1
1
openmpi/contrib/build-mca-comps-outside-of-tree/btl_tcp2.h
Ralph Castain a200e4f865 As per the RFC, bring in the ORTE async progress code and the rewrite of OOB:
*** THIS RFC INCLUDES A MINOR CHANGE TO THE MPI-RTE INTERFACE ***

Note: during the course of this work, it was necessary to completely separate the MPI and RTE progress engines. There were multiple places in the MPI layer where ORTE_WAIT_FOR_COMPLETION was being used. A new OMPI_WAIT_FOR_COMPLETION macro was created (defined in ompi/mca/rte/rte.h) that simply cycles across opal_progress until the provided flag becomes false. Places where the MPI layer blocked waiting for RTE to complete an event have been modified to use this macro.

***************************************************************************************

I am reissuing this RFC because of the time that has passed since its original release. Since its initial release and review, I have debugged it further to ensure it fully supports tests like loop_spawn. It therefore seems ready for merge back to the trunk. Given its prior review, I have set the timeout for one week.

The code is in  https://bitbucket.org/rhc/ompi-oob2


WHAT:    Rewrite of ORTE OOB

WHY:       Support asynchronous progress and a host of other features

WHEN:    Wed, August 21

SYNOPSIS:
The current OOB has served us well, but a number of limitations have been identified over the years. Specifically:

* it is only progressed when called via opal_progress, which can lead to hangs or recursive calls into libevent (which is not supported by that code)

* we've had issues when multiple NICs are available as the code doesn't "shift" messages between transports - thus, all nodes had to be available via the same TCP interface.

* the OOB "unloads" incoming opal_buffer_t objects during the transmission, thus preventing use of OBJ_RETAIN in the code when repeatedly sending the same message to multiple recipients

* there is no failover mechanism across NICs - if the selected NIC (or its attached switch) fails, we are forced to abort

* only one transport (i.e., component) can be "active"


The revised OOB resolves these problems:

* async progress is used for all application processes, with the progress thread blocking in the event library

* each available TCP NIC is supported by its own TCP module. The ability to asynchronously progress each module independently is provided, but not enabled by default (a runtime MCA parameter turns it "on")

* multi-address TCP NICs (e.g., a NIC with both an IPv4 and IPv6 address, or with virtual interfaces) are supported - reachability is determined by comparing the contact info for a peer against all addresses within the range covered by the address/mask pairs for the NIC.

* a message that arrives on one TCP NIC is automatically shifted to whatever NIC that is connected to the next "hop" if that peer cannot be reached by the incoming NIC. If no TCP module will reach the peer, then the OOB attempts to send the message via all other available components - if none can reach the peer, then an "error" is reported back to the RML, which then calls the errmgr for instructions.

* opal_buffer_t now conforms to standard object rules re OBJ_RETAIN as we no longer "unload" the incoming object

* NIC failure is reported to the TCP component, which then tries to resend the message across any other available TCP NIC. If that doesn't work, then the message is given back to the OOB base to try using other components. If all that fails, then the error is reported to the RML, which reports to the errmgr for instructions

* obviously from the above, multiple OOB components (e.g., TCP and UD) can be active in parallel

* the matching code has been moved to the RML (and out of the OOB/TCP component) so it is independent of transport

* routing is done by the individual OOB modules (as opposed to the RML). Thus, both routed and non-routed transports can simultaneously be active

* all blocking send/recv APIs have been removed. Everything operates asynchronously.


