Fixes bug in scheduling of callback by main loop.
If other side of a socket is sending data faster than BIRD is processing, BIRD does not schedule any other callbacks (events, timers, rx/tx callbacks).
This commit is contained in:
parent
2757985709
commit
4323099da9
1 changed files with 13 additions and 2 deletions
|
@ -30,6 +30,12 @@
|
|||
#include "lib/unix.h"
|
||||
#include "lib/sysio.h"
|
||||
|
||||
/* Maximum number of calls of rx/tx handler for one socket in one
|
||||
* select iteration. Should be small enough to not monopolize CPU by
|
||||
* one protocol instance.
|
||||
*/
|
||||
#define MAX_STEPS 4
|
||||
|
||||
/*
|
||||
* Tracked Files
|
||||
*/
|
||||
|
@ -1341,22 +1347,27 @@ io_loop(void)
|
|||
{
|
||||
sock *s = current_sock;
|
||||
int e;
|
||||
int steps = MAX_STEPS;
|
||||
if (FD_ISSET(s->fd, &rd) && s->rx_hook)
|
||||
do
|
||||
{
|
||||
steps--;
|
||||
e = sk_read(s);
|
||||
if (s != current_sock)
|
||||
goto next;
|
||||
}
|
||||
while (e && s->rx_hook);
|
||||
while (e && s->rx_hook && steps);
|
||||
|
||||
steps = MAX_STEPS;
|
||||
if (FD_ISSET(s->fd, &wr))
|
||||
do
|
||||
{
|
||||
steps--;
|
||||
e = sk_write(s);
|
||||
if (s != current_sock)
|
||||
goto next;
|
||||
}
|
||||
while (e);
|
||||
while (e && steps);
|
||||
current_sock = sk_next(s);
|
||||
next: ;
|
||||
}
|
||||
|
|
Loading…
Reference in a new issue