Minki/linux
1
0
Fork 1
mirror of https://github.com/torvalds/linux.git synced 2024-12-30 23:02:08 +00:00
Code Issues Packages Projects Releases Wiki Activity
4d7f693af0
linux/drivers/infiniband/core/iwpm_msg.c

743 lines
24 KiB
C
Raw Normal View History

RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
/*
* Copyright (c) 2014 Intel Corporation. All rights reserved.
* Copyright (c) 2014 Chelsio, Inc. All rights reserved.
*
* This software is available to you under a choice of one of two
* licenses. You may choose to be licensed under the terms of the GNU
* General Public License (GPL) Version 2, available from the file
* COPYING in the main directory of this source tree, or the
* OpenIB.org BSD license below:
*
* Redistribution and use in source and binary forms, with or
* without modification, are permitted provided that the following
* conditions are met:
*
* - Redistributions of source code must retain the above
* copyright notice, this list of conditions and the following
* disclaimer.
*
* - Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials
* provided with the distribution.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include "iwpm_util.h"
RDMA/core: Fix for parsing netlink string attribute The string iwpm_ulib_name is recorded in a nlmsg as a netlink attribute. Without this fix parsing of the nlmsg by the userspace port mapper service fails because of unknown attribute length, causing the port mapper service not to register the client, which has sent the nlmsg. Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Cc: <stable@vger.kernel.org> #v3.16 Reviewed-By: Jason Gunthorpe <jgunthorpe@obsidianresearch.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2015-05-08 21:36:33 +00:00
static const char iwpm_ulib_name[IWPM_ULIBNAME_SIZE] = "iWarpPortMapperUser";
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
static int iwpm_ulib_version = 3;
static int iwpm_user_pid = IWPM_PID_UNDEFINED;
static atomic_t echo_nlmsg_seq;
int iwpm_valid_pid(void)
{
return iwpm_user_pid > 0;
}
/*
* iwpm_register_pid - Send a netlink query to user space
* for the iwarp port mapper pid
*
* nlmsg attributes:
* [IWPM_NLA_REG_PID_SEQ]
* [IWPM_NLA_REG_IF_NAME]
* [IWPM_NLA_REG_IBDEV_NAME]
* [IWPM_NLA_REG_ULIB_NAME]
*/
int iwpm_register_pid(struct iwpm_dev_data *pm_msg, u8 nl_client)
{
struct sk_buff *skb = NULL;
struct iwpm_nlmsg_request *nlmsg_request = NULL;
struct nlmsghdr *nlh;
u32 msg_seq;
const char *err_str = "";
int ret = -EINVAL;
if (!iwpm_valid_client(nl_client)) {
err_str = "Invalid port mapper client";
goto pid_query_error;
}
RDMA/core: Fixes for port mapper client registration Fixes to allow clients to make remove mapping requests, after they have provided the user space service with the mapping information, they are using when the service is restarted. 1) Adding IWPM_REG_VALID, IWPM_REG_INCOMPL and IWPM_REG_UNDEF registration types for the port mapper clients and functions to set/check the registration type. 2) If the port mapper user space service is not available to register the client, then its registration stays IWPM_REG_UNDEF and the registration isn't checked until the service becomes available (no mappings are possible, if the user space service isn't running). 3) After the service is restarted, the user space port mapper pid is set to valid and the client registration is set to IWPM_REG_INCOMPL to allow the client to make remove mapping requests. Signed-off-by: Tatyana Nikolova <Tatyana.E.Nikolova@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Tested-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2015-07-02 17:47:44 +00:00
if (iwpm_check_registration(nl_client, IWPM_REG_VALID) ||
iwpm_user_pid == IWPM_PID_UNAVAILABLE)
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
return 0;
skb = iwpm_create_nlmsg(RDMA_NL_IWPM_REG_PID, &nlh, nl_client);
if (!skb) {
err_str = "Unable to create a nlmsg";
goto pid_query_error;
}
nlh->nlmsg_seq = iwpm_get_nlmsg_seq();
nlmsg_request = iwpm_get_nlmsg_request(nlh->nlmsg_seq, nl_client, GFP_KERNEL);
if (!nlmsg_request) {
err_str = "Unable to allocate netlink request";
goto pid_query_error;
}
msg_seq = atomic_read(&echo_nlmsg_seq);
/* fill in the pid request message */
err_str = "Unable to put attribute of the nlmsg";
ret = ibnl_put_attr(skb, nlh, sizeof(u32), &msg_seq, IWPM_NLA_REG_PID_SEQ);
if (ret)
goto pid_query_error;
iwcm: common code for port mapper moved port mapper related code from drivers into common code Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com> Signed-off-by: Tatyana E. Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Faisal Latif <faisal.latif@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Tested-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2016-02-26 15:18:00 +00:00
ret = ibnl_put_attr(skb, nlh, IFNAMSIZ,
iwpm: crash fix for large connections test During large connection test, there is a crash at wake_up() in the callback as waitq is not yet initialized. Callback can happen before iwpm_wait_complete_req() is called to initialize waitq. To resolve, using signaling semaphore instead of waitq. Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com> Reviewed-by: Tatyana E Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Faisal Latif <faisal.latif@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Tested-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2016-02-26 15:18:05 +00:00
