Date: Thu, 8 Aug 2013 13:23:43 +0400 From: Gleb Smirnoff <glebius@FreeBSD.org> To: Adrian Chadd <adrian@FreeBSD.org> Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org Subject: Re: svn commit: r254082 - head/sys/net80211 Message-ID: <20130808092343.GP20104@FreeBSD.org> In-Reply-To: <201308080509.r7859aWv084743@svn.freebsd.org> References: <201308080509.r7859aWv084743@svn.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Aug 08, 2013 at 05:09:36AM +0000, Adrian Chadd wrote: A> Author: adrian A> Date: Thu Aug 8 05:09:35 2013 A> New Revision: 254082 A> URL: http://svnweb.freebsd.org/changeset/base/254082 A> A> Log: A> Convert net80211 over to using if_transmit for the dispatch from the A> upper layer(s). A> A> This eliminates the if_snd queue from net80211. Yay! A> A> This unfortunately has a few side effects: A> A> * It breaks ALTQ to net80211 for now - sorry everyone, but fixing A> parallelism and eliminating the if_snd queue is more important A> than supporting this broken traffic scheduling model. :-) A> A> * There's no VAP and IC flush methods just yet - I think I'll add A> some NULL methods for now just as placeholders. A> A> * It reduces throughput a little because now net80211 will drop packets A> rather than buffer them if the driver doesn't do its own buffering. A> This will be addressed in the future as I implement per-node software A> queues. Now access to IFF_DRV_OACTIVE bit isn't locked. Is this okay? Ideally drivers should be written in such way that they don't need to set and clear this flag at all. Is it still needed for net80211? -- Totus tuus, Glebius.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20130808092343.GP20104>