Date: Tue, 06 May 2014 10:21:11 -0700 From: Navdeep Parhar <np@FreeBSD.org> To: "Alexander V. Chernikov" <melifaro@FreeBSD.org>, FreeBSD Net <net@freebsd.org>, hackers@freebsd.org Cc: jfv@FreeBSD.org, Adrian Chadd <adrian@freebsd.org>, wollman@freebsd.org, jhb@freebsd.org Subject: Re: Use of contiguous physical memory in ixgbe driver Message-ID: <53691A07.2060304@FreeBSD.org> In-Reply-To: <53690885.1010704@FreeBSD.org> References: <53690885.1010704@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 05/06/14 09:06, Alexander V. Chernikov wrote: > Hello guys. > (bootstrapping people involved in previous version of this topic, sorry > for that) > > There were several problem descriptions/discussions on using 9k+ mbufs > with current allocator in: > if_em: kern/183381 > cxgbe: > http://lists.freebsd.org/pipermail/freebsd-net/2014-February/037834.html > > general one: > http://lists.freebsd.org/pipermail/freebsd-net/2014-January/037673.html I changed cxgbe(4) in response to those discussions (see r263317 for details; r263451 has the man page updates). It still prefers large clusters (if MTU is > 4K) by default but will happily fall back to the 4K zone if it encounters failures when allocating from the larger zones. I also added a knob that you can use to forbid cxgbe(4) from even attempting to allocate from the large zones. So the latest cxgbe(4) should be able to cope just fine at whatever MTU even when the 9K, 16K zones are depleted. Regards, Navdeep
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?53691A07.2060304>