pm_msg->if_name, IWPM_NLA_REG_IF_NAME);
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
if (ret)
goto pid_query_error;
ret = ibnl_put_attr(skb, nlh, IWPM_DEVNAME_SIZE,
pm_msg->dev_name, IWPM_NLA_REG_IBDEV_NAME);
if (ret)
goto pid_query_error;
ret = ibnl_put_attr(skb, nlh, IWPM_ULIBNAME_SIZE,
(char *)iwpm_ulib_name, IWPM_NLA_REG_ULIB_NAME);
if (ret)
goto pid_query_error;
pr_debug("%s: Multicasting a nlmsg (dev = %s ifname = %s iwpm = %s)\n",
__func__, pm_msg->dev_name, pm_msg->if_name, iwpm_ulib_name);
RDMA/netlink: Rename and remove redundant parameter from ibnl_multicast The pointer to netlink header was not used in the ibnl_multicast function, so let's remove it and simplify the function signature. Signed-off-by: Leon Romanovsky <leonro@mellanox.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com>
2017-06-18 12:44:32 +00:00
ret = rdma_nl_multicast(skb, RDMA_NL_GROUP_IWPM, GFP_KERNEL);
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
if (ret) {
skb = NULL; /* skb is freed in the netlink send-op handling */
iwpm_user_pid = IWPM_PID_UNAVAILABLE;
err_str = "Unable to send a nlmsg";
goto pid_query_error;
}
nlmsg_request->req_buffer = pm_msg;
ret = iwpm_wait_complete_req(nlmsg_request);
return ret;
pid_query_error:
pr_info("%s: %s (client = %d)\n", __func__, err_str, nl_client);
if (skb)
dev_kfree_skb(skb);
if (nlmsg_request)
iwpm_free_nlmsg_request(&nlmsg_request->kref);
return ret;
}
/*
* iwpm_add_mapping - Send a netlink add mapping message
* to the port mapper
* nlmsg attributes:
* [IWPM_NLA_MANAGE_MAPPING_SEQ]
* [IWPM_NLA_MANAGE_ADDR]
*/
int iwpm_add_mapping(struct iwpm_sa_data *pm_msg, u8 nl_client)
{
struct sk_buff *skb = NULL;
struct iwpm_nlmsg_request *nlmsg_request = NULL;
struct nlmsghdr *nlh;
u32 msg_seq;
const char *err_str = "";
int ret = -EINVAL;
if (!iwpm_valid_client(nl_client)) {
err_str = "Invalid port mapper client";
goto add_mapping_error;
}
RDMA/core: Fixes for port mapper client registration Fixes to allow clients to make remove mapping requests, after they have provided the user space service with the mapping information, they are using when the service is restarted. 1) Adding IWPM_REG_VALID, IWPM_REG_INCOMPL and IWPM_REG_UNDEF registration types for the port mapper clients and functions to set/check the registration type. 2) If the port mapper user space service is not available to register the client, then its registration stays IWPM_REG_UNDEF and the registration isn't checked until the service becomes available (no mappings are possible, if the user space service isn't running). 3) After the service is restarted, the user space port mapper pid is set to valid and the client registration is set to IWPM_REG_INCOMPL to allow the client to make remove mapping requests. Signed-off-by: Tatyana Nikolova <Tatyana.E.Nikolova@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Tested-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2015-07-02 17:47:44 +00:00
if (!iwpm_valid_pid())
return 0;
if (!iwpm_check_registration(nl_client, IWPM_REG_VALID)) {
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
err_str = "Unregistered port mapper client";
goto add_mapping_error;
}
skb = iwpm_create_nlmsg(RDMA_NL_IWPM_ADD_MAPPING, &nlh, nl_client);
if (!skb) {
err_str = "Unable to create a nlmsg";
goto add_mapping_error;
}
nlh->nlmsg_seq = iwpm_get_nlmsg_seq();
nlmsg_request = iwpm_get_nlmsg_request(nlh->nlmsg_seq, nl_client, GFP_KERNEL);
if (!nlmsg_request) {
err_str = "Unable to allocate netlink request";
goto add_mapping_error;
}
msg_seq = atomic_read(&echo_nlmsg_seq);
/* fill in the add mapping message */
err_str = "Unable to put attribute of the nlmsg";
ret = ibnl_put_attr(skb, nlh, sizeof(u32), &msg_seq,
IWPM_NLA_MANAGE_MAPPING_SEQ);
if (ret)
goto add_mapping_error;
ret = ibnl_put_attr(skb, nlh, sizeof(struct sockaddr_storage),
&pm_msg->loc_addr, IWPM_NLA_MANAGE_ADDR);
if (ret)
goto add_mapping_error;
nlmsg_request->req_buffer = pm_msg;
RDMA/netlink: Rename and remove redundant parameter from ibnl_unicast* Netlink message header is not needed for unicast reply, hence remove it. Signed-off-by: Leon Romanovsky <leonro@mellanox.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com>
2017-06-18 12:35:20 +00:00
ret = rdma_nl_unicast_wait(skb, iwpm_user_pid);
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
if (ret) {
skb = NULL; /* skb is freed in the netlink send-op handling */
iwpm_user_pid = IWPM_PID_UNDEFINED;
err_str = "Unable to send a nlmsg";
goto add_mapping_error;
}
ret = iwpm_wait_complete_req(nlmsg_request);
return ret;
add_mapping_error:
pr_info("%s: %s (client = %d)\n", __func__, err_str, nl_client);
if (skb)
dev_kfree_skb(skb);
if (nlmsg_request)
iwpm_free_nlmsg_request(&nlmsg_request->kref);
return ret;
}
/*
* iwpm_add_and_query_mapping - Send a netlink add and query
* mapping message to the port mapper
* nlmsg attributes:
* [IWPM_NLA_QUERY_MAPPING_SEQ]
* [IWPM_NLA_QUERY_LOCAL_ADDR]
* [IWPM_NLA_QUERY_REMOTE_ADDR]
*/
int iwpm_add_and_query_mapping(struct iwpm_sa_data *pm_msg, u8 nl_client)
{
struct sk_buff *skb = NULL;
struct iwpm_nlmsg_request *nlmsg_request = NULL;
struct nlmsghdr *nlh;
u32 msg_seq;
const char *err_str = "";
int ret = -EINVAL;
if (!iwpm_valid_client(nl_client)) {
err_str = "Invalid port mapper client";
goto query_mapping_error;
}
