Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 22 Jul 1998 17:26:19 -0700
From:      David Greenman <dg@root.com>
To:        zhihuizhang <bf20761@binghamton.edu>
Cc:        Julian Elischer <julian@whistle.com>, hackers <freebsd-hackers@FreeBSD.ORG>
Subject:   Re: Questions on inactive page queue 
Message-ID:  <199807230026.RAA07941@implode.root.com>
In-Reply-To: Your message of "Wed, 22 Jul 1998 12:37:02 EDT." <Pine.SOL.L3.93.980722123228.1433A-100000@ultra10> 

next in thread | previous in thread | raw e-mail | index | archive | help
>> > (1) Why we set a target for the inactive page queue? Who can manipulate
>> > the pages on the inactive page queue? How to guarantee that a page won't
>> > move from the inactive queue while paging out? 
>> > 
>
>David Greenman answered my second question and part of the first one. I
>guess that one of the hardest part of VM is to understand how pages get
>moved around different queues.  In my case, I want to know who could
>possibly access the inactive queue at the same time the pageout is going
>on.

   Accessing the queue and doing something to a page that is on it are
different things. When the system wants to page out a page, the page is
made read-only for all mappings of it and it is then set as busy so that
1) the page-reclaim part of the kernel leaves it alone on subsequant scans,
and 2) so that if a page fault occurs for it, the fault will be suspended
until the system is done paging it out.

-DG

David Greenman
Co-founder/Principal Architect, The FreeBSD Project

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199807230026.RAA07941>