priority of MANGLE and NAT tables
Hello,
Basically i have the following problem: shaping together with SNAT works only in one direction (incoming) .. but i am writing here since i believe it is implementation problem .. not configuration one (but i might have overlooked something).
If i shape using u32 .. the packet gets classified with real addresses, which is too late for outgoing traffic (since the source address is allready rewritten).
So the only chance is using fwmark and classify for that. Fine .. but the problem remains, since there is no hook for mangle table, where the the src addr/port (and dst parts in reply packets) would correspond
to the internal (translated addr/ports) = in PREROUTING .. mangle is before nat (so before rewriting dst addr/port), and there is no mangle hook in POSTROUTING (which would help, since it would be before SNAT).
So my questions are (since any of those would save me):
- what is the reason there is no hook for mangle table in POSTROUTING ?
- is there any reason why NAT is after MANGLE (i.e. has higher numerical priority) ?
I can live with my local change reverting order for MANGLE/SNAT or adding MANGLE hook to the POSTROUTING .. but i would still be interested in your opinion and why it was done this way.
Thanks for any info. Regards,
peter
PS: please CC me, since i am not subscribed
Signature · 2 lines
Peter Kundrat [email protected]
Quoted 2 lines · Peter Kundrat
before rewriting dst addr/port), and there is no mangle hook in POSTROUTING (which would help, since it would be before SNAT).
yes, there is. You must be using a relatively old kernel verison. Think this changed around 2.4.14
Quoted 1 line
peter
Signature · 5 lines
Live long and prosper - Harald Welte / [email protected] http://www.gnumonks.org/ ============================================================================ GCS/E/IT d- s-: a-- C+++ UL++++$ P+++ L++++$ E--- W- N++ o? K- w--- O- M- V-- PS+ PE-- Y+ PGP++ t++ 5-- !X !R tv-- b+++ DI? !D G+ e* h+ r% y+(*)
Quoted 6 lines · Harald Welte
Quoted 2 lines · Peter Kundrat
before rewriting dst addr/port), and there is no mangle hook in POSTROUTING (which would help, since it would be before SNAT).
yes, there is. You must be using a relatively old kernel verison. Think this changed around 2.4.14
In patch-o-matic we also have the interesting match "conntrack" that solves many relevant issues..