[v2, 0/7] Add SD UHS-I and eMMC HS200 support for eSDHC

[v2, 0/7] Add SD UHS-I and eMMC HS200 support for eSDHC

[Yangbo Lu]

It's complicated to support SD UHS-I and eMMC HS200 for eSDHC because
there're many differences between eSDHC and SD/eMMC spec. Several
differences as below must be considered:
1. Peripheral clock must be used instead of platform clock.
    - eSDHC could select peripheral clock or platform clock as its clock
      source. According to RM, UHS-I/HS200 must use peripheral clock since
      it supports higher frequency than platform clock.
    - Patch 1 and patch 6 is to support this.
2. Signal voltage switching requires a control circuit out of eSDHC.
    - eSDHC supports signal voltage switch from 3.3v to 1.8v by
      eSDHC_PROCTL[VOLT_SEL] bit. This bit changes the value of output
      signal SDHC_VS, and there must be a control circuit out of eSDHC
      to change the signal voltage according to SDHC_VS output signal.
    - Patch 2 is to support this.
3. eSDHC uses tuning block for tuning procedure.
    - Tuning clock control register must be configured before tuning.
    - Patch 3 is to support this.
4. Delay is needed between tuning cycles for HS200 tuning.
    - Once a patch removed mdelay between tuning cycles.
      But eSDHC needs it.
    - Patch 4 and patch 5 is to support this.
5. UHS-I/HS200 modes could be enabled in dts node.
    - Patch 7 is to support this.

Please review and merge these patches on mmc git tree if no changes
are required.

Yangbo Lu (7):
  mmc: sdhci-of-esdhc: add peripheral clock support
  mmc: sdhci-of-esdhc: add support for signal voltage switch
  mmc: sdhci-of-esdhc: add tuning support
  mmc: sdhci: add a quirk to restore delay in tuning
  mmc: sdhci-of-esdhc: add delay between tuning cycles
  arm64: dts: ls1046a: add clocks property and compatible for eSDHC node
  arm64: dts: ls1046ardb: add MMC HS200/UHS-1 modes support

 arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts |   8 ++
 arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi    |   3 +-
 drivers/mmc/host/sdhci-esdhc.h                    |   7 +
 drivers/mmc/host/sdhci-of-esdhc.c                 | 167 +++++++++++++++++++++-
 drivers/mmc/host/sdhci.c                          |   3 +-
 drivers/mmc/host/sdhci.h                          |   2 +
 6 files changed, 186 insertions(+), 4 deletions(-)

-- 
2.1.0.27.g96db324

[v2, 0/7] Add SD UHS-I and eMMC HS200 support for eSDHC

[Yangbo Lu]

It's complicated to support SD UHS-I and eMMC HS200 for eSDHC because
there're many differences between eSDHC and SD/eMMC spec. Several
differences as below must be considered:
1. Peripheral clock must be used instead of platform clock.
    - eSDHC could select peripheral clock or platform clock as its clock
      source. According to RM, UHS-I/HS200 must use peripheral clock since
      it supports higher frequency than platform clock.
    - Patch 1 and patch 6 is to support this.
2. Signal voltage switching requires a control circuit out of eSDHC.
    - eSDHC supports signal voltage switch from 3.3v to 1.8v by
      eSDHC_PROCTL[VOLT_SEL] bit. This bit changes the value of output
      signal SDHC_VS, and there must be a control circuit out of eSDHC
      to change the signal voltage according to SDHC_VS output signal.
    - Patch 2 is to support this.
3. eSDHC uses tuning block for tuning procedure.
    - Tuning clock control register must be configured before tuning.
    - Patch 3 is to support this.
4. Delay is needed between tuning cycles for HS200 tuning.
    - Once a patch removed mdelay between tuning cycles.
      But eSDHC needs it.
    - Patch 4 and patch 5 is to support this.
5. UHS-I/HS200 modes could be enabled in dts node.
    - Patch 7 is to support this.

Please review and merge these patches on mmc git tree if no changes
are required.

