(&device->client_data_lock, flags);
+ mutex_lock(&device->sysfs_mutex);
device->reg_state = IB_DEV_UNREGISTERED;
+ mutex_unlock(&device->sysfs_mutex);
}
EXPORT_SYMBOL(ib_unregister_device);
@@ -727,6 +733,10 @@ static int __init ib_core_init(void)
dma_map_sg_hp_wa = 1;
#endif
+ ret = init_mnt_writers();
+ if (ret)
+ printk(KERN_WARNING "Couldn't init mnt_writers\n");
+
ret = ib_sysfs_setup();
if (ret)
printk(KERN_WARNING "Couldn't create InfiniBand device class\n");
diff --git a/drivers/infiniband/core/fmr_pool.c b/drivers/infiniband/core/fmr_pool.c
index bbfc713..4507043 100644
--- a/drivers/infiniband/core/fmr_pool.c
+++ b/drivers/infiniband/core/fmr_pool.c
@@ -29,8 +29,6 @@
* 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.
- *
- * $Id: fmr_pool.c 2730 2005-06-28 16:43:03Z sean.hefty $
*/
#include <linux/errno.h>
@@ -158,8 +156,7 @@ static void ib_fmr_batch_release(struct ib_fmr_pool *pool)
#endif
}
- list_splice(&pool->dirty_list, &unmap_list);
- INIT_LIST_HEAD(&pool->dirty_list);
+ list_splice_init(&pool->dirty_list, &unmap_list);
pool->dirty_len = 0;
spin_unlock_irq(&pool->pool_lock);
@@ -182,8 +179,7 @@ static int ib_fmr_cleanup_thread(void *pool_ptr)
struct ib_fmr_pool *pool = pool_ptr;
do {
- if (pool->dirty_len >= pool->dirty_watermark ||
- atomic_read(&pool->flush_ser) - atomic_read(&pool->req_ser) < 0) {
+ if (atomic_read(&pool->flush_ser) - atomic_read(&pool->req_ser) < 0) {
ib_fmr_batch_release(pool);
atomic_inc(&pool->flush_ser);
@@ -194,8 +190,7 @@ static int ib_fmr_cleanup_thread(void *pool_ptr)
}
set_current_state(TASK_INTERRUPTIBLE);
- if (pool->dirty_len < pool->dirty_watermark &&
- atomic_read(&pool->flush_ser) - atomic_read(&pool->req_ser) >= 0 &&
+ if (atomic_read(&pool->flush_ser) - atomic_read(&pool->req_ser) >= 0 &&
!kthread_should_stop())
schedule();
__set_current_state(TASK_RUNNING);
@@ -308,10 +303,13 @@ struct ib_fmr_pool *ib_create_fmr_pool(struct ib_pd *pd,
.max_maps = pool->max_remaps,
.page_shift = params->page_shift
};
+ int bytes_per_fmr = sizeof *fmr;
+
+ if (pool->cache_bucket)
+ bytes_per_fmr += params->max_pages_per_fmr * sizeof (u64);
for (i = 0; i < params->pool_size; ++i) {
- fmr = kmalloc(sizeof *fmr + params->max_pages_per_fmr * sizeof (u64),
- GFP_KERNEL);
+ fmr = kmalloc(bytes_per_fmr, GFP_KERNEL);
if (!fmr) {
printk(KERN_WARNING PFX "failed to allocate fmr "
"struct for FMR %d\n", i);
@@ -526,8 +524,10 @@ int ib_fmr_pool_unmap(struct ib_pool_fmr *fmr)
list_add_tail(&fmr->list, &pool->free_list);
} else {
list_add_tail(&fmr->list, &pool->dirty_list);
- ++pool->dirty_len;
- wake_up_process(pool->thread);
+ if (++pool->dirty_len >= pool->dirty_watermark) {
+ atomic_inc(&pool->req_ser);
+ wake_up_process(pool->thread);
+ }
}
}
diff --git a/drivers/infiniband/core/iwcm.c b/drivers/infiniband/core/iwcm.c
index 81c9195..8f9509e 100644
--- a/drivers/infiniband/core/iwcm.c
+++ b/drivers/infiniband/core/iwcm.c
@@ -942,8 +942,7 @@ static int iwcm_init_qp_init_attr(struct iwcm_id_private *cm_id_priv,
case IW_CM_STATE_CONN_RECV:
case IW_CM_STATE_ESTABLISHED:
*qp_attr_mask = IB_QP_STATE | IB_QP_ACCESS_FLAGS;
- qp_attr->qp_access_flags = IB_ACCESS_LOCAL_WRITE |
- IB_ACCESS_REMOTE_WRITE|
+ qp_attr->qp_access_flags = IB_ACCESS_REMOTE_WRITE|
IB_ACCESS_REMOTE_READ;
ret = 0;
break;
diff --git a/drivers/infiniband/core/local_sa.c b/drivers/infiniband/core/local_sa.c
index 49ca262..eb62c42 100644
--- a/drivers/infiniband/core/local_sa.c
+++ b/drivers/infiniband/core/local_sa.c
@@ -978,7 +978,7 @@ static struct ib_sa_path_rec *get_next_path(struct ib_sa_attr_iter *iter,
static void report_path(struct work_struct *work)
{
struct sa_path_request *req;
-
+
req = container_of(work, struct sa_path_request, work);
req->callback(0, &req->path_rec, req->context);
ib_sa_client_put(req->client);
@@ -1058,7 +1058,7 @@ int ib_sa_path_rec_get(struct ib_sa_client *client,
req->context = context;
ib_sa_client_get(client);
- queue_work(sa_wq, &req->work);
+ queue_work(sa_wq, &req->work);
*sa_query = ERR_PTR(-EEXIST);
return 0;
diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c
index 4402d03..7934a4c 100644
--- a/drivers/infiniband/core/mad.c
+++ b/drivers/infiniband/core/mad.c
@@ -301,6 +301,16 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
mad_agent_priv->agent.context = context;
mad_agent_priv->agent.qp = port_priv->qp_info[qpn].qp;
mad_agent_priv->agent.port_num = port_num;
+ spin_lock_init(&mad_agent_priv->lock);
+ INIT_LIST_HEAD(&mad_agent_priv->send_list);
+ INIT_LIST_HEAD(&mad_agent_priv->wait_list);
+ INIT_LIST_HEAD(&mad_agent_priv->done_list);
+ INIT_LIST_HEAD(&mad_agent_priv->rmpp_list);
+ INIT_DELAYED_WORK(&mad_agent_priv->timed_work, timeout_sends);
+ INIT_LIST_HEAD(&mad_agent_priv->local_list);
+ INIT_WORK(&mad_agent_priv->local_work, local_completions);
+ atomic_set(&mad_agent_priv->refcount, 1);
+ init_completion(&mad_agent_priv->comp);
spin_lock_irqsave(&port_priv->reg_lock, flags);
mad_agent_priv->agent.hi_tid = ++ib_mad_client_id;
@@ -350,17 +360,6 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
list_add_tail(&mad_agent_priv->agent_list, &port_priv->agent_list);
spin_unlock_irqrestore(&port_priv->reg_lock, flags);
- spin_lock_init(&mad_agent_priv->lock);
- INIT_LIST_HEAD(&mad_agent_priv->send_list);
- INIT_LIST_HEAD(&mad_agent_priv->wait_list);
- INIT_LIST_HEAD(&mad_agent_priv->done_list);
- INIT_LIST_HEAD(&mad_agent_priv->rmpp_list);
- INIT_DELAYED_WORK(&mad_agent_priv->timed_work, timeout_sends);
- INIT_LIST_HEAD(&mad_agent_priv->local_list);
- INIT_WORK(&mad_agent_priv->local_work, local_completions);
- atomic_set(&mad_agent_priv->refcount, 1);
- init_completion(&mad_agent_priv->comp);
-
return &mad_agent_priv->agent;
error4:
@@ -747,9 +746,7 @@ static int handle_outgoing_dr_smp(struct ib_mad_agent_private *mad_agent_priv,
break;
case IB_MAD_RESULT_SUCCESS | IB_MAD_RESULT_CONSUMED:
kmem_cache_free(ib_mad_cache, mad_priv);
- kfree(local);
- ret = 1;
- goto out;
+ break;
case IB_MAD_RESULT_SUCCESS:
/* Treat like an incoming receive MAD */
port_priv = ib_get_mad_port(mad_agent_priv->agent.device,
@@ -760,10 +757,12 @@ static int handle_outgoing_dr_smp(struct ib_mad_agent_private *mad_agent_priv,
&mad_priv->mad.mad);
}
if (!port_priv || !recv_mad_agent) {
+ /*
+ * No receiving agent so drop packet and
+ * generate send completion.
+ */
kmem_cache_free(ib_mad_cache, mad_priv);
- kfree(local);
- ret = 0;
- goto out;
+ break;
}
local->mad_priv = mad_priv;
local->recv_mad_agent = recv_mad_agent;
@@ -1102,7 +1101,9 @@ int ib_post_send_mad(struct ib_mad_send_buf *send_buf,
mad_send_wr->tid = ((struct ib_mad_hdr *) send_buf->mad)->tid;
/* Timeout will be updated after send completes */
mad_send_wr->timeout = msecs_to_jiffies(send_buf->timeout_ms);
- mad_send_wr->retries = send_buf->retries;
+ mad_send_wr->max_retries = send_buf->retries;
+ mad_send_wr->retries_left = send_buf->retries;
+ send_buf->retries = 0;
/* Reference for work request to QP + response */
mad_send_wr->refcount = 1 + (mad_send_wr->timeout > 0);
mad_send_wr->status = IB_WC_SUCCESS;
@@ -1695,9 +1696,8 @@ static inline int rcv_has_same_gid(struct ib_mad_agent_private *mad_agent_priv,
u8 port_num = mad_agent_priv->agent.port_num;
u8 lmc;
- send_resp = ((struct ib_mad *)(wr->send_buf.mad))->
- mad_hdr.method & IB_MGMT_METHOD_RESP;
- rcv_resp = rwc->recv_buf.mad->mad_hdr.method & IB_MGMT_METHOD_RESP;
+ send_resp = ib_response_mad((struct ib_mad *)wr->send_buf.mad);
+ rcv_resp = ib_response_mad(rwc->recv_buf.mad);
if (send_resp == rcv_resp)
/* both requests, or both responses. GIDs different */
@@ -1933,15 +1933,6 @@ local:
if (port_priv->device->process_mad) {
int ret;
- if (!response) {
- printk(KERN_ERR PFX "No memory for response MAD\n");
- /*
- * Is it better to assume that
- * it wouldn't be processed ?
- */
- goto out;
- }
-
ret = port_priv->device->process_mad(port_priv->device, 0,
port_priv->port_num,
wc, &recv->grh,
@@ -2368,7 +2359,7 @@ static void local_completions(struct work_struct *work)
struct ib_mad_local_private *local;
struct ib_mad_agent_private *recv_mad_agent;
unsigned long flags;
- int recv = 0;
+ int free_mad;
struct ib_wc wc;
struct ib_mad_send_wc mad_send_wc;
@@ -2382,14 +2373,15 @@ static void local_completions(struct work_struct *work)
completion_list);
list_del(&local->completion_list);
spin_unlock_irqrestore(&mad_agent_priv->lock, flags);
+ free_mad = 0;
if (local->mad_priv) {
recv_mad_agent = local->recv_mad_agent;
if (!recv_mad_agent) {
printk(KERN_ERR PFX "No receive MAD agent for local completion\n");
+ free_mad = 1;
goto local_send_completion;
}
- recv = 1;
/*
* Defined behavior is to complete response
* before request
@@ -2434,7 +2426,7 @@ local_send_completion:
spin_lock_irqsave(&mad_agent_priv->lock, flags);
atomic_dec(&mad_agent_priv->refcount);
- if (!recv)
+ if (free_mad)
kmem_cache_free(ib_mad_cache, local->mad_priv);
kfree(local);
}
@@ -2445,9 +2437,12 @@ static int retry_send(struct ib_mad_send_wr_private *mad_send_wr)
{
int ret;
- if (!mad_send_wr->retries--)
+ if (!mad_send_wr->retries_left)
return -ETIMEDOUT;
+ mad_send_wr->retries_left--;
+ mad_send_wr->send_buf.retries++;
+
mad_send_wr->timeout = msecs_to_jiffies(mad_send_wr->send_buf.timeout_ms);
if (mad_send_wr->mad_agent_priv->agent.rmpp_version) {
diff --git a/drivers/infiniband/core/mad_priv.h b/drivers/infiniband/core/mad_priv.h
index 9be5cc0..05ce331 100644
--- a/drivers/infiniband/core/mad_priv.h
+++ b/drivers/infiniband/core/mad_priv.h
@@ -30,8 +30,6 @@
* 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.
