Date: Thu, 16 Apr 1998 22:42:14 +0800 From: Peter Wemm <peter@netplex.com.au> To: "Jordan K. Hubbard" <jkh@time.cdrom.com> Cc: Bruce Evans <bde@zeta.org.au>, cvs-all@FreeBSD.ORG, cvs-committers@FreeBSD.ORG, cvs-sys@FreeBSD.ORG, dyson@FreeBSD.ORG Subject: Re: cvs commit: src/sys/kern vfs_subr.c Message-ID: <199804161442.WAA17178@spinner.netplex.com.au> In-Reply-To: Your message of "Thu, 16 Apr 1998 06:55:58 MST." <3518.892734958@time.cdrom.com>
next in thread | previous in thread | raw e-mail | index | archive | help
"Jordan K. Hubbard" wrote: > > > We also lost control of the update interval. > > > > Yes, but it is compile-time tweakable, so those with a desperate need to > > tweak it can still do so, although on a power-of-two basis. > > I think it'd still be nice to have it be sysctl tweakable. If nothing > else, it would give you a quick and easy way of saying "Hey! Stop > waking up the suspended drive in my laptop, dammit!" without having > to go recompile the kernel. Given the present structures, it's not trivial or it'd have been done by now. The problem is that there's a compiled-in power-of-two sized table that represents work to do in 'n' seconds. Even if this were dynamically reallocated and the filesystem sync period was changeable, one would very quickly mess up the interleave of filesystems. Perhaps it'd be better to just have two syncer methods. One that does the interleaving, and the other that does a 'sync the whole damn universe every <n> seconds'. One could switch from the gradual to the all-at-once method but there'd be no point going back. softdep will not run in that mode (not that softdep runs well at present in freebsd anyway...). > Jordan > Cheers, -Peter -- Peter Wemm <peter@netplex.com.au> Netplex Consulting
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199804161442.WAA17178>