• Get rid of PORT_BASE. Replace by AliasRange.

• Factor out the search for a new port. Improves the perfomance a bit.

In D30516#685362, @donner wrote:

Despite the splay is much faster, it can stall during SPLAY_REMOVE for currently unknown reasons.
100% CPU in an infinite loop.

rebase to a buf fixed parent review

• Fix "freeing wrong structure" bug
• Factor out pptpLists

Despite the splay is much faster, it can stall during SPLAY_REMOVE for currently unknown reasons.
100% CPU in an infinite loop.

• Let PPTP use their own data structure.

• Touch links for PPTP on use, too.

• Remove LSNAT from outgoing search.

@lev may you consider testing this patch in your environment? It should speed up you use case.

rebased

• Reorder incoming links by grouping of common search terms.

Rework the whole patch to get rid of failures of the test suite and to
only improve the performance.

• Rename NO_ADDR to ANY_ADDR
• Factor out the common Out filter
• Factor out the common In filter
• Search fully specified links first.
• Separate the partially specified links into a separate data structure.
• Use a (smaller) hash table to speed up the partially link access.
• Clean program flow in _FindLinkIn
• Factor out the outgoing search function.
• Fix small error in termination.
• Factor out Cleanup and Use for links.
• Cleanup the partial find procedure
• Rework the incoming search over full know links.
• Just another early exit of the search loop

In D30283#683929, @lev wrote:

I'm writing this comment to this change almost arbitrary, as it is meta-comments on your (great!) work.

Tests show several broken issues:

2_natout:2_unregistered  ->  failed: 2 checks failed
*** Check failed: 2_natout.c:88: addr_eq(cgn, pip->ip_src) not met
*** Check failed: 2_natout.c:89: addr_eq(pub, pip->ip_src) not met
2_natout:3_cgn  ->  failed: 1 checks failed
*** Check failed: 2_natout.c:122: addr_eq(pub, pip->ip_src) not met
3_natin:4_redirectaddr  ->  failed: 3 checks failed
*** Check failed: 3_natin.c:277: addr_eq(pub, p->ip_src) not met
*** Check failed: 3_natin.c:278: addr_eq(pub, p->ip_src) not met
*** Check failed: 3_natin.c:293: addr_eq(pub, p->ip_src) not met

• Replace per instance global variable timeStamp by a real global one.

• Housekeeping is cheap, so check expire before each use.

The major part of processing time in libalias is gettimeofday(3).

Performance improvement is about 1% if running on full speed.

Does this match you needs, @lev ?

• Add copyright statement

Rebase

Rebase

• rebase to D30412

• Add test for LibAliasRedirectDynamic

• Add test for LSNAT

• Add tests for LibAliasRedirectAddr

• Apply missing refactors

• Add test for change of alias address

Depend on D30412

• Spacing

• Clear state instead of choosing new flows
• Add final variant of PortRedirect
• test overlapping portforwardings
• Refactor UDP tests

• test overlapping portforwardings

• Add final variant of PortRedirect

• Clear state instead of choosing new flows

Factor out the common test infrastructure and start with the first test.

• Select a random flow out of a rolling window in order to simulate overlapping communication.

• Print intermediate timinge

• Attacks come from outside

• Make it more complicated.

Who is going to commit this?
May I? I'd combine both patches into a single MFC action.

Thank you @kp, having this in main allows to test other branches more easily.

In D30335#682354, @lev wrote:

In D30335#682251, @donner wrote:

Does this cover your common use cases, @lev ?

Big part of my use cases is port forwarding into DMZ, both TCP and UDP, with multiple (2000+) streams to same forward.

I did run the tests from D30335 against this patch:

Thank you.

@kp may you have a look again?
Does this cover your common use cases, @lev ?

• Add stress test for exhausting of search data structure