- *
- * $Id: mad_priv.h 5596 2006-03-03 01:00:07Z sean.hefty $
*/
#ifndef __IB_MAD_PRIV_H__
@@ -131,7 +129,8 @@ struct ib_mad_send_wr_private {
struct ib_sge sg_list[IB_MAD_SEND_REQ_MAX_SG];
__be64 tid;
unsigned long timeout;
- int retries;
+ int max_retries;
+ int retries_left;
int retry;
int refcount;
enum ib_wc_status status;
diff --git a/drivers/infiniband/core/mad_rmpp.c b/drivers/infiniband/core/mad_rmpp.c
index d43bc62..3af2b84 100644
--- a/drivers/infiniband/core/mad_rmpp.c
+++ b/drivers/infiniband/core/mad_rmpp.c
@@ -29,8 +29,6 @@
* 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.
- *
- * $Id: mad_rmpp.c 1921 2005-03-02 22:58:44Z sean.hefty $
*/
#include "mad_priv.h"
@@ -135,7 +133,7 @@ static void ack_recv(struct mad_rmpp_recv *rmpp_recv,
msg = ib_create_send_mad(&rmpp_recv->agent->agent, recv_wc->wc->src_qp,
recv_wc->wc->pkey_index, 1, hdr_len,
0, GFP_KERNEL);
- if (!msg)
+ if (IS_ERR(msg))
return;
format_ack(msg, (struct ib_rmpp_mad *) recv_wc->recv_buf.mad, rmpp_recv);
@@ -684,7 +682,7 @@ static void process_rmpp_ack(struct ib_mad_agent_private *agent,
if (seg_num > mad_send_wr->last_ack) {
adjust_last_ack(mad_send_wr, seg_num);
- mad_send_wr->retries = mad_send_wr->send_buf.retries;
+ mad_send_wr->retries_left = mad_send_wr->max_retries;
}
mad_send_wr->newwin = newwin;
if (mad_send_wr->last_ack == mad_send_wr->send_buf.seg_count) {
diff --git a/drivers/infiniband/core/mad_rmpp.h b/drivers/infiniband/core/mad_rmpp.h
index f0616fd..3d336bf 100644
--- a/drivers/infiniband/core/mad_rmpp.h
+++ b/drivers/infiniband/core/mad_rmpp.h
@@ -28,8 +28,6 @@
* 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.
- *
- * $Id: mad_rmpp.h 1921 2005-02-25 22:58:44Z sean.hefty $
*/
#ifndef __MAD_RMPP_H__
diff --git a/drivers/infiniband/core/multicast.c b/drivers/infiniband/core/multicast.c
index dda056c..ca95426 100644
--- a/drivers/infiniband/core/multicast.c
+++ b/drivers/infiniband/core/multicast.c
@@ -73,11 +73,20 @@ struct mcast_device {
};
enum mcast_state {
- MCAST_IDLE,
MCAST_JOINING,
MCAST_MEMBER,
+ MCAST_ERROR,
+};
+
+enum mcast_group_state {
+ MCAST_IDLE,
MCAST_BUSY,
- MCAST_ERROR
+ MCAST_GROUP_ERROR,
+ MCAST_PKEY_EVENT
+};
+
+enum {
+ MCAST_INVALID_PKEY_INDEX = 0xFFFF
};
struct mcast_member;
@@ -93,9 +102,12 @@ struct mcast_group {
struct mcast_member *last_join;
int members[3];
atomic_t refcount;
- enum mcast_state state;
+ enum mcast_group_state state;
struct ib_sa_query *query;
int query_id;
+ u16 pkey_index;
+ u8 leave_state;
+ int retries;
};
struct mcast_member {
@@ -312,6 +324,7 @@ static int send_leave(struct mcast_group *group, u8 leave_state)
rec = group->rec;
rec.join_state = leave_state;
+ group->leave_state = leave_state;
ret = ib_sa_mcmember_rec_query(&sa_client, port->dev->device,
port->port_num, IB_SA_METHOD_DELETE, &rec,
@@ -350,9 +363,19 @@ static int fail_join(struct mcast_group *group, struct mcast_member *member,
static void process_group_error(struct mcast_group *group)
{
struct mcast_member *member;
- int ret;
+ int ret = 0;
+ u16 pkey_index;
+
+ if (group->state == MCAST_PKEY_EVENT)
+ ret = ib_find_pkey(group->port->dev->device,
+ group->port->port_num,
+ be16_to_cpu(group->rec.pkey), &pkey_index);
spin_lock_irq(&group->lock);
+ if (group->state == MCAST_PKEY_EVENT && !ret &&
+ group->pkey_index == pkey_index)
+ goto out;
+
while (!list_empty(&group->active_list)) {
member = list_entry(group->active_list.next,
struct mcast_member, list);
@@ -371,6 +394,7 @@ static void process_group_error(struct mcast_group *group)
}
group->rec.join_state = 0;
+out:
group->state = MCAST_BUSY;
spin_unlock_irq(&group->lock);
}
@@ -387,9 +411,9 @@ static void mcast_work_handler(struct work_struct *work)
retest:
spin_lock_irq(&group->lock);
while (!list_empty(&group->pending_list) ||
- (group->state == MCAST_ERROR)) {
+ (group->state != MCAST_BUSY)) {
- if (group->state == MCAST_ERROR) {
+ if (group->state != MCAST_BUSY) {
spin_unlock_irq(&group->lock);
process_group_error(group);
goto retest;
@@ -466,12 +490,19 @@ static void join_handler(int status, struct ib_sa_mcmember_rec *rec,
void *context)
{
struct mcast_group *group = context;
+ u16 pkey_index = MCAST_INVALID_PKEY_INDEX;
if (status)
process_join_error(group, status);
else {
+ ib_find_pkey(group->port->dev->device, group->port->port_num,
+ be16_to_cpu(rec->pkey), &pkey_index);
+
spin_lock_irq(&group->port->lock);
group->rec = *rec;
+ if (group->state == MCAST_BUSY &&
+ group->pkey_index == MCAST_INVALID_PKEY_INDEX)
+ group->pkey_index = pkey_index;
if (!memcmp(&mgid0, &group->rec.mgid, sizeof mgid0)) {
rb_erase(&group->node, &group->port->table);
mcast_insert(group->port, group, 1);
@@ -486,7 +517,11 @@ static void leave_handler(int status, struct ib_sa_mcmember_rec *rec,
{
struct mcast_group *group = context;
- mcast_work_handler(&group->work);
+ if (status && (group->retries > 0) &&
+ !send_leave(group, group->leave_state))
+ group->retries--;
+ else
+ mcast_work_handler(&group->work);
}
static struct mcast_group *acquire_group(struct mcast_port *port,
@@ -509,8 +544,10 @@ static struct mcast_group *acquire_group(struct mcast_port *port,
if (!group)
return NULL;
+ group->retries = 3;
group->port = port;
group->rec.mgid = *mgid;
+ group->pkey_index = MCAST_INVALID_PKEY_INDEX;
INIT_LIST_HEAD(&group->pending_list);
INIT_LIST_HEAD(&group->active_list);
INIT_WORK(&group->work, mcast_work_handler);
@@ -679,7 +716,8 @@ int ib_init_ah_from_mcmember(struct ib_device *device, u8 port_num,
}
EXPORT_SYMBOL(ib_init_ah_from_mcmember);
-static void mcast_groups_lost(struct mcast_port *port)
+static void mcast_groups_event(struct mcast_port *port,
+ enum mcast_group_state state)
{
struct mcast_group *group;
struct rb_node *node;
@@ -693,7 +731,8 @@ static void mcast_groups_lost(struct mcast_port *port)
atomic_inc(&group->refcount);
queue_work(mcast_wq, &group->work);
}
- group->state = MCAST_ERROR;
+ if (group->state != MCAST_GROUP_ERROR)
+ group->state = state;
spin_unlock(&group->lock);
}
spin_unlock_irqrestore(&port->lock, flags);
@@ -703,16 +742,20 @@ static void mcast_event_handler(struct ib_event_handler *handler,
struct ib_event *event)
{
struct mcast_device *dev;
+ int index;
dev = container_of(handler, struct mcast_device, event_handler);
+ index = event->element.port_num - dev->start_port;
switch (event->event) {
case IB_EVENT_PORT_ERR:
case IB_EVENT_LID_CHANGE:
case IB_EVENT_SM_CHANGE:
case IB_EVENT_CLIENT_REREGISTER:
- mcast_groups_lost(&dev->port[event->element.port_num -
- dev->start_port]);
+ mcast_groups_event(&dev->port[index], MCAST_GROUP_ERROR);
+ break;
+ case IB_EVENT_PKEY_CHANGE:
+ mcast_groups_event(&dev->port[index], MCAST_PKEY_EVENT);
break;
default:
break;
diff --git a/drivers/infiniband/core/namespace.c b/drivers/infiniband/core/namespace.c
new file mode 100644
index 0000000..a33f9e9
--- /dev/null
+++ b/drivers/infiniband/core/namespace.c
@@ -0,0 +1,95 @@
+#include <linux/spinlock_types.h>
+#include <linux/percpu.h>
+#include <linux/mount.h>
+#include <linux/module.h>
+
+struct mnt_writer {
+ /*
+ * If holding multiple instances of this lock, they
+ * must be ordered by cpu number.
+ */
+ spinlock_t lock;
+ struct lock_class_key lock_class; /* compiles out with !lockdep */
+ unsigned long count;
+ struct vfsmount *mnt;
+} ____cacheline_aligned_in_smp;
+static DEFINE_PER_CPU(struct mnt_writer, mnt_writers);
+
+int __init init_mnt_writers(void)
+{
+ int cpu;
+ for_each_possible_cpu(cpu) {
+ struct mnt_writer *writer = &per_cpu(mnt_writers, cpu);
+ spin_lock_init(&writer->lock);
+ lockdep_set_class(&writer->lock, &writer->lock_class);
+ writer->count = 0;
+ }
+ return 0;
+}
+
+static inline void __clear_mnt_count(struct mnt_writer *cpu_writer)
+{
+ if (!cpu_writer->mnt)
+ return;
+ /*
+ * This is in case anyone ever leaves an invalid,
+ * old ->mnt and a count of 0.
+ */
+ if (!cpu_writer->count)
+ return;
+ cpu_writer->count = 0;
+}
+
+static inline void use_cpu_writer_for_mount(struct mnt_writer *cpu_writer,
+ struct vfsmount *mnt)
+{
+ if (cpu_writer->mnt == mnt)
+ return;
+ __clear_mnt_count(cpu_writer);
+ cpu_writer->mnt = mnt;
+}
+
+int mnt_want_write(struct vfsmount *mnt)
+{
+ int ret = 0;
+ struct mnt_writer *cpu_writer;
+
+ cpu_writer = &get_cpu_var(mnt_writers);
+ spin_lock(&cpu_writer->lock);
+ if (__mnt_is_readonly(mnt)) {
+ ret = -EROFS;
+ goto out;
+ }
+ use_cpu_writer_for_mount(cpu_writer, mnt);
+ cpu_writer->count++;
+out:
+ spin_unlock(&cpu_writer->lock);
+ put_cpu_var(mnt_writers);
+ return ret;
+}
+EXPORT_SYMBOL(mnt_want_write);
+
+void mnt_drop_write(struct vfsmount *mnt)
+{
+ struct mnt_writer *cpu_writer;
+
+ cpu_writer = &get_cpu_var(mnt_writers);
+ spin_lock(&cpu_writer->lock);
+
+ use_cpu_writer_for_mount(cpu_writer, mnt);
+ if (cpu_writer->count > 0) {
+ cpu_writer->count--;
+ }
+
+ spin_unlock(&cpu_writer->lock);
+ /*
+ * This could be done right after the spinlock
+ * is taken because the spinlock keeps us on
+ * the cpu, and disables preemption. However,
+ * putting it here bounds the amount that
+ * __mnt_writers can underflow. Without it,
+ * we could theoretically wrap __mnt_writers.
+ */
+ put_cpu_var(mnt_writers);
+}
+EXPORT_SYMBOL(mnt_drop_write);
diff --git a/drivers/infiniband/core/packer.c b/drivers/infiniband/core/packer.c
index c972d72..019bd4b 100644
--- a/drivers/infiniband/core/packer.c
+++ b/drivers/infiniband/core/packer.c
@@ -29,8 +29,6 @@
* 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.
