Hi guys, Just got a good link from a fellow Linux user in my local LUG - whose used traffic shaping for many years, who used it, and says it gave him the possibility to use his ADSL at full speed (both ways) with no latency - which usually occurs otherwise (which is why wondershaper asks you to set max. perf a lot below your actual bandwidth AFAIK). # new features in HTB which enables mpu and overhead - which is very nice for #adsl which is an ATM technology thing, where if you set your rules like this: # #class htb 1:1023 parent 1:102 leaf 1023: prio 3 quantum 1500 rate 95Kbit ceil #475Kbit burst 7b/8 mpu 48b overhead 5b cburst 2Kb/8 mpu 48b overhead 5b level 0 # # it fits in the ATM frames that DSL uses, and does not incur the latency that # otherwise happens when the link is filled. # i do not know if this is actually in 2.6.7 - he has backported HTB to 2.4.20 #because he uses wrtg54 and has to run 2.4.20 because of proprietary modules :( # http://mailman.ds9a.nl/pipermail/lartc/2003q4/010449.html patch at: http://www.mail-archive.com/lartc@mailman.ds9a.nl/msg08542.html Reproducible: Always Steps to Reproduce: 1. 2. 3. Actual Results: # diff -uNr iproute2-2.6.7.20040608.ebuild iproute2-2.6.7.20040608-r1.ebuild --- iproute2-2.6.7.20040608.ebuild 2004-06-30 05:12:54.000000000 +0200 +++ iproute2-2.6.7.20040608-r1.ebuild 2004-07-14 00:38:26.544214928 +0200 @@ -28,6 +28,9 @@ unpack ${A} cd ${S} + # Enable MPU patch from here: http://luxik.cdi.cz/~devik/qos/htb/v3/htb_tc_overhead.diff + # described here: http://mailman.ds9a.nl/pipermail/lartc/2003q4/010449.html + epatch ${FILESDIR}/htb_tc_overhead.diff # Enable HFSC scheduler #45274 if [ ! -z "`grep tc_service_curve ${ROOT}/usr/include/linux/pkt_sched.h`" ] ; then epatch ${FILESDIR}/2.4.7.20020116-hfsc.patch is all I've added to the ebuild - and then the file linked to in description to the files-dir.
Just changed it to enhancement :)
ive e-mailed the upstream author of iproute2 since it still applies pretty cleanly to the latest 2.6.8_pre snapshot
Thanks Spanky - could you let me know what the upstream author says - so I know when to expect the patch included :)
yep, i'll post any response i get here
just got an e-mail back from Stephen Hemminger; he's applied the patch and it'll be in the next release ;)