KNOWN LIMITATIONS:

* although provision is made for component failover as described above, the code for doing so has not been fully implemented yet. At the moment, if all connections for a given peer fail, the errmgr is notified of a "lost connection", which by default results in termination of the job if it was a lifeline

* the IPv6 code is present and compiles, but is not complete. Since the current IPv6 support in the OOB doesn't work anyway, I don't consider this a blocker

* routing is performed at the individual module level, yet the active routed component is selected on a global basis. We probably should update that to reflect that different transports may need/choose to route in different ways

* obviously, not every error path has been tested nor necessarily covered

* determining abnormal termination is more challenging than in the old code as we now potentially have multiple ways of connecting to a process. Ideally, we would declare "connection failed" when *all* transports can no longer reach the process, but that requires some additional (possibly complex) code. For now, the code replicates the old behavior only somewhat modified - i.e., if a module sees its connection fail, it checks to see if it is a lifeline. If so, it notifies the errmgr that the lifeline is lost - otherwise, it notifies the errmgr that a non-lifeline connection was lost.

* reachability is determined solely on the basis of a shared subnet address/mask - more sophisticated algorithms (e.g., the one used in the tcp btl) are required to handle routing via gateways

* the RML needs to assign sequence numbers to each message on a per-peer basis. The receiving RML will then deliver messages in order, thus preventing out-of-order messaging in the case where messages travel across different transports or a message needs to be redirected/resent due to failure of a NIC