- *
- * $Id: packer.c 1349 2004-12-16 21:09:43Z roland $
*/
#include <linux/string.h>
diff --git a/drivers/infiniband/core/sa_query.c b/drivers/infiniband/core/sa_query.c
index a2b1c4a..59ebdd7 100644
--- a/drivers/infiniband/core/sa_query.c
+++ b/drivers/infiniband/core/sa_query.c
@@ -30,8 +30,6 @@
* 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.
- *
- * $Id: sa_query.c 2811 2005-07-06 18:11:43Z halr $
*/
#include <linux/module.h>
@@ -561,6 +559,8 @@ static void update_sm_ah(struct work_struct *work)
}
spin_lock_irq(&port->ah_lock);
+ if (port->sm_ah)
+ kref_put(&port->sm_ah->ref, free_sm_ah);
port->sm_ah = new_ah;
spin_unlock_irq(&port->ah_lock);
@@ -1188,6 +1188,7 @@ static void ib_sa_notice_resp(struct ib_sa_port *port,
struct ib_mad_send_buf *mad_buf;
struct ib_sa_mad *mad;
int ret;
+ unsigned long flags;
mad_buf = ib_create_send_mad(port->notice_agent, 1, 0, 0,
IB_MGMT_SA_HDR, IB_MGMT_SA_DATA,
@@ -1199,11 +1200,16 @@ static void ib_sa_notice_resp(struct ib_sa_port *port,
memcpy(mad, mad_recv_wc->recv_buf.mad, sizeof *mad);
mad->mad_hdr.method = IB_MGMT_METHOD_REPORT_RESP;
- spin_lock_irq(&port->ah_lock);
+ spin_lock_irqsave(&port->ah_lock, flags);
+ if (!port->sm_ah) {
+ spin_unlock_irqrestore(&port->ah_lock, flags);
+ ib_free_send_mad(mad_buf);
+ return;
+ }
kref_get(&port->sm_ah->ref);
mad_buf->context[0] = &port->sm_ah->ref;
mad_buf->ah = port->sm_ah->ah;
- spin_unlock_irq(&port->ah_lock);
+ spin_unlock_irqrestore(&port->ah_lock, flags);
ret = ib_post_send_mad(mad_buf, NULL);
if (ret)
@@ -1392,7 +1398,8 @@ static void ib_sa_remove_one(struct ib_device *device)
for (i = 0; i <= sa_dev->end_port - sa_dev->start_port; ++i) {
ib_unregister_mad_agent(sa_dev->port[i].notice_agent);
ib_unregister_mad_agent(sa_dev->port[i].agent);
- kref_put(&sa_dev->port[i].sm_ah->ref, free_sm_ah);
+ if (sa_dev->port[i].sm_ah)
+ kref_put(&sa_dev->port[i].sm_ah->ref, free_sm_ah);
}
kfree(sa_dev);
diff --git a/drivers/infiniband/core/sysfs.c b/drivers/infiniband/core/sysfs.c
index 70b77ae..1091946 100644
--- a/drivers/infiniband/core/sysfs.c
+++ b/drivers/infiniband/core/sysfs.c
@@ -30,8 +30,6 @@
* 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.
- *
- * $Id: sysfs.c 1349 2004-12-16 21:09:43Z roland $
*/
#include "core_priv.h"
@@ -96,7 +94,7 @@ static ssize_t state_show(struct ib_port *p, struct port_attribute *unused,
char *buf)
{
struct ib_port_attr attr;
- ssize_t ret;
+ ssize_t ret = -ENODEV;
static const char *state_name[] = {
[IB_PORT_NOP] = "NOP",
@@ -107,26 +105,33 @@ static ssize_t state_show(struct ib_port *p, struct port_attribute *unused,
[IB_PORT_ACTIVE_DEFER] = "ACTIVE_DEFER"
};
- ret = ib_query_port(p->ibdev, p->port_num, &attr);
- if (ret)
- return ret;
-
- return sprintf(buf, "%d: %s\n", attr.state,
- attr.state >= 0 && attr.state < ARRAY_SIZE(state_name) ?
- state_name[attr.state] : "UNKNOWN");
+ mutex_lock(&p->ibdev->sysfs_mutex);
+ if (ibdev_is_alive(p->ibdev)) {
+ ret = ib_query_port(p->ibdev, p->port_num, &attr);
+ if (!ret)
+ ret = sprintf(buf, "%d: %s\n", attr.state,
+ attr.state >= 0 &&
+ attr.state < ARRAY_SIZE(state_name) ?
+ state_name[attr.state] : "UNKNOWN");
+ }
+ mutex_unlock(&p->ibdev->sysfs_mutex);
+ return ret;
}
static ssize_t lid_show(struct ib_port *p, struct port_attribute *unused,
char *buf)
{
struct ib_port_attr attr;
- ssize_t ret;
+ ssize_t ret = -ENODEV;
- ret = ib_query_port(p->ibdev, p->port_num, &attr);
- if (ret)
- return ret;
-
- return sprintf(buf, "0x%x\n", attr.lid);
+ mutex_lock(&p->ibdev->sysfs_mutex);
+ if (ibdev_is_alive(p->ibdev)) {
+ ret = ib_query_port(p->ibdev, p->port_num, &attr);
+ if (!ret)
+ ret = sprintf(buf, "0x%x\n", attr.lid);
+ }
+ mutex_unlock(&p->ibdev->sysfs_mutex);
+ return ret;
}
static ssize_t lid_mask_count_show(struct ib_port *p,
@@ -134,52 +139,64 @@ static ssize_t lid_mask_count_show(struct ib_port *p,
char *buf)
{
struct ib_port_attr attr;
- ssize_t ret;
-
- ret = ib_query_port(p->ibdev, p->port_num, &attr);
- if (ret)
- return ret;
+ ssize_t ret = -ENODEV;
- return sprintf(buf, "%d\n", attr.lmc);
+ mutex_lock(&p->ibdev->sysfs_mutex);
+ if (ibdev_is_alive(p->ibdev)) {
+ ret = ib_query_port(p->ibdev, p->port_num, &attr);
+ if (!ret)
+ ret = sprintf(buf, "%d\n", attr.lmc);
+ }
+ mutex_unlock(&p->ibdev->sysfs_mutex);
+ return ret;
}
static ssize_t sm_lid_show(struct ib_port *p, struct port_attribute *unused,
char *buf)
{
struct ib_port_attr attr;
- ssize_t ret;
-
- ret = ib_query_port(p->ibdev, p->port_num, &attr);
- if (ret)
- return ret;
+ ssize_t ret = -ENODEV;
- return sprintf(buf, "0x%x\n", attr.sm_lid);
+ mutex_lock(&p->ibdev->sysfs_mutex);
+ if (ibdev_is_alive(p->ibdev)) {
+ ret = ib_query_port(p->ibdev, p->port_num, &attr);
+ if (!ret)
+ ret = sprintf(buf, "0x%x\n", attr.sm_lid);
+ }
+ mutex_unlock(&p->ibdev->sysfs_mutex);
+ return ret;
}
static ssize_t sm_sl_show(struct ib_port *p, struct port_attribute *unused,
char *buf)
{
struct ib_port_attr attr;
- ssize_t ret;
+ ssize_t ret = -ENODEV;
- ret = ib_query_port(p->ibdev, p->port_num, &attr);
- if (ret)
- return ret;
-
- return sprintf(buf, "%d\n", attr.sm_sl);
+ mutex_lock(&p->ibdev->sysfs_mutex);
+ if (ibdev_is_alive(p->ibdev)) {
+ ret = ib_query_port(p->ibdev, p->port_num, &attr);
+ if (!ret)
+ ret = sprintf(buf, "%d\n", attr.sm_sl);
+ }
+ mutex_unlock(&p->ibdev->sysfs_mutex);
+ return ret;
}
static ssize_t cap_mask_show(struct ib_port *p, struct port_attribute *unused,
char *buf)
{
struct ib_port_attr attr;
- ssize_t ret;
-
- ret = ib_query_port(p->ibdev, p->port_num, &attr);
- if (ret)
- return ret;
+ ssize_t ret = -ENODEV;
- return sprintf(buf, "0x%08x\n", attr.port_cap_flags);
+ mutex_lock(&p->ibdev->sysfs_mutex);
+ if (ibdev_is_alive(p->ibdev)) {
+ ret = ib_query_port(p->ibdev, p->port_num, &attr);
+ if (!ret)
+ ret = sprintf(buf, "0x%08x\n", attr.port_cap_flags);
+ }
+ mutex_unlock(&p->ibdev->sysfs_mutex);
+ return ret;
}
static ssize_t rate_show(struct ib_port *p, struct port_attribute *unused,
@@ -188,24 +205,33 @@ static ssize_t rate_show(struct ib_port *p, struct port_attribute *unused,
struct ib_port_attr attr;
char *speed = "";
int rate;
- ssize_t ret;
-
- ret = ib_query_port(p->ibdev, p->port_num, &attr);
- if (ret)
- return ret;
-
- switch (attr.active_speed) {
- case 2: speed = " DDR"; break;
- case 4: speed = " QDR"; break;
+ ssize_t ret = -ENODEV;
+
+ mutex_lock(&p->ibdev->sysfs_mutex);
+ if (ibdev_is_alive(p->ibdev)) {
+ ret = ib_query_port(p->ibdev, p->port_num, &attr);
+ if (!ret) {
+ switch (attr.active_speed) {
+ case 2: speed = " DDR"; break;
+ case 4: speed = " QDR"; break;
+ }
+
+ rate = 25 * ib_width_enum_to_int(attr.active_width) *
+ attr.active_speed;
+ if (rate < 0) {
+ ret = -EINVAL;
+ goto out;
+ }
+
+ ret = sprintf(buf, "%d%s Gb/sec (%dX%s)\n",
+ rate / 10, rate % 10 ? ".5" : "",
+ ib_width_enum_to_int(attr.active_width),
+ speed);
+ }
}
-
- rate = 25 * ib_width_enum_to_int(attr.active_width) * attr.active_speed;
- if (rate < 0)
- return -EINVAL;
-
- return sprintf(buf, "%d%s Gb/sec (%dX%s)\n",
- rate / 10, rate % 10 ? ".5" : "",
- ib_width_enum_to_int(attr.active_width), speed);
+out:
+ mutex_unlock(&p->ibdev->sysfs_mutex);
+ return ret;
}
static ssize_t phys_state_show(struct ib_port *p, struct port_attribute *unused,
@@ -213,22 +239,26 @@ static ssize_t phys_state_show(struct ib_port *p, struct port_attribute *unused,
{
struct ib_port_attr attr;
- ssize_t ret;
-
- ret = ib_query_port(p->ibdev, p->port_num, &attr);
- if (ret)
- return ret;
-
- switch (attr.phys_state) {
- case 1: return sprintf(buf, "1: Sleep\n");
- case 2: return sprintf(buf, "2: Polling\n");
- case 3: return sprintf(buf, "3: Disabled\n");
- case 4: return sprintf(buf, "4: PortConfigurationTraining\n");
- case 5: return sprintf(buf, "5: LinkUp\n");
- case 6: return sprintf(buf, "6: LinkErrorRecovery\n");
- case 7: return sprintf(buf, "7: Phy Test\n");
- default: return sprintf(buf, "%d: <unknown>\n", attr.phys_state);
+ ssize_t ret = -ENODEV;
+
+ mutex_lock(&p->ibdev->sysfs_mutex);
+ if (ibdev_is_alive(p->ibdev)) {
+ ret = ib_query_port(p->ibdev, p->port_num, &attr);
+ if (!ret) {
+ switch (attr.phys_state) {
+ case 1: ret = sprintf(buf, "1: Sleep\n");
+ case 2: ret = sprintf(buf, "2: Polling\n");
+ case 3: ret = sprintf(buf, "3: Disabled\n");
+ case 4: ret = sprintf(buf, "4: PortConfigurationTraining\n");
+ case 5: ret = sprintf(buf, "5: LinkUp\n");
+ case 6: ret = sprintf(buf, "6: LinkErrorRecovery\n");
+ case 7: ret = sprintf(buf, "7: Phy Test\n");
+ default: ret = sprintf(buf, "%d: <unknown>\n", attr.phys_state);
+ }
+ }
}
+ mutex_unlock(&p->ibdev->sysfs_mutex);
+ return ret;
}
static PORT_ATTR_RO(state);
@@ -258,21 +288,24 @@ static ssize_t show_port_gid(struct ib_port *p, struct port_attribute *attr,
struct port_table_attribute *tab_attr =
container_of(attr, struct port_table_attribute, attr);
union ib_gid gid;
- ssize_t ret;
-
- ret = ib_query_gid(p->ibdev, p->port_num, tab_attr->index, &gid);
- if (ret)
- return ret;
-
- return sprintf(buf, "%04x:%04x:%04x:%04x:%04x:%04x:%04x:%04x\n",
- be16_to_cpu(((__be16 *) gid.raw)[0]),
- be16_to_cpu(((__be16 *) gid.raw)[1]),
- be16_to_cpu(((__be16 *) gid.raw)[2]),
- be16_to_cpu(((__be16 *) gid.raw)[3]),
- be16_to_cpu(((__be16 *) gid.raw)[4]),
- be16_to_cpu(((__be16 *) gid.raw)[5]),
- be16_to_cpu(((__be16 *) gid.raw)[6]),
- be16_to_cpu(((__be16 *) gid.raw)[7]));
+ ssize_t ret = -ENODEV;
+
