From mboxrd@z Thu Jan 1 00:00:00 1970 From: andrew@lunn.ch (Andrew Lunn) Date: Wed, 23 Nov 2016 02:39:40 +0100 Subject: [PATCH RFC] ARM: dts: add support for Turris Omnia In-Reply-To: <1479860851.10840.11@smtp.gmail.com> References: <20161105203841.9661-1-uwe@kleine-koenig.org> <1479126185.15557.5@smtp.gmail.com> <20161114201640.rr32iyjf5a53v33t@perseus.defre.kleine-koenig.org> <20161114202832.GG24546@lunn.ch> <1479586147.10840.0@smtp.gmail.com> <20161120203037.pd5mhqyjeotileve@perseus.defre.kleine-koenig.org> <1479851991.26813.2@smtp.gmail.com> <1479860851.10840.11@smtp.gmail.com> Message-ID: <20161123013940.GM2691@lunn.ch> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wed, Nov 23, 2016 at 01:27:31AM +0100, tomas.hlavacek at nic.cz wrote: > Hi Uwe! > > On Tue, Nov 22, 2016 at 10:59 PM, tomas.hlavacek at nic.cz wrote: > >Anyway I took your patch and tried few things: > >- add pca9538 interrupt-controller > >- add IRQ for 88E1514 PHY - and there is a problem: > ... > > I thought it over and if I am not mistaken this is not going to work > anyway, because pca9538 driver causes the GPIO driver to set > IRQ_NESTED_THREAD, so we can not simply use one of the GPIO expander > pins as IRQ source for 88E1514, because request_irq() on it will > fail ultimately. Actually, the phylib now does uses threaded IRQs, since i implemented interrupt support for the mv88e6xxx driver. Its interrupts require MDIO transactions, so have to be threaded. However, i don't think using interrupts on the pca9538 are reliable. Interrupt support is compile time optional for that driver. It is not clear to me if distributions do compile the driver with interrupts enabled. So it could be the probe fails with OpenWRT, Debian, etc... Andrew From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Lunn Subject: Re: [PATCH RFC] ARM: dts: add support for Turris Omnia Date: Wed, 23 Nov 2016 02:39:40 +0100 Message-ID: <20161123013940.GM2691@lunn.ch> References: <20161105203841.9661-1-uwe@kleine-koenig.org> <1479126185.15557.5@smtp.gmail.com> <20161114201640.rr32iyjf5a53v33t@perseus.defre.kleine-koenig.org> <20161114202832.GG24546@lunn.ch> <1479586147.10840.0@smtp.gmail.com> <20161120203037.pd5mhqyjeotileve@perseus.defre.kleine-koenig.org> <1479851991.26813.2@smtp.gmail.com> <1479860851.10840.11@smtp.gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <1479860851.10840.11-TAvD023jEQEN+BqQ9rBEUg@public.gmane.org> Sender: devicetree-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: tomas.hlavacek-x+rMaJPWets@public.gmane.org Cc: Uwe =?iso-8859-1?Q?Kleine-K=F6nig?= , Mark Rutland , marex-ynQEQJNshbs@public.gmane.org, Jason Cooper , devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Rob Herring , Gregory Clement , linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org, Sebastian Hesselbarth List-Id: devicetree@vger.kernel.org On Wed, Nov 23, 2016 at 01:27:31AM +0100, tomas.hlavacek-x+rMaJPWets@public.gmane.org wrote: > Hi Uwe! > > On Tue, Nov 22, 2016 at 10:59 PM, tomas.hlavacek-x+rMaJPWets@public.gmane.org wrote: > >Anyway I took your patch and tried few things: > >- add pca9538 interrupt-controller > >- add IRQ for 88E1514 PHY - and there is a problem: > ... > > I thought it over and if I am not mistaken this is not going to work > anyway, because pca9538 driver causes the GPIO driver to set > IRQ_NESTED_THREAD, so we can not simply use one of the GPIO expander > pins as IRQ source for 88E1514, because request_irq() on it will > fail ultimately. Actually, the phylib now does uses threaded IRQs, since i implemented interrupt support for the mv88e6xxx driver. Its interrupts require MDIO transactions, so have to be threaded. However, i don't think using interrupts on the pca9538 are reliable. Interrupt support is compile time optional for that driver. It is not clear to me if distributions do compile the driver with interrupts enabled. So it could be the probe fails with OpenWRT, Debian, etc... Andrew -- 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