Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 16 May 2002 01:53:51 +0200
From:      Bernd Walter <ticso@cicely5.cicely.de>
To:        Andrew Gallatin <gallatin@cs.duke.edu>
Cc:        Rich Bud <rbud@fortean.com>, freebsd-alpha@FreeBSD.org
Subject:   Re: natd sig 10 under stable
Message-ID:  <20020515235351.GO4401@cicely5.cicely.de>
In-Reply-To: <15586.55082.170680.378288@grasshopper.cs.duke.edu>
References:  <20020515160407.T28044-100000@nemesis.fortean.com> <15586.55082.170680.378288@grasshopper.cs.duke.edu>

next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, May 15, 2002 at 05:46:18PM -0400, Andrew Gallatin wrote:
> 
> Rich Bud writes:
>  > BTW, Craig Burgess posted about a sig 10 in natd back in february under
>  > different conditions. I didn't see any replies, so i'm guessing this is
>  > the same problem...
> 
> Probably, but this is much more useful information that Craig was ever
> able to supply.  With your information, it may be possible to diagnose
> & fix the problem.
> 
> 
>  > Program received signal SIGBUS, Bus error.
>  > 0x120008614 in AliasHandleQuestion (count=1, q=0x11feb911, pmax=0x11feb926 "", nbtarg=0x11feb838) at /usr/src/lib/libalias/alias_nbt.c:290
>  > 290                     switch ( ntohs(q->type) ) {
>  > (gdb) ptype q
>  > type = struct {
>  >     u_short type;
>  >     u_short class;
>  > } *
>  > q = (NBTNsQuestion *) 0x11feb911, q->class = 1, q->type = 0
>  > 
> 
> This address is unaligned.  Is one of machdep.unaligned_sigbus
> == 1, machdep.unaligned_fix == 0 true?  (or the uac equivalent)

Can you please print pip, *pip, uh, p and nsh in AliasHandleUdpNbtNS()?
I will need to know which part of the packets can be missaligned.
It's much fast if we can realign the whole data part instead of just
each segment.

> If not, perhaps the unaligned fixup handler needs to be enhanced to
> handle a new sort of instruction.  Can you disassemble the code at the
> faulting PC & see what instruction generated the access, please?
> (or enable unaligned printing via uac -p in the shell which starts natd).

Sounds like the most likely reason - especially considering BWX.
I will check out how much work it is to expand our trap.c

-- 
B.Walter              COSMO-Project         http://www.cosmo-project.de
ticso@cicely.de         Usergroup           info@cosmo-project.de


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-alpha" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020515235351.GO4401>