mirrors/bird
mirrors/bird
0
0
Fork 0
mirror of https://gitlab.nic.cz/labs/bird.git synced 2024-11-13 22:58:42 +00:00
Code Releases Activity
e6baff89f8
bird/lib/event.h

76 lines
1.8 KiB
C
Raw Normal View History

Grrr, forgot to commit the event routines themselves :|
1999-02-11 22:18:36 +00:00
/*
* BIRD Library -- Event Processing
*
* (c) 1999 Martin Mares <mj@ucw.cz>
*
* Can be freely distributed and used under the terms of the GNU GPL.
*/
#ifndef _BIRD_EVENT_H_
#define _BIRD_EVENT_H_
#include "lib/resource.h"
Converting the former BFD loop to a universal IO loop and protocol loop. There is a simple universal IO loop, taking care of events, timers and sockets. Primarily, one instance of a protocol should use exactly one IO loop to do all its work, as is now done in BFD. Contrary to previous versions, the loop is now launched and cleaned by the nest/proto.c code, allowing for a protocol to just request its own loop by setting the loop's lock order in config higher than the_bird. It is not supported nor checked if any protocol changed the requested lock order in reconfigure. No protocol should do it at all.
2021-06-19 18:50:18 +00:00
#include "lib/locking.h"
#include <stdatomic.h>
Event lists rewritten to a single linked list In multithreaded environment, we need to pass messages between workers. This is done by queuing events to their respective queues. The double-linked list is not really useful for that as it needs locking everywhere. This commit rewrites the event subsystem to use a single-linked list where events are enqueued by a single atomic instruction and the queue is processed after atomically moving the whole queue aside.
2022-06-24 17:53:34 +00:00
struct birdloop;
Grrr, forgot to commit the event routines themselves :|
1999-02-11 22:18:36 +00:00
typedef struct event {
resource r;
Event handlers no longer return re-queue flag. Instead of using it, just call ev_schedule() on the same handler which should work perfectly now.
2000-04-27 22:28:49 +00:00
void (*hook)(void *);
Grrr, forgot to commit the event routines themselves :|
1999-02-11 22:18:36 +00:00
void *data;
Event lists rewritten to a single linked list In multithreaded environment, we need to pass messages between workers. This is done by queuing events to their respective queues. The double-linked list is not really useful for that as it needs locking everywhere. This commit rewrites the event subsystem to use a single-linked list where events are enqueued by a single atomic instruction and the queue is processed after atomically moving the whole queue aside.
2022-06-24 17:53:34 +00:00
struct event * _Atomic next;
Fixing several race-conditions in event code. After a suggestion by Santiago, I added the direct list pointer into events and the events are now using this value to check whether the route is active or not. Also the whole trick with sentinel node unioned with event list is now gone. For debugging, there is also an internal circular buffer to store what has been recently happening in event code before e.g. a crash happened. By default, this debug is off and must be manually enabled in lib/event.c as it eats quite some time and space.
2022-09-12 09:09:43 +00:00
struct event_list * _Atomic list;
Grrr, forgot to commit the event routines themselves :|
1999-02-11 22:18:36 +00:00
} event;
Fixing several race-conditions in event code. After a suggestion by Santiago, I added the direct list pointer into events and the events are now using this value to check whether the route is active or not. Also the whole trick with sentinel node unioned with event list is now gone. For debugging, there is also an internal circular buffer to store what has been recently happening in event code before e.g. a crash happened. By default, this debug is off and must be manually enabled in lib/event.c as it eats quite some time and space.
2022-09-12 09:09:43 +00:00
typedef struct event_list {
event * _Atomic receiver; /* Event receive list */
event * _Atomic _executor; /* Event execute list */
const char *name;
struct birdloop *loop; /* The executor loop */
Converting the former BFD loop to a universal IO loop and protocol loop. There is a simple universal IO loop, taking care of events, timers and sockets. Primarily, one instance of a protocol should use exactly one IO loop to do all its work, as is now done in BFD. Contrary to previous versions, the loop is now launched and cleaned by the nest/proto.c code, allowing for a protocol to just request its own loop by setting the loop's lock order in config higher than the_bird. It is not supported nor checked if any protocol changed the requested lock order in reconfigure. No protocol should do it at all.
