From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753440AbaIRHsw (ORCPT ); Thu, 18 Sep 2014 03:48:52 -0400 Received: from gw-1.arm.linux.org.uk ([78.32.30.217]:32908 "EHLO pandora.arm.linux.org.uk" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1750983AbaIRHsu (ORCPT ); Thu, 18 Sep 2014 03:48:50 -0400 Date: Thu, 18 Sep 2014 08:48:39 +0100 From: Russell King - ARM Linux To: Daniel Thompson Cc: linaro-kernel@lists.linaro.org, Jason Cooper , patches@linaro.org, Marc Zyngier , Will Deacon , linux-kernel@vger.kernel.org, John Stultz , Thomas Gleixner , Sumit Semwal , linux-arm-kernel@lists.infradead.org Subject: Re: [PATCH 3.17-rc4 v7 4/6] irqchip: gic: Add support for IPI FIQ Message-ID: <20140918074839.GZ12361@n2100.arm.linux.org.uk> References: <1410970218-28847-1-git-send-email-daniel.thompson@linaro.org> <1410970218-28847-5-git-send-email-daniel.thompson@linaro.org> <20140917185138.GX12361@n2100.arm.linux.org.uk> <5419EB27.2040405@linaro.org> <20140917210712.GY12361@n2100.arm.linux.org.uk> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20140917210712.GY12361@n2100.arm.linux.org.uk> User-Agent: Mutt/1.5.19 (2009-01-05) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Sep 17, 2014 at 10:07:13PM +0100, Russell King - ARM Linux wrote: > On Wed, Sep 17, 2014 at 01:12:23PM -0700, Daniel Thompson wrote: > > I may have missed something but this sounds like the expected behaviour > > to me. > > > > Without AckCtl set (GIC_CPU_CTRL bit 2) then it is not possible to > > acknowledge group 1 interrupts from secure mode. Thus in the > > circumstances described above I would expect the system to wedge in the > > interrupt handler because IRQ is raised but the software is not able to > > acknowledge it. > > Setting AckCtl (0x07) results in no change in behaviour - still crashes > when it tries to calibrate the timer. > > > I think there may also problems with leaving CBPR unset. Leaving CBPR > > unset certainly causes a change of behaviour compared to an all group0 > > setup. > > Also tried setting that (0x17), but still the same. Another data point: SDP4430 boots fine with 0x17. The Versatile Express has GIC ID/TYPER register values of: GICC IDR: 0x3901043b GICD TYPER: 0x0000fc62 The SDP4430 has: GICC IDR: 0x3901043b GICD TYPER: 0x0000fc24 so they should in theory be identical GICs, and should behave identically, but they don't. -- FTTC broadband for 0.8mile line: currently at 9.5Mbps down 400kbps up according to speedtest.net.