[Olsr-dev] Floating Point Exceptions with tip of http://gredler.at/hg/olsrd

[Hannes Gredler]

bernd,

pardon - regression introduced by myself -
pls pull the following changeset.

http://gredler.at/hg/olsrd/rev/4efdfbdec587

/hannes

Bernd Petrovitsch wrote:
> Hi!
> 
> I get deterministically a SIGFPE with the tip of
> http://gredler.at/hg/olsrd.
> 
> gdb rells me:
> ----  snip  ----
> -- ALL PLUGINS LOADED! --
> 
> Warning: probably sizeof(pointer) in strscpy(0x8eb0750, gaz, 4)!
> Warning: probably sizeof(pointer) in strscpy(0x8ef08d0, geb, 4)!
> Warning: probably sizeof(pointer) in strscpy(0x8ef5140, geb, 4)!
> 
> Program received signal SIGFPE, Arithmetic exception.
> [Switching to Thread -1208047936 (LWP 18692)]
> 0x0805e90b in olsr_start_timer (rel_time=0, jitter_pct=5 '\005', periodical=OLSR_FALSE, 
>     timer_cb_function=0x7fffffff, context=0x8eacd88, cookie=14) at src/scheduler.c:223
> 223     /**
> (gdb) info locals
> timer = (struct timer_entry *) 0x8ef4758
> (gdb) bt
> #0  0x0805e90b in olsr_start_timer (rel_time=0, jitter_pct=5 '\005', periodical=OLSR_FALSE, 
>     timer_cb_function=0x7fffffff, context=0x8eacd88, cookie=14) at src/scheduler.c:223
> #1  0x0805f0d0 in olsr_set_timer (timer_ptr=0x8eace18, rel_time=100, jitter_pct=255 '�', 
>     periodical=OLSR_FALSE, timer_cb_function=0x80578b4 <olsr_expire_mid_entries>, context=0x8eacd88, 
>     cookie=14) at src/scheduler.c:707
> #2  0x08057989 in olsr_expire_mid_entries (context=0x8eacd88) at src/mid_set.c:116
> #3  0x0805eda2 in olsr_walk_timers (last_run=0x807a460) at src/scheduler.c:406
> #4  0x0805ef4f in olsr_scheduler () at src/scheduler.c:146
> #5  0x080575a0 in main (argc=<value optimized out>, argv=<value optimized out>) at src/main.c:422
> (gdb) print *timer
> $1 = {timer_list = {next = 0x0, prev = 0x0}, timer_clock = 796816560, timer_period = 0, timer_cookie = 10, 
>   timer_jitter_pct = 5 '\005', timer_flags = 0 '\0', timer_random = 1901243200, 
>   timer_cb = 0x805ff3e <olsr_expire_tc_edge_gc>, timer_cb_context = 0x8eafb68}
> ----  snip  ----
> I'm probably just blind that I don't see any possible problem (and have
> no time ATM to investigate more).
> 
> 	Bernd