Date: Tue, 28 Jun 2022 23:38:15 +0000 From: bugzilla-noreply@freebsd.org To: jail@FreeBSD.org Subject: [Bug 255830] dummynet(4) queues/pipes do not work inside of a VNET jail Message-ID: <bug-255830-29815-IhWlnCT4bJ@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-255830-29815@https.bugs.freebsd.org/bugzilla/> References: <bug-255830-29815@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D255830 Peter Much <pmc@citylink.dinoex.sub.org> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |pmc@citylink.dinoex.sub.org --- Comment #9 from Peter Much <pmc@citylink.dinoex.sub.org> --- I am seeing these messages too, and no, Joshua, You're not the only one try= ing such things with dummynet. A few comments: * dummynet did never malfunction, it did simply reinsert the shaped packets into the host's ipfw instead of the jail's one - because there was only one dummynet on the system. The issue was more serios: every jail could manipul= ate the hosts pipes parameters, so it was necessary to run jails in securelevel= =3D3. * This appears to be fixed in 13.1. Kudos to the author! * What has sadly not been fixed alongside is the same issue with ng_ipfw, (Should be simpler, because netgraph itself is already VIMAGE-able) I once tried the bufferbloat thing, ran into the very same issue, considere= d it too difficult to fix on my own, and then moved the router into a bhyve inst= ead. There timing issues were more challenging, but the shaping apparently did w= ork. Currently it doesn't anymore. Thanks for Your configuration data, that might help me when looking into that issue occasionally - maybe the bhyve has to leave again. Now I tried to do the ng_ipfw fix myself, but with horror then noticed the beforementioned error-messages, and another one that looks worse: Apr 22 15:34:06 <kern.crit> edge kernel: Freed UMA keg (IPFW counters) was = not empty (1 items). Lost 20 pages of memory. Jun 15 20:41:45 <kern.crit> edge kernel: Freed UMA keg (IPFW counters) was = not empty (2 items). Lost 40 pages of memory. Jun 26 17:06:23 <kern.crit> edge kernel: [55226] Freed UMA keg (IPFW counte= rs) was not empty (3 items). Lost 60 pages of memory. Jun 28 22:42:20 <kern.crit> edge kernel: [248183] Freed UMA keg (IPFW count= ers) was not empty (3 items). Lost 40 pages of memory. But this one also obviousely exists since upgrade to 13.1 (RC2), and seems = to happen at shutdown after some uptime; it cannot be related to my recent hacking.=20 (I do massive and frequent ruleset updates during runtime.) --=20 You are receiving this mail because: You are the assignee for the bug.=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-255830-29815-IhWlnCT4bJ>