Yangbo Lu (7):
  mmc: sdhci-of-esdhc: add peripheral clock support
  mmc: sdhci-of-esdhc: add support for signal voltage switch
  mmc: sdhci-of-esdhc: add tuning support
  mmc: sdhci: add a quirk to restore delay in tuning
  mmc: sdhci-of-esdhc: add delay between tuning cycles
  arm64: dts: ls1046a: add clocks property and compatible for eSDHC node
  arm64: dts: ls1046ardb: add MMC HS200/UHS-1 modes support

 arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts |   8 ++
 arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi    |   3 +-
 drivers/mmc/host/sdhci-esdhc.h                    |   7 +
 drivers/mmc/host/sdhci-of-esdhc.c                 | 167 +++++++++++++++++++++-
 drivers/mmc/host/sdhci.c                          |   3 +-
 drivers/mmc/host/sdhci.h                          |   2 +
 6 files changed, 186 insertions(+), 4 deletions(-)

-- 
2.1.0.27.g96db324

[v2, 1/7] mmc: sdhci-of-esdhc: add peripheral clock support

[Yangbo Lu]

eSDHC could select peripheral clock or platform clock as clock source by
the PCS bit of eSDHC Control Register, and this bit couldn't be reset by
software reset for all. In default, the platform clock is used. But we have
to use peripheral clock since it has a higher frequency to support eMMC
HS200 mode and SD UHS-I mode. This patch is to add peripheral clock support
and use it instead of platform clock if it's declared in eSDHC dts node.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
---
Changes for v2:
	- None
---
 drivers/mmc/host/sdhci-esdhc.h    |  1 +
 drivers/mmc/host/sdhci-of-esdhc.c | 70 +++++++++++++++++++++++++++++++++++++--
 2 files changed, 69 insertions(+), 2 deletions(-)

diff --git a/drivers/mmc/host/sdhci-esdhc.h b/drivers/mmc/host/sdhci-esdhc.h
index ece8b37..5343fc0 100644
--- a/drivers/mmc/host/sdhci-esdhc.h
+++ b/drivers/mmc/host/sdhci-esdhc.h
@@ -54,6 +54,7 @@
 
 /* Control Register for DMA transfer */
 #define ESDHC_DMA_SYSCTL		0x40c
+#define ESDHC_PERIPHERAL_CLK_SEL	0x00080000
 #define ESDHC_DMA_SNOOP			0x00000040
 
 #endif /* _DRIVERS_MMC_SDHCI_ESDHC_H */
diff --git a/drivers/mmc/host/sdhci-of-esdhc.c b/drivers/mmc/host/sdhci-of-esdhc.c
index d3aa671..84865b0 100644
--- a/drivers/mmc/host/sdhci-of-esdhc.c
+++ b/drivers/mmc/host/sdhci-of-esdhc.c
@@ -19,6 +19,7 @@
 #include <linux/delay.h>
 #include <linux/module.h>
 #include <linux/sys_soc.h>
+#include <linux/clk.h>
 #include <linux/mmc/host.h>
 #include "sdhci-pltfm.h"
 #include "sdhci-esdhc.h"
@@ -30,6 +31,7 @@ struct sdhci_esdhc {
 	u8 vendor_ver;
 	u8 spec_ver;
 	bool quirk_incorrect_hostver;
+	unsigned int peripheral_clock;
 };
 
 /**
@@ -414,15 +416,25 @@ static int esdhc_of_enable_dma(struct sdhci_host *host)
 static unsigned int esdhc_of_get_max_clock(struct sdhci_host *host)
 {
 	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
 
-	return pltfm_host->clock;
+	if (esdhc->peripheral_clock)
+		return esdhc->peripheral_clock;
+	else
+		return pltfm_host->clock;
 }
 
 static unsigned int esdhc_of_get_min_clock(struct sdhci_host *host)
 {
 	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+	unsigned int clock;
 
