On Thu, 8 Oct 2020 at 12:58, Adrian Hunter <adrian.hunter@xxxxxxxxx> wrote:
>
> On 8/10/20 12:27 pm, Ulf Hansson wrote:
> > On Thu, 8 Oct 2020 at 04:12, <muhammad.husaini.zulkifli@xxxxxxxxx> wrote:
> >>
> >> From: Muhammad Husaini Zulkifli <muhammad.husaini.zulkifli@xxxxxxxxx>
> >>
> >> Voltage switching sequence is needed to support UHS-1 interface.
> >> There are 2 places to control the voltage.
> >> 1) By setting the AON register using firmware driver calling
> >> system-level platform management layer (SMC) to set the register.
> >> 2) By controlling the GPIO expander value to drive either 1.8V or 3.3V
> >> for power mux input.
> >>
> >> Signed-off-by: Muhammad Husaini Zulkifli <muhammad.husaini.zulkifli@xxxxxxxxx>
> >> Reviewed-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxx>
> >> Reviewed-by: Adrian Hunter <adrian.hunter@xxxxxxxxx>
> >> ---
> >> drivers/mmc/host/sdhci-of-arasan.c | 126 +++++++++++++++++++++++++++++
> >> 1 file changed, 126 insertions(+)
> >>
> >> diff --git a/drivers/mmc/host/sdhci-of-arasan.c b/drivers/mmc/host/sdhci-of-arasan.c
> >> index 46aea6516133..ea2467b0073d 100644
> >> --- a/drivers/mmc/host/sdhci-of-arasan.c
> >> +++ b/drivers/mmc/host/sdhci-of-arasan.c
> >> @@ -16,6 +16,7 @@
> >> */
> >>
> >> #include <linux/clk-provider.h>
> >> +#include <linux/gpio/consumer.h>
> >> #include <linux/mfd/syscon.h>
> >> #include <linux/module.h>
> >> #include <linux/of_device.h>
> >> @@ -23,6 +24,7 @@
> >> #include <linux/regmap.h>
> >> #include <linux/of.h>
> >> #include <linux/firmware/xlnx-zynqmp.h>
> >> +#include <linux/firmware/intel/keembay_firmware.h>
> >>
> >> #include "cqhci.h"
> >> #include "sdhci-pltfm.h"
> >> @@ -136,6 +138,7 @@ struct sdhci_arasan_clk_data {
> >> * @soc_ctl_base: Pointer to regmap for syscon for soc_ctl registers.
> >> * @soc_ctl_map: Map to get offsets into soc_ctl registers.
> >> * @quirks: Arasan deviations from spec.
> >> + * @uhs_gpio: Pointer to the uhs gpio.
> >> */
> >> struct sdhci_arasan_data {
> >> struct sdhci_host *host;
> >> @@ -150,6 +153,7 @@ struct sdhci_arasan_data {
> >> struct regmap *soc_ctl_base;
> >> const struct sdhci_arasan_soc_ctl_map *soc_ctl_map;
> >> unsigned int quirks;
> >> + struct gpio_desc *uhs_gpio;
> >>
> >> /* Controller does not have CD wired and will not function normally without */
> >> #define SDHCI_ARASAN_QUIRK_FORCE_CDTEST BIT(0)
> >> @@ -361,6 +365,112 @@ static int sdhci_arasan_voltage_switch(struct mmc_host *mmc,
> >> return -EINVAL;
> >> }
> >>
> >> +static int sdhci_arasan_keembay_voltage_switch(struct mmc_host *mmc,
> >> + struct mmc_ios *ios)
> >> +{
> >> + struct sdhci_host *host = mmc_priv(mmc);
> >> + struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
> >> + struct sdhci_arasan_data *sdhci_arasan = sdhci_pltfm_priv(pltfm_host);
> >> + u16 ctrl_2, clk;
> >> + int ret;
> >> +
> >> + switch (ios->signal_voltage) {
> >> + case MMC_SIGNAL_VOLTAGE_180:
> >> + clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
> >> + clk &= ~SDHCI_CLOCK_CARD_EN;
> >> + sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
> >> +
> >> + clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
> >> + if (clk & SDHCI_CLOCK_CARD_EN)
> >> + return -EAGAIN;
> >> +
> >> + sdhci_writeb(host, SDHCI_POWER_ON | SDHCI_POWER_180,
> >> + SDHCI_POWER_CONTROL);
> >> +
> >> + /*
> >> + * Set VDDIO_B voltage to Low for 1.8V
> >> + * which is controlling by GPIO Expander.
> >> + */
> >> + gpiod_set_value_cansleep(sdhci_arasan->uhs_gpio, 0);
> >> +
> >> + /*
> >> + * This is like a final gatekeeper. Need to ensure changed voltage
> >> + * is settled before and after turn on this bit.
> >> + */
> >> + usleep_range(1000, 1100);
> >> +
> >> + ret = keembay_sd_voltage_selection(KEEMBAY_SET_1V8_VOLT);
> >> + if (ret)
> >> + return ret;
> >> +
> >> + usleep_range(1000, 1100);
> >
> > No, sorry, but I don't like this.
> >
> > This looks like a GPIO regulator with an extension of using the
> > keembay_sd_voltage_selection() thingy. I think you can model these
> > things behind a regulator and hook it up as a vqmmc supply in DT
> > instead. BTW, this is the common way we deal with these things for mmc
> > host drivers.
>
> It seemed to me that would just result in calling regulator API instead of
> GPIO API but the flow above would otherwise be unchanged i.e. no benefit
>
To me, the benefit is about avoiding platform specific code in drivers
- but also about consistency. For I/O signal voltage, the common
method here, is to model this as a GPIO regulator. This means we can
use these available helpers from the core:
mmc_regulator_set_vqmmc()
mmc_regulator_get_supply()
Kind regards
Uffe
