On 19/09/18 13:04, Dominik DL6ER wrote:
> Hey Simon,
>
> On 19.09.2018 13:27, Simon Kelley wrote:
>> when rd is not set, never answer
>> from the cache, but always forward the query. That would allow dig
>> +trace to work.
>>
>> Does hat seem sensible?
>
> Yes, that seems useful.
>
> Best,
>
Hello Simon,
> If you look, lots of things are different between the two logs. In the
second one,
> dhcpcd is doing routing table changes, for instance. That could explain
why dnsmasq
> gives up trying to confirm SLAAC addresses because it gets transient "no
route to host"
> returns. (see
Hey Simon,
On 19.09.2018 13:27, Simon Kelley wrote:
> when rd is not set, never answer
> from the cache, but always forward the query. That would allow dig
> +trace to work.
>
> Does hat seem sensible?
Yes, that seems useful.
Best,
Dominik
___
On Wed, Sep 19, 2018 at 1:44 PM Simon Kelley wrote:
> This all makes me slightly uneasy. I think the "out of memory"
> explanation for the crashes you are seeing is not a good one.
No, I agree. I have compiled an OpenWRT image without the fix and
installed it on my device, and I am trying to
On 19/09/18 08:59, Kristian Evensen wrote:
> Hi Simon,
>
> Thanks for a quick reply.
>
> On Wed, Sep 19, 2018 at 12:23 AM Simon Kelley wrote:
>> Thanks for the report. The obvious explanation is that whine_malloc() is
>> returning NULL, and the code should handle that. whine_malloc only
>>
On 19/09/18 11:09, Bernard CLABOTS wrote:
> Thanks a lot for this answer.
>
> Indeed, it is a special case as we have a simple two way Request/ACK,
> this is also what is seen with some implementations when quickly
> unplugging/re-plugging the cable, it is legal AFAIK.
>
> I also agree on the
The change in question causes dnsmasq to always return SERVFAIL for
queries without the "use recursion" bit set.
The relevant quote in the reference
http://cs.unc.edu/~fabian/course_papers/cache_snooping.pdf
is this:
Recommendation 2: secondly, and most importantly, non-authoritative
requests
Thanks a lot for this answer.
Indeed, it is a special case as we have a simple two way Request/ACK, this is
also what is seen with some implementations when quickly unplugging/re-plugging
the cable, it is legal AFAIK.
I also agree on the necessity to be efficient in case of loss of the lease dB.
Dear list members,
I expect "dig +trace" to show a trace of the delegation path from the
root name servers for the name being looked up. This behavior is broken
since commit 087eb76140725f8f1892ba6f251ea052d3440966
> On 19 Sep 2018, at 08:59, Kristian Evensen wrote:
>
> Hi Simon,
>
> Thanks for a quick reply.
>
> On Wed, Sep 19, 2018 at 12:23 AM Simon Kelley wrote:
>> Thanks for the report. The obvious explanation is that whine_malloc() is
>> returning NULL, and the code should handle that.
Hi Simon,
Thanks for a quick reply.
On Wed, Sep 19, 2018 at 12:23 AM Simon Kelley wrote:
> Thanks for the report. The obvious explanation is that whine_malloc() is
> returning NULL, and the code should handle that. whine_malloc only
> returns NULL if the system cannot allocate any more memory,
11 matches
Mail list logo