-	return pltfm_host->clock / 256 / 16;
+	if (esdhc->peripheral_clock)
+		clock = esdhc->peripheral_clock;
+	else
+		clock = pltfm_host->clock;
+	return clock / 256 / 16;
 }
 
 static void esdhc_of_set_clock(struct sdhci_host *host, unsigned int clock)
@@ -512,6 +524,33 @@ static void esdhc_pltfm_set_bus_width(struct sdhci_host *host, int width)
 	sdhci_writel(host, ctrl, ESDHC_PROCTL);
 }
 
+static void esdhc_clock_enable(struct sdhci_host *host, bool enable)
+{
+	u32 val;
+	u32 timeout;
+
+	val = sdhci_readl(host, ESDHC_SYSTEM_CONTROL);
+
+	if (enable)
+		val |= ESDHC_CLOCK_SDCLKEN;
+	else
+		val &= ~ESDHC_CLOCK_SDCLKEN;
+
+	sdhci_writel(host, val, ESDHC_SYSTEM_CONTROL);
+
+	timeout = 20;
+	val = ESDHC_CLOCK_STABLE;
+	while (!(sdhci_readl(host, ESDHC_PRSSTAT) & val)) {
+		if (timeout == 0) {
+			pr_err("%s: Internal clock never stabilised.\n",
+				mmc_hostname(host->mmc));
+			break;
+		}
+		timeout--;
+		mdelay(1);
+	}
+}
+
 static void esdhc_reset(struct sdhci_host *host, u8 mask)
 {
 	sdhci_reset(host, mask);
@@ -610,6 +649,9 @@ static void esdhc_init(struct platform_device *pdev, struct sdhci_host *host)
 {
 	struct sdhci_pltfm_host *pltfm_host;
 	struct sdhci_esdhc *esdhc;
+	struct device_node *np;
+	struct clk *clk;
+	u32 val;
 	u16 host_ver;
 
 	pltfm_host = sdhci_priv(host);
@@ -623,6 +665,30 @@ static void esdhc_init(struct platform_device *pdev, struct sdhci_host *host)
 		esdhc->quirk_incorrect_hostver = true;
 	else
 		esdhc->quirk_incorrect_hostver = false;
+
+	np = pdev->dev.of_node;
+	clk = of_clk_get(np, 0);
+	if (!IS_ERR(clk)) {
+		/*
+		 * esdhc->peripheral_clock would be assigned with a value
+		 * which is eSDHC base clock when use periperal clock.
+		 * For ls1046a, the clock value got by common clk API is
+		 * peripheral clock while the eSDHC base clock is 1/2
+		 * peripheral clock.
+		 */
+		if (of_device_is_compatible(np, "fsl,ls1046a-esdhc"))
+			esdhc->peripheral_clock = clk_get_rate(clk) / 2;
+		else
+			esdhc->peripheral_clock = clk_get_rate(clk);
+	}
+
+	if (esdhc->peripheral_clock) {
+		esdhc_clock_enable(host, false);
+		val = sdhci_readl(host, ESDHC_DMA_SYSCTL);
+		val |= ESDHC_PERIPHERAL_CLK_SEL;
+		sdhci_writel(host, val, ESDHC_DMA_SYSCTL);
+		esdhc_clock_enable(host, true);
+	}
 }
 
 static int sdhci_esdhc_probe(struct platform_device *pdev)
-- 
2.1.0.27.g96db324

[v2, 1/7] mmc: sdhci-of-esdhc: add peripheral clock support

[Yangbo Lu]

eSDHC could select peripheral clock or platform clock as clock source by
the PCS bit of eSDHC Control Register, and this bit couldn't be reset by
software reset for all. In default, the platform clock is used. But we have
to use peripheral clock since it has a higher frequency to support eMMC
HS200 mode and SD UHS-I mode. This patch is to add peripheral clock support
and use it instead of platform clock if it's declared in eSDHC dts node.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
---
Changes for v2:
	- None
---
 drivers/mmc/host/sdhci-esdhc.h    |  1 +
 drivers/mmc/host/sdhci-of-esdhc.c | 70 +++++++++++++++++++++++++++++++++++++--
 2 files changed, 69 insertions(+), 2 deletions(-)

diff --git a/drivers/mmc/host/sdhci-esdhc.h b/drivers/mmc/host/sdhci-esdhc.h
index ece8b37..5343fc0 100644
--- a/drivers/mmc/host/sdhci-esdhc.h
+++ b/drivers/mmc/host/sdhci-esdhc.h
@@ -54,6 +54,7 @@
 
