[systemd-devel] [PATCH v5] Add sabridge for socket activation of traditional daemons

[David Strauss]

On Tue, Oct 15, 2013 at 5:33 PM, Zbigniew Jędrzejewski-Szmek
<zbyszek at in.waw.pl> wrote:
> So, is the bump to 16kB based on benchmarking? If you could post any
> results that would be great. Also, how does the original threaded aproach
> compare to the sd_event based one?

The 16kB bump is based on using the same buffer size as libcurl. Any
network library that established can't be using a choice that's *too*
wrong. The 4kB size was only based on someone else's suggestion to go
larger than the default MTU. My original 1kB size was based on no
real-world data at all.