2021-06-19 18:50:18 +00:00
} event_list;
Grrr, forgot to commit the event routines themselves :|
1999-02-11 22:18:36 +00:00
extern event_list global_event_list;
Rate-limit scheduling of work-events In general, events are code handling some some condition, which is scheduled when such condition happened and executed independently from I/O loop. Work-events are a subgroup of events that are scheduled repeatedly until some (often significant) work is done (e.g. feeding routes to protocol). All scheduled events are executed during each I/O loop iteration. Separate work-events from regular events to a separate queue and rate limit their execution to a fixed number per I/O loop iteration. That should prevent excess latency when many work-events are scheduled at one time (e.g. simultaneous reload of many BGP sessions).
2021-03-12 14:35:56 +00:00
extern event_list global_work_list;
Grrr, forgot to commit the event routines themselves :|
1999-02-11 22:18:36 +00:00
event *ev_new(pool *);
Event handlers no longer return re-queue flag. Instead of using it, just call ev_schedule() on the same handler which should work perfectly now.
2000-04-27 22:28:49 +00:00
void ev_run(event *);
Event lists rewritten to a single linked list In multithreaded environment, we need to pass messages between workers. This is done by queuing events to their respective queues. The double-linked list is not really useful for that as it needs locking everywhere. This commit rewrites the event subsystem to use a single-linked list where events are enqueued by a single atomic instruction and the queue is processed after atomically moving the whole queue aside.
2022-06-24 17:53:34 +00:00
void ev_init_list(event_list *, struct birdloop *loop, const char *name);
void ev_enqueue(event_list *, event *);
#define ev_send ev_enqueue
Converting the former BFD loop to a universal IO loop and protocol loop. There is a simple universal IO loop, taking care of events, timers and sockets. Primarily, one instance of a protocol should use exactly one IO loop to do all its work, as is now done in BFD. Contrary to previous versions, the loop is now launched and cleaned by the nest/proto.c code, allowing for a protocol to just request its own loop by setting the loop's lock order in config higher than the_bird. It is not supported nor checked if any protocol changed the requested lock order in reconfigure. No protocol should do it at all.
2021-06-19 18:50:18 +00:00
#define ev_send_loop(l, e) ev_send(birdloop_event_list((l)), (e))
#define ev_schedule(e) ({ ASSERT_THE_BIRD_LOCKED; if (!ev_active((e))) ev_send(&global_event_list, (e)); })
#define ev_schedule_work(e) ({ ASSERT_THE_BIRD_LOCKED; if (!ev_active((e))) ev_send(&global_work_list, (e)); })
Grrr, forgot to commit the event routines themselves :|
1999-02-11 22:18:36 +00:00
void ev_postpone(event *);
Rate-limit scheduling of work-events In general, events are code handling some some condition, which is scheduled when such condition happened and executed independently from I/O loop. Work-events are a subgroup of events that are scheduled repeatedly until some (often significant) work is done (e.g. feeding routes to protocol). All scheduled events are executed during each I/O loop iteration. Separate work-events from regular events to a separate queue and rate limit their execution to a fixed number per I/O loop iteration. That should prevent excess latency when many work-events are scheduled at one time (e.g. simultaneous reload of many BGP sessions).
2021-03-12 14:35:56 +00:00
int ev_run_list_limited(event_list *, uint);
Event lists rewritten to a single linked list In multithreaded environment, we need to pass messages between workers. This is done by queuing events to their respective queues. The double-linked list is not really useful for that as it needs locking everywhere. This commit rewrites the event subsystem to use a single-linked list where events are enqueued by a single atomic instruction and the queue is processed after atomically moving the whole queue aside.
