Date: Sat, 17 Mar 2012 18:49:51 -0500 From: Alan Cox <alc@rice.edu> To: Nathan Whitehorn <nwhitehorn@freebsd.org> Cc: Alan Cox <alc@FreeBSD.org>, svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org Subject: Re: svn commit: r233097 - head/sys/amd64/amd64 Message-ID: <4F65231F.1000101@rice.edu> In-Reply-To: <A64E77A8-291B-4DD9-9271-015DE8069F77@freebsd.org> References: <201203172204.q2HM4xBH088986@svn.freebsd.org> <A64E77A8-291B-4DD9-9271-015DE8069F77@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 03/17/2012 17:14, Nathan Whitehorn wrote: > > > On Mar 17, 2012, at 5:04 PM, Alan Cox wrote: > >> Author: alc >> Date: Sat Mar 17 22:04:58 2012 >> New Revision: 233097 >> URL: http://svn.freebsd.org/changeset/base/233097 >> >> Log: >> With the changes over the past year to how accesses to the page's dirty >> field are synchronized, there is no need for pmap_protect() to acquire >> the page queues lock unless it is going to access the pv lists. >> >> Reviewed by: kib > > Under what circumstances does the page queue lock actually have to be > held? You need it when you are manipulating the pvo_vlink lists. That said, we are at the point where the PPC/AIM pmap could define its own PVO list mutex and not (ab)use the page queues mutex for this purpose. Just do a straightforward, mechanical substitution, and you should be fine. > It looks like from this that I can remove the page queue lock from the > PPC/AIM pmap_protect() completely, but I don't have a good sense of it. Yes, you can this instant remove the acquisition and release of the page queues lock from this function.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4F65231F.1000101>