Date: Mon, 8 Dec 2003 09:04:09 -0800 From: Marcel Moolenaar <marcel@xcllnt.net> To: Bruce Evans <bde@zeta.org.au> Cc: cvs-all@FreeBSD.org Subject: Re: cvs commit: src/usr.bin/vmstat vmstat.c Message-ID: <20031208170409.GA3666@dhcp01.pn.xcllnt.net> In-Reply-To: <20031208225505.H6229@gamplex.bde.org> References: <200312080757.hB87vvTA048651@repoman.freebsd.org> <20031208225505.H6229@gamplex.bde.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Dec 08, 2003 at 11:13:53PM +1100, Bruce Evans wrote: > On Sun, 7 Dec 2003, Marcel Moolenaar wrote: > > > marcel 2003/12/07 23:57:57 PST > > > > FreeBSD src repository > > > > Modified files: > > usr.bin/vmstat vmstat.c > > Log: > > Unbreak vmstat -i on ia64: > > o nintr and inamlen must by of type size_t, not int, > > o Remove now unnecessary casts, > > o Handle the aflag differently, because the intr. names have a > > fixed width and almost always have trailing spaces. > > The fixed width and padding with spaces is a new kernel bug on i386's. > Was it broken before on ia64's? vmstat -i worked (and works again) on ia64. The interface remains broken. > I don't like the aflag handling and removed most of it. Interrupts > with nonzero counts should always be displayed (like they used to be), > and there is little need for different levels of allness. Genuine interrupt sources with zero counts should always be displayed, but since the interface is broken, we don't know if an entry is genuine or not. > This patch depends on other changes to print the rate more precisely. As long as we divide the count by the uptime, there's no way we can ever have any form of precision. This is ok for the one-shot output, but the repeated output should base the rate on the difference between the new and the old values, divided by the repeat interval. -- Marcel Moolenaar USPA: A-39004 marcel@xcllnt.net
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20031208170409.GA3666>