Date: Thu, 21 Dec 2000 11:18:28 -0800 From: Alfred Perlstein <bright@wintelcom.net> To: John Baldwin <jhb@FreeBSD.org> Cc: Julian Elischer <julian@elischer.org>, smp@FreeBSD.org, archie@FreeBSD.org Subject: Re: looking for locking advice.. Message-ID: <20001221111828.X19572@fw.wintelcom.net> In-Reply-To: <XFMail.001221111202.jhb@FreeBSD.org>; from jhb@FreeBSD.org on Thu, Dec 21, 2000 at 11:12:02AM -0800 References: <20001221105710.U19572@fw.wintelcom.net> <XFMail.001221111202.jhb@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
* John Baldwin <jhb@FreeBSD.org> [001221 11:12] wrote: > > On 21-Dec-00 Alfred Perlstein wrote: > > * John Baldwin <jhb@FreeBSD.org> [001221 10:47] wrote: > >> > >> On 21-Dec-00 Alfred Perlstein wrote: > >> > I wouldn't worry all too much about the distiction between interrupt > >> > execution state versus non-interrupt. Sure device nodes want to complete > >> > as soon as possible, however eventually we'll have _multiple_ software > >> > interrupt threads running on each processor so latency is reduced. > >> > >> > ps ax | grep 'swi[0-9]' | wc -l > >> 7 > >> > >> We already have multiple software interrupt threads that can run > >> concurrently. > >> :) > > > > But not in the same class though, right? > > > > You can't more than one network software inetrrupt running atm, no? > > You could if you wanted. Just create multiple threads via sinthand_add() and > schedule them with sched_swi(). (Yes, this is another namespace I'm going to > fix0r). *nod*, we'll really want to run at least one of each interrupt class on each cpu most likely otherwise the network stack remains serialized based on only having a single scheduleable interrupt thread. -- -Alfred Perlstein - [bright@wintelcom.net|alfred@freebsd.org] "I have the heart of a child; I keep it in a jar on my desk." To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20001221111828.X19572>