| Message ID | 1453197165-20194-1-git-send-email-shawn.lin@rock-chips.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Hi Shawn, Am Dienstag, 19. Januar 2016, 17:52:45 schrieb Shawn Lin: > We find rk3368 fail to enumerate sd card since > the default state is gpio function. Meahwhile, lots of > rockchip platform dts file assign pinctrl to dw_mmc but actually > dw_mmc never use it. Fortunately, those platforms' default io > state is sdmmc function, but no always right for all. > > Cc: Heiko Stuebner <heiko@sntech.de> > Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com> I do believe any issue you see, must come from somewhere else, because the driver-core does in drivers/base/dd.c really_probe(): pinctrl_bind_pins() which in turn already does the needed pinctrl state handling. So that default-state should be set up even before the dwmmc probe function is called. The sharing between sdmmc and uart2 pins is the same on the rk3036 and so far Caesar seems to have been able to switch the pin-state to sdmmc without touching the dw_mmc driver itself. Heiko > --- > > drivers/mmc/host/dw_mmc-rockchip.c | 15 +++++++++++++++ > 1 file changed, 15 insertions(+) > > diff --git a/drivers/mmc/host/dw_mmc-rockchip.c > b/drivers/mmc/host/dw_mmc-rockchip.c index d9c92f3..308b424 100644 > --- a/drivers/mmc/host/dw_mmc-rockchip.c > +++ b/drivers/mmc/host/dw_mmc-rockchip.c > @@ -9,6 +9,7 @@ > > #include <linux/module.h> > #include <linux/platform_device.h> > +#include <linux/pinctrl/consumer.h> > #include <linux/clk.h> > #include <linux/mmc/host.h> > #include <linux/mmc/dw_mmc.h> > @@ -21,6 +22,8 @@ > #define RK3288_CLKGEN_DIV 2 > > struct dw_mci_rockchip_priv_data { > + struct pinctrl *pinctrl; > + struct pinctrl_state *pins_default; > struct clk *drv_clk; > struct clk *sample_clk; > int default_sample_phase; > @@ -78,6 +81,10 @@ static void dw_mci_rk3288_set_ios(struct dw_mci *host, > struct mmc_ios *ios) /* Make sure we use phases which we can enumerate > with */ > if (!IS_ERR(priv->sample_clk)) > clk_set_phase(priv->sample_clk, priv->default_sample_phase); > + > + /* Make sure we use correct pinctrl which we can enumerate with */ > + if (!IS_ERR(priv->pinctrl) && !IS_ERR(priv->pins_default)) > + pinctrl_select_state(priv->pinctrl, priv->pins_default); > } > > #define NUM_PHASES 360 > @@ -223,6 +230,14 @@ static int dw_mci_rk3288_parse_dt(struct dw_mci > *host) if (IS_ERR(priv->sample_clk)) > dev_dbg(host->dev, "ciu_sample not available\n"); > > + priv->pinctrl = devm_pinctrl_get(host->dev); > + if (!IS_ERR(priv->pinctrl)) { > + priv->pins_default = pinctrl_lookup_state(priv->pinctrl, > + "default"); > + if (IS_ERR(priv->pins_default)) > + dev_dbg(host->dev, "pinctrl not available\n"); > + } > + > host->priv = priv; > > return 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 2016/1/19 18:07, Heiko Stuebner wrote: > Hi Shawn, > > Am Dienstag, 19. Januar 2016, 17:52:45 schrieb Shawn Lin: >> We find rk3368 fail to enumerate sd card since >> the default state is gpio function. Meahwhile, lots of >> rockchip platform dts file assign pinctrl to dw_mmc but actually >> dw_mmc never use it. Fortunately, those platforms' default io >> state is sdmmc function, but no always right for all. >> >> Cc: Heiko Stuebner <heiko@sntech.de> >> Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com> > > I do believe any issue you see, must come from somewhere else, because the > driver-core does in drivers/base/dd.c really_probe(): > pinctrl_bind_pins() yes, pinctrl_bind_pins does turn pinctrl state. But I try debug a little more and seems somewhere else switch the state before dw_mmc probe. Now I *guess* that may relate to our ATF or MCU firmware since rk3368's sdmmc io can also be used as mcu_jtag for system debug. I need to double check that point, thanks. Hi Caesar, does your patch able to consider adding SDMMC/CPU_JTAG/MCU_JTAG managment for rk3368 as well? > which in turn already does the needed pinctrl state handling. > > So that default-state should be set up even before the dwmmc probe function > is called. > > The sharing