 /* Control Register for DMA transfer */
 #define ESDHC_DMA_SYSCTL		0x40c
+#define ESDHC_PERIPHERAL_CLK_SEL	0x00080000
 #define ESDHC_DMA_SNOOP			0x00000040
 
 #endif /* _DRIVERS_MMC_SDHCI_ESDHC_H */
diff --git a/drivers/mmc/host/sdhci-of-esdhc.c b/drivers/mmc/host/sdhci-of-esdhc.c
index d3aa671..84865b0 100644
--- a/drivers/mmc/host/sdhci-of-esdhc.c
+++ b/drivers/mmc/host/sdhci-of-esdhc.c
@@ -19,6 +19,7 @@
 #include <linux/delay.h>
 #include <linux/module.h>
 #include <linux/sys_soc.h>
+#include <linux/clk.h>
 #include <linux/mmc/host.h>
 #include "sdhci-pltfm.h"
 #include "sdhci-esdhc.h"
@@ -30,6 +31,7 @@ struct sdhci_esdhc {
 	u8 vendor_ver;
 	u8 spec_ver;
 	bool quirk_incorrect_hostver;
+	unsigned int peripheral_clock;
 };
 
 /**
@@ -414,15 +416,25 @@ static int esdhc_of_enable_dma(struct sdhci_host *host)
 static unsigned int esdhc_of_get_max_clock(struct sdhci_host *host)
 {
 	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
 
-	return pltfm_host->clock;
+	if (esdhc->peripheral_clock)
+		return esdhc->peripheral_clock;
+	else
+		return pltfm_host->clock;
 }
 
 static unsigned int esdhc_of_get_min_clock(struct sdhci_host *host)
 {
 	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+	unsigned int clock;
 
-	return pltfm_host->clock / 256 / 16;
+	if (esdhc->peripheral_clock)
+		clock = esdhc->peripheral_clock;
+	else
+		clock = pltfm_host->clock;
+	return clock / 256 / 16;
 }
 
 static void esdhc_of_set_clock(struct sdhci_host *host, unsigned int clock)
@@ -512,6 +524,33 @@ static void esdhc_pltfm_set_bus_width(struct sdhci_host *host, int width)
 	sdhci_writel(host, ctrl, ESDHC_PROCTL);
 }
 
+static void esdhc_clock_enable(struct sdhci_host *host, bool enable)
+{
+	u32 val;
+	u32 timeout;
+
+	val = sdhci_readl(host, ESDHC_SYSTEM_CONTROL);
+
+	if (enable)
+		val |= ESDHC_CLOCK_SDCLKEN;
+	else
+		val &= ~ESDHC_CLOCK_SDCLKEN;
+
+	sdhci_writel(host, val, ESDHC_SYSTEM_CONTROL);
+
+	timeout = 20;
+	val = ESDHC_CLOCK_STABLE;
+	while (!(sdhci_readl(host, ESDHC_PRSSTAT) & val)) {
+		if (timeout == 0) {
+			pr_err("%s: Internal clock never stabilised.\n",
+				mmc_hostname(host->mmc));
+			break;
+		}
+		timeout--;
+		mdelay(1);
+	}
+}
+
 static void esdhc_reset(struct sdhci_host *host, u8 mask)
 {
 	sdhci_reset(host, mask);
@@ -610,6 +649,9 @@ static void esdhc_init(struct platform_device *pdev, struct sdhci_host *host)
 {
 	struct sdhci_pltfm_host *pltfm_host;
 	struct sdhci_esdhc *esdhc;
+	struct device_node *np;
+	struct clk *clk;
+	u32 val;
 	u16 host_ver;
 