2022-06-24 17:53:34 +00:00
#define ev_run_list(l) ev_run_list_limited((l), ~0)
Grrr, forgot to commit the event routines themselves :|
1999-02-11 22:18:36 +00:00
Converting the former BFD loop to a universal IO loop and protocol loop. There is a simple universal IO loop, taking care of events, timers and sockets. Primarily, one instance of a protocol should use exactly one IO loop to do all its work, as is now done in BFD. Contrary to previous versions, the loop is now launched and cleaned by the nest/proto.c code, allowing for a protocol to just request its own loop by setting the loop's lock order in config higher than the_bird. It is not supported nor checked if any protocol changed the requested lock order in reconfigure. No protocol should do it at all.
2021-06-19 18:50:18 +00:00
#define LEGACY_EVENT_LIST(l) (((l) == &global_event_list) || ((l) == &global_work_list))
Fininshing integrated OSPF.
2014-11-03 09:42:55 +00:00
static inline int
ev_active(event *e)
{
Fixing several race-conditions in event code. After a suggestion by Santiago, I added the direct list pointer into events and the events are now using this value to check whether the route is active or not. Also the whole trick with sentinel node unioned with event list is now gone. For debugging, there is also an internal circular buffer to store what has been recently happening in event code before e.g. a crash happened. By default, this debug is off and must be manually enabled in lib/event.c as it eats quite some time and space.
2022-09-12 09:09:43 +00:00
return atomic_load_explicit(&e->list, memory_order_acquire) != NULL;
Event lists rewritten to a single linked list In multithreaded environment, we need to pass messages between workers. This is done by queuing events to their respective queues. The double-linked list is not really useful for that as it needs locking everywhere. This commit rewrites the event subsystem to use a single-linked list where events are enqueued by a single atomic instruction and the queue is processed after atomically moving the whole queue aside.
2022-06-24 17:53:34 +00:00
}
Converting the former BFD loop to a universal IO loop and protocol loop. There is a simple universal IO loop, taking care of events, timers and sockets. Primarily, one instance of a protocol should use exactly one IO loop to do all its work, as is now done in BFD. Contrary to previous versions, the loop is now launched and cleaned by the nest/proto.c code, allowing for a protocol to just request its own loop by setting the loop's lock order in config higher than the_bird. It is not supported nor checked if any protocol changed the requested lock order in reconfigure. No protocol should do it at all.
2021-06-19 18:50:18 +00:00
Event lists rewritten to a single linked list In multithreaded environment, we need to pass messages between workers. This is done by queuing events to their respective queues. The double-linked list is not really useful for that as it needs locking everywhere. This commit rewrites the event subsystem to use a single-linked list where events are enqueued by a single atomic instruction and the queue is processed after atomically moving the whole queue aside.
2022-06-24 17:53:34 +00:00
static inline event_list *
ev_get_list(event *e)
{
Fixing several race-conditions in event code. After a suggestion by Santiago, I added the direct list pointer into events and the events are now using this value to check whether the route is active or not. Also the whole trick with sentinel node unioned with event list is now gone. For debugging, there is also an internal circular buffer to store what has been recently happening in event code before e.g. a crash happened. By default, this debug is off and must be manually enabled in lib/event.c as it eats quite some time and space.
2022-09-12 09:09:43 +00:00
return atomic_load_explicit(&e->list, memory_order_acquire);
Fininshing integrated OSPF.
2014-11-03 09:42:55 +00:00
}
Lib: Add and use ev_new_init()
2018-10-01 13:55:23 +00:00
static inline event*
ev_new_init(pool *p, void (*hook)(void *), void *data)
{
event *e = ev_new(p);
e->hook = hook;
e->data = data;
return e;
}
Fininshing integrated OSPF.
2014-11-03 09:42:55 +00:00
Grrr, forgot to commit the event routines themselves :|
1999-02-11 22:18:36 +00:00
#endif
Copy Permalink