RDMA/core: Fixes for port mapper client registration Fixes to allow clients to make remove mapping requests, after they have provided the user space service with the mapping information, they are using when the service is restarted. 1) Adding IWPM_REG_VALID, IWPM_REG_INCOMPL and IWPM_REG_UNDEF registration types for the port mapper clients and functions to set/check the registration type. 2) If the port mapper user space service is not available to register the client, then its registration stays IWPM_REG_UNDEF and the registration isn't checked until the service becomes available (no mappings are possible, if the user space service isn't running). 3) After the service is restarted, the user space port mapper pid is set to valid and the client registration is set to IWPM_REG_INCOMPL to allow the client to make remove mapping requests. Signed-off-by: Tatyana Nikolova <Tatyana.E.Nikolova@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Tested-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2015-07-02 17:47:44 +00:00
if (!iwpm_valid_pid())
return 0;
if (!iwpm_check_registration(nl_client, IWPM_REG_VALID)) {
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
err_str = "Unregistered port mapper client";
goto query_mapping_error;
}
ret = -ENOMEM;
skb = iwpm_create_nlmsg(RDMA_NL_IWPM_QUERY_MAPPING, &nlh, nl_client);
if (!skb) {
err_str = "Unable to create a nlmsg";
goto query_mapping_error;
}
nlh->nlmsg_seq = iwpm_get_nlmsg_seq();
nlmsg_request = iwpm_get_nlmsg_request(nlh->nlmsg_seq,
nl_client, GFP_KERNEL);
if (!nlmsg_request) {
err_str = "Unable to allocate netlink request";
goto query_mapping_error;
}
msg_seq = atomic_read(&echo_nlmsg_seq);
/* fill in the query message */
err_str = "Unable to put attribute of the nlmsg";
ret = ibnl_put_attr(skb, nlh, sizeof(u32), &msg_seq,
IWPM_NLA_QUERY_MAPPING_SEQ);
if (ret)
goto query_mapping_error;
ret = ibnl_put_attr(skb, nlh, sizeof(struct sockaddr_storage),
&pm_msg->loc_addr, IWPM_NLA_QUERY_LOCAL_ADDR);
if (ret)
goto query_mapping_error;
ret = ibnl_put_attr(skb, nlh, sizeof(struct sockaddr_storage),
&pm_msg->rem_addr, IWPM_NLA_QUERY_REMOTE_ADDR);
if (ret)
goto query_mapping_error;
nlmsg_request->req_buffer = pm_msg;
RDMA/netlink: Rename and remove redundant parameter from ibnl_unicast* Netlink message header is not needed for unicast reply, hence remove it. Signed-off-by: Leon Romanovsky <leonro@mellanox.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com>
2017-06-18 12:35:20 +00:00
ret = rdma_nl_unicast_wait(skb, iwpm_user_pid);
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
if (ret) {
skb = NULL; /* skb is freed in the netlink send-op handling */
err_str = "Unable to send a nlmsg";
goto query_mapping_error;
}
ret = iwpm_wait_complete_req(nlmsg_request);
return ret;
query_mapping_error:
pr_info("%s: %s (client = %d)\n", __func__, err_str, nl_client);
if (skb)
dev_kfree_skb(skb);
if (nlmsg_request)
iwpm_free_nlmsg_request(&nlmsg_request->kref);
return ret;
}
/*
* iwpm_remove_mapping - Send a netlink remove mapping message
* to the port mapper
* nlmsg attributes:
* [IWPM_NLA_MANAGE_MAPPING_SEQ]
* [IWPM_NLA_MANAGE_ADDR]
*/
int iwpm_remove_mapping(struct sockaddr_storage *local_addr, u8 nl_client)
{
struct sk_buff *skb = NULL;
struct nlmsghdr *nlh;
u32 msg_seq;
const char *err_str = "";
int ret = -EINVAL;
if (!iwpm_valid_client(nl_client)) {
err_str = "Invalid port mapper client";
goto remove_mapping_error;
}
RDMA/core: Fixes for port mapper client registration Fixes to allow clients to make remove mapping requests, after they have provided the user space service with the mapping information, they are using when the service is restarted. 1) Adding IWPM_REG_VALID, IWPM_REG_INCOMPL and IWPM_REG_UNDEF registration types for the port mapper clients and functions to set/check the registration type. 2) If the port mapper user space service is not available to register the client, then its registration stays IWPM_REG_UNDEF and the registration isn't checked until the service becomes available (no mappings are possible, if the user space service isn't running). 3) After the service is restarted, the user space port mapper pid is set to valid and the client registration is set to IWPM_REG_INCOMPL to allow the client to make remove mapping requests. Signed-off-by: Tatyana Nikolova <Tatyana.E.Nikolova@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Tested-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2015-07-02 17:47:44 +00:00
if (!iwpm_valid_pid())
return 0;
if (iwpm_check_registration(nl_client, IWPM_REG_UNDEF)) {
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
err_str = "Unregistered port mapper client";
goto remove_mapping_error;
}
skb = iwpm_create_nlmsg(RDMA_NL_IWPM_REMOVE_MAPPING, &nlh, nl_client);
if (!skb) {
ret = -ENOMEM;
err_str = "Unable to create a nlmsg";
goto remove_mapping_error;
}
msg_seq = atomic_read(&echo_nlmsg_seq);
nlh->nlmsg_seq = iwpm_get_nlmsg_seq();
err_str = "Unable to put attribute of the nlmsg";
ret = ibnl_put_attr(skb, nlh, sizeof(u32), &msg_seq,
IWPM_NLA_MANAGE_MAPPING_SEQ);
if (ret)
goto remove_mapping_error;
ret = ibnl_put_attr(skb, nlh, sizeof(struct sockaddr_storage),
local_addr, IWPM_NLA_MANAGE_ADDR);
if (ret)
goto remove_mapping_error;
RDMA/netlink: Rename and remove redundant parameter from ibnl_unicast* Netlink message header is not needed for unicast reply, hence remove it. Signed-off-by: Leon Romanovsky <leonro@mellanox.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com>
