From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752020AbdLFUWz (ORCPT ); Wed, 6 Dec 2017 15:22:55 -0500 Received: from lelnx194.ext.ti.com ([198.47.27.80]:28630 "EHLO lelnx194.ext.ti.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751500AbdLFUWx (ORCPT ); Wed, 6 Dec 2017 15:22:53 -0500 Subject: Re: [PATCH v2 16/19] ASoC: tlv320aic31xx: Add short circuit detection support From: "Andrew F. Davis" To: Mark Brown CC: Liam Girdwood , Rob Herring , Mark Rutland , =?UTF-8?Q?Beno=c3=aet_Cousson?= , Tony Lindgren , , , References: <20171129213300.20021-1-afd@ti.com> <20171129213300.20021-17-afd@ti.com> <20171201133930.xzfjjj6wzkbtk6c4@sirena.org.uk> <4d956cbc-829c-0d81-2980-aa2a79f36660@ti.com> Message-ID: Date: Wed, 6 Dec 2017 14:22:39 -0600 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.4.0 MIME-Version: 1.0 In-Reply-To: <4d956cbc-829c-0d81-2980-aa2a79f36660@ti.com> Content-Type: text/plain; charset="windows-1252" Content-Language: en-US Content-Transfer-Encoding: 7bit X-EXCLAIMER-MD-CONFIG: e1e8a2fd-e40a-4ac6-ac9b-f7e9cc9ee180 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 12/01/2017 09:32 AM, Andrew F. Davis wrote: > On 12/01/2017 07:39 AM, Mark Brown wrote: >> On Wed, Nov 29, 2017 at 03:32:57PM -0600, Andrew F. Davis wrote: >> >>> +static irqreturn_t aic31xx_irq(int irq, void *data) >>> +{ >>> + struct aic31xx_priv *aic31xx = (struct aic31xx_priv *)data; >> >> There is no need to cast away from void, doing so can only mask bugs. >> >>> + ret = regmap_read(aic31xx->regmap, AIC31XX_INTRDACFLAG, &value); >>> + if (ret) { >>> + dev_err(dev, "Failed to read interrupt mask: %d\n", ret); >>> + return IRQ_NONE; >>> + } >>> + >>> + if (value & AIC31XX_HPLSCDETECT) >>> + dev_err(dev, "Short circuit on Left output is detected\n"); >>> + if (value & AIC31XX_HPRSCDETECT) >>> + dev_err(dev, "Short circuit on Right output is detected\n"); >>> + >>> + return IRQ_HANDLED; >> >> This will report the interrupt as handled even if we didn't see an >> interrupt we understand which will break shared interrupt lines. At the >> very least we should log other interrupt sources numerically. >> > > Okay, I think I can make that work by checking if no bits are set in the > interrupt regs and returning early if not, IRQ_NONE. > This turned out to be more difficult than I expected, plus if I do handle an interrupt it doesn't mean the other device did not right? So this wouldn't fix shared lines as far as I can tell, but I don't register it as shared so this should be fine. As for your other suggestion of "log other interrupt sources numerically", could you explain this or point to an example of what you mean? >>> + if (aic31xx->irq > 0) { >>> + regmap_update_bits(aic31xx->regmap, AIC31XX_GPIO1, >>> + AIC31XX_GPIO1_FUNC_MASK, >>> + AIC31XX_GPIO1_INT1 << >>> + AIC31XX_GPIO1_FUNC_SHIFT); >> >> Is the interrupt only available on GPIO1? >> > > Some devices can route this to GPIO2 IIRC. > > I'm not sure how that would be supported, I think we would need to add > interrupt names to DT so users could specify which gpio they wired their > IRQ lines to. > > interrupt = <&host 23>; > interrupt-name = "gpio2"; > > or similar? >