From mboxrd@z Thu Jan 1 00:00:00 1970 From: Felipe Balbi Subject: Re: [PATCH 3/4] net: ethernet: cpsw: split out IRQ handler Date: Fri, 2 Jan 2015 21:02:38 -0600 Message-ID: <20150103030238.GA19959@saruman> References: <1420222228-31949-1-git-send-email-balbi@ti.com> <1420222228-31949-4-git-send-email-balbi@ti.com> <20150102185535.GB4920@saruman> <20150102190350.GC4920@saruman> Reply-To: Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="OgqxwSJOaUobr8KG" Return-path: Content-Disposition: inline In-Reply-To: Sender: netdev-owner@vger.kernel.org To: Dave Taht Cc: Felipe Balbi , "netdev@vger.kernel.org" , Linux OMAP Mailing List List-Id: linux-omap@vger.kernel.org --OgqxwSJOaUobr8KG Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Hi, On Fri, Jan 02, 2015 at 02:56:36PM -0800, Dave Taht wrote: > On Fri, Jan 2, 2015 at 11:03 AM, Felipe Balbi wrote: > > Hi, > > > > (please use reply-all to keep mailing lists in Cc, also avoid > > top-posting) >=20 > I am trying not to read netdev right now... and failing, obviously. oops :-) > > On Fri, Jan 02, 2015 at 10:58:29AM -0800, Dave Taht wrote: > >> The beaglebone only has a 100mbit phy, so you aren't going to get more > >> than that. > > > > very true :-) Still, with AM437x SK which is definitely GigE, I'm > > getting 201Mbits/sec. > > > >> (so do a lot of IoT devices). > >> > >> So you have the two patches that went by on BQL and on NAPI for the be= agle? > > > > no, got any pointers ? >=20 > the relevant thread was "am335x: cpsw: phy ignores max-speed setting" >=20 > and the initial very small BQL enablement patch was here: >=20 > https://patchwork.ozlabs.org/patch/407640/ I'll test it out, sure. > (it needed a saner treatment of a failure to dma something in > cpsw_tx_packet_submit - the patch as is has also been part of nelsons > trees for the beaglebone for a while) >=20 > But it was rightly pointed out later in the thread that this change >=20 > +#define CPSW_POLL_WEIGHT 16 >=20 > made for the biggest part of the improvement, and someone else on the > thread proposed handling that more dynamically for 100mbit phys with > another patch (that I can't find at the moment) >=20 > ... but the root cause of the excessive latency in this driver was the > single tx/rx dma queue, which you are addressing in your patch set. I still think there's a lot of work pending for CPSW, the think slows to a crawl and takes a lot of CPU for something that should be mostly handled by DMA. I can very easily get 85% CPU usage with iperf. > So if you glop on more of the above, mo better, perhaps you will win > bigger. >=20 > I will try to slice out some time to boot up a beagle on net-next next we= ek. my patches aren't applied yet, however. cheers --=20 balbi --OgqxwSJOaUobr8KG Content-Type: application/pgp-signature; name="signature.asc" Content-Description: Digital signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iQIcBAEBAgAGBQJUp1vOAAoJEIaOsuA1yqREm1gP/0z4KoxkQp9/N3Y4IVLT44j3 c5voy1pNgWiJn/DCGG0TdCq4PM9J3uKaTnM8wInBH38vOKW0nFSzHlSYUf8SoEd0 k5EUhp53r259RDu1gTGNmxAV8R3aBBQga0luQX/U2wKRYeh1H0exlKxCWuaHSe3v 7p6WdwLsgx204pind2IEnWTG/DqOfP5UKnFCOIE2p5ZB4Vo2O8562us36p0SXFXX esATQT51TQIp3JMPm7DcCBxQE6KfAF3vRNMdApSuMgctyC/gNjsbwuzfLT0z4FUd sBT5YpB6kr2SyXdMxKnHmFPAdiBb7CxRNdo5DmdIO577o6AX/5Ym49tbWV5tdCzd 4ryMIoI7FZAxvt6lzc3QVHqlrZuD/zPsjJ8484P6O3k9Uhhv+8lOrnnUcFV1TEvr CWwUV3b/ReCxX8EHKB5nw5Iej/Kfhi2L2h7EFmEgB+Q6TMBbCIVTSZQxmr8H21ZO J+F7xOk6SgvR2Z1r8eNWJZP/06XxXP95LF17tPQb5uR8IxLpUdxpWuIXwAgE6qOn ZQ7RgbevJE3/qVMqMpSIWtMmJUEcWQ1wnE7YW27WJr2/N8GEb7uPA8T9+83RPsZ/ IVD2HqNFdw0FjeTysi9hnpFc44Iq0eYuP4HHTd2AajCNNwnpZa/RQgidoeFbuiEk 5R0mCdgsvA3wgl5t3Qu4 =zIb8 -----END PGP SIGNATURE----- --OgqxwSJOaUobr8KG--