From mboxrd@z Thu Jan 1 00:00:00 1970 From: Marc Kleine-Budde Subject: Re: [PATCH 3.19-rc3] c_can: SIE disabled when berr-reporting is off to reduce irq flood Date: Sun, 18 Jan 2015 19:56:05 +0100 Message-ID: <54BC01C5.2070803@pengutronix.de> References: <9c72f211-becc-4c0f-94f6-0700dfb1195e@GRBSR0089.marel.net> <1735533.0yOonAfCy1@heinz> <54B472B2.4010300@optusnet.com.au> <20150113153243.26859.47218@shannon> <54BBFCAF.2090603@pengutronix.de> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha512; protocol="application/pgp-signature"; boundary="IF7vwLpnvBnJ0VGN4AamrA05BoWoqOC4e" Return-path: Received: from metis.ext.pengutronix.de ([92.198.50.35]:39694 "EHLO metis.ext.pengutronix.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751011AbbARS4R (ORCPT ); Sun, 18 Jan 2015 13:56:17 -0500 In-Reply-To: Sender: linux-can-owner@vger.kernel.org List-ID: To: Viktor Babrian Cc: Andri Yngvason , Tom Evans , linux-can@vger.kernel.org This is an OpenPGP/MIME signed message (RFC 4880 and 3156) --IF7vwLpnvBnJ0VGN4AamrA05BoWoqOC4e Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable On 01/18/2015 07:52 PM, Viktor Babrian wrote: >=20 >> NACK. The transitions are important for some applications. I've an not= >> yet mainlined series that limits the bus errors. I'm going to port the= m >> to the current kernel tomorrow. Can you test it? >=20 > This controller simply does not have a warn->pass state transition=20 > interrupt. The only way to accurately (right on time) report this=20 > transition is to switch on all error interrupts (ie the status interrup= t)=20 > that cause interrupt flood in some scenarios e.g. when frame is not ack= ed=20 > by other parties. > If we don't turn on the status interrupt, the warn->pass state transiti= on=20 > can be detected later on a successful rx/tx event (or when entering bus= =20 > off etc). >=20 > I have learned that the warn->pass transition interrupt is also missing= =20 > from other controllers as well (like Flexcan). Interrupt flood caused b= y=20 > enabling error interrupts regardless of the berr-reporting state is an = > issue there too. >=20 > Also note that when the system is flooded by these error interrupts,=20 > system performace degrades so much that serial terminal echo becomes=20 > annoyingly long. When this happens, I wonder how usable the "accurate" = > state transition reports are anyway. As I said, they are vital to some applications. The bus error limiting series will follow. Marc --=20 Pengutronix e.K. | Marc Kleine-Budde | Industrial Linux Solutions | Phone: +49-231-2826-924 | Vertretung West/Dortmund | Fax: +49-5121-206917-5555 | Amtsgericht Hildesheim, HRA 2686 | http://www.pengutronix.de | --IF7vwLpnvBnJ0VGN4AamrA05BoWoqOC4e Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQIcBAEBCgAGBQJUvAHLAAoJECte4hHFiupUvqwP/3xI5F8wK8QA9Jj5uZ6lTC9I ags0+3J3QcFwgXfcwIm4z0xEwMKqVgl8ObHi2nY6smFjMGjWpTXHeP7034XgWzfm dES2K8Fxx/sZF7Q/nIXBT4oVkTFikTPOEy0wNAV4bfTYfCD9gk0DFfv90b9DmUo0 wt/jZhbEh/qWN0uwikMbgRsrdXXbAZ81KxAqG5O4UoIxYb1QV0K4FOsdoJzlu5g9 Hp0Xk3dMY+HeFEvSy4qlwcbSUKBXTGg76HFFEzZdzspueX4W4bjDQzA0jEpCf1uO jiotzV00st0fLdKX3FmfK4wRP81MtXdEs9glZgxtdU0luJZomQmBGsUfo5U2n8o8 r31cKPrSHr5fT1N5nur0h94EeDpo4793kt9W2jipiyrIzn4KokXCmXChgM+bQwP5 twi9VRDJWjxSFxHmQsxsAI5Udh6kNiu1wcVze4O5rtPZhhFF6T4M1TSyW669Mz/4 kYIb2X5axTttD7583MlqBYLE2ANtg+CjdEeBQUTfkDxPf3kNmvSAvtjB/VN4h/97 m8FBMvjR0w7o1ngDNvqNfCLuvthiFlkncb8hG10QTVndCyBVJPlMQKv0vcwhCpu/ bqBpYNiEDrNrlb9DjS42n11x+EO5GCCnxGdB4oRRRP4Yc+9J1jiZ9mwZ3yI+hf9C eYFd1n7KdjQX+9UDzgL1 =jjXC -----END PGP SIGNATURE----- --IF7vwLpnvBnJ0VGN4AamrA05BoWoqOC4e--