a200e4f865
*** 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.
176 строки
6.6 KiB
C
176 строки
6.6 KiB
C
/*
|
|
* Copyright (c) 2004-2005 The Trustees of Indiana University and Indiana
|
|
* University Research and Technology
|
|
* Corporation. All rights reserved.
|
|
* Copyright (c) 2004-2005 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) 2012-2013 Los Alamos National Security, LLC. All rights
|
|
* reserved.
|
|
* $COPYRIGHT$
|
|
*
|
|
* Additional copyrights may follow
|
|
*
|
|
* $HEADER$
|
|
*/
|
|
/** @file:
|
|
*
|
|
* the oob framework
|
|
*/
|
|
|
|
#ifndef _MCA_OOB_BASE_H_
|
|
#define _MCA_OOB_BASE_H_
|
|
|
|
#include "orte_config.h"
|
|
|
|
#ifdef HAVE_UNISTD_H
|
|
#include <unistd.h>
|
|
#endif
|
|
#ifdef HAVE_SYS_UIO_H
|
|
#include <sys/uio.h>
|
|
#endif
|
|
#ifdef HAVE_NET_UIO_H
|
|
#include <net/uio.h>
|
|
#endif
|
|
|
|
#include "opal/class/opal_bitmap.h"
|
|
#include "opal/class/opal_hash_table.h"
|
|
#include "opal/class/opal_list.h"
|
|
#include "opal/mca/mca.h"
|
|
#include "opal/mca/event/event.h"
|
|
|
|
#include "orte/mca/oob/oob.h"
|
|
|
|
BEGIN_C_DECLS
|
|
|
|
/*
|
|
* Convenience Typedef
|
|
*/
|
|
typedef struct {
|
|
char *include;
|
|
char *exclude;
|
|
opal_list_t components;
|
|
opal_list_t actives;
|
|
int max_uri_length;
|
|
opal_hash_table_t peers;
|
|
bool use_module_threads;
|
|
} orte_oob_base_t;
|
|
ORTE_DECLSPEC extern orte_oob_base_t orte_oob_base;
|
|
|
|
typedef struct {
|
|
opal_object_t super;
|
|
mca_oob_base_component_t *component;
|
|
opal_bitmap_t addressable;
|
|
} orte_oob_base_peer_t;
|
|
OBJ_CLASS_DECLARATION(orte_oob_base_peer_t);
|
|
|
|
/* MCA framework */
|
|
ORTE_DECLSPEC extern mca_base_framework_t orte_oob_base_framework;
|
|
ORTE_DECLSPEC int orte_oob_base_select(void);
|
|
|
|
/* Access the OOB internal functions via set of event-based macros
|
|
* for inserting messages and other commands into the
|
|
* OOB event base. This ensures that all OOB operations occur
|
|
* asynchronously in a thread-safe environment.
|
|
* Note that this doesn't mean that messages will be *sent*
|
|
* in order as that depends on the specific transport being
|
|
* used, when that module's event base indicates the transport
|
|
* is available, etc.
|
|
*/
|
|
typedef struct {
|
|
opal_object_t super;
|
|
opal_event_t ev;
|
|
orte_rml_send_t *msg;
|
|
} orte_oob_send_t;
|
|
OBJ_CLASS_DECLARATION(orte_oob_send_t);
|
|
|
|
/* All OOB sends are based on iovec's and are async as the RML
|
|
* acts as the initial interface to prepare all communications.
|
|
* The send_nb function will enter the message into the OOB
|
|
* base, which will then check to see if a transport for the
|
|
* intended target has already been assigned. If so, the message
|
|
* is immediately placed into that module's event base for
|
|
* transmission. If not, the function will loop across all available
|
|
* components until one identifies that it has a module capable
|
|
* of reaching the target.
|
|
*/
|
|
typedef void (*mca_oob_send_callback_fn_t)(int status,
|
|
struct iovec *iov,
|
|
int count, void *cbdata);
|
|
|
|
ORTE_DECLSPEC void orte_oob_base_send_nb(int fd, short args, void *cbdata);
|
|
#define ORTE_OOB_SEND(m) \
|
|
do { \
|
|
orte_oob_send_t *cd; \
|
|
cd = OBJ_NEW(orte_oob_send_t); \
|
|
cd->msg = (m); \
|
|
opal_event_set(orte_event_base, &cd->ev, -1, \
|
|
OPAL_EV_WRITE, \
|
|
orte_oob_base_send_nb, cd); \
|
|
opal_event_set_priority(&cd->ev, ORTE_MSG_PRI); \
|
|
opal_event_active(&cd->ev, OPAL_EV_WRITE, 1); \
|
|
}while(0);
|
|
|
|
/* Our contact info is actually subject to change as transports
|
|
* can fail at any time. So a request to obtain our URI requires
|
|
* that we get a snapshot in time. Since the request always comes
|
|
* thru the rml, and we share that event base, we can just cycle
|
|
* across the components to collect the info.
|
|
*
|
|
* During initial wireup, we can only transfer contact info on the daemon
|
|
* command line. This limits what we can send to a string representation of
|
|
* the actual contact info, which gets sent in a uri-like form. Not every
|
|
* oob module can support this transaction, so this function will loop
|
|
* across all oob components/modules, letting each add to the uri string if
|
|
* it supports bootstrap operations. An error will be returned in the cbfunc
|
|
* if NO component can successfully provide a contact.
|
|
*
|
|
* Note: since there is a limit to what an OS will allow on a cmd line, we
|
|
* impose a limit on the length of the resulting uri via an MCA param. The
|
|
* default value of -1 implies unlimited - however, users with large numbers
|
|
* of interfaces on their nodes may wish to restrict the size.
|
|
*
|
|
* Since all components define their address info at component start,
|
|
* it is unchanged and does not require acess via event
|
|
*/
|
|
#define ORTE_OOB_GET_URI(u) orte_oob_base_get_addr(u)
|
|
ORTE_DECLSPEC void orte_oob_base_get_addr(char **uri);
|
|
|
|
/**
|
|
* Extract initial contact information from a string uri
|
|
*
|
|
* During initial wireup, we can only transfer contact info on the daemon
|
|
* command line. This limits what we can send to a string representation of
|
|
* the actual contact info, which gets sent in a uri-like form. Not every
|
|
* oob module can support this transaction, so this function will loop
|
|
* across all oob components/modules, letting each look at the uri and extract
|
|
* info from it if it can.
|
|
*/
|
|
typedef struct {
|
|
opal_object_t super;
|
|
opal_event_t ev;
|
|
char *uri;
|
|
} mca_oob_uri_req_t;
|
|
OBJ_CLASS_DECLARATION(mca_oob_uri_req_t);
|
|
|
|
#define ORTE_OOB_SET_URI(u) \
|
|
do { \
|
|
mca_oob_uri_req_t *rq; \
|
|
rq = OBJ_NEW(mca_oob_uri_req_t); \
|
|
rq->uri = strdup((u)); \
|
|
opal_event_set(orte_event_base, &(rq)->ev, -1, \
|
|
OPAL_EV_WRITE, \
|
|
orte_oob_base_set_addr, (rq)); \
|
|
opal_event_set_priority(&(rq)->ev, ORTE_MSG_PRI); \
|
|
opal_event_active(&(rq)->ev, OPAL_EV_WRITE, 1); \
|
|
}while(0);
|
|
ORTE_DECLSPEC void orte_oob_base_set_addr(int fd, short args, void *cbdata);
|
|
|
|
END_C_DECLS
|
|
#endif
|
|
|