Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 9 Feb 2001 10:14:21 -0500 (EST)
From:      Andrew Gallatin <gallatin@cs.duke.edu>
To:        freebsd-current@FreeBSD.ORG
Cc:        Dag-Erling Smorgrav <des@ofug.org>, Julian Elischer <julian@elischer.org>, Josef Karthauser <joe@tao.org.uk>, Robert Watson <rwatson@FreeBSD.ORG>, Brian Somers <brian@Awfulhak.org>, Bruce Evans <bde@zeta.org.au>, freebsd-current@FreeBSD.ORG, freebsd-hackers@FreeBSD.ORG
Subject:   Re: What's changed recently with vmware/linuxemu/file I/O
Message-ID:  <14980.2381.849005.313177@grasshopper.cs.duke.edu>
In-Reply-To: <14979.29437.518299.842853@grasshopper.cs.duke.edu>
References:  <xzpsnlqmh0o.fsf@flood.ping.uio.no> <Pine.NEB.3.96L.1010207220443.19807J-100000@fledge.watson.org> <20010208113519.A789@tao.org.uk> <3A828C2C.F7CDA809@elischer.org> <xzp4ry5gve0.fsf@flood.ping.uio.no> <14979.29437.518299.842853@grasshopper.cs.duke.edu>

next in thread | previous in thread | raw e-mail | index | archive | help

Andrew Gallatin writes:
 > 
 > Dag-Erling Smorgrav writes:
 >  > Julian Elischer <julian@elischer.org> writes:
 >  > > I believe that vmware mmaps a region of memory and then somehow syncs 
 >  > > it to disk. (It is certainly doing something like it here).
 >  > 
 >  > Theory: VMWare mmaps a region of memory corresponding to the virtual
 >  > machine's "physical" RAM, then touches every page during startup.
 >  > Unless some form of clustering is done, this causes 16384 write
 >  > operations for a 64 MB virtual machine...
 >  > 
 > 
 > Pretty much.  But the issue is that this should never hit the disk
 > unless we're under memory pressure because it is mapped MAP_NOSYNC
 > (actually the file is unlinked prior to the mmap() and a heuristic in
 > vm_mmap() detects this and sets MAP_NOSYNC).

I take it back.  At least with the latest version of vmware, it is
apparently not mapped MAP_NOSYNC.  I think they've moved from
mmap'ing a file in $TMPDIR to just using the CONFIG.std save/resume
file.  Perhaps this is only if you have resumed from a suspended
state... I haven't checked that out yet.

At any rate, hacking linux_mmap to ad MAP_NOSYNC to mmaped files, in
combination with yesterdays patch, appears to improve
perf. considerably. 

Drew

------------------------------------------------------------------------------
Andrew Gallatin, Sr Systems Programmer	http://www.cs.duke.edu/~gallatin
Duke University				Email: gallatin@cs.duke.edu
Department of Computer Science		Phone: (919) 660-6590



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?14980.2381.849005.313177>