+ mutex_lock(&p->ibdev->sysfs_mutex);
+ if (ibdev_is_alive(p->ibdev)) {
+ ret = ib_query_gid(p->ibdev, p->port_num, tab_attr->index, &gid);
+ if (!ret)
+ ret = sprintf(buf, "%04x:%04x:%04x:%04x:%04x:%04x:%04x:%04x\n",
+ be16_to_cpu(((__be16 *) gid.raw)[0]),
+ be16_to_cpu(((__be16 *) gid.raw)[1]),
+ be16_to_cpu(((__be16 *) gid.raw)[2]),
+ be16_to_cpu(((__be16 *) gid.raw)[3]),
+ be16_to_cpu(((__be16 *) gid.raw)[4]),
+ be16_to_cpu(((__be16 *) gid.raw)[5]),
+ be16_to_cpu(((__be16 *) gid.raw)[6]),
+ be16_to_cpu(((__be16 *) gid.raw)[7]));
+ }
+ mutex_unlock(&p->ibdev->sysfs_mutex);
+ return ret;
}
static ssize_t show_port_pkey(struct ib_port *p, struct port_attribute *attr,
@@ -281,13 +314,16 @@ static ssize_t show_port_pkey(struct ib_port *p, struct port_attribute *attr,
struct port_table_attribute *tab_attr =
container_of(attr, struct port_table_attribute, attr);
u16 pkey;
- ssize_t ret;
-
- ret = ib_query_pkey(p->ibdev, p->port_num, tab_attr->index, &pkey);
- if (ret)
- return ret;
+ ssize_t ret = -ENODEV;
- return sprintf(buf, "0x%04x\n", pkey);
+ mutex_lock(&p->ibdev->sysfs_mutex);
+ if (ibdev_is_alive(p->ibdev)) {
+ ret = ib_query_pkey(p->ibdev, p->port_num, tab_attr->index, &pkey);
+ if (!ret)
+ ret = sprintf(buf, "0x%04x\n", pkey);
+ }
+ mutex_unlock(&p->ibdev->sysfs_mutex);
+ return ret;
}
#define PORT_PMA_ATTR(_name, _counter, _width, _offset) \
@@ -310,6 +346,12 @@ static ssize_t show_pma_counter(struct ib_port *p, struct port_attribute *attr,
if (!p->ibdev->process_mad)
return sprintf(buf, "N/A (no PMA)\n");
+ mutex_lock(&p->ibdev->sysfs_mutex);
+ if (!ibdev_is_alive(p->ibdev)) {
+ ret = -ENODEV;
+ goto out;
+ }
+
in_mad = kzalloc(sizeof *in_mad, GFP_KERNEL);
out_mad = kmalloc(sizeof *out_mad, GFP_KERNEL);
if (!in_mad || !out_mad) {
@@ -356,7 +398,7 @@ static ssize_t show_pma_counter(struct ib_port *p, struct port_attribute *attr,
out:
kfree(in_mad);
kfree(out_mad);
-
+ mutex_unlock(&p->ibdev->sysfs_mutex);
return ret;
}
@@ -376,6 +418,12 @@ static PORT_PMA_ATTR(port_xmit_data , 12, 32, 192);
static PORT_PMA_ATTR(port_rcv_data , 13, 32, 224);
static PORT_PMA_ATTR(port_xmit_packets , 14, 32, 256);
static PORT_PMA_ATTR(port_rcv_packets , 15, 32, 288);
+/*
+ * There is no bit allocated for port_xmit_wait in the CounterSelect field
+ * (IB spec). However, since this bit is ignored when reading
+ * (show_pma_counter), the _counter field of port_xmit_wait can be set to zero.
+ */
+static PORT_PMA_ATTR(port_xmit_wait , 0, 32, 320);
static struct attribute *pma_attrs[] = {
&port_pma_attr_symbol_error.attr.attr,
@@ -394,6 +442,7 @@ static struct attribute *pma_attrs[] = {
&port_pma_attr_port_rcv_data.attr.attr,
&port_pma_attr_port_xmit_packets.attr.attr,
&port_pma_attr_port_rcv_packets.attr.attr,
+ &port_pma_attr_port_xmit_wait.attr.attr,
NULL
};
@@ -511,19 +560,10 @@ static int add_port(struct ib_device *device, int port_num)
p->ibdev = device;
p->port_num = port_num;
- p->kobj.ktype = &port_type;
-
- p->kobj.parent = kobject_get(&device->ports_parent);
- if (!p->kobj.parent) {
- ret = -EBUSY;
- goto err;
- }
-
- ret = kobject_set_name(&p->kobj, "%d", port_num);
- if (ret)
- goto err_put;
- ret = kobject_register(&p->kobj);
+ ret = kobject_init_and_add(&p->kobj, &port_type,
+ kobject_get(device->ports_parent),
+ "%d", port_num);
if (ret)
goto err_put;
@@ -552,6 +592,7 @@ static int add_port(struct ib_device *device, int port_num)
list_add_tail(&p->kobj.entry, &device->port_list);
+ kobject_uevent(&p->kobj, KOBJ_ADD);
return 0;
err_free_pkey:
@@ -573,9 +614,7 @@ err_remove_pma:
sysfs_remove_group(&p->kobj, &pma_group);
err_put:
- kobject_put(&device->ports_parent);
-
-err:
+ kobject_put(device->ports_parent);
kfree(p);
return ret;
}
@@ -600,20 +639,20 @@ static ssize_t show_sys_image_guid(struct class_device *cdev, char *buf)
{
struct ib_device *dev = container_of(cdev, struct ib_device, class_dev);
struct ib_device_attr attr;
- ssize_t ret;
-
- if (!ibdev_is_alive(dev))
- return -ENODEV;
-
- ret = ib_query_device(dev, &attr);
- if (ret)
- return ret;
-
- return sprintf(buf, "%04x:%04x:%04x:%04x\n",
- be16_to_cpu(((__be16 *) &attr.sys_image_guid)[0]),
- be16_to_cpu(((__be16 *) &attr.sys_image_guid)[1]),
- be16_to_cpu(((__be16 *) &attr.sys_image_guid)[2]),
- be16_to_cpu(((__be16 *) &attr.sys_image_guid)[3]));
+ ssize_t ret = -ENODEV;
+
+ mutex_lock(&dev->sysfs_mutex);
+ if (ibdev_is_alive(dev)) {
+ ret = ib_query_device(dev, &attr);
+ if (!ret)
+ ret = sprintf(buf, "%04x:%04x:%04x:%04x\n",
+ be16_to_cpu(((__be16 *) &attr.sys_image_guid)[0]),
+ be16_to_cpu(((__be16 *) &attr.sys_image_guid)[1]),
+ be16_to_cpu(((__be16 *) &attr.sys_image_guid)[2]),
+ be16_to_cpu(((__be16 *) &attr.sys_image_guid)[3]));
+ }
+ mutex_unlock(&dev->sysfs_mutex);
+ return ret;
}
static ssize_t show_node_guid(struct class_device *cdev, char *buf)
@@ -642,24 +681,26 @@ static ssize_t set_node_desc(struct class_device *cdev, const char *buf,
{
struct ib_device *dev = container_of(cdev, struct ib_device, class_dev);
struct ib_device_modify desc = {};
- int ret;
+ int ret = -ENODEV;
if (!dev->modify_device)
return -EIO;
memcpy(desc.node_desc, buf, min_t(int, count, 64));
- ret = ib_modify_device(dev, IB_DEVICE_MODIFY_NODE_DESC, &desc);
- if (ret)
- return ret;
-
- return count;
+ mutex_lock(&dev->sysfs_mutex);
+ if (ibdev_is_alive(dev)) {
+ ret = ib_modify_device(dev, IB_DEVICE_MODIFY_NODE_DESC, &desc);
+ if (!ret)
+ ret = count;
+ }
+ mutex_unlock(&dev->sysfs_mutex);
+ return ret;
}
static CLASS_DEVICE_ATTR(node_type, S_IRUGO, show_node_type, NULL);
static CLASS_DEVICE_ATTR(sys_image_guid, S_IRUGO, show_sys_image_guid, NULL);
static CLASS_DEVICE_ATTR(node_guid, S_IRUGO, show_node_guid, NULL);
-static CLASS_DEVICE_ATTR(node_desc, S_IRUGO | S_IWUSR, show_node_desc,
- set_node_desc);
+static CLASS_DEVICE_ATTR(node_desc, S_IRUGO | S_IWUSR, show_node_desc, set_node_desc);
static struct class_device_attribute *ib_class_attributes[] = {
&class_device_attr_node_type,
@@ -674,6 +715,124 @@ static struct class ib_class = {
.uevent = ib_device_uevent,
};
+/* Show a given an attribute in the statistics group */
+static ssize_t show_protocol_stat(struct class_device *cdev,
+ char *buf,
+ unsigned offset)
+{
+ struct ib_device *dev = container_of(cdev, struct ib_device, class_dev);
+ union rdma_protocol_stats stats;
+ ssize_t ret = -ENODEV;
+
+ mutex_lock(&dev->sysfs_mutex);
+ if (ibdev_is_alive(dev)) {
+ ret = dev->get_protocol_stats(dev, &stats);
+ if (!ret)
+ ret = sprintf(buf, "%llu\n",
+ (unsigned long long)
+ ((u64 *) &stats)[offset]);
+ }
+ mutex_unlock(&dev->sysfs_mutex);
+ return ret;
+}
+
+/* generate a read-only iwarp statistics attribute */
+#define IW_STATS_ENTRY(name) \
+static ssize_t show_##name(struct class_device *cdev, \
+ char *buf) \
+{ \
+ return show_protocol_stat(cdev, buf, \
+ offsetof(struct iw_protocol_stats, name) / \
+ sizeof (u64)); \
+} \
+static CLASS_DEVICE_ATTR(name, S_IRUGO, show_##name, NULL)
+
+IW_STATS_ENTRY(ipInReceives);
+IW_STATS_ENTRY(ipInHdrErrors);
+IW_STATS_ENTRY(ipInTooBigErrors);
+IW_STATS_ENTRY(ipInNoRoutes);
+IW_STATS_ENTRY(ipInAddrErrors);
+IW_STATS_ENTRY(ipInUnknownProtos);
+IW_STATS_ENTRY(ipInTruncatedPkts);
+IW_STATS_ENTRY(ipInDiscards);
+IW_STATS_ENTRY(ipInDelivers);
+IW_STATS_ENTRY(ipOutForwDatagrams);
+IW_STATS_ENTRY(ipOutRequests);
+IW_STATS_ENTRY(ipOutDiscards);
+IW_STATS_ENTRY(ipOutNoRoutes);
+IW_STATS_ENTRY(ipReasmTimeout);
+IW_STATS_ENTRY(ipReasmReqds);
+IW_STATS_ENTRY(ipReasmOKs);
+IW_STATS_ENTRY(ipReasmFails);
+IW_STATS_ENTRY(ipFragOKs);
+IW_STATS_ENTRY(ipFragFails);
+IW_STATS_ENTRY(ipFragCreates);
+IW_STATS_ENTRY(ipInMcastPkts);
+IW_STATS_ENTRY(ipOutMcastPkts);
+IW_STATS_ENTRY(ipInBcastPkts);
+IW_STATS_ENTRY(ipOutBcastPkts);
+IW_STATS_ENTRY(tcpRtoAlgorithm);
+IW_STATS_ENTRY(tcpRtoMin);
+IW_STATS_ENTRY(tcpRtoMax);
+IW_STATS_ENTRY(tcpMaxConn);
+IW_STATS_ENTRY(tcpActiveOpens);
+IW_STATS_ENTRY(tcpPassiveOpens);
+IW_STATS_ENTRY(tcpAttemptFails);
+IW_STATS_ENTRY(tcpEstabResets);
+IW_STATS_ENTRY(tcpCurrEstab);
+IW_STATS_ENTRY(tcpInSegs);
+IW_STATS_ENTRY(tcpOutSegs);
+IW_STATS_ENTRY(tcpRetransSegs);
+IW_STATS_ENTRY(tcpInErrs);
+IW_STATS_ENTRY(tcpOutRsts);
+
+static struct attribute *iw_proto_stats_attrs[] = {
+ &class_device_attr_ipInReceives.attr,
+ &class_device_attr_ipInHdrErrors.attr,
+ &class_device_attr_ipInTooBigErrors.attr,
+ &class_device_attr_ipInNoRoutes.attr,
+ &class_device_attr_ipInAddrErrors.attr,
+ &class_device_attr_ipInUnknownProtos.attr,
+ &class_device_attr_ipInTruncatedPkts.attr,
+ &class_device_attr_ipInDiscards.attr,
+ &class_device_attr_ipInDelivers.attr,
+ &class_device_attr_ipOutForwDatagrams.attr,
+ &class_device_attr_ipOutRequests.attr,
+ &class_device_attr_ipOutDiscards.attr,
+ &class_device_attr_ipOutNoRoutes.attr,
+ &class_device_attr_ipReasmTimeout.attr,
+ &class_device_attr_ipReasmReqds.attr,
+ &class_device_attr_ipReasmOKs.attr,
+ &class_device_attr_ipReasmFails.attr,
+ &class_device_attr_ipFragOKs.attr,
+ &class_device_attr_ipFragFails.attr,
+ &class_device_attr_ipFragCreates.attr,
+ &class_device_attr_ipInMcastPkts.attr,
+ &class_device_attr_ipOutMcastPkts.attr,
+ &class_device_attr_ipInBcastPkts.attr,
+ &class_device_attr_ipOutBcastPkts.attr,
+ &class_device_attr_tcpRtoAlgorithm.attr,
+ &class_device_attr_tcpRtoMin.attr,
+ &class_device_attr_tcpRtoMax.attr,
+ &class_device_attr_tcpMaxConn.attr,
+ &class_device_attr_tcpActiveOpens.attr,
+ &class_device_attr_tcpPassiveOpens.attr,