This commit was SVN r29058.
2013-08-22 16:37:40 +00:00

335 строки
10 KiB
C

/*
* Copyright (c) 2004-2007 The Trustees of Indiana University and Indiana
* University Research and Technology
* Corporation. All rights reserved.
* Copyright (c) 2004-2009 The University of Tennessee and The University
* of Tennessee Research Foundation. All rights
* reserved.
* Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
* University of Stuttgart. All rights reserved.
* Copyright (c) 2004-2005 The Regents of the University of California.
* All rights reserved.
* Copyright (c) 2011 Cisco Systems, Inc. All rights reserved.
* $COPYRIGHT$
*
* Additional copyrights may follow
*
* $HEADER$
*/
/**
* @file
*/
#ifndef MCA_BTL_TCP_H
#define MCA_BTL_TCP_H
#include "ompi_config.h"
#ifdef HAVE_SYS_TYPES_H
#include <sys/types.h>
#endif
#ifdef HAVE_SYS_SOCKET_H
#include <sys/socket.h>
#endif
#ifdef HAVE_NETINET_IN_H
#include <netinet/in.h>
#endif
/* Open MPI includes */
#include "opal/mca/event/event.h"
#include "ompi/class/ompi_free_list.h"
#include "ompi/mca/btl/btl.h"
#include "ompi/mca/btl/base/base.h"
#include "ompi/mca/mpool/mpool.h"
#include "ompi/mca/btl/btl.h"
#include "opal/class/opal_hash_table.h"
#define MCA_BTL_TCP_STATISTICS 0
BEGIN_C_DECLS
/**
* TCP BTL component.
*/
struct mca_btl_tcp2_component_t {
mca_btl_base_component_2_0_0_t super; /**< base BTL component */
uint32_t tcp_addr_count; /**< total number of addresses */
uint32_t tcp_num_btls; /**< number of hcas available to the TCP component */
uint32_t tcp_num_links; /**< number of logical links per physical device */
struct mca_btl_tcp2_module_t **tcp_btls; /**< array of available BTL modules */
struct mca_btl_tcp2_proc_t* tcp_local; /**< local proc struct */
int tcp_free_list_num; /**< initial size of free lists */
int tcp_free_list_max; /**< maximum size of free lists */
int tcp_free_list_inc; /**< number of elements to alloc when growing free lists */
int tcp_endpoint_cache; /**< amount of cache on each endpoint */
opal_hash_table_t tcp_procs; /**< hash table of tcp proc structures */
opal_list_t tcp_events; /**< list of pending tcp events */
opal_mutex_t tcp_lock; /**< lock for accessing module state */
opal_event_t tcp_recv_event; /**< recv event for IPv4 listen socket */
int tcp_listen_sd; /**< IPv4 listen socket for incoming connection requests */
unsigned short tcp_listen_port; /**< IPv4 listen port */
int32_t tcp_port_min; /**< IPv4 minimum port */
int32_t tcp_port_range; /**< IPv4 port range */
#if OPAL_ENABLE_IPV6
opal_event_t tcp6_recv_event; /**< recv event for IPv6 listen socket */
int tcp6_listen_sd; /**< IPv6 listen socket for incoming connection requests */
unsigned short tcp6_listen_port; /**< IPv6 listen port */
int32_t tcp6_port_min; /**< IPv4 minimum port */
int32_t tcp6_port_range; /**< IPv4 port range */
#endif
/* Port range restriction */
char* tcp_if_include; /**< comma seperated list of interface to include */
char* tcp_if_exclude; /**< comma seperated list of interface to exclude */
int tcp_sndbuf; /**< socket sndbuf size */
int tcp_rcvbuf; /**< socket rcvbuf size */
int tcp_disable_family; /**< disabled AF_family */
/* free list of fragment descriptors */
ompi_free_list_t tcp_frag_eager;
ompi_free_list_t tcp_frag_max;
ompi_free_list_t tcp_frag_user;
/* Do we want to use TCP_NODELAY? */
int tcp_use_nodelay;
};
typedef struct mca_btl_tcp2_component_t mca_btl_tcp2_component_t;
OMPI_MODULE_DECLSPEC extern mca_btl_tcp2_component_t mca_btl_tcp2_component;
/**
* BTL Module Interface
*/
struct mca_btl_tcp2_module_t {
mca_btl_base_module_t super; /**< base BTL interface */
uint16_t tcp_ifkindex; /** <BTL kernel interface index */
#if 0
int tcp_ifindex; /**< PTL interface index */
#endif
struct sockaddr_storage tcp_ifaddr; /**< PTL interface address */
uint32_t tcp_ifmask; /**< PTL interface netmask */
opal_list_t tcp_endpoints;
#if MCA_BTL_TCP_STATISTICS
size_t tcp_bytes_sent;
size_t tcp_bytes_recv;
size_t tcp_send_handler;
#endif
};
typedef struct mca_btl_tcp2_module_t mca_btl_tcp2_module_t;
extern mca_btl_tcp2_module_t mca_btl_tcp2_module;
#define CLOSE_THE_SOCKET(socket) close(socket)
/**
* Register TCP component parameters with the MCA framework
*/
extern int mca_btl_tcp2_component_open(void);
/**
* Any final cleanup before being unloaded.
*/
extern int mca_btl_tcp2_component_close(void);
/**
* TCP component initialization.
*
* @param num_btl_modules (OUT) Number of BTLs returned in BTL array.
* @param allow_multi_user_threads (OUT) Flag indicating wether BTL supports user threads (TRUE)
