It looks like we'll need:
- one share timers queue for the rare tasks that need to wake up
anywhere
- one private timers queue per thread
- one global queue per group
- one local queue per thread
And may be we can simply get rid of any global/shared run queue as
we don't seem to have any task bound to a subset of threads.
This macro was used both for binding and for lookups. When binding tasks
or FDs, using all_threads_mask instead is better as it will later be per
group. For lookups, ~0UL always does the job. Thus in practice the macro
was already almost not used anymore since the rest of the code could run
fine with a constant of all ones there.
Instead of having a global mask of all the profiled threads, let's have
one flag per thread in each thread's flags. They are never accessed more
than one at a time an are better located inside the threads' contexts for
both performance and scalability.
The pools currently have plenty of options (and some usefull ones were
even lost with the modern design), but most of them could be categorized
along a few use cases, namely, performance, reliability, debuggability.
This document explores various ways to try to combine them and their
effect in a less complex way for the long term.
A number of outdated docs dating 2012 about buffers implementation
and management were totally irrelevant to the current code (and even
to most 1.8 code as well). These docs have all been removed so that
only the up to date documentation remains.
This was the first transparent proxy technology supported by haproxy
circa 2005 but it was obsoleted in 2007 by Tproxy 4.0 which removed a
lot of the earlier versions' shortcomings and was finally merged into
the kernel. Since nobody has been using cttproxy for many years now
and nobody has even just tried to compile the files, it's time to
remove it. The doc was updated as well.
These ones were design notes and ideas collected during the 1.5
development phase lying on my development machine. There might still
be some value in keeping them for future reference since they mention
certain corner cases.
Also add some thoughts about the existing and new design.
Note: an earlier design used the names "head" and "tail" for both sides
of the buffer, but it appears awkward as these words may be understood
in two forms (feed by head, output by tail, or make the newcomers wait
at the tail of the queue). Also there were already a few functions in the
code making use of either terminology. So better avoid this terminology
and use "input" and "output" instead.
Ilya Shipitsin
Willy Tarreau
Joseph Herlant