+ &class_device_attr_tcpAttemptFails.attr,
+ &class_device_attr_tcpEstabResets.attr,
+ &class_device_attr_tcpCurrEstab.attr,
+ &class_device_attr_tcpInSegs.attr,
+ &class_device_attr_tcpOutSegs.attr,
+ &class_device_attr_tcpRetransSegs.attr,
+ &class_device_attr_tcpInErrs.attr,
+ &class_device_attr_tcpOutRsts.attr,
+ NULL
+};
+
+static struct attribute_group iw_stats_group = {
+ .name = "proto_stats",
+ .attrs = iw_proto_stats_attrs,
+};
+
int ib_device_register_sysfs(struct ib_device *device)
{
struct class_device *class_dev = &device->class_dev;
@@ -697,17 +856,12 @@ int ib_device_register_sysfs(struct ib_device *device)
goto err_unregister;
}
- device->ports_parent.parent = kobject_get(&class_dev->kobj);
- if (!device->ports_parent.parent) {
- ret = -EBUSY;
- goto err_unregister;
- }
- ret = kobject_set_name(&device->ports_parent, "ports");
- if (ret)
- goto err_put;
- ret = kobject_register(&device->ports_parent);
- if (ret)
+ device->ports_parent = kobject_create_and_add("ports",
+ kobject_get(&class_dev->kobj));
+ if (!device->ports_parent) {
+ ret = -ENOMEM;
goto err_put;
+ }
if (device->node_type == RDMA_NODE_IB_SWITCH) {
ret = add_port(device, 0);
@@ -721,6 +875,12 @@ int ib_device_register_sysfs(struct ib_device *device)
}
}
+ if (device->node_type == RDMA_NODE_RNIC && device->get_protocol_stats) {
+ ret = sysfs_create_group(&class_dev->kobj, &iw_stats_group);
+ if (ret)
+ goto err_put;
+ }
+
return 0;
err_put:
@@ -738,7 +898,7 @@ err_put:
}
}
- kobject_put(&class_dev->kobj);
+ kobject_unregister(&class_dev->kobj);
err_unregister:
class_device_unregister(class_dev);
@@ -761,7 +921,9 @@ void ib_device_unregister_sysfs(struct ib_device *device)
kobject_unregister(p);
}
- kobject_unregister(&device->ports_parent);
+ kobject_put(device->ports_parent);
+ /* WA for memory leak */
+ kfree(device->ports_parent);
class_device_unregister(&device->class_dev);
}
diff --git a/drivers/infiniband/core/ucm.c b/drivers/infiniband/core/ucm.c
index 424983f..1c967e1 100644
--- a/drivers/infiniband/core/ucm.c
+++ b/drivers/infiniband/core/ucm.c
@@ -29,8 +29,6 @@
* 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.
- *
- * $Id: ucm.c 4311 2005-12-05 18:42:01Z sean.hefty $
*/
#include <linux/completion.h>
@@ -106,6 +104,9 @@ enum {
IB_UCM_MAX_DEVICES = 32
};
+/* ib_cm and ib_user_cm modules share /sys/class/infiniband_cm */
+extern struct class cm_class;
+
#define IB_UCM_BASE_DEV MKDEV(IB_UCM_MAJOR, IB_UCM_BASE_MINOR)
static void ib_ucm_add_one(struct ib_device *device);
@@ -1152,6 +1153,14 @@ static unsigned int ib_ucm_poll(struct file *filp,
return mask;
}
+/*
+ * ib_ucm_open() does not need the BKL:
+ *
+ * - no global state is referred to;
+ * - there is no ioctl method to race against;
+ * - no further module initialization is required for open to work
+ * after the device is registered.
+ */
static int ib_ucm_open(struct inode *inode, struct file *filp)
{
struct ib_ucm_file *file;
@@ -1199,7 +1208,7 @@ static int ib_ucm_close(struct inode *inode, struct file *filp)
return 0;
}
-static void ib_ucm_release_class_dev(struct class_device *class_dev)
+static void ucm_release_class_dev(struct class_device *class_dev)
{
struct ib_ucm_device *dev;
@@ -1217,11 +1226,6 @@ static const struct file_operations ucm_fops = {
.poll = ib_ucm_poll,
};
-static struct class ucm_class = {
- .name = "infiniband_cm",
- .release = ib_ucm_release_class_dev
-};
-
static ssize_t show_ibdev(struct class_device *class_dev, char *buf)
{
struct ib_ucm_device *dev;
@@ -1257,9 +1261,10 @@ static void ib_ucm_add_one(struct ib_device *device)
if (cdev_add(&ucm_dev->dev, IB_UCM_BASE_DEV + ucm_dev->devnum, 1))
goto err;
- ucm_dev->class_dev.class = &ucm_class;
+ ucm_dev->class_dev.class = &cm_class;
ucm_dev->class_dev.dev = device->dma_device;
ucm_dev->class_dev.devt = ucm_dev->dev.dev;
+ ucm_dev->class_dev.release = ucm_release_class_dev;
snprintf(ucm_dev->class_dev.class_id, BUS_ID_SIZE, "ucm%d",
ucm_dev->devnum);
if (class_device_register(&ucm_dev->class_dev))
@@ -1306,40 +1311,34 @@ static int __init ib_ucm_init(void)
"infiniband_cm");
if (ret) {
printk(KERN_ERR "ucm: couldn't register device number\n");
- goto err;
+ goto error1;
}
- ret = class_register(&ucm_class);
- if (ret) {
- printk(KERN_ERR "ucm: couldn't create class infiniband_cm\n");
- goto err_chrdev;
- }
-
- ret = class_create_file(&ucm_class, &class_attr_abi_version);
+ ret = class_create_file(&cm_class, &class_attr_abi_version);
if (ret) {
printk(KERN_ERR "ucm: couldn't create abi_version attribute\n");
- goto err_class;
+ goto error2;
}
ret = ib_register_client(&ucm_client);
if (ret) {
printk(KERN_ERR "ucm: couldn't register client\n");
- goto err_class;
+ goto error3;
}
return 0;
-err_class:
- class_unregister(&ucm_class);
-err_chrdev:
+error3:
+ class_remove_file(&cm_class, &class_attr_abi_version);
+error2:
unregister_chrdev_region(IB_UCM_BASE_DEV, IB_UCM_MAX_DEVICES);
-err:
+error1:
return ret;
}
static void __exit ib_ucm_cleanup(void)
{
ib_unregister_client(&ucm_client);
- class_unregister(&ucm_class);
+ class_remove_file(&cm_class, &class_attr_abi_version);
unregister_chrdev_region(IB_UCM_BASE_DEV, IB_UCM_MAX_DEVICES);
idr_destroy(&ctx_id_table);
}
diff --git a/drivers/infiniband/core/ucma.c b/drivers/infiniband/core/ucma.c
index ef52dbb..31f3671 100644
--- a/drivers/infiniband/core/ucma.c
+++ b/drivers/infiniband/core/ucma.c
@@ -31,6 +31,7 @@
*/
#include <linux/completion.h>
+#include <linux/file.h>
#include <linux/mutex.h>
#include <linux/poll.h>
#include <linux/idr.h>
@@ -80,9 +81,7 @@ struct ucma_multicast {
u64 uid;
struct list_head list;
- struct sockaddr addr;
- u8 pad[sizeof(struct sockaddr_in6) -
- sizeof(struct sockaddr)];
+ struct sockaddr_storage addr;
};
struct ucma_event {
@@ -602,18 +601,18 @@ static ssize_t ucma_query_route(struct ucma_file *file,
return PTR_ERR(ctx);
memset(&resp, 0, sizeof resp);
- addr = &ctx->cm_id->route.addr.src_addr;
+ addr = (struct sockaddr *) &ctx->cm_id->route.addr.src_addr;
memcpy(&resp.src_addr, addr, addr->sa_family == AF_INET ?
sizeof(struct sockaddr_in) :
sizeof(struct sockaddr_in6));
- addr = &ctx->cm_id->route.addr.dst_addr;
+ addr = (struct sockaddr *) &ctx->cm_id->route.addr.dst_addr;
memcpy(&resp.dst_addr, addr, addr->sa_family == AF_INET ?
sizeof(struct sockaddr_in) :
sizeof(struct sockaddr_in6));
if (!ctx->cm_id->device)
goto out;
- resp.node_guid = ctx->cm_id->device->node_guid;
+ resp.node_guid = (__force __u64) ctx->cm_id->device->node_guid;
resp.port_num = ctx->cm_id->port_num;
switch (rdma_node_get_transport(ctx->cm_id->device->node_type)) {
case RDMA_TRANSPORT_IB:
@@ -905,14 +904,14 @@ static ssize_t ucma_join_multicast(struct ucma_file *file,
mutex_lock(&file->mut);
mc = ucma_alloc_multicast(ctx);
- if (IS_ERR(mc)) {
- ret = PTR_ERR(mc);
+ if (mc == NULL) {
+ ret = -ENOMEM;
goto err1;
}
mc->uid = cmd.uid;
memcpy(&mc->addr, &cmd.addr, sizeof cmd.addr);
- ret = rdma_join_multicast(ctx->cm_id, &mc->addr, mc);
+ ret = rdma_join_multicast(ctx->cm_id, (struct sockaddr *) &mc->addr, mc);
if (ret)
goto err2;
@@ -928,7 +927,7 @@ static ssize_t ucma_join_multicast(struct ucma_file *file,
return 0;
err3:
- rdma_leave_multicast(ctx->cm_id, &mc->addr);
+ rdma_leave_multicast(ctx->cm_id, (struct sockaddr *) &mc->addr);
ucma_cleanup_mc_events(mc);
err2:
mutex_lock(&mut);
@@ -974,7 +973,7 @@ static ssize_t ucma_leave_multicast(struct ucma_file *file,
goto out;
}
- rdma_leave_multicast(mc->ctx->cm_id, &mc->addr);
+ rdma_leave_multicast(mc->ctx->cm_id, (struct sockaddr *) &mc->addr);
mutex_lock(&mc->ctx->file->mut);
ucma_cleanup_mc_events(mc);
list_del(&mc->list);
@@ -991,6 +990,96 @@ out:
return ret;
}
+static void ucma_lock_files(struct ucma_file *file1, struct ucma_file *file2)
+{
+ /* Acquire mutex's based on pointer comparison to prevent deadlock. */
+ if (file1 < file2) {
+ mutex_lock(&file1->mut);
+ mutex_lock(&file2->mut);
+ } else {
+ mutex_lock(&file2->mut);
+ mutex_lock(&file1->mut);
+ }
+}
+
+static void ucma_unlock_files(struct ucma_file *file1, struct ucma_file *file2)
+{
+ if (file1 < file2) {
+ mutex_unlock(&file2->mut);
+ mutex_unlock(&file1->mut);
+ } else {
+ mutex_unlock(&file1->mut);
+ mutex_unlock(&file2->mut);
+ }
+}
+
+static void ucma_move_events(struct ucma_context *ctx, struct ucma_file *file)
+{
+ struct ucma_event *uevent, *tmp;
+
+ list_for_each_entry_safe(uevent, tmp, &ctx->file->event_list, list)
+ if (uevent->ctx == ctx)
+ list_move_tail(&uevent->list, &file->event_list);
+}
+
+static ssize_t ucma_migrate_id(struct ucma_file *new_file,
+ const char __user *inbuf,
+ int in_len, int out_len)
+{
+ struct rdma_ucm_migrate_id cmd;
+ struct rdma_ucm_migrate_resp resp;
+ struct ucma_context *ctx;
+ struct file *filp;
+ struct ucma_file *cur_file;
+ int ret = 0;
+
+ if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
+ return -EFAULT;
+
+ /* Get current fd to protect against it being closed */
+ filp = fget(cmd.fd);
+ if (!filp)
+ return -ENOENT;
+
+ /* Validate current fd and prevent destruction of id. */
+ ctx = ucma_get_ctx(filp->private_data, cmd.id);
+ if (IS_ERR(ctx)) {
+ ret = PTR_ERR(ctx);
+ goto file_put;
+ }
+
+ cur_file = ctx->file;
+ if (cur_file == new_file) {
+ resp.events_reported = ctx->events_reported;
+ goto response;
+ }
+
+ /*
+ * Migrate events between fd's, maintaining order, and avoiding new
+ * events being added before existing events.
