Date: Wed, 6 Nov 2002 11:20:13 -0500 From: Barney Wolff <barney@tp.databus.com> To: Petri Helenius <pete@he.iki.fi> Cc: Barney Wolff <barney@tp.databus.com>, freebsd-net@FreeBSD.ORG Subject: Re: bpf Message-ID: <20021106162013.GA60489@tp.databus.com> In-Reply-To: <05af01c2856a$f103c960$862a40c1@PHE> References: <03fe01c28386$186fed80$862a40c1@PHE> <20021103225402.GA28812@tp.databus.com> <05af01c2856a$f103c960$862a40c1@PHE>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Nov 06, 2002 at 10:03:06AM +0200, Petri Helenius wrote: > > The specific problem with bpf is that one might have a half-full buffer > of captured data when the select timeout hits. In that case the select > returns with no FDs ready while I think it really should return with > the bpf fd. (and if you look at the code on sys/net/bpf.c, there is > timeout handling towards that goal) but I?ve yet to figure out where it goes > wrong. > > IMO, at timeout the code should check if bd_slen > 0 and if yes, > do ROTATE_BUFFERS and bpf_wakeup() The bpf timeout has nothing to do with the select timeout. Two separate facilities. > The functionality I?m looking for is to get all the packets accumulated, say > in 1 second in single read regardless of if I got a buffer?s full of data. Then do a select with no FDs just to get a periodic wakeup. Then at wakeup do the bpf read, having set nonblocking and immediate mode so you don't get stuck in the read. -- Barney Wolff http://www.databus.com/bwresume.pdf I'm available by contract or FT, in the NYC metro area or via the 'Net. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20021106162013.GA60489>