From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753067AbcEMMTd (ORCPT ); Fri, 13 May 2016 08:19:33 -0400 Received: from eusmtp01.atmel.com ([212.144.249.242]:9283 "EHLO eusmtp01.atmel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751139AbcEMMTc (ORCPT ); Fri, 13 May 2016 08:19:32 -0400 Date: Fri, 13 May 2016 14:19:32 +0200 From: Ludovic Desroches To: Adrian Hunter CC: Ludovic Desroches , , , , , Kevin Liu , Jialing Fu , Jisheng Zhang Subject: Re: [RESEND PATCH] mmc: sdhci: fix wakeup configuration Message-ID: <20160513121932.GS27024@odux.rfo.atmel.com> Mail-Followup-To: Adrian Hunter , ulf.hansson@linaro.org, linux-mmc@vger.kernel.org, linux-kernel@vger.kernel.org, nicolas.ferre@atmel.com, Kevin Liu , Jialing Fu , Jisheng Zhang References: <1463131622-6380-1-git-send-email-ludovic.desroches@atmel.com> <5735BC71.4020500@intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Disposition: inline In-Reply-To: <5735BC71.4020500@intel.com> User-Agent: Mutt/1.6.0 (2016-04-01) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, May 13, 2016 at 02:37:21PM +0300, Adrian Hunter wrote: > + cc some Marvell people because they added this code > > On 13/05/16 12:27, Ludovic Desroches wrote: > > Activating wakeup event is not enough to get a wakeup signal. The > > corresponding events have to be enabled in the Interrupt Status Enable > > Register too. > > That seems to follow the specification. Did you find that you actually > needed this change to get it to work? > Yes, I din't wake up on card event without this patch. > > > > Signed-off-by: Ludovic Desroches > > --- > > Hi, > > > > I just updated sdhci_enable_irq_wakeups() not sdhci_disable_irq_wakeups() > > because I don't think it is necessary to configure SDHCI_INT_ENABLE at this > > step, it will be done with sdhci_init() or sdhci_enable_card_detection(). > > OK, but that should be in the commit message and commented in the code too. I'll add it. > > > > > While I was writing this patch, several questions came to my mind: > > - Is the naming correct? wakeup signal is not an irq. 'enable_irq_wakeups' can > > be a bit confusing. > > I don't support renaming things unless the names are really really bad. > These names are OK. > > > - If we want to wakeup from irq, we may have to set SDHCI_INT_ENABLE and > > SDHCI_SIGNAL_ENABLE and not rely on a previous configuration, isn't it? > > I imagine the card interrupt will occur when it is enabled. I don't know > about card insert / remove. > > What works for you? > I only use wakeup events and not irqs in system PM (patches for sdhci-of-at91 are not sent yet, I am waiting for the inclusion of other patches). What I mean is that if we want to wake up from irqs, it could be safer and clearer to set explicitely SDHCI_INT_ENABLE and SDHCI_SIGNAL_ENABLE in sdhci_enable_irq_wakeups(). It is just a thought and not needed for this patch. > > > > Regards > > > > Ludovic > > > > drivers/mmc/host/sdhci.c | 26 ++++++++++++++++++-------- > > 1 file changed, 18 insertions(+), 8 deletions(-) > > > > diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c > > index b284924..6fc69ed 100644 > > --- a/drivers/mmc/host/sdhci.c > > +++ b/drivers/mmc/host/sdhci.c > > @@ -2638,18 +2638,28 @@ static irqreturn_t sdhci_thread_irq(int irq, void *dev_id) > > \*****************************************************************************/ > > > > #ifdef CONFIG_PM > > +/* > > + * To enable wakeup events, the corresponding events have to be enabled in > > + * the Interrupt Status Enable register too. See 'Table 1-6: Wakeup Signal > > + * Table' in the SD Host Controller Standard Specification. > > + */ > > void sdhci_enable_irq_wakeups(struct sdhci_host *host) > > { > > - u8 val; > > - u8 mask = SDHCI_WAKE_ON_INSERT | SDHCI_WAKE_ON_REMOVE > > - | SDHCI_WAKE_ON_INT; > > + u8 wakeup_val; > > + u8 wakeup_mask = SDHCI_WAKE_ON_INSERT | SDHCI_WAKE_ON_REMOVE | > > + SDHCI_WAKE_ON_INT; > > Please don't rename variables, or tidy-up code that you are not changing. > I thought it will be better to have wakeup_val and irq_val instead of val and irq_val. That's why I renamed it while introducing irq_val. > > + u32 irq_val = SDHCI_INT_CARD_INSERT | SDHCI_INT_CARD_REMOVE | > > + SDHCI_INT_CARD_INT; > > > > - val = sdhci_readb(host, SDHCI_WAKE_UP_CONTROL); > > - val |= mask ; > > These 2 line are actually unchanged. > > > + wakeup_val = sdhci_readb(host, SDHCI_WAKE_UP_CONTROL); > > + wakeup_val |= wakeup_mask; > > /* Avoid fake wake up */ > > - if (host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION) > > - val &= ~(SDHCI_WAKE_ON_INSERT | SDHCI_WAKE_ON_REMOVE); > > - sdhci_writeb(host, val, SDHCI_WAKE_UP_CONTROL); > > + if (host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION) { > > + wakeup_val &= ~(SDHCI_WAKE_ON_INSERT | SDHCI_WAKE_ON_REMOVE); > > + irq_val &= ~(SDHCI_INT_CARD_INSERT | SDHCI_INT_CARD_REMOVE); > > + } > > + sdhci_writeb(host, wakeup_val, SDHCI_WAKE_UP_CONTROL); > > + sdhci_writel(host, irq_val, SDHCI_INT_ENABLE); > > } > > EXPORT_SYMBOL_GPL(sdhci_enable_irq_wakeups); > > > > >