 	pltfm_host = sdhci_priv(host);
@@ -623,6 +665,30 @@ static void esdhc_init(struct platform_device *pdev, struct sdhci_host *host)
 		esdhc->quirk_incorrect_hostver = true;
 	else
 		esdhc->quirk_incorrect_hostver = false;
+
+	np = pdev->dev.of_node;
+	clk = of_clk_get(np, 0);
+	if (!IS_ERR(clk)) {
+		/*
+		 * esdhc->peripheral_clock would be assigned with a value
+		 * which is eSDHC base clock when use periperal clock.
+		 * For ls1046a, the clock value got by common clk API is
+		 * peripheral clock while the eSDHC base clock is 1/2
+		 * peripheral clock.
+		 */
+		if (of_device_is_compatible(np, "fsl,ls1046a-esdhc"))
+			esdhc->peripheral_clock = clk_get_rate(clk) / 2;
+		else
+			esdhc->peripheral_clock = clk_get_rate(clk);
+	}
+
+	if (esdhc->peripheral_clock) {
+		esdhc_clock_enable(host, false);
+		val = sdhci_readl(host, ESDHC_DMA_SYSCTL);
+		val |= ESDHC_PERIPHERAL_CLK_SEL;
+		sdhci_writel(host, val, ESDHC_DMA_SYSCTL);
+		esdhc_clock_enable(host, true);
+	}
 }
 
 static int sdhci_esdhc_probe(struct platform_device *pdev)
-- 
2.1.0.27.g96db324

[v2, 2/7] mmc: sdhci-of-esdhc: add support for signal voltage switch

[Yangbo Lu]

eSDHC supports signal voltage switch from 3.3v to 1.8v by
eSDHC_PROCTL[VOLT_SEL] bit. This bit changes the value of output
signal SDHC_VS, and there must be a control circuit out of eSDHC
to change the signal voltage according to SDHC_VS output signal.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
---
Changes for v2:
	- Used Adrain's method to support voltage switching:
          host->mmc_host_ops.start_signal_voltage_switch =
                     esdhc_signal_voltage_switch;
---
 drivers/mmc/host/sdhci-esdhc.h    |  1 +
 drivers/mmc/host/sdhci-of-esdhc.c | 74 +++++++++++++++++++++++++++++++++++++++
 2 files changed, 75 insertions(+)

diff --git a/drivers/mmc/host/sdhci-esdhc.h b/drivers/mmc/host/sdhci-esdhc.h
index 5343fc0..6869567 100644
--- a/drivers/mmc/host/sdhci-esdhc.h
+++ b/drivers/mmc/host/sdhci-esdhc.h
@@ -37,6 +37,7 @@
 
 /* Protocol Control Register */
 #define ESDHC_PROCTL			0x28
+#define ESDHC_VOLT_SEL			0x00000400
 #define ESDHC_CTRL_4BITBUS		(0x1 << 1)
 #define ESDHC_CTRL_8BITBUS		(0x2 << 1)
 #define ESDHC_CTRL_BUSWIDTH_MASK	(0x3 << 1)
diff --git a/drivers/mmc/host/sdhci-of-esdhc.c b/drivers/mmc/host/sdhci-of-esdhc.c
index 84865b0..9ab6943 100644
--- a/drivers/mmc/host/sdhci-of-esdhc.c
+++ b/drivers/mmc/host/sdhci-of-esdhc.c
@@ -16,6 +16,7 @@
 #include <linux/err.h>
 #include <linux/io.h>
 #include <linux/of.h>
+#include <linux/of_address.h>
 #include <linux/delay.h>
 #include <linux/module.h>
 #include <linux/sys_soc.h>
@@ -559,6 +560,76 @@ static void esdhc_reset(struct sdhci_host *host, u8 mask)
 	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
 }
 
