Permalink
Commits on May 15, 2012
  1. https://www.bufferbloat.net/issues/124

    http://ecn.csail.mit.edu
    adding the patch to test path ECN - provided by Steve Bauer - MIT
    Compiled And basic-tested on cerowrt.
    K committed May 15, 2012
Commits on May 14, 2012
  1. Latest batch of codel related changes

    Still need real uci support
    Dave Taht committed May 14, 2012
Commits on May 10, 2012
Commits on May 1, 2012
  1. Altered polipo to pull from github

    CeroWrt Build committed May 1, 2012
  2. Use read only git access for gnugol

    CeroWrt Build committed May 1, 2012
Commits on Apr 29, 2012
  1. Netperf to head

    Dave Taht committed Apr 29, 2012
Commits on Apr 26, 2012
  1. Quagga: Another big endian fix

    Thx denis!
    Dave Taht committed Apr 26, 2012
  2. quagga: fix Big Endian bug - update quagga to head

    A big endian bug crept into recent builds of quagga. This fixes it,
    so far as I know.
    Dave Taht committed Apr 26, 2012
Commits on Apr 23, 2012
  1. More cleanups

    Dave Taht committed Apr 23, 2012
  2. Fix credits again

    Dave Taht committed Apr 23, 2012
  3. A couple errors crept into the credits file

    Dave Taht committed Apr 23, 2012
  4. update version of cero onboard doc

    Still have to review the doc, but the credits file was tedious...
    Dave Taht committed Apr 23, 2012
  5. Updated credits file

    Dave Taht committed Apr 23, 2012
  6. Added mildly updated credits file

    I need to go through the big lists, but I wanted to make sure
    I'd got the new people I'd remembered best
    Dave Taht committed Apr 23, 2012
  7. Retired out of tree version of iptables/iproute2

    Dave Taht committed Apr 23, 2012
Commits on Apr 20, 2012
  1. update netperf to head

    Dave Taht committed Apr 20, 2012
Commits on Apr 18, 2012
Commits on Apr 16, 2012
  1. update to quagga head

    ospf bug fixed
    Dave Taht committed Apr 16, 2012
Commits on Apr 14, 2012
  1. Make named + dnssec startup mildly more robust.

    If this doesn't work more of the time, I'm disabling dnssec entirely.
    
    It turned out that most of my problems are on first boot,
    where you are still rewriting flash and you can't save certain conf
    files
    Dave Taht committed Apr 14, 2012
  2. gpsd: update to 3.5 released version

    Still intend to track it via git, but this is 3.5.
    
    Can hopefully produce a set of patches to get support into openwrt
    mainline, but that needs scons, etc.
    Dave Taht committed Apr 14, 2012
  3. gpsd: deletes get queued up differently

    Not for the first time, I get bit by git.
    Dave Taht committed Apr 14, 2012
  4. Add UCLIBC workaround for gpsd 3.5

    Pushed patch up to eric, so I figure on my next pull,
    this will be in mainline.
    Dave Taht committed Apr 14, 2012
Commits on Apr 12, 2012
  1. Quagga-RE-0.99.17.9 release

    Dave Taht committed Apr 12, 2012
  2. Near-final gpsd 3.5 test

    Dave Taht committed Apr 12, 2012
Commits on Apr 11, 2012
  1. ccnx: fix bustage

    sdwalker committed Apr 11, 2012
Commits on Apr 10, 2012
  1. Update gpsd to head

    Dave Taht committed Apr 10, 2012
  2. update quagga to head - ospf fixes

    Dave Taht committed Apr 10, 2012
Commits on Apr 9, 2012
  1. Update quagga to head

    Dave Taht committed Apr 9, 2012
  2. Improved bloat.sh script

    Dave Taht committed Apr 9, 2012
Commits on Apr 5, 2012
  1. Update gpsd to git head in prep for 3.5 release

    Dave Taht committed Apr 5, 2012
Commits on Apr 4, 2012
  1. Makefile

    Dave Taht committed Apr 4, 2012
  2. Relentless TCP

    Matt Mathis, 2008
    
    Relentless TCP is designed to maintain a standing queue at some network
    bottleneck.  It does so without regards to other traffic, and thus it is not
    generally fair to other network users.  It MUST NOT be used unless the network
    is designed to actively control it to protect other flows and the Internet at
    large from its aggressiveness.
    
    To protect other flows, the network carring Relentless TCP traffic has to
    implement Scavenger Service, Fair Queuing, Approximate Fair Dropping, or some
    other capacity allocation algorithm.
    
    To protect the Internet at large the network carring Relentless TCP traffic
    has to be a physically or logically isolated from the rest of the Internet.
    Be espically warry of the potential for dynamic routing to choose an
    alternative path that can not adiquiately contol Relentless TCP.
    
    To minimize the risks associated with Relentless TCP, it has to be enabled on a
    per-connection basis.
    Dave Taht committed Apr 4, 2012
Commits on Apr 3, 2012
  1. named: caching the roots has too many dependencies

    While the performance benefit of caching the roots seems good...
    
    on boot, this leads to cascade failures while dnssec is not working.
    bind ends up throttling the connections to the roots, or vice versa
    and it can take minutes to actually get anywhere, during which time
    ntp is kicking the nameserver and vice versa, and sometimes
    we never get anywhere.
    Dave Taht committed Apr 3, 2012