Support SO_MAX_PACING_RATE.
authorSteinar H. Gunderson <sgunderson@bigfoot.com>
Sat, 28 Sep 2013 22:37:43 +0000 (00:37 +0200)
committerSteinar H. Gunderson <sgunderson@bigfoot.com>
Sat, 28 Sep 2013 22:37:43 +0000 (00:37 +0200)
commitce0d42a14da5686366d0c73fa9b152e90619ff91
tree9f5f0b8bd842d6f29d86d905f0a6785e271064e3
parent4ba4a9f0eac9777ef2d5071e980ae0d1ba22d048
Support SO_MAX_PACING_RATE.

SO_MAX_PACING_RATE is the newfangled socket option from Eric Dumazet,
used with the new fq packet scheduler in Linux. It allows you to set
a max rate for the socket (presumably a stricter upper bound than the
RTT-based estimate from the kernel), delivering pacing without having
to resort to the relatively complex mark setup. It seems to enter
the Linux kernel in 3.13 at the earliest; not unlikely even later.

In time, fwmark will be deprecated, but the implementation of TCP pacing in
Linux is still a bit shaky (especially with not-always-filling applications
like streaming), so fwmark will stay the primary solution for now.
15 files changed:
README
client.cpp
config.cpp
config.h
cubemap.1
cubemap.config.sample
main.cpp
server.cpp
server.h
serverpool.cpp
serverpool.h
stream.cpp
stream.h
udpstream.cpp
udpstream.h