Hi Simon, I have a question about commit 6656790f in regard changes to tcp_request function.
commit 6656790f2498f2a0b21086bc4ab47a2e38429a7c
Author: Simon Kelley <si...@thekelleys.org.uk>
Date: Tue Jan 7 20:46:33 2025 +0000
Handle queries with EDNS client subnet fields better.
The call to add_edns0_config moved up, even before the "find_pseudoheader"
condition that sets have_pseudoheader. To my understanding add_edns0_config
prepares the original request to be forwarded (with regard to EDNS).
Since the function adds EDNS to the packet, the original query is altered even
before it has been processed and decided whether to handle locally or forward,
ie. it invalidates "have_pseudoheader".
The general logic in receive_query differs from tcp_request:
1. checks
2. find_pseudoheader
3. flags
4. answer : {disallowed,auth, saved_question + request (from cache)} => send
5. forward if not (local) answered
vs
1. checks
2. saved_question
3. find_pseudoheader
4. flags
5. answer : {disallowed,auth,request (from cache)} => send
6. forward if not (local) answered
Kind regards,
Tijs
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Dnsmasq-discuss mailing list Dnsmasq-discuss@lists.thekelleys.org.uk https://lists.thekelleys.org.uk/cgi-bin/mailman/listinfo/dnsmasq-discuss
