From owner-cvs-all Tue Oct 16 20:48:20 2001 Delivered-To: cvs-all@freebsd.org Received: from mailman.zeta.org.au (mailman.zeta.org.au [203.26.10.16]) by hub.freebsd.org (Postfix) with ESMTP id 51C5337B403; Tue, 16 Oct 2001 20:48:12 -0700 (PDT) Received: from bde.zeta.org.au (bde.zeta.org.au [203.2.228.102]) by mailman.zeta.org.au (8.9.3/8.8.7) with ESMTP id NAA17508; Wed, 17 Oct 2001 13:48:00 +1000 Date: Wed, 17 Oct 2001 13:47:10 +1000 (EST) From: Bruce Evans X-X-Sender: To: Dag-Erling Smorgrav Cc: Matt Dillon , , Subject: Re: cvs commit: src/sys/vm vnode_pager.c In-Reply-To: Message-ID: <20011017132416.E13949-100000@delplex.bde.org> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On 16 Oct 2001, Dag-Erling Smorgrav wrote: > I did 'kgmon -rb ; sync ; sync ; sync ; kgmon -hpr'. The results are > attached. There seems to be a locking problem that obscures the real problem (unless it is the problem). Try it without INVARIANTS or WITNESS. > BTW, it should be possible to make profiling a normal kernel option > (with a little help from 'makeoptions' like we do for debugging). > It'd take a few changes to conf/{files,options} and the Makefile > skeletons, but it would allow us to remove quite a bit of special- > case code from config(8). Well, it's a special option ("profile 2" in NOTES is equivalent to -pp on the command line). This is just as convenient in config files as "options profiling 2". There is only a little bit of special-case code to support enabling profiling in config. There is some special-case code for debugging to, and the extras for profiling are just a couple of lines in lang.l and config.y to support the keyword, and 4 extra lines in mkmakefile.c to print a bit more when profiling is enabled. I think the printfs for both profiling and debugging can be handled by ifdefs on PROFLEVEL and DEBUGLEVEL if they aren't already, but this would currently increase the amount of code from one set of printfs in config to ${N_MACHINE_ARCH} sets of conditionals in Makefile.${MACHINE}. We should separate out the common part of Makefile.${MACHINE} before cleaning this up. Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message