[OLSR-users] Revised LQ code

[aaron]

On Sat, 22 Jan 2005, Thomas Lopatic wrote:

ok, i just fetched the cvs version and compiled it with -pg -g for i386

Here is what I got immediately after startup (gdb olsrd ; set args -d 1; 
run):


--- 16:03:19.59 ------------------------------------------------- TOPOLOGY

Source IP addr   Dest IP addr     LQ     ILQ    ETX

Program received signal SIGSEGV, Segmentation fault.
0x401a95d0 in plugin_io () from /usr/lib/olsrd_dot_draw.so.0.3
(gdb) bt
#0  0x401a95d0 in plugin_io () from /usr/lib/olsrd_dot_draw.so.0.3
#1  0x080545bc in __gmon_start__ ()
#2  0x0805650a in __gmon_start__ ()
#3  0x0804ff66 in __gmon_start__ ()
#4  0x08050166 in __gmon_start__ ()
#5  0x08055a5c in __gmon_start__ ()
#6  0x0805570d in __gmon_start__ ()
#7  0x08059d8c in __gmon_start__ ()
#8  0x08059592 in __gmon_start__ ()
#9  0x0805106d in __gmon_start__ ()
#10 0x400597f8 in __libc_start_main () from /lib/tls/libc.so.6
#11 0x4017bfcc in ?? () from /lib/tls/libc.so.6

was dot_draw pthread enabled ?

Same segfault if compiled without -pg



> Hi there,
>
> The CVS repository now contains revised link quality code. Routing table 
> creation should now use substantially less CPU cycles, as lookups are now 
> done via balanced binary trees instead of linear lists.
>
> Moreover, the ETX metric is now also used for HNA entries.
>
> I have up to now only been able to test the new code on my test network, 
> which consists of merely three nodes. So, I'd be happy if anyone could test 
> this on a single node (just in case of any unexpected problems) in a large 
> network.
>
> I a few hours, the changes should have been replicated to the public 
> SourceForge CVS repository. If you want to test, make sure that your checked 
> out copy of the source code contains version 1.22 of lq_route.c.
>
> -Thomas
>
>
> _______________________________________________
> olsr-users mailing list
> (spam-protected)
> https://www.olsr.org/mailman/listinfo/olsr-users
>