between sdmmc and uart2 pins is the same on the rk3036 and so > far Caesar seems to have been able to switch the pin-state to sdmmc without > touching the dw_mmc driver itself. > > > Heiko > > >> --- >> >> drivers/mmc/host/dw_mmc-rockchip.c | 15 +++++++++++++++ >> 1 file changed, 15 insertions(+) >> >> diff --git a/drivers/mmc/host/dw_mmc-rockchip.c >> b/drivers/mmc/host/dw_mmc-rockchip.c index d9c92f3..308b424 100644 >> --- a/drivers/mmc/host/dw_mmc-rockchip.c >> +++ b/drivers/mmc/host/dw_mmc-rockchip.c >> @@ -9,6 +9,7 @@ >> >> #include <linux/module.h> >> #include <linux/platform_device.h> >> +#include <linux/pinctrl/consumer.h> >> #include <linux/clk.h> >> #include <linux/mmc/host.h> >> #include <linux/mmc/dw_mmc.h> >> @@ -21,6 +22,8 @@ >> #define RK3288_CLKGEN_DIV 2 >> >> struct dw_mci_rockchip_priv_data { >> + struct pinctrl *pinctrl; >> + struct pinctrl_state *pins_default; >> struct clk *drv_clk; >> struct clk *sample_clk; >> int default_sample_phase; >> @@ -78,6 +81,10 @@ static void dw_mci_rk3288_set_ios(struct dw_mci *host, >> struct mmc_ios *ios) /* Make sure we use phases which we can enumerate >> with */ >> if (!IS_ERR(priv->sample_clk)) >> clk_set_phase(priv->sample_clk, priv->default_sample_phase); >> + >> + /* Make sure we use correct pinctrl which we can enumerate with */ >> + if (!IS_ERR(priv->pinctrl) && !IS_ERR(priv->pins_default)) >> + pinctrl_select_state(priv->pinctrl, priv->pins_default); >> } >> >> #define NUM_PHASES 360 >> @@ -223,6 +230,14 @@ static int dw_mci_rk3288_parse_dt(struct dw_mci >> *host) if (IS_ERR(priv->sample_clk)) >> dev_dbg(host->dev, "ciu_sample not available\n"); >> >> + priv->pinctrl = devm_pinctrl_get(host->dev); >> + if (!IS_ERR(priv->pinctrl)) { >> + priv->pins_default = pinctrl_lookup_state(priv->pinctrl, >> + "default"); >> + if (IS_ERR(priv->pins_default)) >> + dev_dbg(host->dev, "pinctrl not available\n"); >> + } >> + >> host->priv = priv; >> >> return 0; > > > >
? 2016?01?19? 19:07, Shawn Lin ??: > On 2016/1/19 18:07, Heiko Stuebner wrote: >> Hi Shawn, >> >> Am Dienstag, 19. Januar 2016, 17:52:45 schrieb Shawn Lin: >>> We find rk3368 fail to enumerate sd card since >>> the default state is gpio function. Meahwhile, lots of >>> rockchip platform dts file assign pinctrl to dw_mmc but actually >>> dw_mmc never use it. Fortunately, those platforms' default io >>> state is sdmmc function, but no always right for all. >>> >>> Cc: Heiko Stuebner <heiko@sntech.de> >>> Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com> >> >> I do believe any issue you see, must come from somewhere else, >> because the >> driver-core does in drivers/base/dd.c really_probe(): >> pinctrl_bind_pins() > > yes, pinctrl_bind_pins does turn pinctrl state. But I try debug a little > more and seems somewhere else switch the state before dw_mmc probe. > Now I *guess* that may relate to our ATF or MCU firmware since > rk3368's sdmmc io can also be used as mcu_jtag for system debug. I > need to double check that point, thanks. > Yep, maybe that's related to the firmware setting. I guess you can repro this issue on RK3368 platform. At least until now, the RK3036, RK3288....that's work well. > Hi Caesar, > does your patch able to consider adding SDMMC/CPU_JTAG/MCU_JTAG > managment for rk3368 as well? > That's no related, I will add the sdmmc node and check it on tomorrow if the time can be availabled. >> which in turn already does the needed pinctrl state handling. >> >> So that default-state should be set up even before the dwmmc probe >> function >> is called. >> >> The sharing between sdmmc and uart2 pins is the same on the rk3036 >> and so >> far Caesar seems to have been able to switch the pin-state to sdmmc >> without >> touching the dw_mmc driver itself. >> Yep, that's working on rk3036 kylin board. --- Thanks, Caesar -- 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 2016/1/19 19:45, Caesar Wang wrote: > > ? 2016?01?19? 