Date: Sun, 14 Jul 2002 09:47:56 -0700 From: Luigi Rizzo <rizzo@icir.org> To: David Miller <dmiller@sparks.net> Cc: hackers@FreeBSD.ORG Subject: Re: setsockopt() weirdness Message-ID: <20020714094756.G74633@iguana.icir.org> In-Reply-To: <Pine.BSF.4.21.0207141241480.15972-100000@search.sparks.net>; from dmiller@sparks.net on Sun, Jul 14, 2002 at 12:43:51PM -0400 References: <20020714085832.C74633@iguana.icir.org> <Pine.BSF.4.21.0207141241480.15972-100000@search.sparks.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, Jul 14, 2002 at 12:43:51PM -0400, David Miller wrote: > On Sun, 14 Jul 2002, Luigi Rizzo wrote: > > > On Sun, Jul 14, 2002 at 11:49:46AM -0400, David Miller wrote: > > ... > > > HZ is set to 5000; the machine is intended to process several tens of > > > thousands of very small packets per second, and interrupt processing was a > > > big problem. > > > > why don't you use "options DEVICE_POLLING" then :) > > I am. Perhaps I understood wrong, but I thought that HZ controlled the > maximum latency when polling? oh yes, but HZ=5000 sounds a bit on the high side... do you really need a max latency of 200us ?? Plus, you still poll in the idle loop, so unless the box is overloaded, your average latency will still be shorted than 1/HZ. (all of this is of course irrelevant for the original subject:) cheers luigi To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020714094756.G74633>