2017-06-18 12:35:20 +00:00
ret = rdma_nl_unicast_wait(skb, iwpm_user_pid);
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
if (ret) {
skb = NULL; /* skb is freed in the netlink send-op handling */
iwpm_user_pid = IWPM_PID_UNDEFINED;
err_str = "Unable to send a nlmsg";
goto remove_mapping_error;
}
iwpm_print_sockaddr(local_addr,
"remove_mapping: Local sockaddr:");
return 0;
remove_mapping_error:
pr_info("%s: %s (client = %d)\n", __func__, err_str, nl_client);
if (skb)
dev_kfree_skb_any(skb);
return ret;
}
/* netlink attribute policy for the received response to register pid request */
static const struct nla_policy resp_reg_policy[IWPM_NLA_RREG_PID_MAX] = {
[IWPM_NLA_RREG_PID_SEQ] = { .type = NLA_U32 },
[IWPM_NLA_RREG_IBDEV_NAME] = { .type = NLA_STRING,
.len = IWPM_DEVNAME_SIZE - 1 },
[IWPM_NLA_RREG_ULIB_NAME] = { .type = NLA_STRING,
.len = IWPM_ULIBNAME_SIZE - 1 },
[IWPM_NLA_RREG_ULIB_VER] = { .type = NLA_U16 },
[IWPM_NLA_RREG_PID_ERR] = { .type = NLA_U16 }
};
/*
* iwpm_register_pid_cb - Process a port mapper response to
* iwpm_register_pid()
*/
int iwpm_register_pid_cb(struct sk_buff *skb, struct netlink_callback *cb)
{
struct iwpm_nlmsg_request *nlmsg_request = NULL;
struct nlattr *nltb[IWPM_NLA_RREG_PID_MAX];
struct iwpm_dev_data *pm_msg;
char *dev_name, *iwpm_name;
u32 msg_seq;
u8 nl_client;
u16 iwpm_version;
const char *msg_type = "Register Pid response";
if (iwpm_parse_nlmsg(cb, IWPM_NLA_RREG_PID_MAX,
resp_reg_policy, nltb, msg_type))
return -EINVAL;
msg_seq = nla_get_u32(nltb[IWPM_NLA_RREG_PID_SEQ]);
nlmsg_request = iwpm_find_nlmsg_request(msg_seq);
if (!nlmsg_request) {
pr_info("%s: Could not find a matching request (seq = %u)\n",
__func__, msg_seq);
return -EINVAL;
}
pm_msg = nlmsg_request->req_buffer;
nl_client = nlmsg_request->nl_client;
dev_name = (char *)nla_data(nltb[IWPM_NLA_RREG_IBDEV_NAME]);
iwpm_name = (char *)nla_data(nltb[IWPM_NLA_RREG_ULIB_NAME]);
iwpm_version = nla_get_u16(nltb[IWPM_NLA_RREG_ULIB_VER]);
/* check device name, ulib name and version */
if (strcmp(pm_msg->dev_name, dev_name) ||
strcmp(iwpm_ulib_name, iwpm_name) ||
iwpm_version != iwpm_ulib_version) {
pr_info("%s: Incorrect info (dev = %s name = %s version = %d)\n",
__func__, dev_name, iwpm_name, iwpm_version);
nlmsg_request->err_code = IWPM_USER_LIB_INFO_ERR;
goto register_pid_response_exit;
}
iwpm_user_pid = cb->nlh->nlmsg_pid;
atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq);
pr_debug("%s: iWarp Port Mapper (pid = %d) is available!\n",
__func__, iwpm_user_pid);
if (iwpm_valid_client(nl_client))
RDMA/core: Fixes for port mapper client registration Fixes to allow clients to make remove mapping requests, after they have provided the user space service with the mapping information, they are using when the service is restarted. 1) Adding IWPM_REG_VALID, IWPM_REG_INCOMPL and IWPM_REG_UNDEF registration types for the port mapper clients and functions to set/check the registration type. 2) If the port mapper user space service is not available to register the client, then its registration stays IWPM_REG_UNDEF and the registration isn't checked until the service becomes available (no mappings are possible, if the user space service isn't running). 3) After the service is restarted, the user space port mapper pid is set to valid and the client registration is set to IWPM_REG_INCOMPL to allow the client to make remove mapping requests. Signed-off-by: Tatyana Nikolova <Tatyana.E.Nikolova@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Tested-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2015-07-02 17:47:44 +00:00
iwpm_set_registration(nl_client, IWPM_REG_VALID);
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
register_pid_response_exit:
nlmsg_request->request_done = 1;
/* always for found nlmsg_request */
kref_put(&nlmsg_request->kref, iwpm_free_nlmsg_request);
barrier();
iwpm: crash fix for large connections test During large connection test, there is a crash at wake_up() in the callback as waitq is not yet initialized. Callback can happen before iwpm_wait_complete_req() is called to initialize waitq. To resolve, using signaling semaphore instead of waitq. Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com> Reviewed-by: Tatyana E Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Faisal Latif <faisal.latif@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Tested-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2016-02-26 15:18:05 +00:00
up(&nlmsg_request->sem);
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
return 0;
}
/* netlink attribute policy for the received response to add mapping request */
static const struct nla_policy resp_add_policy[IWPM_NLA_RMANAGE_MAPPING_MAX] = {
[IWPM_NLA_MANAGE_MAPPING_SEQ] = { .type = NLA_U32 },
[IWPM_NLA_MANAGE_ADDR] = { .len = sizeof(struct sockaddr_storage) },
[IWPM_NLA_MANAGE_MAPPED_LOC_ADDR] = { .len = sizeof(struct sockaddr_storage) },
[IWPM_NLA_RMANAGE_MAPPING_ERR] = { .type = NLA_U16 }
};
/*
* iwpm_add_mapping_cb - Process a port mapper response to
* iwpm_add_mapping()
*/
int iwpm_add_mapping_cb(struct sk_buff *skb, struct netlink_callback *cb)
{
struct iwpm_sa_data *pm_msg;
struct iwpm_nlmsg_request *nlmsg_request = NULL;
struct nlattr *nltb[IWPM_NLA_RMANAGE_MAPPING_MAX];
struct sockaddr_storage *local_sockaddr;
