Re: [PATCH v2] mmc: mxs-mmc: Introduce regulator support

[Ulf Hansson]

On Mon, 28 Jan 2019 at 11:20, Martin Kepplinger <martink@xxxxxxxxx> wrote:
>
> From: Martin Kepplinger <martin.kepplinger@xxxxxxxxxxxxx>
>
> This adds support for explicitly switching the mmc's power on and off
> which is needed for example for WL1837 wifi controllers. ip link set wlan0 down
> doesn't turn off the VMMC regulator which leads to hangs when loading firmware.

Isn't there a GPIO that needs to be toggled and separate clk for the
WiFi chip that needs to be enabled as well?

>
> Tested on i.MX28.
>
> Signed-off-by: Martin Kepplinger <martin.kepplinger@xxxxxxxxxxxxx>
> ---
>
> again, this isn't new. it's (rebased and simplified)
> https://patchwork.kernel.org/patch/4365751/
>
> Thanks Robin for your input!
>
> revision history
> ----------------
> v1: was just a question why this hasn't gone in earlier.
>
>
>  drivers/mmc/host/mxs-mmc.c | 31 +++++++++++++++++++++++--------
>  1 file changed, 23 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/mmc/host/mxs-mmc.c b/drivers/mmc/host/mxs-mmc.c
> index add1e70195ea..fdaca0fcec99 100644
> --- a/drivers/mmc/host/mxs-mmc.c
> +++ b/drivers/mmc/host/mxs-mmc.c
> @@ -66,11 +66,13 @@ struct mxs_mmc_host {
>         struct mmc_request              *mrq;
>         struct mmc_command              *cmd;
>         struct mmc_data                 *data;
> +       struct regulator                *vmmc;

You don't need this here as this is already part of the generic struct
mmc_host (via the struct mmc_supply).

>
>         unsigned char                   bus_width;
>         spinlock_t                      lock;
>         int                             sdio_irq_en;
>         bool                            broken_cd;
> +       unsigned char                   power_mode;

Ditto.

>  };
>
>  static int mxs_mmc_get_cd(struct mmc_host *mmc)
> @@ -517,6 +519,24 @@ static void mxs_mmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
>         else
>                 host->bus_width = 0;
>
> +       if (host->vmmc && ios->power_mode != host->power_mode) {
> +               switch (ios->power_mode) {
> +               case MMC_POWER_OFF:
> +                       if (regulator_disable(host->vmmc))
> +                               dev_err(mmc_dev(host->mmc),
> +                                       "Failed to disable vmmc regulator\n");

You should use mmc_regulator_set_ocr() instead.

> +                       break;
> +               case MMC_POWER_UP:
> +                       if (regulator_enable(host->vmmc))

Ditto.

> +                               dev_err(mmc_dev(host->mmc),
> +                                       "Failed to enable vmmc regulator\n");
> +                       break;
> +               default:
> +                       break;
> +               }
> +               host->power_mode = ios->power_mode;
> +       }
> +
>         if (ios->clock)
>                 mxs_ssp_set_clk_rate(&host->ssp, ios->clock);
>  }
> @@ -613,16 +633,11 @@ static int mxs_mmc_probe(struct platform_device *pdev)
>
>         host->mmc = mmc;
>         host->sdio_irq_en = 0;
> +       host->power_mode = MMC_POWER_OFF;

This isn't needed. The state is already controlled by the mmc core.

>
>         reg_vmmc = devm_regulator_get(&pdev->dev, "vmmc");

You should use mmc_regulator_get_supply() instead.

> -       if (!IS_ERR(reg_vmmc)) {
> -               ret = regulator_enable(reg_vmmc);
> -               if (ret) {
> -                       dev_err(&pdev->dev,
> -                               "Failed to enable vmmc regulator: %d\n", ret);
> -                       goto out_mmc_free;
> -               }
> -       }
> +       if (!IS_ERR(reg_vmmc))
> +               host->vmmc = reg_vmmc;
>
>         ssp->clk = devm_clk_get(&pdev->dev, NULL);
>         if (IS_ERR(ssp->clk)) {
> --
> 2.20.1
>

Kind regards
Uffe