22 Bře
2024
22 Bře
'24
20:35
Michael Grimm via knot-resolver-users <knot-resolver-users(a)lists.nic.cz> wrote:
I did give it a try, knot-resolver-v6.0.6 [1], created a local port, stole the Makefile
from knot-resolver (5.7.1), and failed further down meson.build ...
FYI: In the meantime I succeeded in compiling knot-resolver-v6.0.7 as of today's
commit:
MWN> git log -1mer-waases|root> git log -1
commit 561b40762ea45a6332a0b108b19f1649b5a5cfd7 (HEAD -> 6.0, origin/HEAD, origin/6.0)
A local knot-resolver-current port is up compiling knot-resolver-v6.0.7 and handling all
possible dependencies.
kresd is up and running fine.
BUT:
1) I needed to patch
'manager/knot_resolver_manager/kresd_controller/supervisord/plugin/notifymodule.c'
to complete successful compilation.
2) The reason is FreeBSD's different structures and credentials regarding sockets.
3) I don't know if I did this in a proper way.
4) I cannot test that socket patched in 1), yet, because I cannot get kresctl run (python
modules loading or ...?)
5) But I can use knotc-5.7.1 to retrieve statistics from a socket defined by:
'net.listen('/var/run/kresd/control/kresd.sock', nil, { kind =
'control' })',
but I don't know if that patch in 1) addresses the very same socket and if my test
is relevant at all?
That patch mentioned in 1) addresses:
#) include sys/ucred.h
#) use LOCAL_PEERCRED instead of SO_PASSCRED
#) use 'struct xucred' instead of 'struct ucred'
#) use SCM_CREDS instead of SCM_CREDENTIALS
#) use cred.cr_pid instead of cred.pid as returned value
If you are interested in getting hold on a 'git diff' patch, please let me know
how (mail, git).
Regards,
Michael
Michael Grimm via knot-resolver-users <knot-resolver-users(a)lists.nic.cz> wrote:
> Vladimír Čunát via knot-resolver-users <knot-resolver-users(a)lists.nic.cz>
wrote:
Some of these aspects get improved in versions >=
6.0. Control of the GC and kresd processes is integrated inside, also stats aggregation
and there's new kresctl command (and user config gets rewritten completely, from Lua
to YAML). But honestly we don't do any BSD ourselves, so I don't expect you want
to package and try it at this point…
Perhaps I will ;-)