From mboxrd@z Thu Jan 1 00:00:00 1970 From: Vincent Pelletier Subject: Re: leds-gpio on x86 Date: Wed, 19 Aug 2015 00:56:35 +0200 Message-ID: <20150819005635.1daee673@x2> References: <20150806173558.GZ1540@lahna.fi.intel.com> <20150806201820.5ca0764d@x2> <20150807105745.GB1540@lahna.fi.intel.com> <20150808140656.6b81d3b9@x2> <20150811120038.GF1541@lahna.fi.intel.com> <20150811194256.08816c63@x2> <20150812125314.GS1541@lahna.fi.intel.com> <20150815123640.45a5b9bf@x2> <20150817230823.61de477a@x2> <20150818093226.3db0b553@x2> <20150818090216.GG1552@lahna.fi.intel.com> <20150818133850.42bb31cf@x2> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Return-path: Received: from mail-wi0-f182.google.com ([209.85.212.182]:33248 "EHLO mail-wi0-f182.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750710AbbHRW4j (ORCPT ); Tue, 18 Aug 2015 18:56:39 -0400 Received: by wijp15 with SMTP id p15so112204369wij.0 for ; Tue, 18 Aug 2015 15:56:37 -0700 (PDT) In-Reply-To: <20150818133850.42bb31cf@x2> Sender: linux-leds-owner@vger.kernel.org List-Id: linux-leds@vger.kernel.org To: Mika Westerberg Cc: linux-leds@vger.kernel.org, Simon Guinot Hello, I did a first pass on gpio-f7188x module, and pushed several commits in my current working branch: https://github.com/vpelletier/linux/commits/ts651 This commit contains the relevant part of my changes: gpio: gpio-f7188x: Use mutex for access serialisation. (currently: https://github.com/vpelletier/linux/commit/0fecf37bcee67d577f7e1e8e58afc463f452bcff ) but "of course", it does not work: kernel Oops as soon as I trigger the concurrent access which caused the original issue. With a single active core, it works fine of course. Sadly, the traceback is longer than on-screen info, despite being at 1920x1080 resolution with default console font. I recall reading about a project to display kernel errors as qr-codes, did it get somewhere ? Also, watchdog kicks in a few seconds after, complaining about hard lock on (at least) one core, and scrolling even more. I suspect both cores are hard locked, and have so far no idea what I did wrong: - each acquire is followed by a release, with no code path around it - gpio_device.can_sleep is true (as was the case before my patch) ...would I have somehow poked at another bug ? Regards, -- Vincent Pelletier