* @param have_hidden_threads (OUT) Flag indicating wether BTL uses threads (TRUE)
*/
extern mca_btl_base_module_t** mca_btl_tcp2_component_init(
int *num_btl_modules,
bool allow_multi_user_threads,
bool have_hidden_threads
);
/**
* TCP component control.
*/
int mca_btl_tcp2_component_control(
int param,
void* value,
size_t size
);
/**
* TCP component progress.
*/
extern int mca_btl_tcp2_component_progress(void);
/**
* Cleanup any resources held by the BTL.
*
* @param btl BTL instance.
* @return OMPI_SUCCESS or error status on failure.
*/
extern int mca_btl_tcp2_finalize(
struct mca_btl_base_module_t* btl
);
/**
* PML->BTL notification of change in the process list.
*
* @param btl (IN)
* @param nprocs (IN) Number of processes
* @param procs (IN) Set of processes
* @param peers (OUT) Set of (optional) peer addressing info.
* @param peers (IN/OUT) Set of processes that are reachable via this BTL.
* @return OMPI_SUCCESS or error status on failure.
*
*/
extern int mca_btl_tcp2_add_procs(
struct mca_btl_base_module_t* btl,
size_t nprocs,
struct ompi_proc_t **procs,
struct mca_btl_base_endpoint_t** peers,
opal_bitmap_t* reachable
);
/**
* PML->BTL notification of change in the process list.
*
* @param btl (IN) BTL instance
* @param nproc (IN) Number of processes.
* @param procs (IN) Set of processes.
* @param peers (IN) Set of peer data structures.
* @return Status indicating if cleanup was successful
*
*/
extern int mca_btl_tcp2_del_procs(
struct mca_btl_base_module_t* btl,
size_t nprocs,
struct ompi_proc_t **procs,
struct mca_btl_base_endpoint_t** peers
);
/**
* Initiate an asynchronous send.
*
* @param btl (IN) BTL module
* @param endpoint (IN) BTL addressing information
* @param descriptor (IN) Description of the data to be transfered
* @param tag (IN) The tag value used to notify the peer.
*/
extern int mca_btl_tcp2_send(
struct mca_btl_base_module_t* btl,
struct mca_btl_base_endpoint_t* btl_peer,
struct mca_btl_base_descriptor_t* descriptor,
mca_btl_base_tag_t tag
);
/**
* Initiate an asynchronous put.
*
* @param btl (IN) BTL module
* @param endpoint (IN) BTL addressing information
* @param descriptor (IN) Description of the data to be transferred
*/
extern int mca_btl_tcp2_put(
struct mca_btl_base_module_t* btl,
struct mca_btl_base_endpoint_t* btl_peer,
struct mca_btl_base_descriptor_t* decriptor
);
/**
* Initiate an asynchronous get.
*
* @param btl (IN) BTL module
* @param endpoint (IN) BTL addressing information
* @param descriptor (IN) Description of the data to be transferred
*/
extern int mca_btl_tcp2_get(
struct mca_btl_base_module_t* btl,
struct mca_btl_base_endpoint_t* btl_peer,
struct mca_btl_base_descriptor_t* decriptor
);
/**
* Allocate a descriptor with a segment of the requested size.
* Note that the BTL layer may choose to return a smaller size
* if it cannot support the request.
*
* @param btl (IN) BTL module
* @param size (IN) Request segment size.
*/
extern mca_btl_base_descriptor_t* mca_btl_tcp2_alloc(
struct mca_btl_base_module_t* btl,
struct mca_btl_base_endpoint_t* endpoint,
uint8_t order,
size_t size,
uint32_t flags);
/**
* Return a segment allocated by this BTL.
*
* @param btl (IN) BTL module
* @param descriptor (IN) Allocated descriptor.
*/
extern int mca_btl_tcp2_free(
struct mca_btl_base_module_t* btl,
mca_btl_base_descriptor_t* des);
/**
* Prepare a descriptor for send/rdma using the supplied
* convertor. If the convertor references data that is contigous,
* the descriptor may simply point to the user buffer. Otherwise,
* this routine is responsible for allocating buffer space and
* packing if required.
*
* @param btl (IN) BTL module
* @param endpoint (IN) BTL peer addressing
* @param convertor (IN) Data type convertor
* @param reserve (IN) Additional bytes requested by upper layer to precede user data
* @param size (IN/OUT) Number of bytes to prepare (IN), number of bytes actually prepared (OUT)
*/
mca_btl_base_descriptor_t* mca_btl_tcp2_prepare_src(
struct mca_btl_base_module_t* btl,
struct mca_btl_base_endpoint_t* peer,
struct mca_mpool_base_registration_t*,
struct opal_convertor_t* convertor,
uint8_t order,
size_t reserve,
size_t* size,
uint32_t flags
);
extern mca_btl_base_descriptor_t* mca_btl_tcp2_prepare_dst(
struct mca_btl_base_module_t* btl,
struct mca_btl_base_endpoint_t* peer,
struct mca_mpool_base_registration_t*,
struct opal_convertor_t* convertor,
uint8_t order,
size_t reserve,
size_t* size,
uint32_t flags);
/**
* Fault Tolerance Event Notification Function
* @param state Checkpoint Stae
* @return OMPI_SUCCESS or failure status
*/
int mca_btl_tcp2_ft_event(int state);
END_C_DECLS
#endif