+/* The SCFG, Supplemental Configuration Unit, provides SoC specific
+ * configuration and status registers for the device. There is a
+ * SDHC IO VSEL control register on SCFG for some platforms. It's
+ * used to support SDHC IO voltage switching.
+ */
+static const struct of_device_id scfg_device_ids[] = {
+	{ .compatible = "fsl,t1040-scfg", },
+	{ .compatible = "fsl,ls1012a-scfg", },
+	{ .compatible = "fsl,ls1046a-scfg", },
+	{}
+};
+
+/* SDHC IO VSEL control register definition */
+#define SCFG_SDHCIOVSELCR	0x408
+#define SDHCIOVSELCR_TGLEN	0x80000000
+#define SDHCIOVSELCR_VSELVAL	0x60000000
+#define SDHCIOVSELCR_SDHC_VS	0x00000001
+
+static int esdhc_signal_voltage_switch(struct mmc_host *mmc,
+				       struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct device_node *scfg_node;
+	void __iomem *scfg_base = NULL;
+	u32 sdhciovselcr;
+	u32 val;
+
+	/*
+	 * Signal Voltage Switching is only applicable for Host Controllers
+	 * v3.00 and above.
+	 */
+	if (host->version < SDHCI_SPEC_300)
+		return 0;
+
+	val = sdhci_readl(host, ESDHC_PROCTL);
+
+	switch (ios->signal_voltage) {
+	case MMC_SIGNAL_VOLTAGE_330:
+		val &= ~ESDHC_VOLT_SEL;
+		sdhci_writel(host, val, ESDHC_PROCTL);
+		return 0;
+	case MMC_SIGNAL_VOLTAGE_180:
+		scfg_node = of_find_matching_node(NULL, scfg_device_ids);
+		if (scfg_node)
+			scfg_base = of_iomap(scfg_node, 0);
+		if (scfg_base) {
+			sdhciovselcr = SDHCIOVSELCR_TGLEN |
+				       SDHCIOVSELCR_VSELVAL;
+			iowrite32be(sdhciovselcr,
+				scfg_base + SCFG_SDHCIOVSELCR);
+
+			val |= ESDHC_VOLT_SEL;
+			sdhci_writel(host, val, ESDHC_PROCTL);
+			mdelay(5);
+
+			sdhciovselcr = SDHCIOVSELCR_TGLEN |
+				       SDHCIOVSELCR_SDHC_VS;
+			iowrite32be(sdhciovselcr,
+				scfg_base + SCFG_SDHCIOVSELCR);
+			iounmap(scfg_base);
+		} else {
+			val |= ESDHC_VOLT_SEL;
+			sdhci_writel(host, val, ESDHC_PROCTL);
+		}
+		return 0;
+	default:
+		return 0;
+	}
+}
+
 #ifdef CONFIG_PM_SLEEP
 static u32 esdhc_proctl;
 static int esdhc_of_suspend(struct device *dev)
@@ -708,6 +779,9 @@ static int sdhci_esdhc_probe(struct platform_device *pdev)
 		host = sdhci_pltfm_init(pdev, &sdhci_esdhc_be_pdata,
 					sizeof(struct sdhci_esdhc));
 
+	host->mmc_host_ops.start_signal_voltage_switch =
+		esdhc_signal_voltage_switch;
+
 	if (IS_ERR(host))
 		return PTR_ERR(host);
 
-- 
2.1.0.27.g96db324

[v2, 2/7] mmc: sdhci-of-esdhc: add support for signal voltage switch

[Yangbo Lu]

eSDHC supports signal voltage switch from 3.3v to 1.8v by
eSDHC_PROCTL[VOLT_SEL] bit. This bit changes the value of output
signal SDHC_VS, and there must be a control circuit out of eSDHC
to change the signal voltage according to SDHC_VS output signal.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
---
Changes for v2:
	- Used Adrain's method to support voltage switching:
          host->mmc_host_ops.start_signal_voltage_switch =
                     esdhc_signal_voltage_switch;
---
 drivers/mmc/host/sdhci-esdhc.h    |  1 +
 drivers/mmc/host/sdhci-of-esdhc.c | 74 +++++++++++++++++++++++++++++++++++++++
 2 files changed, 75 insertions(+)

diff --git a/drivers/mmc/host/sdhci-esdhc.h b/drivers/mmc/host/sdhci-esdhc.h
index 5343fc0..6869567 100644
--- a/drivers/mmc/host/sdhci-esdhc.h
+++ b/drivers/mmc/host/sdhci-esdhc.h
@@ -37,6 +37,7 @@
 
