warnings from backtrace

[Gary E. Miller]

Yo Hal!

On Thu, 01 Jun 2017 21:45:25 -0700
Hal Murray <hmurray at megapathdsl.net> wrote:

> > As explained in the email: force a seccomp failure, see if the
> > output looks better to you.  The output should be improved a bit,
> > if you have seccomp but no execinfo.h  Improved with a backtrace if
> > you have execinfo.h.   
> 
> I changed the
>   #ifdef ENABLE_DNS_LOOKUP
> to
>   #ifdef xENABLE_DNS_LOOKUP
> 
>  1 Jun 21:39:02 ntpd[12595]: sandbox: seccomp enabled.
>  1 Jun 21:39:03 ntpd[12595]: dns_probe: trap, cast_flags:8, flags:1
> Bad system call

That means that catchTrap() was itself trapped.

Please run strace on it and send result.  Yet another syscall that
needs to be whitelisted.  Add that syscall to the whitelist and you'll
prolly work fine.

> We should come up with some official way to trigger a backtrace.

Already have it, the assert() macros.  That is what I am trying to
fix.  See my email earlier tonight on that subject.

> Should we add a new command line switch?  Or hack in something like
> generate a backtrace on --version or --help if debug is non-zero.

See my email earlier tonight on the subject.

At a minimum --enable-debug-gdb is needed right now.  That is
necessary to get the symbol table.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
	gem at rellim.com  Tel:+1 541 382 8588

	    Veritas liberabit vos. -- Quid est veritas?
    "If you can’t measure it, you can’t improve it." - Lord Kelvin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 488 bytes
Desc: OpenPGP digital signature
URL: <https://lists.ntpsec.org/pipermail/devel/attachments/20170601/c5f26537/attachment.bin>