Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 11 Jan 2007 11:22:09 -0500
From:      John Baldwin <jhb@freebsd.org>
To:        Peter Jeremy <peterjeremy@optushome.com.au>
Cc:        "Brad L. Chisholm" <blc@bsdwins.com>, freebsd-hackers@freebsd.org
Subject:   Re: Kernel hang on 6.x
Message-ID:  <200701111122.09936.jhb@freebsd.org>
In-Reply-To: <20070111092153.GG833@turion.vk2pj.dyndns.org>
References:  <20070110215207.GA85834@bsdone.bsdwins.com> <200701102211.39412.jhb@freebsd.org> <20070111092153.GG833@turion.vk2pj.dyndns.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thursday 11 January 2007 04:21, Peter Jeremy wrote:
> On Wed, 2007-Jan-10 22:11:38 -0500, John Baldwin wrote:
> >64 / 14.  That gives a result of 153405586.  However, you really want to
> >round this up to a multiple of 288 (because the kernel rounds it down to
> >a multiple of 288), so I'd use a value of at least 153405792.
> 
> Looking at the code, it seems that each SWAPMETA object manages 16 pages.

Up to 16 pages.  Not all 16 slots are always used apparently (based on 
empirical evidence where I've locked up a box that had enough swap_zone items 
for all of swap space assuming each object mapped 16 pages and it still 
deadlocked).  I think it depends on the access pattern as to how full the 
various objects are.

-- 
John Baldwin



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200701111122.09936.jhb>