Re: [PATCH] mmc: mmc_spi: Add Card Detect comments and fix CD GPIO case

From: Magnus Damm
Date: Fri Feb 26 2016 - 03:50:25 EST


On Tue, Feb 16, 2016 at 1:06 PM, Magnus Damm <magnus.damm@xxxxxxxxx> wrote:
> From: Magnus Damm <damm+renesas@xxxxxxxxxxxxx>
>
> This patch adds comments to describe the various ways card detection
> may be implemented and when polling is needed. Also a fix is included
> to prevent the MMC SPI driver from polling when a CD GPIO that supports
> interrupts is specified using the gpios DT property (case 1).
>
> Without this patch the DT node below results in the following output:
>
> spi_gpio: spi-gpio { /* SD2 @ CN12 */
> compatible = "spi-gpio";
> #address-cells = <1>;
> #size-cells = <0>;
> gpio-sck = <&gpio6 16 GPIO_ACTIVE_HIGH>;
> gpio-mosi = <&gpio6 17 GPIO_ACTIVE_HIGH>;
> gpio-miso = <&gpio6 18 GPIO_ACTIVE_HIGH>;
> num-chipselects = <1>;
> cs-gpios = <&gpio6 21 GPIO_ACTIVE_LOW>;
> status = "okay";
>
> spi@0 {
> compatible = "mmc-spi-slot";
> reg = <0>;
> voltage-ranges = <3200 3400>;
> spi-max-frequency = <25000000>;
> gpios = <&gpio6 22 GPIO_ACTIVE_LOW>; /* CD */
> };
> };
>
> # dmesg | grep mmc
> mmc_spi spi32766.0: SD/MMC host mmc0, no WP, no poweroff, cd polling
> mmc0: host does not support reading read-only switch, assuming write-enable
> mmc0: new SDHC card on SPI
> mmcblk0: mmc0:0000 SU04G 3.69 GiB
> mmcblk0: p1
>
> With this patch applied the "cd polling" portion above disappears.
>
> Signed-off-by: Magnus Damm <damm+renesas@xxxxxxxxxxxxx>
> ---
>
> Developed on top of renesas-drivers-2016-02-09-v4.5-rc3 which
> simply put is v4.5-rc3 + selected linux-next trees.
>
> drivers/mmc/host/mmc_spi.c | 14 ++++++++++++++
> 1 file changed, 14 insertions(+)
>
> --- 0001/drivers/mmc/host/mmc_spi.c
> +++ work/drivers/mmc/host/mmc_spi.c 2016-02-16 12:47:47.900513000 +0900
> @@ -1437,11 +1437,25 @@ static int mmc_spi_probe(struct spi_devi
> if (status != 0)
> goto fail_add_host;
>
> + /* Card detection in case of DT is supported in several ways:
> + * 1) GPIO with interrupt support using "gpios" DT property.
> + * 2) GPIO without interrupt support using "gpios" DT property.
> + * 3) Interrupt only using "interrupt-parent/interrupts" DT properties.
> + * 4) Both "gpios" and "interrupt-parent/interrupts" DT properties.
> + * 5) Nothing specified.
> + * For case 2 and 5 above polling is required.
> + */
> if (host->pdata && host->pdata->flags & MMC_SPI_USE_CD_GPIO) {
> status = mmc_gpio_request_cd(mmc, host->pdata->cd_gpio,
> host->pdata->cd_debounce);
> if (status != 0)
> goto fail_add_host;
> +
> + /* The platform has a CD GPIO signal that may support
> + * interrupts, so let mmc_gpiod_request_cd_irq() decide
> + * if polling is needed or not.
> + */
> + mmc->caps &= ~MMC_CAP_NEEDS_POLL;
> mmc_gpiod_request_cd_irq(mmc);
> }
>

Did anyone run into the same issue that CD polling is enabled even
though the GPIO supports interrupt?

Perhaps it would be better to break this out into two patches - one
for the fix and one for the documentation bits?

Thanks,

/ magnus