Date: Wed, 29 Jan 2014 11:21:37 -0800 From: John-Mark Gurney <jmg@funkthat.com> To: Garrett Wollman <wollman@csail.mit.edu> Cc: freebsd-net@freebsd.org Subject: Re: Big physically contiguous mbuf clusters Message-ID: <20140129192137.GF93141@funkthat.com> In-Reply-To: <21225.20047.947384.390241@khavrinen.csail.mit.edu> References: <21225.20047.947384.390241@khavrinen.csail.mit.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
Garrett Wollman wrote this message on Wed, Jan 29, 2014 at 13:54 -0500: > Resolved: that mbuf clusters longer than one page ought not be > supported. There is too much physical-memory fragmentation for them > to be of use on a moderately active server. 9k mbufs are especially > bad, since in the fragmented case they waste 3k per allocation. I agree, but I am split on removing the code as there are still very broken controllers that may require them, though in those cases, it might be helpful to have a tunable that lets you set how many jumbo frames are allocated at boot, and these pages are never released back to the system... We definately need to fix all the drivers that use MJUM9BYTES which apparently are quite a few: http://fxr.watson.org/fxr/ident?im=excerpts;i=MJUM9BYTES -- John-Mark Gurney Voice: +1 415 225 5579 "All that I will do, has been done, All that I have, has not."
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20140129192137.GF93141>