Date: Sun, 30 Nov 2014 22:03:13 +0300 From: Gleb Smirnoff <glebius@FreeBSD.org> To: Adrian Chadd <adrian@freebsd.org> Cc: "svn-src-head@freebsd.org" <svn-src-head@freebsd.org>, "svn-src-all@freebsd.org" <svn-src-all@freebsd.org>, "src-committers@freebsd.org" <src-committers@freebsd.org> Subject: Re: svn commit: r275326 - in head: sys/dev/cxgbe/tom sys/kern sys/netinet sys/sys usr.bin/bluetooth/btsockstat usr.bin/netstat usr.bin/systat Message-ID: <20141130190313.GC47144@FreeBSD.org> In-Reply-To: <CAJ-VmoniCHqV7FO97RWa%2BJxhWmSfy09LPY3gZW%2BQ9wwJfL74JA@mail.gmail.com> References: <201411301252.sAUCqYXm055601@svn.freebsd.org> <CAJ-VmoniCHqV7FO97RWa%2BJxhWmSfy09LPY3gZW%2BQ9wwJfL74JA@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Adrian, On Sun, Nov 30, 2014 at 09:55:05AM -0800, Adrian Chadd wrote: A> The whole idea of an mbuf whose data isn't quite there yet makes A> debugging issues rather amusing, as now you may have VM issues to A> debug at the same time as you're debugging network related stuff. I don't see how VM issues are related. The mbufs are referencing allocated, mapped and wired pages via sf_bufs. The same way as they did it since 1998. The only difference is that pages aren't populated with data. No VM changes happen to the pages while the mbufs that reference them sit in the socket buffer. A> I really think this could've been done without all the back-handed VM A> work. The mbufs and IO buffers both have completion function calls; it A> would've been much less intrusive to do it that way. If you can describe the alternative design at least in some detail, I am listening. -- Totus tuus, Glebius.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20141130190313.GC47144>