struct sockaddr_storage *mapped_sockaddr;
const char *msg_type;
u32 msg_seq;
msg_type = "Add Mapping response";
if (iwpm_parse_nlmsg(cb, IWPM_NLA_RMANAGE_MAPPING_MAX,
resp_add_policy, nltb, msg_type))
return -EINVAL;
atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq);
msg_seq = nla_get_u32(nltb[IWPM_NLA_MANAGE_MAPPING_SEQ]);
nlmsg_request = iwpm_find_nlmsg_request(msg_seq);
if (!nlmsg_request) {
pr_info("%s: Could not find a matching request (seq = %u)\n",
__func__, msg_seq);
return -EINVAL;
}
pm_msg = nlmsg_request->req_buffer;
local_sockaddr = (struct sockaddr_storage *)
nla_data(nltb[IWPM_NLA_MANAGE_ADDR]);
mapped_sockaddr = (struct sockaddr_storage *)
nla_data(nltb[IWPM_NLA_MANAGE_MAPPED_LOC_ADDR]);
if (iwpm_compare_sockaddr(local_sockaddr, &pm_msg->loc_addr)) {
nlmsg_request->err_code = IWPM_USER_LIB_INFO_ERR;
goto add_mapping_response_exit;
}
if (mapped_sockaddr->ss_family != local_sockaddr->ss_family) {
pr_info("%s: Sockaddr family doesn't match the requested one\n",
__func__);
nlmsg_request->err_code = IWPM_USER_LIB_INFO_ERR;
goto add_mapping_response_exit;
}
memcpy(&pm_msg->mapped_loc_addr, mapped_sockaddr,
sizeof(*mapped_sockaddr));
iwpm_print_sockaddr(&pm_msg->loc_addr,
"add_mapping: Local sockaddr:");
iwpm_print_sockaddr(&pm_msg->mapped_loc_addr,
"add_mapping: Mapped local sockaddr:");
add_mapping_response_exit:
nlmsg_request->request_done = 1;
/* always for found request */
kref_put(&nlmsg_request->kref, iwpm_free_nlmsg_request);
barrier();
iwpm: crash fix for large connections test During large connection test, there is a crash at wake_up() in the callback as waitq is not yet initialized. Callback can happen before iwpm_wait_complete_req() is called to initialize waitq. To resolve, using signaling semaphore instead of waitq. Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com> Reviewed-by: Tatyana E Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Faisal Latif <faisal.latif@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Tested-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2016-02-26 15:18:05 +00:00
up(&nlmsg_request->sem);
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
return 0;
}
RDMA/core: Enable the iWarp Port Mapper to provide the actual address of the connecting peer to its clients Add functionality to enable the port mapper on the passive side to provide to its clients the actual (non-mapped) ip/tcp address information of the connecting peer 1) Adding remote_info_cb() to process the address info of the connecting peer The address info is provided by the user space port mapper service when the connection is initiated by the peer 2) Adding a hash list to store the remote address info 3) Adding functionality to add/remove the remote address info After the info has been provided to the port mapper client, it is removed from the hash list Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2015-04-21 20:28:10 +00:00
/* netlink attribute policy for the response to add and query mapping request
* and response with remote address info */
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
static const struct nla_policy resp_query_policy[IWPM_NLA_RQUERY_MAPPING_MAX] = {
[IWPM_NLA_QUERY_MAPPING_SEQ] = { .type = NLA_U32 },
[IWPM_NLA_QUERY_LOCAL_ADDR] = { .len = sizeof(struct sockaddr_storage) },
[IWPM_NLA_QUERY_REMOTE_ADDR] = { .len = sizeof(struct sockaddr_storage) },
[IWPM_NLA_RQUERY_MAPPED_LOC_ADDR] = { .len = sizeof(struct sockaddr_storage) },
[IWPM_NLA_RQUERY_MAPPED_REM_ADDR] = { .len = sizeof(struct sockaddr_storage) },
[IWPM_NLA_RQUERY_MAPPING_ERR] = { .type = NLA_U16 }
};
/*
* iwpm_add_and_query_mapping_cb - Process a port mapper response to
* iwpm_add_and_query_mapping()
*/
int iwpm_add_and_query_mapping_cb(struct sk_buff *skb,
struct netlink_callback *cb)
{
struct iwpm_sa_data *pm_msg;
struct iwpm_nlmsg_request *nlmsg_request = NULL;
struct nlattr *nltb[IWPM_NLA_RQUERY_MAPPING_MAX];
struct sockaddr_storage *local_sockaddr, *remote_sockaddr;
struct sockaddr_storage *mapped_loc_sockaddr, *mapped_rem_sockaddr;
const char *msg_type;
u32 msg_seq;
u16 err_code;
msg_type = "Query Mapping response";
if (iwpm_parse_nlmsg(cb, IWPM_NLA_RQUERY_MAPPING_MAX,
resp_query_policy, nltb, msg_type))
return -EINVAL;
atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq);
msg_seq = nla_get_u32(nltb[IWPM_NLA_QUERY_MAPPING_SEQ]);
nlmsg_request = iwpm_find_nlmsg_request(msg_seq);
if (!nlmsg_request) {
pr_info("%s: Could not find a matching request (seq = %u)\n",
__func__, msg_seq);
RDMA/core: Fix indentation Make indentation consistent. Detected by smatch. Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com> Cc: Tatyana Nikolova <Tatyana.E.Nikolova@intel.com> Cc: Steve Wise <swise@opengridcomputing.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: Sagi Grimberg <sagi@gimberg.me> Signed-off-by: Doug Ledford <dledford@redhat.com>
2016-06-03 19:08:00 +00:00
return -EINVAL;
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
}
pm_msg = nlmsg_request->req_buffer;
local_sockaddr = (struct sockaddr_storage *)
nla_data(nltb[IWPM_NLA_QUERY_LOCAL_ADDR]);
remote_sockaddr = (struct sockaddr_storage *)
nla_data(nltb[IWPM_NLA_QUERY_REMOTE_ADDR]);
mapped_loc_sockaddr = (struct sockaddr_storage *)
nla_data(nltb[IWPM_NLA_RQUERY_MAPPED_LOC_ADDR]);
mapped_rem_sockaddr = (struct sockaddr_storage *)
nla_data(nltb[IWPM_NLA_RQUERY_MAPPED_REM_ADDR]);