 /* Protocol Control Register */
 #define ESDHC_PROCTL			0x28
+#define ESDHC_VOLT_SEL			0x00000400
 #define ESDHC_CTRL_4BITBUS		(0x1 << 1)
 #define ESDHC_CTRL_8BITBUS		(0x2 << 1)
 #define ESDHC_CTRL_BUSWIDTH_MASK	(0x3 << 1)
diff --git a/drivers/mmc/host/sdhci-of-esdhc.c b/drivers/mmc/host/sdhci-of-esdhc.c
index 84865b0..9ab6943 100644
--- a/drivers/mmc/host/sdhci-of-esdhc.c
+++ b/drivers/mmc/host/sdhci-of-esdhc.c
@@ -16,6 +16,7 @@
 #include <linux/err.h>
 #include <linux/io.h>
 #include <linux/of.h>
+#include <linux/of_address.h>
 #include <linux/delay.h>
 #include <linux/module.h>
 #include <linux/sys_soc.h>
@@ -559,6 +560,76 @@ static void esdhc_reset(struct sdhci_host *host, u8 mask)
 	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
 }
 
+/* The SCFG, Supplemental Configuration Unit, provides SoC specific
+ * configuration and status registers for the device. There is a
+ * SDHC IO VSEL control register on SCFG for some platforms. It's
+ * used to support SDHC IO voltage switching.
+ */
+static const struct of_device_id scfg_device_ids[] = {
+	{ .compatible = "fsl,t1040-scfg", },
+	{ .compatible = "fsl,ls1012a-scfg", },
+	{ .compatible = "fsl,ls1046a-scfg", },
+	{}
+};
+
+/* SDHC IO VSEL control register definition */
+#define SCFG_SDHCIOVSELCR	0x408
+#define SDHCIOVSELCR_TGLEN	0x80000000
+#define SDHCIOVSELCR_VSELVAL	0x60000000
+#define SDHCIOVSELCR_SDHC_VS	0x00000001
+
+static int esdhc_signal_voltage_switch(struct mmc_host *mmc,
+				       struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct device_node *scfg_node;
+	void __iomem *scfg_base = NULL;
+	u32 sdhciovselcr;
+	u32 val;
+
+	/*
+	 * Signal Voltage Switching is only applicable for Host Controllers
+	 * v3.00 and above.
+	 */
+	if (host->version < SDHCI_SPEC_300)
+		return 0;
+
+	val = sdhci_readl(host, ESDHC_PROCTL);
+
+	switch (ios->signal_voltage) {
+	case MMC_SIGNAL_VOLTAGE_330:
+		val &= ~ESDHC_VOLT_SEL;
+		sdhci_writel(host, val, ESDHC_PROCTL);
+		return 0;
+	case MMC_SIGNAL_VOLTAGE_180:
+		scfg_node = of_find_matching_node(NULL, scfg_device_ids);
+		if (scfg_node)
+			scfg_base = of_iomap(scfg_node, 0);
+		if (scfg_base) {
+			sdhciovselcr = SDHCIOVSELCR_TGLEN |
+				       SDHCIOVSELCR_VSELVAL;
+			iowrite32be(sdhciovselcr,
+				scfg_base + SCFG_SDHCIOVSELCR);
+
+			val |= ESDHC_VOLT_SEL;
+			sdhci_writel(host, val, ESDHC_PROCTL);
+			mdelay(5);
+
+			sdhciovselcr = SDHCIOVSELCR_TGLEN |
+				       SDHCIOVSELCR_SDHC_VS;
+			iowrite32be(sdhciovselcr,
+				scfg_base + SCFG_SDHCIOVSELCR);
+			iounmap(scfg_base);
+		} else {
+			val |= ESDHC_VOLT_SEL;
+			sdhci_writel(host, val, ESDHC_PROCTL);
+		}
+		return 0;
+	default:
+		return 0;
+	}
+}
+
 #ifdef CONFIG_PM_SLEEP
 static u32 esdhc_proctl;
 static int esdhc_of_suspend(struct device *dev)
@@ -708,6 +779,9 @@ static int sdhci_esdhc_probe(struct platform_device *pdev)
 		host = sdhci_pltfm_init(pdev, &sdhci_esdhc_be_pdata,
 					sizeof(struct sdhci_esdhc));
 
