Recent changes to bugs

#17 new protocol, rawudp

Corey Minyard — Tue, 03 Oct 2017 12:59:35 -0000

status: open --> closed
Group: --> v1.0_(example)

#20 Fedora 18 - no listen port created

Corey Minyard — Tue, 28 Jun 2016 19:34:29 -0000

status: open --> closed-works-for-me
Group: --> v1.0_(example)

Fedora 18 - no listen port created

Paul — Fri, 15 Mar 2013 14:00:49 -0000

When running ser2net on Fedora 18 I am only seeing ipv6 ports being opened.

When I try and connect to the ipv4 equivalent (1028 in this case) ser2net produces a segmentation fault and exits

Fedora 18 - no listen port created

Paul — Fri, 15 Mar 2013 14:00:49 -0000

Ticket 20 has been modified: Fedora 18 - no listen port created
Edited By: Corey Minyard (cminyard)
Status updated: u'open' => u'closed-works-for-me'
milestone updated: '' => u'v1.0(example)'

cannot start if ipv6 module is not loaded

michube — Tue, 29 Jan 2013 19:46:48 -0000

error:
pi@raspberrypi ~/ser2net-2.8-rc3 $ sudo ser2net -n -d -u -p 333
ser2net[2563]: Error on line 3, Unable to create TCP socket
ser2net[2563]: Unable to create TCP socket: Address family not supported by protocol
Unable to open control port, see syslog: 333

i found a similar issue here: http://www.spinics.net/lists/linux-nfs/msg21046.html

workarround: i inserted the ipv6 module, then it works.

Segmentation fault on conenct

Anonymous — Wed, 26 Sep 2012 14:29:25 -0000

I've compiled under Debian the latst 2.8 rc2 and it crash everytime I try to open a network conenction to the serial port. It won't crash when connect to the control port, only for serial port. Crash regardless of port or type or anything. After connect, it just wait for half a second or so and then crash. This happened even if no device is connected to that serial port.

The previous 2.7 version works on same machine, must be a new bug added with this rc2. I haven't tested rc1.

Any solution?

new protocol, rawudp

rspadim — Mon, 28 Nov 2011 17:34:54 -0000

hi guys,
could we add another raw protocol?
this time is: rawudp
it´s like raw
but instead TCP socket, use a UDP socket, and send all received data from serial to last computer that sent something, for example
computer connect to serial from ip 172.16.0.1 port 1234
all received data from serial is sent back to 172.16.0.1 por 1234
ok it must send something before ser2net know what´s ip and port is 'connected' since udp is connectionless
maybe another parameter could be added when no connection is done, send to a default ip/port previous configured (at ser2net.conf or command line)

why? many microcontrollers have udp-serial converter, but i can´t test it with linux machines, since i don´t have one here to test (i´m just writing code with UDP, but my final user test without i have tested, if i could test it before send program to user could be very nice)

new protocol, rawudp

rspadim — Mon, 28 Nov 2011 17:34:54 -0000

Ticket 17 has been modified: new protocol, rawudp
Edited By: Corey Minyard (cminyard)
Status updated: u'open' => u'closed'
milestone updated: '' => u'v1.0(example)'

ARM - SIGBUS

Anonymous — Fri, 12 Aug 2011 01:26:56 -0000

hi guys, i found a bug when compilling with some ARM microprocessors, at uint32_t convertion it get a SIGBUS error from kernel

the problem is at line near 2606 of file dataxfer.c

it's:
*((uint32_t *) (outopt+2)) = htonl(val);

and should be rewrited as:

uint32_t tmp_v;
tmp_v = htonl(val);
outopt[2] = tmp_v & 0xFF000000 >> 24;
outopt[3] = tmp_v & 0xFF0000 >> 16;
outopt[4] = tmp_v & 0xFF00 >> 8;
outopt[5] = tmp_v & 0xFF;

check this link to understand the problem:
http://lists.freedesktop.org/archives/avahi/2005-September/000281.html

LCK.. file is not FHS compliant

Marc Haber — Mon, 02 Jun 2008 11:51:04 -0000

Hi,

ser2net writes its process name to the LCK file locking the serial port, while the FHS specifies that lock files need to be in the HDB UUCP lock file format. The HDB format is to store the process identifier (PID) as a ten byte ASCII decimal number, with a trailing newline. For example, if process 1230 holds a lock file, it would contain the eleven characters: space, space, space, space, space, space, one, two, three, zero, and newline.

To me, the process name does not look to be compliant.

Or, are there multiple standards competing with each other?

Greetings
Marc