Date: Fri, 12 Dec 2008 12:40:42 +0900 From: Pyun YongHyeon <pyunyh@gmail.com> To: Josh Carroll <josh.carroll@gmail.com> Cc: Steve Franks <bahamasfranks@gmail.com>, current-list freebsd <freebsd-current@freebsd.org> Subject: Re: Call for testers: Atheros AR8121(L1E)/AR8113/AR8114(L2E) ethernet Message-ID: <20081212034042.GL46707@cdnetworks.co.kr> In-Reply-To: <8cb6106e0812111923l15f1f715g6f20f5925e1d471a@mail.gmail.com> References: <4910C055.8000505@janh.de> <20081105013558.GA99795@cdnetworks.co.kr> <20081203090658.GJ9639@cdnetworks.co.kr> <37502393@bb.ipt.ru> <20081206023016.GF22093@cdnetworks.co.kr> <539c60b90812081127s4ffb509fnea9d44d4298da666@mail.gmail.com> <8cb6106e0812081252j2b0c8e78g4dcecf8d3770c269@mail.gmail.com> <8cb6106e0812101745l54b23a08k7fbeddeb605f88ea@mail.gmail.com> <20081212020539.GI46707@cdnetworks.co.kr> <8cb6106e0812111923l15f1f715g6f20f5925e1d471a@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Dec 11, 2008 at 10:23:50PM -0500, Josh Carroll wrote: > > Would you show me the output of "sysctl dev.ale.0.stats"? > > Here is the output shortly after a reboot, before I try to stream anything: > > dev.ale.0.stats.rx.good_frames: 1843 > dev.ale.0.stats.rx.good_bcast_frames: 9 > dev.ale.0.stats.rx.good_mcast_frames: 0 > dev.ale.0.stats.rx.pause_frames: 0 > dev.ale.0.stats.rx.control_frames: 0 > dev.ale.0.stats.rx.crc_errs: 0 [...] > And after trying to stream (I let it struggle along for about 30-40 seconds): > > dev.ale.0.stats.rx.good_frames: 4350 > dev.ale.0.stats.rx.good_bcast_frames: 35 > dev.ale.0.stats.rx.good_mcast_frames: 0 > dev.ale.0.stats.rx.pause_frames: 3636 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ > dev.ale.0.stats.rx.control_frames: 0 > dev.ale.0.stats.rx.crc_errs: 0 [...] I guess it's caused by flow-control frames. The flow-control feature is disabled in most drivers as mii(4) layer still lacks the feture. ale(4) has flow-control support code but it is in disabled state. When mii(4) is ready to handle flow-controls ale(4) may work better. em(4) does not rely on mii(4) layer so it implemented flow-controls in driver. You can check flow-control satus of em(4) with "sysctl dev.em.0.stats=1"(See XON/XOFF). -- Regards, Pyun YongHyeon
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20081212034042.GL46707>