[Dnsmasq-discuss] Segfault in DNSSEC code

[Simon Kelley]

On 15/04/14 22:39, Manish Singh wrote:
> I've run across a segfault in the DNSSEC code when resolving a domain, when
> DNSSEC builtin but turned off:
> 
> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread 0x7f3d178fe700 (LWP 10762)]
> 0x0000000000407e26 in extract_name (header=0x1001272, plen=46,
>     pp=0x7fffdc948590, name=0x0, isExtract=1, extrabytes=4) at rfc1035.c:27
> 27        *cp = 0;
> (gdb) bt
> #0  0x0000000000407e26 in extract_name (header=0x1001272, plen=46,
>     pp=0x7fffdc948590, name=0x0, isExtract=1, extrabytes=4) at rfc1035.c:27
> #1  0x0000000000455419 in hash_questions (header=0x1001272, plen=46,
> name=0x0)
>     at dnssec.c:2284
> #2  0x0000000000421160 in tcp_request (confd=11, now=1397591659,
>     local_addr=0x7fffdc9487b0, netmask=..., auth_dns=0) at forward.c:1745
> #3  0x00000000004295e7 in check_dns_listeners (set=0x7fffdc948920,
>     now=1397591659) at dnsmasq.c:1591
> #4  0x0000000000427c88 in main (argc=10, argv=0x7fffdc948c38) at
> dnsmasq.c:955
> 
> daemon->keyname is eventually passed into extract_name, but it is NULL
> since the code that initializes is guarded by an
> option_bool(OPT_DNSSEC_VALID) check.
> 
> I don't really know enough about DNSSEC to ascertain why this code path got
> triggered when it shouldn't be.

It's a copy-n-paste error, I think. In any case
s/daemon->keyname/daemon->namebuff/
fixes it. I've just committed the fix to git.

Thanks for that, it's a nasty bug, probably worth of an early release to
fix it.


Cheers,


Simon.


> 
> -Manish
> 
> 
> 
> _______________________________________________
> Dnsmasq-discuss mailing list
> Dnsmasq-discuss at lists.thekelleys.org.uk
> http://lists.thekelleys.org.uk/mailman/listinfo/dnsmasq-discuss
>