+ */
+ ucma_lock_files(cur_file, new_file);
+ mutex_lock(&mut);
+
+ list_move_tail(&ctx->list, &new_file->ctx_list);
+ ucma_move_events(ctx, new_file);
+ ctx->file = new_file;
+ resp.events_reported = ctx->events_reported;
+
+ mutex_unlock(&mut);
+ ucma_unlock_files(cur_file, new_file);
+
+response:
+ if (copy_to_user((void __user *)(unsigned long)cmd.response,
+ &resp, sizeof(resp)))
+ ret = -EFAULT;
+
+ ucma_put_ctx(ctx);
+file_put:
+ fput(filp);
+ return ret;
+}
+
static ssize_t (*ucma_cmd_table[])(struct ucma_file *file,
const char __user *inbuf,
int in_len, int out_len) = {
@@ -1012,6 +1101,7 @@ static ssize_t (*ucma_cmd_table[])(struct ucma_file *file,
[RDMA_USER_CM_CMD_NOTIFY] = ucma_notify,
[RDMA_USER_CM_CMD_JOIN_MCAST] = ucma_join_multicast,
[RDMA_USER_CM_CMD_LEAVE_MCAST] = ucma_leave_multicast,
+ [RDMA_USER_CM_CMD_MIGRATE_ID] = ucma_migrate_id
};
static ssize_t ucma_write(struct file *filp, const char __user *buf,
@@ -1056,6 +1146,14 @@ static unsigned int ucma_poll(struct file *filp, struct poll_table_struct *wait)
return mask;
}
+/*
+ * ucma_open() does not need the BKL:
+ *
+ * - no global state is referred to;
+ * - there is no ioctl method to race against;
+ * - no further module initialization is required for open to work
+ * after the device is registered.
+ */
static int ucma_open(struct inode *inode, struct file *filp)
{
struct ucma_file *file;
diff --git a/drivers/infiniband/core/ud_header.c b/drivers/infiniband/core/ud_header.c
index 997c07d..9d74fb8 100644
--- a/drivers/infiniband/core/ud_header.c
+++ b/drivers/infiniband/core/ud_header.c
@@ -29,8 +29,6 @@
* 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.
- *
- * $Id: ud_header.c 1349 2004-12-16 21:09:43Z roland $
*/
#include <linux/errno.h>
@@ -243,6 +241,36 @@ void ib_ud_header_init(int payload_bytes,
EXPORT_SYMBOL(ib_ud_header_init);
/**
+ * ib_lrh_header_pack - Pack LRH header struct into wire format
+ * @lrh:unpacked LRH header struct
+ * @buf:Buffer to pack into
+ *
+ * ib_lrh_header_pack() packs the LRH header structure @lrh into
+ * wire format in the buffer @buf.
+ */
+int ib_lrh_header_pack(struct ib_unpacked_lrh *lrh, void *buf)
+{
+ ib_pack(lrh_table, ARRAY_SIZE(lrh_table), lrh, buf);
+ return 0;
+}
+EXPORT_SYMBOL(ib_lrh_header_pack);
+
+/**
+ * ib_lrh_header_unpack - Unpack LRH structure from wire format
+ * @lrh:unpacked LRH header struct
+ * @buf:Buffer to pack into
+ *
+ * ib_lrh_header_unpack() unpacks the LRH header structure from
+ * wire format (in buf) into @lrh.
+ */
+int ib_lrh_header_unpack(void *buf, struct ib_unpacked_lrh *lrh)
+{
+ ib_unpack(lrh_table, ARRAY_SIZE(lrh_table), buf, lrh);
+ return 0;
+}
+EXPORT_SYMBOL(ib_lrh_header_unpack);
+
+/**
* ib_ud_header_pack - Pack UD header struct into wire format
* @header:UD header struct
* @buf:Buffer to pack into
diff --git a/drivers/infiniband/core/umem.c b/drivers/infiniband/core/umem.c
index dd4b7ad..bc0245c 100644
--- a/drivers/infiniband/core/umem.c
+++ b/drivers/infiniband/core/umem.c
@@ -30,8 +30,6 @@
* 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.
- *
- * $Id: uverbs_mem.c 2743 2005-06-28 22:27:59Z roland $
*/
#include <linux/mm.h>
@@ -42,10 +40,6 @@
#include "uverbs.h"
-static int allow_weak_ordering;
-module_param(allow_weak_ordering, bool, 0444);
-MODULE_PARM_DESC(allow_weak_ordering, "Allow weak ordering for data registered memory");
-
#define IB_UMEM_MAX_PAGE_CHUNK \
((PAGE_SIZE - offsetof(struct ib_umem_chunk, page_list)) / \
((void *) &((struct ib_umem_chunk *) 0)->page_list[1] - \
@@ -101,21 +95,14 @@ static void dma_unmap_sg_ia64(struct ib_device *ibdev,
#endif
-static void __ib_umem_release(struct ib_device *dev, struct ib_umem *umem, int dirty, int dmasync)
+static void __ib_umem_release(struct ib_device *dev, struct ib_umem *umem, int dirty)
{
struct ib_umem_chunk *chunk, *tmp;
int i;
- DEFINE_DMA_ATTRS(attrs);
-
- if (dmasync)
- dma_set_attr(DMA_ATTR_WRITE_BARRIER, &attrs);
- else if (allow_weak_ordering)
- dma_set_attr(DMA_ATTR_WEAK_ORDERING, &attrs);
-
list_for_each_entry_safe(chunk, tmp, &umem->chunk_list, list) {
- ib_dma_unmap_sg_attrs(dev, chunk->page_list,
- chunk->nents, DMA_BIDIRECTIONAL, &attrs);
+ ib_dma_unmap_sg(dev, chunk->page_list,
+ chunk->nents, DMA_BIDIRECTIONAL);
for (i = 0; i < chunk->nents; ++i) {
struct page *page = sg_page(&chunk->page_list[i]);
@@ -136,7 +123,7 @@ static void __ib_umem_release(struct ib_device *dev, struct ib_umem *umem, int d
* @access: IB_ACCESS_xxx flags for memory being pinned
* @dmasync: flush in-flight DMA when the memory region is written
*/
-static struct ib_umem *__ib_umem_get(struct ib_ucontext *context, unsigned long addr,
+struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr,
size_t size, int access, int dmasync)
{
struct ib_umem *umem;
@@ -154,9 +141,6 @@ static struct ib_umem *__ib_umem_get(struct ib_ucontext *context, unsigned long
if (dmasync)
dma_set_attr(DMA_ATTR_WRITE_BARRIER, &attrs);
- else if (allow_weak_ordering)
- dma_set_attr(DMA_ATTR_WEAK_ORDERING, &attrs);
-
if (!can_do_mlock())
return ERR_PTR(-EPERM);
@@ -214,7 +198,7 @@ static struct ib_umem *__ib_umem_get(struct ib_ucontext *context, unsigned long
ret = 0;
while (npages) {
ret = get_user_pages(current, current->mm, cur_base,
- min_t(int, npages,
+ min_t(unsigned long, npages,
PAGE_SIZE / sizeof (struct page *)),
1, !umem->writable, page_list, vma_list);
@@ -268,7 +252,7 @@ static struct ib_umem *__ib_umem_get(struct ib_ucontext *context, unsigned long
out:
if (ret < 0) {
- __ib_umem_release(context->device, umem, 0, dmasync);
+ __ib_umem_release(context->device, umem, 0);
kfree(umem);
} else
current->mm->locked_vm = locked;
@@ -280,21 +264,8 @@ out:
return ret < 0 ? ERR_PTR(ret) : umem;
}
-
-struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr,
- size_t size, int access)
-{
- return __ib_umem_get(context, addr, size, access, 0);
-}
EXPORT_SYMBOL(ib_umem_get);
-struct ib_umem *ib_umem_get_dmasync(struct ib_ucontext *context, unsigned long addr,
- size_t size, int access)
-{
- return __ib_umem_get(context, addr, size, access, 1);
-}
-EXPORT_SYMBOL(ib_umem_get_dmasync);
-
static void ib_umem_account(struct work_struct *work)
{
struct ib_umem *umem = container_of(work, struct ib_umem, work);
@@ -310,13 +281,13 @@ static void ib_umem_account(struct work_struct *work)
* ib_umem_release - release memory pinned with ib_umem_get
* @umem: umem struct to release
*/
-static void ib_umem_release_attr(struct ib_umem *umem, int dmasync)
+void ib_umem_release(struct ib_umem *umem)
{
struct ib_ucontext *context = umem->context;
struct mm_struct *mm;
unsigned long diff;
- __ib_umem_release(umem->context->device, umem, 1, dmasync);
+ __ib_umem_release(umem->context->device, umem, 1);
mm = get_task_mm(current);
if (!mm) {
@@ -351,19 +322,8 @@ static void ib_umem_release_attr(struct ib_umem *umem, int dmasync)
mmput(mm);
kfree(umem);
}
-
-void ib_umem_release(struct ib_umem *umem)
-{
- return ib_umem_release_attr(umem, 0);
-}
EXPORT_SYMBOL(ib_umem_release);
-void ib_umem_release_dmasync(struct ib_umem *umem)
-{
- return ib_umem_release_attr(umem, 1);
-}
-EXPORT_SYMBOL(ib_umem_release_dmasync);
-
int ib_umem_page_count(struct ib_umem *umem)
{
struct ib_umem_chunk *chunk;
diff --git a/drivers/infiniband/core/user_mad.c b/drivers/infiniband/core/user_mad.c
index b53eac4..895b231 100644
--- a/drivers/infiniband/core/user_mad.c
+++ b/drivers/infiniband/core/user_mad.c
@@ -2,6 +2,7 @@
* Copyright (c) 2004 Topspin Communications. All rights reserved.
* Copyright (c) 2005 Voltaire, Inc. All rights reserved.
* Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
+ * Copyright (c) 2008 Cisco. 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
@@ -30,8 +31,6 @@
* 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.
