The accounting was just broken and overly messy anyway.

The new way of transitioning to layer callbacks - done because of
portability (mainly to aarch64) - is a bit expensive.  This is a simple
way of recovering that cost.  Merge 603a24fc regressed speed a bit.

The watchdog module now can be loaded without systemd, has customisable
callbacks, and can do real DNS queries and check their results.

I'm really sorry; I didn't notice and it only hit parts that
*apparently* aren't tested normally.  Only 32-bit systems would be
affected, due to the structure only changing ABI on 32-bit systems.

Refusing to answer queries without RD bit makes it harder
to read what data is present in resolver's cache.

Somehow I didn't notice this field when adding ::add_selected.
We probably never put anything into answer's ADDITIONAL,
so noone's noticed a problem until now.

Vladimír Čunát authored 5 years ago and Petr Špaček committed 5 years ago

Don't stash a packet with mismatching QNAME+QTYPE.
When receiving an NXDOMAIN or NODATA packet in an insecure zone,
it would get cached with KR_RANK_INSECURE regardless of mismatch
in QNAME.  If the 0x20 pattern was preserved in the fake QNAME,
such packet would then be used to answer queries with matching QNAME,
even if there's no proof that this QNAME is insecure.

Also the log now uses the same format query UID format as elsewhere.

Petr Špaček authored 5 years ago and Vladimír Čunát committed 5 years ago

Regression in 19ef77d1 (unreleased): Default value IN for qclass
parameter was missing.  Also consistently use A as default for qtype.

Thanks to changes in this branch the functions are called with correctly
typed parameters already, so these weird casts can be deleted.

We're very much depending on LuaJIT, for years already.
The removed stuff seems very unlikely to be useful in any LuaJIT
in forseeable future.  The Lua language is effectively forked.

... without light userdata and yet it should be efficient.
Also with checkout layer, but I have no nice way of testing that one.

In particular, throw errors when used in weird ways,
instead of the usual "return nil" semantics.
That might be surprising to some lua users.

...when making calls from lua to C modules; use luaJIT FFI instead.
This eliminates some cases of lua_pushlightuserdata().

On many places we've been assuming that there's only a single worker,
but we still often didn't utilize the property well.  To get the pointer
we used various ways, all even untyped:
 - __worker global variable in lua
 - uv_default_loop()->data
 - kr_request::daemon_context

Now we instead simply define a global typed pointer the_worker.

Nitpick: also worker_{init,deinit}() are reordered to correspond
to the order of the fields, etc.

Now we hopefully won't need to touch it for a long time.

It gets handy for debugging, especially when looking at dufferent
callback "instances".

It would be pointless anyway as it cannot provide more information.

As far as I understand the debug.sethook('call') descrption we are not
going to have any locals except function parameters so we can simplify
the code.

Vladimír Čunát authored 5 years ago and Petr Špaček committed 5 years ago

"Unfortunately", for FFI-bound C functions there it doesn't hold that
missing parameters would be converted to nil/NULL.
Still, this function seems unlikely to have been used outside the repo.

Vladimír Čunát authored 5 years ago and Petr Špaček committed 5 years ago

It caused crashes somewhere inside LuaJIT or the JIT-ed code.

This will allow modules like view etc. work in the same way with packets
received over HTTP.

First version which actually works with Firefox DoH in default
configuration.

Limitations:
- does not support HTTP GET method
- headers for HTTP cache are not generated
- error handling is largely missing
- no tests
- ACLs will not work, modules do not see source IP address of the HTTP
endpoint