Ok1zia at 10:36, 3 January 2010

2010-01-03T10:36:18Z

← Older revision		Revision as of 10:36, 3 January 2010
Line 44:		Line 44:
	19018 ? Ss 0:00 /usr/bin/ax25d -l -n		19018 ? Ss 0:00 /usr/bin/ax25d -l -n

	Patched source is here: [[File:ax25d-zia.c]]		Patched source is here: [[File:ax25d-zia.c]] and patch [[File:ax25d-zia.diff]]. All is against ax25d from ax25-tools 0.0.8.

Ok1zia at 10:33, 3 January 2010

2010-01-03T10:33:09Z

← Older revision		Revision as of 10:33, 3 January 2010
Line 44:		Line 44:
	19018 ? Ss 0:00 /usr/bin/ax25d -l -n		19018 ? Ss 0:00 /usr/bin/ax25d -l -n

	Patched source is here: [[ax25d-zia.c]]		Patched source is here: [[File:ax25d-zia.c]]

Ok1zia: Created page with 'On OK0NAG, the ax25 after some days crashed. In this article you can find workaround. Ax25d exited with these lines in /var/log/syslog: Dec 30 23:49:07 nagano ax25d[680]: AX.2...'

2010-01-03T10:28:33Z

Created page with 'On OK0NAG, the ax25 after some days crashed. In this article you can find workaround. Ax25d exited with these lines in /var/log/syslog: Dec 30 23:49:07 nagano ax25d[680]: AX.2...'

New page

On OK0NAG, the ax25 after some days crashed. In this article you can find workaround.

Ax25d exited with these lines in /var/log/syslog:

Dec 30 23:49:07 nagano ax25d[680]: AX.25 OK1HMP-0 (ax0) client
Dec 30 23:49:07 nagano ax25d[718]: AX.25 OK1HMP-0 (ax0) client
Dec 30 23:49:07 nagano ax25d[17984]: accept error Software caused connection abort, closing socket on port ax0

I looked into source code and found this code:

i = TRUE;
ioctl(paxl->fd, FIONBIO, &i);

addrlen = sizeof(struct full_sockaddr_ax25);
new = accept(paxl->fd, (struct sockaddr *)&sockaddr, &addrlen);

i = FALSE;
ioctl(paxl->fd, FIONBIO, &i);

if (new < 0) {
if (errno == EWOULDBLOCK)
continue; /* It's gone ??? */

if (Logging)
syslog(LOG_ERR, "accept error %m, closing socket on port %s", paxl->port);
close(paxl->fd);
paxl->fd = -1;
continue;
}

Problem is in accept, which returns error. I don't know why. I think If I would ignore it, the deamon can stay in endless loop. I've found (hopefully) better solution - the daemon will not fork. When this error occurs, it exites and will be respawned again. It must be removed from init scripts:

update-rc.d -f ax25d remove

Simplest way to respawn it is init. Add similar line to /etc/inittab:

9:23:respawn:/usr/bin/zax25d -l -n

And tell about change:
telinit q

Check if process is running:
# ps xa|grep ax25
19018 ? Ss 0:00 /usr/bin/ax25d -l -n

Patched source is here: [[ax25d-zia.c]]

Ax25d patch - Revision history

Ok1zia at 10:36, 3 January 2010

Ok1zia at 10:33, 3 January 2010

Ok1zia: Created page with 'On OK0NAG, the ax25 after some days crashed. In this article you can find workaround. Ax25d exited with these lines in /var/log/syslog: Dec 30 23:49:07 nagano ax25d[680]: AX.2...'