err_code = nla_get_u16(nltb[IWPM_NLA_RQUERY_MAPPING_ERR]);
if (err_code == IWPM_REMOTE_QUERY_REJECT) {
pr_info("%s: Received a Reject (pid = %u, echo seq = %u)\n",
__func__, cb->nlh->nlmsg_pid, msg_seq);
nlmsg_request->err_code = IWPM_REMOTE_QUERY_REJECT;
}
if (iwpm_compare_sockaddr(local_sockaddr, &pm_msg->loc_addr) ||
iwpm_compare_sockaddr(remote_sockaddr, &pm_msg->rem_addr)) {
pr_info("%s: Incorrect local sockaddr\n", __func__);
nlmsg_request->err_code = IWPM_USER_LIB_INFO_ERR;
goto query_mapping_response_exit;
}
if (mapped_loc_sockaddr->ss_family != local_sockaddr->ss_family ||
mapped_rem_sockaddr->ss_family != remote_sockaddr->ss_family) {
pr_info("%s: Sockaddr family doesn't match the requested one\n",
__func__);
nlmsg_request->err_code = IWPM_USER_LIB_INFO_ERR;
goto query_mapping_response_exit;
}
memcpy(&pm_msg->mapped_loc_addr, mapped_loc_sockaddr,
sizeof(*mapped_loc_sockaddr));
memcpy(&pm_msg->mapped_rem_addr, mapped_rem_sockaddr,
sizeof(*mapped_rem_sockaddr));
iwpm_print_sockaddr(&pm_msg->loc_addr,
"query_mapping: Local sockaddr:");
iwpm_print_sockaddr(&pm_msg->mapped_loc_addr,
"query_mapping: Mapped local sockaddr:");
iwpm_print_sockaddr(&pm_msg->rem_addr,
"query_mapping: Remote sockaddr:");
iwpm_print_sockaddr(&pm_msg->mapped_rem_addr,
"query_mapping: Mapped remote sockaddr:");
query_mapping_response_exit:
nlmsg_request->request_done = 1;
/* always for found request */
kref_put(&nlmsg_request->kref, iwpm_free_nlmsg_request);
barrier();
iwpm: crash fix for large connections test During large connection test, there is a crash at wake_up() in the callback as waitq is not yet initialized. Callback can happen before iwpm_wait_complete_req() is called to initialize waitq. To resolve, using signaling semaphore instead of waitq. Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com> Reviewed-by: Tatyana E Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Faisal Latif <faisal.latif@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Tested-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2016-02-26 15:18:05 +00:00
up(&nlmsg_request->sem);
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
return 0;
}
RDMA/core: Enable the iWarp Port Mapper to provide the actual address of the connecting peer to its clients Add functionality to enable the port mapper on the passive side to provide to its clients the actual (non-mapped) ip/tcp address information of the connecting peer 1) Adding remote_info_cb() to process the address info of the connecting peer The address info is provided by the user space port mapper service when the connection is initiated by the peer 2) Adding a hash list to store the remote address info 3) Adding functionality to add/remove the remote address info After the info has been provided to the port mapper client, it is removed from the hash list Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2015-04-21 20:28:10 +00:00
/*
* iwpm_remote_info_cb - Process a port mapper message, containing
* the remote connecting peer address info
*/
int iwpm_remote_info_cb(struct sk_buff *skb, struct netlink_callback *cb)
{
struct nlattr *nltb[IWPM_NLA_RQUERY_MAPPING_MAX];
struct sockaddr_storage *local_sockaddr, *remote_sockaddr;
struct sockaddr_storage *mapped_loc_sockaddr, *mapped_rem_sockaddr;
struct iwpm_remote_info *rem_info;
const char *msg_type;
u8 nl_client;
int ret = -EINVAL;
msg_type = "Remote Mapping info";
if (iwpm_parse_nlmsg(cb, IWPM_NLA_RQUERY_MAPPING_MAX,
resp_query_policy, nltb, msg_type))
return ret;
nl_client = RDMA_NL_GET_CLIENT(cb->nlh->nlmsg_type);
if (!iwpm_valid_client(nl_client)) {
pr_info("%s: Invalid port mapper client = %d\n",
__func__, nl_client);
return ret;
}
atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq);
local_sockaddr = (struct sockaddr_storage *)
nla_data(nltb[IWPM_NLA_QUERY_LOCAL_ADDR]);
remote_sockaddr = (struct sockaddr_storage *)
nla_data(nltb[IWPM_NLA_QUERY_REMOTE_ADDR]);
mapped_loc_sockaddr = (struct sockaddr_storage *)
nla_data(nltb[IWPM_NLA_RQUERY_MAPPED_LOC_ADDR]);
mapped_rem_sockaddr = (struct sockaddr_storage *)
nla_data(nltb[IWPM_NLA_RQUERY_MAPPED_REM_ADDR]);
if (mapped_loc_sockaddr->ss_family != local_sockaddr->ss_family ||
mapped_rem_sockaddr->ss_family != remote_sockaddr->ss_family) {
pr_info("%s: Sockaddr family doesn't match the requested one\n",
__func__);
return ret;
}
rem_info = kzalloc(sizeof(struct iwpm_remote_info), GFP_ATOMIC);
if (!rem_info) {
ret = -ENOMEM;
return ret;
}
memcpy(&rem_info->mapped_loc_sockaddr, mapped_loc_sockaddr,
sizeof(struct sockaddr_storage));
memcpy(&rem_info->remote_sockaddr, remote_sockaddr,
sizeof(struct sockaddr_storage));
memcpy(&rem_info->mapped_rem_sockaddr, mapped_rem_sockaddr,
sizeof(struct sockaddr_storage));
rem_info->nl_client = nl_client;
iwpm_add_remote_info(rem_info);
iwpm_print_sockaddr(local_sockaddr,
"remote_info: Local sockaddr:");
iwpm_print_sockaddr(mapped_loc_sockaddr,
"remote_info: Mapped local sockaddr:");
iwpm_print_sockaddr(remote_sockaddr,
"remote_info: Remote sockaddr:");
iwpm_print_sockaddr(mapped_rem_sockaddr,
"remote_info: Mapped remote sockaddr:");
return ret;
}
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
/* netlink attribute policy for the received request for mapping info */
static const struct nla_policy resp_mapinfo_policy[IWPM_NLA_MAPINFO_REQ_MAX] = {
[IWPM_NLA_MAPINFO_ULIB_NAME] = { .type = NLA_STRING,