19:07, Shawn Lin ??: >> On 2016/1/19 18:07, Heiko Stuebner wrote: >>> Hi Shawn, >>> >>> Am Dienstag, 19. Januar 2016, 17:52:45 schrieb Shawn Lin: >>>> We find rk3368 fail to enumerate sd card since >>>> the default state is gpio function. Meahwhile, lots of >>>> rockchip platform dts file assign pinctrl to dw_mmc but actually >>>> dw_mmc never use it. Fortunately, those platforms' default io >>>> state is sdmmc function, but no always right for all. >>>> >>>> Cc: Heiko Stuebner <heiko@sntech.de> >>>> Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com> >>> >>> I do believe any issue you see, must come from somewhere else, >>> because the >>> driver-core does in drivers/base/dd.c really_probe(): >>> pinctrl_bind_pins() >> >> yes, pinctrl_bind_pins does turn pinctrl state. But I try debug a little >> more and seems somewhere else switch the state before dw_mmc probe. >> Now I *guess* that may relate to our ATF or MCU firmware since >> rk3368's sdmmc io can also be used as mcu_jtag for system debug. I >> need to double check that point, thanks. >> > > Yep, maybe that's related to the firmware setting. > I guess you can repro this issue on RK3368 platform. > At least until now, the RK3036, RK3288....that's work well. > Hi Caesar, Then I check this issue with MCU firmware guys and it does force pin setting into mcu_jtag function? Now I can see only 3368 apply MCU firmware, right? Another thing, seems old version firmware didn't break the sdmmc function. I think it's better to slove this problem in firmware side, not Linux. So, abandon this patch! > >> Hi Caesar, >> does your patch able to consider adding SDMMC/CPU_JTAG/MCU_JTAG >> managment for rk3368 as well? >> > > That's no related, I will add the sdmmc node and check it on tomorrow > if the time can be availabled. > > >>> which in turn already does the needed pinctrl state handling. >>> >>> So that default-state should be set up even before the dwmmc probe >>> function >>> is called. >>> >>> The sharing between sdmmc and uart2 pins is the same on the rk3036 >>> and so >>> far Caesar seems to have been able to switch the pin-state to sdmmc >>> without >>> touching the dw_mmc driver itself. >>> > > Yep, that's working on rk3036 kylin board. > > > --- > Thanks, > Caesar > > >
diff --git a/drivers/mmc/host/dw_mmc-rockchip.c b/drivers/mmc/host/dw_mmc-rockchip.c index d9c92f3..308b424 100644 --- a/drivers/mmc/host/dw_mmc-rockchip.c +++ b/drivers/mmc/host/dw_mmc-rockchip.c @@ -9,6 +9,7 @@ #include <linux/module.h> #include <linux/platform_device.h> +#include <linux/pinctrl/consumer.h> #include <linux/clk.h> #include <linux/mmc/host.h> #include <linux/mmc/dw_mmc.h> @@ -21,6 +22,8 @@ #define RK3288_CLKGEN_DIV 2 struct dw_mci_rockchip_priv_data { + struct pinctrl *pinctrl; + struct pinctrl_state *pins_default; struct clk *drv_clk; struct clk *sample_clk; int default_sample_phase; @@ -78,6 +81,10 @@ static void dw_mci_rk3288_set_ios(struct dw_mci *host, struct mmc_ios *ios) /* Make sure we use phases which we can enumerate with */ if (!IS_ERR(priv->sample_clk)) clk_set_phase(priv->sample_clk, priv->default_sample_phase); + + /* Make sure we use correct pinctrl which we can enumerate with */ + if (!IS_ERR(priv->pinctrl) && !IS_ERR(priv->pins_default)) + pinctrl_select_state(priv->pinctrl, priv->pins_default); } #define NUM_PHASES 360 @@ -223,6 +230,14 @@ static int dw_mci_rk3288_parse_dt(struct dw_mci *host) if (IS_ERR(priv->sample_clk)) dev_dbg(host->dev, "ciu_sample not available\n"); + priv->pinctrl = devm_pinctrl_get(host->dev); + if (!IS_ERR(priv->pinctrl)) { + priv->pins_default = pinctrl_lookup_state(priv->pinctrl, + "default"); + if (IS_ERR(priv->pins_default)) + dev_dbg(host->dev, "pinctrl not available\n"); + } + host->priv = priv; return 0;
We find rk3368 fail to enumerate sd card since the default state is gpio function. Meahwhile, lots of rockchip platform dts file assign pinctrl to dw_mmc but actually dw_mmc never use it. Fortunately, those platforms' default io state is sdmmc function, but no always right for all. Cc: Heiko Stuebner <heiko@sntech.de> Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com> --- drivers/mmc/host/dw_mmc-rockchip.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+)