Date: Sat, 18 Nov 2000 00:02:37 -0800 (PST) From: Matt Dillon <dillon@earth.backplane.com> To: David Greenman <dg@root.com> Cc: Tor Egge <tegge@FreeBSD.ORG>, cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG Subject: Re: cvs commit: src/sys/kern vfs_cluster.c Message-ID: <200011180802.eAI82bP79942@earth.backplane.com> References: <200011180045.QAA04915@implode.root.com>
next in thread | previous in thread | raw e-mail | index | archive | help
:>tegge 2000/11/17 15:40:08 PST :> :> Modified files: :> sys/kern vfs_cluster.c :> Log: :> Don't attempt to cluster write buffers where the VMIO flag isn't set. : : Huh? Why wouldn't you want to do that? : :-DG : :David Greenman :Co-founder, The FreeBSD Project - http://www.freebsd.org I would appreciate it if any VFS/BIO related patches were Cc'd to both Kirk and I prior to being committed. Both Kirk and I have been doing an extensive audit of the VFS/BIO subsystem trying to track down a serious bitmap corruption bug. In regards to this particular commit, it looks like a big NOP. A buffer accessible via a [gb]incore() lookup will either have B_MALLOC set or B_VMIO set. If B_MALLOC isn't set then B_VMIO will be. There really isn't any need to check for B_VMIO. Rather then integrate in a check, if you are paranoid you should probably put a KASSERT() in instead (KASSERT if B_VMIO is not set). The clustering code does indeed only work with VMIO-backed buffers, because those are the only buffers which properly set the b_pages[] array which is how the clustering code gloms buffers together. -Matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200011180802.eAI82bP79942>