Date: Wed, 24 Jan 2007 05:49:31 -0800 From: Luigi Rizzo <rizzo@icir.org> To: Randall Stewart <rrs@cisco.com> Cc: freebsd-net <freebsd-net@freebsd.org> Subject: Re: mbuf patch with sysctl suggestions too Message-ID: <20070124054931.B56550@xorpc.icir.org> In-Reply-To: <45B7631A.3070001@cisco.com>; from rrs@cisco.com on Wed, Jan 24, 2007 at 08:46:02AM -0500 References: <45B679F3.3080407@cisco.com> <20070124051050.A56064@xorpc.icir.org> <45B7631A.3070001@cisco.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Jan 24, 2007 at 08:46:02AM -0500, Randall Stewart wrote: > Luigi Rizzo wrote: > > On Tue, Jan 23, 2007 at 04:11:15PM -0500, Randall Stewart wrote: > >> Hi all: > >> > >> Here is iteration 2 of the mbuf patch with limits I > >> proposed. > >> > >> Also note the changes for sysctl stuff that Lugi suggested. > >> Please let me know what you think :-) > > > > ... > >> + newnmbjclusters = nmbjumbop; > >> + error = sysctl_int_checked(oidp, &newnmbjclusters, nmbjumbop, > >> + SYSCTL_NO_LIMIT, req); > > > > A few things here: > > - i don't see much of a point in defining SYSCTL_NO_LIMIT; > > UINT32_MAX would do perfectly there, and i think it is easier > > to understand than SYSCTL_NO_LIMIT (which looks like a flag). > > > > ok > > - here and in other places you do not allow decresaing the value > > (by putting min = nmbjumbop etc.), and i am not sure why. > > I understand a reasonable lower bound, but i guess the worst > > that can happen, when you reduce the limit to something above > > the current allocation, is that nothing is allocated until > > you go again below the limit, right ? > > Well.. no I believe someone (was in Lin) mentioned that > you can get a live-lock if you allow a reduction.. and > thus the mbuf clusters were NOT allowed to be reduced.. maybe... but then this is definitely worth putting a note explaining why. cheers luigi
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20070124054931.B56550>