Date: Sun, 12 Jan 2003 03:51:14 -0800 (PST) From: Kelly Yancey <kbyanc@posi.net> To: Trish Lynch <trish@bsdunix.net> Cc: "Daniel C. Sobral" <dcs@newsguy.com>, Peter Wemm <peter@wemm.org>, <current@FreeBSD.ORG>, <kbyanc@FreeBSD.ORG> Subject: Re: Serious issues with kqueue on sockets on CURRENT. Message-ID: <20030112034405.F27352-100000@gateway.posi.net> In-Reply-To: <20030111194549.V758-100000@femme>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, 11 Jan 2003, Trish Lynch wrote:
> On Sat, 11 Jan 2003, Daniel C. Sobral wrote:
>
> > Peter, reverting the revisions below *does* fix the problem. Tim has an
> > alternative patch, though. At any rate, it seems kbyanc's solution was
> > overly simplistic. But things are broken either way, and I'm not sure
> > Tim's patch doesn't result in the kind of situation rev 1.134 tried to
> > fix, nor if his patch actually gets all cases of the bug that results
> > from 1.134.
> >
> > At any rate, I think that not receiving any event (after 1.134) is worse
> > than receiving and event claim to have more bytes than are actually
> > available (pre 1.134). It's not just Juli who have this problem.
> > AilleCat, for instance, once she heard on irc that kq had a problem,
> > tracked the problem *she* was having to the same place.
> >
>
> Yes, this is correct, some events weren't being triggered and now, with
> reverting back (with some of the current changes like the aesthetic change
> to soo_kqfilter instead of sokqfilter,) now our application that relies
> upon kqueue for scheduling runs about twice as fast....
>
> Maxim gave me a patch that accomplishes exactly what I did by hand... but
> it also leaves it in a state that it was before 1.134 where there were
> some problems that were supposed to be fixed in 1.134 and after... however
> IMO its *less* broken :)
>
> Anyway, since my understanding of this is much less than anyone else I'm
> inclined to go with whatever solution actually makes the events trigger
> for us :)
>
> I'm not a kernel programmer, nor will I ever be, I just know that
> reverting uipc_socket.c did solve some major problems I was having :)
>
> -Trish
>
I'm sorry, I'm afraid I am not familiar with the issue being discussed. Is
there a PR I can reference for more information? Exactly what events aren't
being received? Being as the logic for when to return a kevent as of
uipc_socket.c:1.136 is exactly the same as before (just the data value in the
kevent is different), I can't see how any events could *not* be returned that
weren't returned before. But then again, without knowing the symptoms you are
seeing, I can't say for sure.
Kelly
--
Kelly Yancey -- kbyanc@{posi.net,FreeBSD.org}
Visit the BSD driver database: http://www.posi.net/freebsd/drivers/
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030112034405.F27352-100000>