From owner-freebsd-smp Fri May 18 20:49:30 2001 Delivered-To: freebsd-smp@freebsd.org Received: from superconductor.rush.net (superconductor.rush.net [208.9.155.8]) by hub.freebsd.org (Postfix) with ESMTP id BE9E937B422; Fri, 18 May 2001 20:49:27 -0700 (PDT) (envelope-from bright@superconductor.rush.net) Received: (from bright@localhost) by superconductor.rush.net (8.11.2/8.11.2) id f4J3nL916036; Fri, 18 May 2001 23:49:21 -0400 (EDT) Date: Fri, 18 May 2001 23:49:20 -0400 From: Alfred Perlstein To: Andrew Gallatin Cc: smp@FreeBSD.ORG, jake@FreeBSD.ORG, jhb@FreeBSD.ORG, bp@FreeBSD.ORG Subject: Re: please review, that pesky vm lock Message-ID: <20010518234920.T7118@superconductor.rush.net> References: <20010517041116.O12365@superconductor.rush.net> <20010518043550.O7118@superconductor.rush.net> <15109.58028.650986.78190@grasshopper.cs.duke.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0us In-Reply-To: <15109.58028.650986.78190@grasshopper.cs.duke.edu>; from gallatin@cs.duke.edu on Fri, May 18, 2001 at 11:04:12PM -0400 X-all-your-base: are belong to us. Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org * Andrew Gallatin [010518 23:04] wrote: > > Alfred Perlstein writes: > > * Alfred Perlstein [010517 04:11] wrote: > > > The giant vm lock is pretty stable. The patch is here: > > > http://people.freebsd.org/~alfred/vm.diff > > > > > > > Updated patch is at the above url, several things were fixed: > > Two questions: > > Are you planning to make a pass through the MD code of other archs > such as alpha & ia64 prior to commit? If not, please give me a heads > up of at least 24-48 hours before you commit so that I can do alpha. > I'd rather you did it, but I'm willing to take it on if nobody else > is... I'm sorry, I had the patchset up for quite a while and committed it shortly before getting your message. As far as the alpha/ia64 the fixes should be trivial, you basically have: .) a couple of places where you'll catch assertions unless the vm lock is held, .) the chance to remove aquiring Giant from trap() .) a couple of places where you'll need to include sys/mutex.h to squash compiler errors. > Also, a number of files in the patch appear to have the addition of > #include as their only change (coda/coda_namecache.c, > linprocfs_misc.c, etc). Is this intentional? Yes, because of inline functions in the vm headers they require that certain defines are present for the assertions and structures used in those inlines. -- -Alfred Perlstein [alfred@freebsd.org] Instead of asking why a piece of software is using "1970s technology," start asking why software is ignoring 30 years of accumulated wisdom. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message