From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?UTF-8?B?UmFmYcWCIE1pxYJlY2tp?= Subject: Re: [PATCH] ARM: BCM5301X: Add back handler ignoring external imprecise aborts Date: Mon, 31 Oct 2016 23:16:20 +0100 Message-ID: References: <20161029111229.26875-1-zajec5@gmail.com> <6e252703-0010-83ae-fa14-57352add007a@hauke-m.de> <18db27c7-1055-ff3c-0b0b-eeeaf5f3f5e0@gmail.com> <3847db9f-a8e1-b7fe-c2fa-19ea893bae5f@broadcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: <3847db9f-a8e1-b7fe-c2fa-19ea893bae5f-dY08KVG/lbpWk0Htik3J/w@public.gmane.org> Sender: devicetree-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Scott Branden Cc: Florian Fainelli , Hauke Mehrtens , "linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org" , Arnd Bergmann , Bjorn Helgaas , Lucas Stach , Jon Mason , Mark Rutland , Ray Jui , Scott Branden , bcm-kernel-feedback-list , "devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org" , Linux PCI , =?UTF-8?B?UmFmYcWCIE1pxYJlY2tp?= List-Id: devicetree@vger.kernel.org On 31 October 2016 at 23:05, Scott Branden wro= te: > On 16-10-31 02:56 PM, Florian Fainelli wrote: >> - fixups the aborts in the kernel, look where they come from, by using >> some bit of magic, looking at PCIe registers and whatnot (provided that >> is even possible), not too hard, but can take a while, and is error pron= e > > Option 4 sounds like the proper solution - check the range the abort is d= ue > to the PCI device enumeration and only ignore those aborts. This was already suggested by Arnd: http://lists.infradead.org/pipermail/linux-arm-kernel/2016-April/422873.htm= l and Ray was checking some internal datasheets, but I'm afraid he never found info on checking if error was caused by PCIe controller. Maybe we could think about some BCM5301X API (extra symbol exported by arch code) for starting ignoring aborts and stopping that. We could ignore aborts during scanning only but honestly it sounds like a bit hacky solution to me. >> I can certainly advocate that option 3 is the one that gives a working >> device, and this is what matters from a distribution perspective like >> LEDE/OpenWrt. >> > Since I don't use BCM5301x option 3 is fine by me. So for it seems like the best solution to me, but I'm open for changes if someone points another that is clean & better one. --=20 Rafa=C5=82 -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-oi0-f68.google.com ([209.85.218.68]:33387 "EHLO mail-oi0-f68.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S947422AbcJaWQV (ORCPT ); Mon, 31 Oct 2016 18:16:21 -0400 MIME-Version: 1.0 In-Reply-To: <3847db9f-a8e1-b7fe-c2fa-19ea893bae5f@broadcom.com> References: <20161029111229.26875-1-zajec5@gmail.com> <6e252703-0010-83ae-fa14-57352add007a@hauke-m.de> <18db27c7-1055-ff3c-0b0b-eeeaf5f3f5e0@gmail.com> <3847db9f-a8e1-b7fe-c2fa-19ea893bae5f@broadcom.com> From: =?UTF-8?B?UmFmYcWCIE1pxYJlY2tp?= Date: Mon, 31 Oct 2016 23:16:20 +0100 Message-ID: Subject: Re: [PATCH] ARM: BCM5301X: Add back handler ignoring external imprecise aborts To: Scott Branden Cc: Florian Fainelli , Hauke Mehrtens , "linux-arm-kernel@lists.infradead.org" , Arnd Bergmann , Bjorn Helgaas , Lucas Stach , Jon Mason , Mark Rutland , Ray Jui , Scott Branden , bcm-kernel-feedback-list , "devicetree@vger.kernel.org" , Linux PCI , =?UTF-8?B?UmFmYcWCIE1pxYJlY2tp?= Content-Type: text/plain; charset=UTF-8 Sender: linux-pci-owner@vger.kernel.org List-ID: On 31 October 2016 at 23:05, Scott Branden wro= te: > On 16-10-31 02:56 PM, Florian Fainelli wrote: >> - fixups the aborts in the kernel, look where they come from, by using >> some bit of magic, looking at PCIe registers and whatnot (provided that >> is even possible), not too hard, but can take a while, and is error pron= e > > Option 4 sounds like the proper solution - check the range the abort is d= ue > to the PCI device enumeration and only ignore those aborts. This was already suggested by Arnd: http://lists.infradead.org/pipermail/linux-arm-kernel/2016-April/422873.htm= l and Ray was checking some internal datasheets, but I'm afraid he never found info on checking if error was caused by PCIe controller. Maybe we could think about some BCM5301X API (extra symbol exported by arch code) for starting ignoring aborts and stopping that. We could ignore aborts during scanning only but honestly it sounds like a bit hacky solution to me. >> I can certainly advocate that option 3 is the one that gives a working >> device, and this is what matters from a distribution perspective like >> LEDE/OpenWrt. >> > Since I don't use BCM5301x option 3 is fine by me. So for it seems like the best solution to me, but I'm open for changes if someone points another that is clean & better one. --=20 Rafa=C5=82 From mboxrd@z Thu Jan 1 00:00:00 1970 From: zajec5@gmail.com (=?UTF-8?B?UmFmYcWCIE1pxYJlY2tp?=) Date: Mon, 31 Oct 2016 23:16:20 +0100 Subject: [PATCH] ARM: BCM5301X: Add back handler ignoring external imprecise aborts In-Reply-To: <3847db9f-a8e1-b7fe-c2fa-19ea893bae5f@broadcom.com> References: <20161029111229.26875-1-zajec5@gmail.com> <6e252703-0010-83ae-fa14-57352add007a@hauke-m.de> <18db27c7-1055-ff3c-0b0b-eeeaf5f3f5e0@gmail.com> <3847db9f-a8e1-b7fe-c2fa-19ea893bae5f@broadcom.com> Message-ID: To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On 31 October 2016 at 23:05, Scott Branden wrote: > On 16-10-31 02:56 PM, Florian Fainelli wrote: >> - fixups the aborts in the kernel, look where they come from, by using >> some bit of magic, looking at PCIe registers and whatnot (provided that >> is even possible), not too hard, but can take a while, and is error prone > > Option 4 sounds like the proper solution - check the range the abort is due > to the PCI device enumeration and only ignore those aborts. This was already suggested by Arnd: http://lists.infradead.org/pipermail/linux-arm-kernel/2016-April/422873.html and Ray was checking some internal datasheets, but I'm afraid he never found info on checking if error was caused by PCIe controller. Maybe we could think about some BCM5301X API (extra symbol exported by arch code) for starting ignoring aborts and stopping that. We could ignore aborts during scanning only but honestly it sounds like a bit hacky solution to me. >> I can certainly advocate that option 3 is the one that gives a working >> device, and this is what matters from a distribution perspective like >> LEDE/OpenWrt. >> > Since I don't use BCM5301x option 3 is fine by me. So for it seems like the best solution to me, but I'm open for changes if someone points another that is clean & better one. -- Rafa?