| Message ID | 1457605858-30907-1-git-send-email-ludovic.desroches@atmel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Thu, Mar 10, 2016 at 11:30:58AM +0100, Ludovic Desroches wrote: > It was impossible to wake-up on card detect event because when sdhci > controller is runtime suspend, it is assumed that all the clocks are > disabled so we can't get irqs. > If the device is removable and there is no gpio to manage the card > detection then card detection polling is used. > > Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com> > --- > > Hi Ulf, Adrian, > > Following the discussion, I need to fix my issue. I think we could both agree > on this patch that I see more as temporary workaround. > > I will try to change the muxing of the card detect pio in order to no more > use the sdhci controller to manage it but a gpio. If it do not work then > I may send another patch in the spirit of the previous one (keeping one clock > enabled and not calling sdhci_runtime_suspend_host()). I have changed the muxing for the card detect signal to get it on a gpio instead of the card detect pio (peripheral io) of the controller. I thought it can work because we have a 'force card detect' bit. Unfortunately, it is resetted when we perform a software reset. Moreover I need to keep the card detect pio, because it used to perform fast startup from ultra low power modes. Maybe it could work by using an extra pin but it invovles to do hardawre changes on the boards and needing two pins for one signal is not acceptable. I'll do more tests to use runtime pm without polling for card detection and without using sdhci_runtime_suspend_host(). > > Regards > > drivers/mmc/host/sdhci-of-at91.c | 16 ++++++++++++++++ > 1 file changed, 16 insertions(+) > > diff --git a/drivers/mmc/host/sdhci-of-at91.c b/drivers/mmc/host/sdhci-of-at91.c > index 35c02fc..dac8508 100644 > --- a/drivers/mmc/host/sdhci-of-at91.c > +++ b/drivers/mmc/host/sdhci-of-at91.c > @@ -18,6 +18,7 @@ > #include <linux/err.h> > #include <linux/io.h> > #include <linux/mmc/host.h> > +#include <linux/mmc/slot-gpio.h> > #include <linux/module.h> > #include <linux/of.h> > #include <linux/of_device.h> > @@ -204,6 +205,21 @@ static int sdhci_at91_probe(struct platform_device *pdev) > if (ret) > goto pm_runtime_disable; > > + /* > + * When calling sdhci_runtime_suspend_host(), the sdhci layer makes > + * the assumption that all the clocks of the controller are disabled. > + * It means we can't get irq from it when it is runtime suspended. > + * For that reason, it is not planned to wake-up on a card detect irq > + * from the controller. > + * If we want to use runtime PM and to be able to wake-up on card > + * insertion, we have to use a GPIO for the card detection or we can > + * use polling for the card detection. Be aware that using polling > + * will resume/suspend the controller between each attempt. > + */ > + if (!(host->mmc->caps & MMC_CAP_NONREMOVABLE) && > + IS_ERR_VALUE(mmc_gpio_get_cd(host->mmc))) > + host->mmc->caps |= MMC_CAP_NEEDS_POLL; > + > pm_runtime_put_autosuspend(&pdev->dev); > > return 0; > -- > 2.5.0 > -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 10 March 2016 at 11:30, Ludovic Desroches <ludovic.desroches@atmel.com> wrote: > It was impossible to wake-up on card detect event because when sdhci > controller is runtime suspend, it is assumed that all the clocks are > disabled so we can't get irqs. > If the device is removable and there is no gpio to manage the card > detection then card detection polling is used. > > Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com> > --- > > Hi Ulf, Adrian, > > Following the discussion, I need to fix my issue. I think we could both agree > on this patch that I see more as temporary workaround. > > I will try to change the muxing of the card detect pio in order to no more > use the sdhci controller to manage it but a gpio. If it do not work then > I may send