segfault in vtimer update shortterm

[mehlis]

should this be possible?

Program received signal SIGSEGV, Segmentation fault.

0x0804f385 in update_shortterm () at vtimer.c:65
65      hwtimer_next_absolute = shortterm_queue_root.next->priority;
Missing separate debuginfos, use: debuginfo-install glibc-2.17-11.fc19.i686
(gdb) bt
#0  0x0804f385 in update_shortterm () at vtimer.c:65
#1  0x0804f4c0 in vtimer_callback (ptr=0x0) at vtimer.c:142
#2  0x080610c9 in multiplexer (source=3) at hwtimer.c:48
#3  0x080628d4 in hwtimer_isr_timer () at hwtimer_cpu.c:145
#4  0x0804993b in native_irq_handler () at irq_cpu.c:294
#5  0x490fdcbb in makecontext () from /lib/libc.so.6
#6  0x00000000 in ?? ()

[LudwigKnuepfer]

This is a not reliably reproducible error... I'm hoping to understand it someday. Not sure if vtimer or native hwtimer is to blame.

[LudwigKnuepfer]

Might be related to system time (clock).

[LudwigKnuepfer]

It appears that waiting a bit (up to about two minutes) makes the immediate problem go away.. (hence the system clock guess).

PS:
Most probable explanation:
The bug gets bored easily when not being the center of attention and it will go haunt someone else ;-)

[mehlis]

a starting point for debugging can be:

  static int update_shortterm(void)
  {
+    if (!shortterm_queue_root.next) {
+        DEBUG("!shortterm_queue_root.next");
+        return 0;
+    }
+

[LudwigKnuepfer]

#374

[mehlis]

fixed in #374