hi... roger i am CCing you because a) i think the problems are via-rhines fault in 2.6.16 b) this software reproducibly triggers something have not seen before i need people who test the new fragmentation code in netjack-0.12. i think i am seeing kernel problems, because there are patches going on for the via-rhine driver. it looks like the huge packet load triggers some sort of bug in the via-rhine network driver. Because after using netjack for some seconds, the link is broken. (this is reproducible with 2.6.16) "ifconfig eth0 down; ifconfig eth0 up" fixes it again. i dont think that the results i am getting are reliable. So i am asking you to test it. please check out http://netjack.sf.net/netjack-0.12test1.tar.bz2 compile as usual with: bash# scons jack_source_dir=<jack_source> then on the master with normal jackd running: bash# jacknet_client -p <other_machine> -P 24 -C 24 on the <other_machine> bash# jackd -R -d net -C 24 -P 24 please note that -r and -p on jackd -d net are no more needed and autodetected. i would also be interested in OSX results of the old netjack-0.11 for OSX compilation add "with-alsa=0" -- torben Hohn http://galan.sourceforge.net -- The graphical Audio language