- *
- * $Id: user_mad.c 5596 2006-03-03 01:00:07Z sean.hefty $
*/
#include <linux/module.h>
@@ -42,12 +41,12 @@
#include <linux/cdev.h>
#include <linux/dma-mapping.h>
#include <linux/poll.h>
-#include <linux/rwsem.h>
+#include <linux/mutex.h>
#include <linux/kref.h>
#include <linux/compat.h>
+#include <linux/semaphore.h>
#include <asm/uaccess.h>
-#include <asm/semaphore.h>
#include <rdma/ib_mad.h>
#include <rdma/ib_user_mad.h>
@@ -94,7 +93,7 @@ struct ib_umad_port {
struct class_device *sm_class_dev;
struct semaphore sm_sem;
- struct rw_semaphore mutex;
+ struct mutex file_mutex;
struct list_head file_list;
struct ib_device *ib_dev;
@@ -110,11 +109,11 @@ struct ib_umad_device {
};
struct ib_umad_file {
+ struct mutex mutex;
struct ib_umad_port *port;
struct list_head recv_list;
struct list_head send_list;
struct list_head port_list;
- spinlock_t recv_lock;
spinlock_t send_lock;
wait_queue_head_t recv_wait;
struct ib_mad_agent *agent[IB_UMAD_MAX_AGENTS];
@@ -156,7 +155,7 @@ static int hdr_size(struct ib_umad_file *file)
sizeof (struct ib_user_mad_hdr_old);
}
-/* caller must hold port->mutex at least for reading */
+/* caller must hold file->mutex */
static struct ib_mad_agent *__get_agent(struct ib_umad_file *file, int id)
{
return file->agents_dead ? NULL : file->agent[id];
@@ -168,32 +167,30 @@ static int queue_packet(struct ib_umad_file *file,
{
int ret = 1;
- down_read(&file->port->mutex);
+ mutex_lock(&file->mutex);
for (packet->mad.hdr.id = 0;
packet->mad.hdr.id < IB_UMAD_MAX_AGENTS;
packet->mad.hdr.id++)
if (agent == __get_agent(file, packet->mad.hdr.id)) {
- spin_lock_irq(&file->recv_lock);
list_add_tail(&packet->list, &file->recv_list);
- spin_unlock_irq(&file->recv_lock);
wake_up_interruptible(&file->recv_wait);
ret = 0;
break;
}
- up_read(&file->port->mutex);
+ mutex_unlock(&file->mutex);
return ret;
}
static void dequeue_send(struct ib_umad_file *file,
struct ib_umad_packet *packet)
- {
+{
spin_lock_irq(&file->send_lock);
list_del(&packet->list);
spin_unlock_irq(&file->send_lock);
- }
+}
static void send_handler(struct ib_mad_agent *agent,
struct ib_mad_send_wc *send_wc)
@@ -341,10 +338,10 @@ static ssize_t ib_umad_read(struct file *filp, char __user *buf,
if (count < hdr_size(file))
return -EINVAL;
- spin_lock_irq(&file->recv_lock);
+ mutex_lock(&file->mutex);
while (list_empty(&file->recv_list)) {
- spin_unlock_irq(&file->recv_lock);
+ mutex_unlock(&file->mutex);
if (filp->f_flags & O_NONBLOCK)
return -EAGAIN;
@@ -353,13 +350,13 @@ static ssize_t ib_umad_read(struct file *filp, char __user *buf,
!list_empty(&file->recv_list)))
return -ERESTARTSYS;
- spin_lock_irq(&file->recv_lock);
+ mutex_lock(&file->mutex);
}
packet = list_entry(file->recv_list.next, struct ib_umad_packet, list);
list_del(&packet->list);
- spin_unlock_irq(&file->recv_lock);
+ mutex_unlock(&file->mutex);
if (packet->recv_wc)
ret = copy_recv_mad(file, buf, packet, count);
@@ -368,9 +365,9 @@ static ssize_t ib_umad_read(struct file *filp, char __user *buf,
if (ret < 0) {
/* Requeue packet */
- spin_lock_irq(&file->recv_lock);
+ mutex_lock(&file->mutex);
list_add(&packet->list, &file->recv_list);
- spin_unlock_irq(&file->recv_lock);
+ mutex_unlock(&file->mutex);
} else {
if (packet->recv_wc)
ib_free_recv_mad(packet->recv_wc);
@@ -481,7 +478,7 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf,
goto err;
}
- down_read(&file->port->mutex);
+ mutex_lock(&file->mutex);
agent = __get_agent(file, packet->mad.hdr.id);
if (!agent) {
@@ -577,7 +574,7 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf,
if (ret)
goto err_send;
- up_read(&file->port->mutex);
+ mutex_unlock(&file->mutex);
return count;
err_send:
@@ -587,7 +584,7 @@ err_msg:
err_ah:
ib_destroy_ah(ah);
err_up:
- up_read(&file->port->mutex);
+ mutex_unlock(&file->mutex);
err:
kfree(packet);
return ret;
@@ -613,11 +610,12 @@ static int ib_umad_reg_agent(struct ib_umad_file *file, void __user *arg,
{
struct ib_user_mad_reg_req ureq;
struct ib_mad_reg_req req;
- struct ib_mad_agent *agent;
+ struct ib_mad_agent *agent = NULL;
int agent_id;
int ret;
- down_write(&file->port->mutex);
+ mutex_lock(&file->port->file_mutex);
+ mutex_lock(&file->mutex);
if (!file->port->ib_dev) {
ret = -EPIPE;
@@ -666,13 +664,13 @@ found:
send_handler, recv_handler, file);
if (IS_ERR(agent)) {
ret = PTR_ERR(agent);
+ agent = NULL;
goto out;
}
if (put_user(agent_id,
(u32 __user *) (arg + offsetof(struct ib_user_mad_reg_req, id)))) {
ret = -EFAULT;
- ib_unregister_mad_agent(agent);
goto out;
}
@@ -690,7 +688,13 @@ found:
ret = 0;
out:
- up_write(&file->port->mutex);
+ mutex_unlock(&file->mutex);
+
+ if (ret && agent)
+ ib_unregister_mad_agent(agent);
+
+ mutex_unlock(&file->port->file_mutex);
+
return ret;
}
@@ -703,7 +707,8 @@ static int ib_umad_unreg_agent(struct ib_umad_file *file, u32 __user *arg)
if (get_user(id, arg))
return -EFAULT;
- down_write(&file->port->mutex);
+ mutex_lock(&file->port->file_mutex);
+ mutex_lock(&file->mutex);
if (id < 0 || id >= IB_UMAD_MAX_AGENTS || !__get_agent(file, id)) {
ret = -EINVAL;
@@ -714,11 +719,13 @@ static int ib_umad_unreg_agent(struct ib_umad_file *file, u32 __user *arg)
file->agent[id] = NULL;
out:
- up_write(&file->port->mutex);
+ mutex_unlock(&file->mutex);
if (agent)
ib_unregister_mad_agent(agent);
+ mutex_unlock(&file->port->file_mutex);
+
return ret;
}
@@ -726,12 +733,12 @@ static long ib_umad_enable_pkey(struct ib_umad_file *file)
{
int ret = 0;
- down_write(&file->port->mutex);
+ mutex_lock(&file->mutex);
if (file->already_used)
ret = -EINVAL;
else
file->use_pkey_index = 1;
- up_write(&file->port->mutex);
+ mutex_unlock(&file->mutex);
return ret;
}
@@ -768,6 +775,19 @@ static long ib_umad_compat_ioctl(struct file *filp, unsigned int cmd,
}
#endif
+/*
+ * ib_umad_open() does not need the BKL:
+ *
+ * - umad_port[] accesses are protected by port_lock, the
+ * ib_umad_port structures are properly reference counted, and
+ * everything else is purely local to the file being created, so
+ * races against other open calls are not a problem;
+ * - the ioctl method does not affect any global state outside of the
+ * file structure being operated on;
+ * - the port is added to umad_port[] as the last part of module
+ * initialization so the open method will either immediately run
+ * -ENXIO, or all required initialization will be done.
+ */
static int ib_umad_open(struct inode *inode, struct file *filp)
{
struct ib_umad_port *port;
@@ -783,7 +803,7 @@ static int ib_umad_open(struct inode *inode, struct file *filp)
if (!port)
return -ENXIO;
- down_write(&port->mutex);
+ mutex_lock(&port->file_mutex);
if (!port->ib_dev) {
ret = -ENXIO;
@@ -797,7 +817,7 @@ static int ib_umad_open(struct inode *inode, struct file *filp)
goto out;
}
- spin_lock_init(&file->recv_lock);
+ mutex_init(&file->mutex);
spin_lock_init(&file->send_lock);
INIT_LIST_HEAD(&file->recv_list);
INIT_LIST_HEAD(&file->send_list);
@@ -809,7 +829,7 @@ static int ib_umad_open(struct inode *inode, struct file *filp)
list_add_tail(&file->port_list, &port->file_list);
out:
- up_write(&port->mutex);
+ mutex_unlock(&port->file_mutex);
return ret;
}
@@ -821,7 +841,8 @@ static int ib_umad_close(struct inode *inode, struct file *filp)
int already_dead;
int i;
- down_write(&file->port->mutex);
+ mutex_lock(&file->port->file_mutex);
+ mutex_lock(&file->mutex);
already_dead = file->agents_dead;
file->agents_dead = 1;
@@ -834,14 +855,14 @@ static int ib_umad_close(struct inode *inode, struct file *filp)
list_del(&file->port_list);
- downgrade_write(&file->port->mutex);
+ mutex_unlock(&file->mutex);
if (!already_dead)
for (i = 0; i < IB_UMAD_MAX_AGENTS; ++i)
if (file->agent[i])
ib_unregister_mad_agent(file->agent[i]);
- up_read(&file->port->mutex);
+ mutex_unlock(&file->port->file_mutex);
kfree(file);
kref_put(&dev->ref, ib_umad_release_dev);
@@ -914,10 +935,10 @@ static int ib_umad_sm_close(struct inode *inode, struct file *filp)
};
int ret = 0;
- down_write(&port->mutex);
+ mutex_lock(&port->file_mutex);
if (port->ib_dev)
ret = ib_modify_port(port->ib_dev, port->port_num, 0, &props);
- up_write(&port->mutex);
+ mutex_unlock(&port->file_mutex);
up(&port->sm_sem);
@@ -981,7 +1002,7 @@ static int ib_umad_init_port(struct ib_device *device, int port_num,
port->ib_dev = device;
port->port_num = port_num;
init_MUTEX(&port->sm_sem);
- init_rwsem(&port->mutex);
+ mutex_init(&port->file_mutex);
INIT_LIST_HEAD(&port->file_list);
port->dev = cdev_alloc();
@@ -1052,6 +1073,7 @@ err_cdev:
static void ib_umad_kill_port(struct ib_umad_port *port)
{
struct ib_umad_file *file;
+ int already_dead;
int id;
class_set_devdata(port->class_dev, NULL);
@@ -1067,42 +1089,22 @@ static void ib_umad_kill_port(struct ib_umad_port *port)
umad_port[port->dev_num] = NULL;
spin_unlock(&port_lock);
- down_write(&port->mutex);
+ mutex_lock(&port->file_mutex);
port->ib_dev = NULL;
- /*
- * Now go through the list of files attached to this port and
- * unregister all of their MAD agents. We need to hold
- * port->mutex while doing this to avoid racing with
- * ib_umad_close(), but we can't hold the mutex for writing
- * while calling ib_unregister_mad_agent(), since that might
- * deadlock by calling back into queue_packet(). So we
- * downgrade our lock to a read lock, and then drop and
- * reacquire the write lock for the next iteration.
- *
- * We do list_del_init() on the file's list_head so that the
- * list_del in ib_umad_close() is still OK, even after the
- * file is removed from the list.
- */
- while (!list_empty(&port->file_list)) {
- file = list_entry(port->file_list.next, struct ib_umad_file,
- port_list);
-
+ list_for_each_entry(file, &port->file_list, port_list) {
+ mutex_lock(&file->mutex);
+ already_dead = file->agents_dead;
file->agents_dead = 1;
- list_del_init(&file->port_list);
-
- downgrade_write(&port->mutex);
+ mutex_unlock(&file->mutex);
for (id = 0; id < IB_UMAD_MAX_AGENTS; ++id)
if (file->agent[id])
ib_unregister_mad_agent(file->agent[id]);
-
- up_read(&port->mutex);
- down_write(&port->mutex);
}
- up_write(&port->mutex);
+ mutex_unlock(&port->file_mutex);
clear_bit(port->dev_num, dev_map);
}
diff --git a/drivers/infiniband/core/uverbs.h b/drivers/infiniband/core/uverbs.h
index c75eb6c..d62fcad 100644
--- a/drivers/infiniband/core/uverbs.h
+++ b/drivers/infiniband/core/uverbs.h
@@ -32,8 +32,6 @@
* 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.
- *
- * $Id: uverbs.h 2559 2005-06-06 19:43:16Z roland $
*/
#ifndef UVERBS_H
@@ -81,13 +79,13 @@ struct ib_uverbs_device {
struct ib_uverbs_event_file {
struct kref ref;
- struct file *file;
struct ib_uverbs_file *uverbs_file;
spinlock_t lock;
- int is_async;
wait_queue_head_t poll_wait;
struct fasync_struct *async_queue;
struct list_head event_list;
+ int is_async;
+ int is_closed;
};
struct ib_uverbs_file {
diff --git a/drivers/infiniband/core/uverbs_cmd.c b/drivers/infiniband/core/uverbs_cmd.c
index 9e98cec..56feab6 100644
--- a/drivers/infiniband/core/uverbs_cmd.c
+++ b/drivers/infiniband/core/uverbs_cmd.c
@@ -31,8 +31,6 @@
* 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.
