[En-Nut-Discussion] Bug in timer.c?

Harald Kipp harald.kipp at egnite.de
Tue Jul 12 15:41:24 CEST 2005


Ole,

Many thanks for providing this info. No idea yet, but
it may help.

Fact is, that two people reported long term instabilities
with a 3.9.8 preview, mainly based on the latest CVS.
I had not been able to track this down and hoped for
additional reports like yours. :-) Thus, 3.9.8 had
been delayed, although 3.9.7 contains the nasty UDP/ICMP
bug.

Harald

At 15:09 12.07.2005 +0200, Ole Reinhardt wrote:
>Hi all,
>
>while hacking on my rfcomm code I struggled over a bug in the timer
>handling.
>
>rfcomm uses a oneshot timer.
>
>I use the unix emulation, but I suppose this also could happen with the
>avr code?
>
>That the gdb output:
>
>[rfcomm-cmd]$inquiry sync
>Starting sync inquiry for 5 sec:
>Devices: 1
>Device         bt_addr cod        RSSI
>[0]: 00:02:72:40:42:b0 0x003e0100 0
>[rfcomm-cmd]$rfsession 0 1
>RFCOMM connect to 00:02:72:40:42:b0 Channel 1
>L2CAP:Created connection for handle: 1 at index 0
>RFCOMM: L2CAP connect . Unique handle: 0021
>
>Program received signal SIGSEGV, Segmentation fault.
>[Switching to Thread -1260528720 (LWP 18658)]
>0x0804a77b in NutTimerProcessElapsed () at timer.c:324
>324                     nexttn->tn_next->tn_ticks_left +=
>nexttn->tn_ticks_left;(gdb) bt
>#0  0x0804a77b in NutTimerProcessElapsed () at timer.c:324
>#1  0x0804bab5 in NutThreadResume () at thread.c:318
>#2  0x0804bcbc in NutThreadYield () at thread.c:401
>#3  0x0804b325 in NutEventPost (qhp=0x8068404) at event.c:437
>#4  0x0804f80d in _bt_l2cap_free_packet (pkt=0x80686cc) at
>bt_l2cap.c:368
>#5  0x08052051 in _bt_l2cap_handle_pkt (con_handle=1, pkt=0x80686cc,
>len=16)
>     at bt_l2cap.c:1604
>#6  0x0805246a in BT_L2CAP (arg=0x0) at bt_l2cap.c:1709
>#7  0x0804b484 in NutThreadEntry (arg=0x80694ec) at thread.c:102
>#8  0xb7f222bb in start_thread () from /lib/tls/libpthread.so.0
>#9  0xb7e9700e in clone () from /lib/tls/libc.so.6
>(gdb)
>
>Anybody knows more about this bug?
>
>I use the latest code from CVS (btnut and NutOS as well)
>
>Regards,
>
>Ole Reinhardt




More information about the En-Nut-Discussion mailing list