.len = IWPM_ULIBNAME_SIZE - 1 },
[IWPM_NLA_MAPINFO_ULIB_VER] = { .type = NLA_U16 }
};
/*
* iwpm_mapping_info_cb - Process a port mapper request for mapping info
*/
int iwpm_mapping_info_cb(struct sk_buff *skb, struct netlink_callback *cb)
{
struct nlattr *nltb[IWPM_NLA_MAPINFO_REQ_MAX];
const char *msg_type = "Mapping Info response";
u8 nl_client;
char *iwpm_name;
u16 iwpm_version;
int ret = -EINVAL;
if (iwpm_parse_nlmsg(cb, IWPM_NLA_MAPINFO_REQ_MAX,
resp_mapinfo_policy, nltb, msg_type)) {
pr_info("%s: Unable to parse nlmsg\n", __func__);
return ret;
}
iwpm_name = (char *)nla_data(nltb[IWPM_NLA_MAPINFO_ULIB_NAME]);
iwpm_version = nla_get_u16(nltb[IWPM_NLA_MAPINFO_ULIB_VER]);
if (strcmp(iwpm_ulib_name, iwpm_name) ||
iwpm_version != iwpm_ulib_version) {
pr_info("%s: Invalid port mapper name = %s version = %d\n",
__func__, iwpm_name, iwpm_version);
return ret;
}
nl_client = RDMA_NL_GET_CLIENT(cb->nlh->nlmsg_type);
if (!iwpm_valid_client(nl_client)) {
pr_info("%s: Invalid port mapper client = %d\n",
__func__, nl_client);
return ret;
}
RDMA/core: Fixes for port mapper client registration Fixes to allow clients to make remove mapping requests, after they have provided the user space service with the mapping information, they are using when the service is restarted. 1) Adding IWPM_REG_VALID, IWPM_REG_INCOMPL and IWPM_REG_UNDEF registration types for the port mapper clients and functions to set/check the registration type. 2) If the port mapper user space service is not available to register the client, then its registration stays IWPM_REG_UNDEF and the registration isn't checked until the service becomes available (no mappings are possible, if the user space service isn't running). 3) After the service is restarted, the user space port mapper pid is set to valid and the client registration is set to IWPM_REG_INCOMPL to allow the client to make remove mapping requests. Signed-off-by: Tatyana Nikolova <Tatyana.E.Nikolova@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Tested-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2015-07-02 17:47:44 +00:00
iwpm_set_registration(nl_client, IWPM_REG_INCOMPL);
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq);
RDMA/core: Fixes for port mapper client registration Fixes to allow clients to make remove mapping requests, after they have provided the user space service with the mapping information, they are using when the service is restarted. 1) Adding IWPM_REG_VALID, IWPM_REG_INCOMPL and IWPM_REG_UNDEF registration types for the port mapper clients and functions to set/check the registration type. 2) If the port mapper user space service is not available to register the client, then its registration stays IWPM_REG_UNDEF and the registration isn't checked until the service becomes available (no mappings are possible, if the user space service isn't running). 3) After the service is restarted, the user space port mapper pid is set to valid and the client registration is set to IWPM_REG_INCOMPL to allow the client to make remove mapping requests. Signed-off-by: Tatyana Nikolova <Tatyana.E.Nikolova@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Tested-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2015-07-02 17:47:44 +00:00
iwpm_user_pid = cb->nlh->nlmsg_pid;
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
if (!iwpm_mapinfo_available())
return 0;
pr_debug("%s: iWarp Port Mapper (pid = %d) is available!\n",
RDMA/core: Fixes for port mapper client registration Fixes to allow clients to make remove mapping requests, after they have provided the user space service with the mapping information, they are using when the service is restarted. 1) Adding IWPM_REG_VALID, IWPM_REG_INCOMPL and IWPM_REG_UNDEF registration types for the port mapper clients and functions to set/check the registration type. 2) If the port mapper user space service is not available to register the client, then its registration stays IWPM_REG_UNDEF and the registration isn't checked until the service becomes available (no mappings are possible, if the user space service isn't running). 3) After the service is restarted, the user space port mapper pid is set to valid and the client registration is set to IWPM_REG_INCOMPL to allow the client to make remove mapping requests. Signed-off-by: Tatyana Nikolova <Tatyana.E.Nikolova@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Tested-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2015-07-02 17:47:44 +00:00
__func__, iwpm_user_pid);
ret = iwpm_send_mapinfo(nl_client, iwpm_user_pid);
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
return ret;
}
/* netlink attribute policy for the received mapping info ack */
static const struct nla_policy ack_mapinfo_policy[IWPM_NLA_MAPINFO_NUM_MAX] = {
[IWPM_NLA_MAPINFO_SEQ] = { .type = NLA_U32 },
[IWPM_NLA_MAPINFO_SEND_NUM] = { .type = NLA_U32 },
[IWPM_NLA_MAPINFO_ACK_NUM] = { .type = NLA_U32 }
};
/*
* iwpm_ack_mapping_info_cb - Process a port mapper ack for
* the provided mapping info records
*/
int iwpm_ack_mapping_info_cb(struct sk_buff *skb, struct netlink_callback *cb)
{
struct nlattr *nltb[IWPM_NLA_MAPINFO_NUM_MAX];
u32 mapinfo_send, mapinfo_ack;
const char *msg_type = "Mapping Info Ack";
if (iwpm_parse_nlmsg(cb, IWPM_NLA_MAPINFO_NUM_MAX,
ack_mapinfo_policy, nltb, msg_type))
return -EINVAL;
mapinfo_send = nla_get_u32(nltb[IWPM_NLA_MAPINFO_SEND_NUM]);