- *
- * $Id: uverbs_cmd.c 2708 2005-06-24 17:27:21Z roland $
*/
#include <linux/file.h>
@@ -919,7 +917,7 @@ ssize_t ib_uverbs_poll_cq(struct ib_uverbs_file *file,
resp->wc[i].opcode = wc[i].opcode;
resp->wc[i].vendor_err = wc[i].vendor_err;
resp->wc[i].byte_len = wc[i].byte_len;
- resp->wc[i].imm_data = (__u32 __force) wc[i].imm_data;
+ resp->wc[i].ex.imm_data = (__u32 __force) wc[i].ex.imm_data;
resp->wc[i].qp_num = wc[i].qp->qp_num;
resp->wc[i].src_qp = wc[i].src_qp;
resp->wc[i].wc_flags = wc[i].wc_flags;
@@ -1463,7 +1461,6 @@ ssize_t ib_uverbs_post_send(struct ib_uverbs_file *file,
next->num_sge = user_wr->num_sge;
next->opcode = user_wr->opcode;
next->send_flags = user_wr->send_flags;
- next->imm_data = (__be32 __force) user_wr->imm_data;
if (is_ud) {
next->wr.ud.ah = idr_read_ah(user_wr->wr.ud.ah,
@@ -1476,14 +1473,24 @@ ssize_t ib_uverbs_post_send(struct ib_uverbs_file *file,
next->wr.ud.remote_qkey = user_wr->wr.ud.remote_qkey;
} else {
switch (next->opcode) {
- case IB_WR_RDMA_WRITE:
case IB_WR_RDMA_WRITE_WITH_IMM:
+ next->ex.imm_data =
+ (__be32 __force) user_wr->ex.imm_data;
+ case IB_WR_RDMA_WRITE:
case IB_WR_RDMA_READ:
next->wr.rdma.remote_addr =
user_wr->wr.rdma.remote_addr;
next->wr.rdma.rkey =
user_wr->wr.rdma.rkey;
break;
+ case IB_WR_SEND_WITH_IMM:
+ next->ex.imm_data =
+ (__be32 __force) user_wr->ex.imm_data;
+ break;
+ case IB_WR_SEND_WITH_INV:
+ next->ex.invalidate_rkey =
+ user_wr->ex.invalidate_rkey;
+ break;
case IB_WR_ATOMIC_CMP_AND_SWP:
case IB_WR_ATOMIC_FETCH_AND_ADD:
next->wr.atomic.remote_addr =
diff --git a/drivers/infiniband/core/uverbs_main.c b/drivers/infiniband/core/uverbs_main.c
index cfa567b..68b75e7 100644
--- a/drivers/infiniband/core/uverbs_main.c
+++ b/drivers/infiniband/core/uverbs_main.c
@@ -32,8 +32,6 @@
* 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.
- *
- * $Id: uverbs_main.c 2733 2005-06-28 19:14:34Z roland $
*/
#include <linux/module.h>
@@ -352,7 +350,7 @@ static int ib_uverbs_event_close(struct inode *inode, struct file *filp)
struct ib_uverbs_event *entry, *tmp;
spin_lock_irq(&file->lock);
- file->file = NULL;
+ file->is_closed = 1;
list_for_each_entry_safe(entry, tmp, &file->event_list, list) {
if (entry->counter)
list_del(&entry->obj_list);
@@ -390,7 +388,7 @@ void ib_uverbs_comp_handler(struct ib_cq *cq, void *cq_context)
return;
spin_lock_irqsave(&file->lock, flags);
- if (!file->file) {
+ if (file->is_closed) {
spin_unlock_irqrestore(&file->lock, flags);
return;
}
@@ -423,7 +421,7 @@ static void ib_uverbs_async_handler(struct ib_uverbs_file *file,
unsigned long flags;
spin_lock_irqsave(&file->async_file->lock, flags);
- if (!file->async_file->file) {
+ if (file->async_file->is_closed) {
spin_unlock_irqrestore(&file->async_file->lock, flags);
return;
}
@@ -509,6 +507,7 @@ struct file *ib_uverbs_alloc_event_file(struct ib_uverbs_file *uverbs_file,
ev_file->uverbs_file = uverbs_file;
ev_file->async_queue = NULL;
ev_file->is_async = is_async;
+ ev_file->is_closed = 0;
*fd = get_unused_fd();
if (*fd < 0) {
@@ -522,8 +521,6 @@ struct file *ib_uverbs_alloc_event_file(struct ib_uverbs_file *uverbs_file,
goto err_fd;
}
- ev_file->file = filp;
-
/*
* fops_get() can't fail here, because we're coming from a
* system call on a uverbs file, which will already have a
@@ -616,6 +613,18 @@ static int ib_uverbs_mmap(struct file *filp, struct vm_area_struct *vma)
return file->device->ib_dev->mmap(file->ucontext, vma);
}
+/*
+ * ib_uverbs_open() does not need the BKL:
+ *
+ * - dev_table[] accesses are protected by map_lock, the
+ * ib_uverbs_device structures are properly reference counted, and
+ * everything else is purely local to the file being created, so
+ * races against other open calls are not a problem;
+ * - there is no ioctl method to race against;
+ * - the device is added to dev_table[] as the last part of module
+ * initialization, the open method will either immediately run
+ * -ENXIO, or all required initialization will be done.
+ */
static int ib_uverbs_open(struct inode *inode, struct file *filp)
{
struct ib_uverbs_device *dev;
@@ -657,7 +666,6 @@ err_module:
err:
kref_put(&dev->ref, ib_uverbs_release_dev);
-
return ret;
}
diff --git a/drivers/infiniband/core/verbs.c b/drivers/infiniband/core/verbs.c
index 84709ed..a7da9be 100644
--- a/drivers/infiniband/core/verbs.c
+++ b/drivers/infiniband/core/verbs.c
@@ -34,8 +34,6 @@
* 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.
- *
- * $Id: verbs.c 1349 2004-12-16 21:09:43Z roland $
*/
#include <linux/errno.h>
@@ -248,7 +246,9 @@ int ib_modify_srq(struct ib_srq *srq,
struct ib_srq_attr *srq_attr,
enum ib_srq_attr_mask srq_attr_mask)
{
- return srq->device->modify_srq(srq, srq_attr, srq_attr_mask, NULL);
+ return srq->device->modify_srq ?
+ srq->device->modify_srq(srq, srq_attr, srq_attr_mask, NULL) :
+ -ENOSYS;
}
EXPORT_SYMBOL(ib_modify_srq);
@@ -315,7 +315,6 @@ static const struct {
} qp_state_table[IB_QPS_ERR + 1][IB_QPS_ERR + 1] = {
[IB_QPS_RESET] = {
[IB_QPS_RESET] = { .valid = 1 },
- [IB_QPS_ERR] = { .valid = 1 },
[IB_QPS_INIT] = {
.valid = 1,
.req_param = {
@@ -679,6 +678,9 @@ struct ib_mr *ib_reg_phys_mr(struct ib_pd *pd,
{
struct ib_mr *mr;
+ if (!pd->device->reg_phys_mr)
+ return ERR_PTR(-ENOSYS);
+
mr = pd->device->reg_phys_mr(pd, phys_buf_array, num_phys_buf,
mr_access_flags, iova_start);
@@ -750,6 +752,52 @@ int ib_dereg_mr(struct ib_mr *mr)
}
EXPORT_SYMBOL(ib_dereg_mr);
+struct ib_mr *ib_alloc_fast_reg_mr(struct ib_pd *pd, int max_page_list_len)
+{
+ struct ib_mr *mr;
+
+ if (!pd->device->alloc_fast_reg_mr)
+ return ERR_PTR(-ENOSYS);
+
+ mr = pd->device->alloc_fast_reg_mr(pd, max_page_list_len);
+
+ if (!IS_ERR(mr)) {
+ mr->device = pd->device;
+ mr->pd = pd;
+ mr->uobject = NULL;
+ atomic_inc(&pd->usecnt);
+ atomic_set(&mr->usecnt, 0);
+ }
+
+ return mr;
+}
+EXPORT_SYMBOL(ib_alloc_fast_reg_mr);
+
+struct ib_fast_reg_page_list *ib_alloc_fast_reg_page_list(struct ib_device *device,
+ int max_page_list_len)
+{
+ struct ib_fast_reg_page_list *page_list;
+
+ if (!device->alloc_fast_reg_page_list)
+ return ERR_PTR(-ENOSYS);
+
+ page_list = device->alloc_fast_reg_page_list(device, max_page_list_len);
+
+ if (!IS_ERR(page_list)) {
+ page_list->device = device;
+ page_list->max_page_list_len = max_page_list_len;
+ }
+
+ return page_list;
+}
+EXPORT_SYMBOL(ib_alloc_fast_reg_page_list);
+
+void ib_free_fast_reg_page_list(struct ib_fast_reg_page_list *page_list)
+{
+ page_list->device->free_fast_reg_page_list(page_list);
+}
+EXPORT_SYMBOL(ib_free_fast_reg_page_list);
+
/* Memory windows */
struct ib_mw *ib_alloc_mw(struct ib_pd *pd)
diff --git a/drivers/infiniband/core/writeback.c b/drivers/infiniband/core/writeback.c
new file mode 100644
index 0000000..7f8ebbe
--- /dev/null
+++ b/drivers/infiniband/core/writeback.c
@@ -0,0 +1,106 @@
+#include <linux/fs.h>
+#include <linux/pagemap.h>
+#include <linux/pagevec.h>
+#include <linux/writeback.h>
+#include <linux/mpage.h>
+#include <linux/module.h>
+
+int backport_write_cache_pages(struct address_space *mapping,
+ struct writeback_control *wbc, backport_writepage_t writepage,
+ void *data)
+{
+ struct backing_dev_info *bdi = mapping->backing_dev_info;
+ int ret = 0;
+ int done = 0;
+ struct pagevec pvec;
+ int nr_pages;
+ pgoff_t index;
+ pgoff_t end; /* Inclusive */
+ int scanned = 0;
+ int range_whole = 0;
+ long nr_to_write = wbc->nr_to_write;
+
+ if (wbc->nonblocking && bdi_write_congested(bdi)) {
+ wbc->encountered_congestion = 1;
+ return 0;
+ }
+
+ pagevec_init(&pvec, 0);
+ if (wbc->range_cyclic) {
+ index = mapping->writeback_index; /* Start from prev offset */
+ end = -1;
+ } else {
+ index = wbc->range_start >> PAGE_CACHE_SHIFT;
+ end = wbc->range_end >> PAGE_CACHE_SHIFT;
+ if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
+ range_whole = 1;
+ scanned = 1;
+ }
+retry:
+ while (!done && (index <= end) &&
+ (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
+ PAGECACHE_TAG_DIRTY,
+ min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1))) {
+ unsigned i;
+
+ scanned = 1;
+ for (i = 0; i < nr_pages; i++) {
+ struct page *page = pvec.pages[i];
+
+ /*
+ * At this point we hold neither mapping->tree_lock nor
+ * lock on the page itself: the page may be truncated or
+ * invalidated (changing page->mapping to NULL), or even
+ * swizzled back from swapper_space to tmpfs file
+ * mapping
+ */
+ lock_page(page);
+
+ if (unlikely(page->mapping != mapping)) {
+ unlock_page(page);
+ continue;
+ }
+
+ if (!wbc->range_cyclic && page->index > end) {
+ done = 1;
+ unlock_page(page);
+ continue;
+ }
+
+ if (wbc->sync_mode != WB_SYNC_NONE)
+ wait_on_page_writeback(page);
+
+ if (PageWriteback(page) ||
+ !clear_page_dirty_for_io(page)) {
+ unlock_page(page);
+ continue;
+ }
+
+ ret = (*writepage)(page, wbc, data);
+
+ if (unlikely(ret == AOP_WRITEPAGE_ACTIVATE)) {
+ unlock_page(page);
+ ret = 0;
+ }
+ if (ret || (--nr_to_write <= 0))
+ done = 1;
+ if (wbc->nonblocking && bdi_write_congested(bdi)) {
+ wbc->encountered_congestion = 1;
+ done = 1;
+ }
+ }
+ pagevec_release(&pvec);
+ cond_resched();
+ }
+ if (!scanned && !done) {
+ /*
+ * We hit the last page and there is more work to be done: wrap
+ * back to the start of the file
+ */
+ scanned = 1;
+ index = 0;
+ goto retry;
+ }
+ return ret;
+}
+EXPORT_SYMBOL(backport_write_cache_pages);
distrib
>
Scientific%20Linux
>
5x
>
x86_64
>
by-pkgid
>
fc11cd6e1c513a17304da94a5390f3cd
>
files
>
2691
