From mboxrd@z Thu Jan 1 00:00:00 1970 From: Lowell Gilbert References: <44eg2nuhee.fsf@be-well.ilk.org> <44a8dbueie.fsf@be-well.ilk.org> Date: Mon, 14 Nov 2016 16:56:30 -0500 In-Reply-To: (Philippe Gerum's message of "Mon, 14 Nov 2016 10:09:41 +0100") Message-ID: <44inrphgj5.fsf@lowell-desk.lan> MIME-Version: 1.0 Content-Type: text/plain Subject: Re: [Xenomai] irqchip missing? List-Id: Discussions about the Xenomai project List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: xenomai@xenomai.org Philippe Gerum writes: > On 11/07/2016 11:11 PM, Lowell Gilbert wrote: >> Lowell Gilbert writes: >> >>> I have a system based on a Cortex A9. When I was running under Xenomai >>> 2.6, I was able to grab interrupts in RTDM. A few months ago, I updated >>> from a 3.16 Linux kernel and 2.6.4 Xenomai to a 4.1 kernel with 3.0.2 >>> Xenomai. I'm trying to grab the same IRQ now, and the irqchip doesn't >>> seem to have a set_affinity callback, so I get a kernel dump. >> >> ipipe_set_irq_affinity() is being called with a GPIO chip, instead of >> the GIC that I expect. Why would that be? >> > > ipipe_set_irq_affinity() receives an interrupt number, not an IC > descriptor. Are you referring to the irq_set_affinity() handler from the > irq_chip descriptor instead? Please be specific about the code you refer > to (call site). In ipipe_set_irq_affinity(), I checked irq_get_chip(irq)->name. That gave me "gpio-dwapb" rather than "GIC". There is no irq_set_affinity driver on that chip, and so it crashes. I haven't figured out how the chip table gets set up. Be well.