> On 31 May 2025, at 19:50, Laurent Bercot <ska-skaware_at_skarnet.org> wrote:
>
> That sounds slower than it should be. My router is a 1.6 GHz Atom E620,
> with 2 GiB of RAM as well, and it establishes connections in under half
> a second. Is there any delay you can identify, or is it really all hot
> CPU computations?
Never really thought of that. I'm not sure how to profile that part specifically.
Maybe I can write a more explicit command line (I typically just just s6-tlsserver)
that includes perf?
Those numbers come from curl and Safari mostly, so if it's actually too slow to
be true maybe it's because something on the client side is making the wrong
estimates?
That said, the link is also slow, the NIC stopped working so I'm using a USB
wireless adapter, so I assumed the rountrips were the main issue.
I can try to profile that part while testing the TaskWarrior stuff, I'm sending
requests anyway.
> It's in the latest Alpine stable release, 3.22. It's also been in edge
> for a while (I generally push new versions of my software to Alpine edge
> shortly after cutting them.)
Nice. Time to upgrade :)
> Neat. However, be aware that s6-tls[cd]io is often used in chains of
> commands: the ciphertext end could very well be two pipes, rather than a
> socket; and kTLS only supports TCP sockets. For tipidee, it will work
> because s6-tlsd-io's ciphertext end is indeed the socket, but there are
> settings where it won't be usable (e.g. smtpd-starttls-proxy).
Yes. I misspoke, what I meant by a version of s6-tlsio was a very limited
replacement for my uses, which are mostly tipidee centered.
But if you think it's worth it, I might try to patch it and send the patch the real one.
I assumed you kept it out for portability.
> If establishing a session uses non-negligible time compared to using
> the session, then indeed you want to reuse the same connection as much
> as possible, and non-NPH would be the way. Now that tipideed streams CGI
> results, you should be able to use it, and as a bonus it does zero-copy.
Yes, that sounds like the sensible thing to do. That said, I hardcoded a bunch of
stuff, so I'll change it after I'm sure the actual program is working.
Received on Sun Jun 01 2025 - 01:16:58 CEST