another patch in the spirit of the previous one (keeping one clock > enabled and not calling sdhci_runtime_suspend_host()). > > Regards > > drivers/mmc/host/sdhci-of-at91.c | 16 ++++++++++++++++ > 1 file changed, 16 insertions(+) > > diff --git a/drivers/mmc/host/sdhci-of-at91.c b/drivers/mmc/host/sdhci-of-at91.c > index 35c02fc..dac8508 100644 > --- a/drivers/mmc/host/sdhci-of-at91.c > +++ b/drivers/mmc/host/sdhci-of-at91.c > @@ -18,6 +18,7 @@ > #include <linux/err.h> > #include <linux/io.h> > #include <linux/mmc/host.h> > +#include <linux/mmc/slot-gpio.h> > #include <linux/module.h> > #include <linux/of.h> > #include <linux/of_device.h> > @@ -204,6 +205,21 @@ static int sdhci_at91_probe(struct platform_device *pdev) > if (ret) > goto pm_runtime_disable; > > + /* > + * When calling sdhci_runtime_suspend_host(), the sdhci layer makes > + * the assumption that all the clocks of the controller are disabled. > + * It means we can't get irq from it when it is runtime suspended. > + * For that reason, it is not planned to wake-up on a card detect irq > + * from the controller. > + * If we want to use runtime PM and to be able to wake-up on card > + * insertion, we have to use a GPIO for the card detection or we can > + * use polling for the card detection. Be aware that using polling > + * will resume/suspend the controller between each attempt. > + */ > + if (!(host->mmc->caps & MMC_CAP_NONREMOVABLE) && > + IS_ERR_VALUE(mmc_gpio_get_cd(host->mmc))) > + host->mmc->caps |= MMC_CAP_NEEDS_POLL; Perhaps also clearing SDHCI_QUIRK_BROKEN_CARD_DETECTION, as this doesn't play well if this is set via parsing the DTS. > + > pm_runtime_put_autosuspend(&pdev->dev); > > return 0; > -- > 2.5.0 > Otherwise this looks okay to me. Kind regards Uffe -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/mmc/host/sdhci-of-at91.c b/drivers/mmc/host/sdhci-of-at91.c index 35c02fc..dac8508 100644 --- a/drivers/mmc/host/sdhci-of-at91.c +++ b/drivers/mmc/host/sdhci-of-at91.c @@ -18,6 +18,7 @@ #include <linux/err.h> #include <linux/io.h> #include <linux/mmc/host.h> +#include <linux/mmc/slot-gpio.h> #include <linux/module.h> #include <linux/of.h> #include <linux/of_device.h> @@ -204,6 +205,21 @@ static int sdhci_at91_probe(struct platform_device *pdev) if (ret) goto pm_runtime_disable; + /* + * When calling sdhci_runtime_suspend_host(), the sdhci layer makes + * the assumption that all the clocks of the controller are disabled. + * It means we can't get irq from it when it is runtime suspended. + * For that reason, it is not planned to wake-up on a card detect irq + * from the controller. + * If we want to use runtime PM and to be able to wake-up on card + * insertion, we have to use a GPIO for the card detection or we can + * use polling for the card detection. Be aware that using polling + * will resume/suspend the controller between each attempt. + */ + if (!(host->mmc->caps & MMC_CAP_NONREMOVABLE) && + IS_ERR_VALUE(mmc_gpio_get_cd(host->mmc))) + host->mmc->caps |= MMC_CAP_NEEDS_POLL; + pm_runtime_put_autosuspend(&pdev->dev); return 0;
It was impossible to wake-up on card detect event because when sdhci controller is runtime suspend, it is assumed that all the clocks are disabled so we can't get irqs. If the device is removable and there is no gpio to manage the card detection then card detection polling is used. Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com> --- Hi Ulf, Adrian, Following the discussion, I need to fix my issue. I think we could both agree on this patch that I see more as temporary workaround. I will try to change the muxing of the card detect pio in order to no more use the sdhci controller to manage it but a gpio. If it do not work then I may send another patch in the spirit of the previous one (keeping one clock enabled and not calling sdhci_runtime_suspend_host()). Regards drivers/mmc/host/sdhci-of-at91.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+)