mapinfo_ack = nla_get_u32(nltb[IWPM_NLA_MAPINFO_ACK_NUM]);
if (mapinfo_ack != mapinfo_send)
pr_info("%s: Invalid mapinfo number (sent = %u ack-ed = %u)\n",
__func__, mapinfo_send, mapinfo_ack);
atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq);
return 0;
}
/* netlink attribute policy for the received port mapper error message */
static const struct nla_policy map_error_policy[IWPM_NLA_ERR_MAX] = {
[IWPM_NLA_ERR_SEQ] = { .type = NLA_U32 },
[IWPM_NLA_ERR_CODE] = { .type = NLA_U16 },
};
/*
* iwpm_mapping_error_cb - Process a port mapper error message
*/
int iwpm_mapping_error_cb(struct sk_buff *skb, struct netlink_callback *cb)
{
struct iwpm_nlmsg_request *nlmsg_request = NULL;
int nl_client = RDMA_NL_GET_CLIENT(cb->nlh->nlmsg_type);
struct nlattr *nltb[IWPM_NLA_ERR_MAX];
u32 msg_seq;
u16 err_code;
const char *msg_type = "Mapping Error Msg";
if (iwpm_parse_nlmsg(cb, IWPM_NLA_ERR_MAX,
map_error_policy, nltb, msg_type))
return -EINVAL;
msg_seq = nla_get_u32(nltb[IWPM_NLA_ERR_SEQ]);
err_code = nla_get_u16(nltb[IWPM_NLA_ERR_CODE]);
pr_info("%s: Received msg seq = %u err code = %u client = %d\n",
__func__, msg_seq, err_code, nl_client);
/* look for nlmsg_request */
nlmsg_request = iwpm_find_nlmsg_request(msg_seq);
if (!nlmsg_request) {
/* not all errors have associated requests */
pr_debug("Could not find matching req (seq = %u)\n", msg_seq);
return 0;
}
atomic_set(&echo_nlmsg_seq, cb->nlh->nlmsg_seq);
nlmsg_request->err_code = err_code;
nlmsg_request->request_done = 1;
/* always for found request */
kref_put(&nlmsg_request->kref, iwpm_free_nlmsg_request);
barrier();
iwpm: crash fix for large connections test During large connection test, there is a crash at wake_up() in the callback as waitq is not yet initialized. Callback can happen before iwpm_wait_complete_req() is called to initialize waitq. To resolve, using signaling semaphore instead of waitq. Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com> Reviewed-by: Tatyana E Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Faisal Latif <faisal.latif@intel.com> Reviewed-by: Steve Wise <swise@opengridcomputing.com> Tested-by: Steve Wise <swise@opengridcomputing.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
2016-02-26 15:18:05 +00:00
up(&nlmsg_request->sem);
RDMA/core: Add support for iWARP Port Mapper user space service This patch adds iWARP Port Mapper (IWPM) Version 2 support. The iWARP Port Mapper implementation is based on the port mapper specification section in the Sockets Direct Protocol paper - http://www.rdmaconsortium.org/home/draft-pinkerton-iwarp-sdp-v1.0.pdf Existing iWARP RDMA providers use the same IP address as the native TCP/IP stack when creating RDMA connections. They need a mechanism to claim the TCP ports used for RDMA connections to prevent TCP port collisions when other host applications use TCP ports. The iWARP Port Mapper provides a standard mechanism to accomplish this. Without this service it is possible for RDMA application to bind/listen on the same port which is already being used by native TCP host application. If that happens the incoming TCP connection data can be passed to the RDMA stack with error. The iWARP Port Mapper solution doesn't contain any changes to the existing network stack in the kernel space. All the changes are contained with the infiniband tree and also in user space. The iWARP Port Mapper service is implemented as a user space daemon process. Source for the IWPM service is located at http://git.openfabrics.org/git?p=~tnikolova/libiwpm-1.0.0/.git;a=summary The iWARP driver (port mapper client) sends to the IWPM service the local IP address and TCP port it has received from the RDMA application, when starting a connection. The IWPM service performs a socket bind from user space to get an available TCP port, called a mapped port, and communicates it back to the client. In that sense, the IWPM service is used to map the TCP port, which the RDMA application uses to any port available from the host TCP port space. The mapped ports are used in iWARP RDMA connections to avoid collisions with native TCP stack which is aware that these ports are taken. When an RDMA connection using a mapped port is terminated, the client notifies the IWPM service, which then releases the TCP port. The message exchange between the IWPM service and the iWARP drivers (between user space and kernel space) is implemented using netlink sockets. 1) Netlink interface functions are added: ibnl_unicast() and ibnl_mulitcast() for sending netlink messages to user space 2) The signature of the existing ibnl_put_msg() is changed to be more generic 3) Two netlink clients are added: RDMA_NL_NES, RDMA_NL_C4IW corresponding to the two iWarp drivers - nes and cxgb4 which use the IWPM service 4) Enums are added to enumerate the attributes in the netlink messages, which are exchanged between the user space IWPM service and the iWARP drivers Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com> Signed-off-by: Steve Wise <swise@opengridcomputing.com> Reviewed-by: PJ Waskiewicz <pj.waskiewicz@solidfire.com> [ Fold in range checking fixes and nlh_next removal as suggested by Dan Carpenter and Steve Wise. Fix sparse endianness in hash. - Roland ] Signed-off-by: Roland Dreier <roland@purestorage.com>
2014-03-26 22:07:35 +00:00
return 0;
}
Reference in New Issue Copy Permalink