+	host->mmc_host_ops.start_signal_voltage_switch =
+		esdhc_signal_voltage_switch;
+
 	if (IS_ERR(host))
 		return PTR_ERR(host);
 
-- 
2.1.0.27.g96db324

[v2, 3/7] mmc: sdhci-of-esdhc: add tuning support

[Yangbo Lu]

eSDHC uses tuning block for tuning procedure. So the tuning
block control register must be configured properly before tuning.

Signed-off-by: Yangbo Lu <yangbo.lu-3arQi8VN3Tc@public.gmane.org>
---
Changes for v3:
	- Replaced old function for mmc_host_ops.execute_tuning with
	  esdhc_execute_tuning to support eSDHC tuning.
---
 drivers/mmc/host/sdhci-esdhc.h    |  5 +++++
 drivers/mmc/host/sdhci-of-esdhc.c | 21 +++++++++++++++++++++
 2 files changed, 26 insertions(+)

diff --git a/drivers/mmc/host/sdhci-esdhc.h b/drivers/mmc/host/sdhci-esdhc.h
index 6869567..c4bbd74 100644
--- a/drivers/mmc/host/sdhci-esdhc.h
+++ b/drivers/mmc/host/sdhci-esdhc.h
@@ -53,9 +53,14 @@
 #define ESDHC_CLOCK_HCKEN		0x00000002
 #define ESDHC_CLOCK_IPGEN		0x00000001
 
+/* Tuning Block Control Register */
+#define ESDHC_TBCTL			0x120
+#define ESDHC_TB_EN			0x00000004
+
 /* Control Register for DMA transfer */
 #define ESDHC_DMA_SYSCTL		0x40c
 #define ESDHC_PERIPHERAL_CLK_SEL	0x00080000
+#define ESDHC_FLUSH_ASYNC_FIFO		0x00040000
 #define ESDHC_DMA_SNOOP			0x00000040
 
 #endif /* _DRIVERS_MMC_SDHCI_ESDHC_H */
diff --git a/drivers/mmc/host/sdhci-of-esdhc.c b/drivers/mmc/host/sdhci-of-esdhc.c
index 9ab6943..8007188 100644
--- a/drivers/mmc/host/sdhci-of-esdhc.c
+++ b/drivers/mmc/host/sdhci-of-esdhc.c
@@ -630,6 +630,25 @@ static int esdhc_signal_voltage_switch(struct mmc_host *mmc,
 	}
 }
 
+static int esdhc_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	u32 val;
+
+	/* Use tuning block for tuning procedure */
+	esdhc_clock_enable(host, false);
+	val = sdhci_readl(host, ESDHC_DMA_SYSCTL);
+	val |= ESDHC_FLUSH_ASYNC_FIFO;
+	sdhci_writel(host, val, ESDHC_DMA_SYSCTL);
+
+	val = sdhci_readl(host, ESDHC_TBCTL);
+	val |= ESDHC_TB_EN;
+	sdhci_writel(host, val, ESDHC_TBCTL);
+	esdhc_clock_enable(host, true);
+
+	return sdhci_execute_tuning(mmc, opcode);
+}
+
 #ifdef CONFIG_PM_SLEEP
 static u32 esdhc_proctl;
 static int esdhc_of_suspend(struct device *dev)
@@ -782,6 +801,8 @@ static int sdhci_esdhc_probe(struct platform_device *pdev)
 	host->mmc_host_ops.start_signal_voltage_switch =
 		esdhc_signal_voltage_switch;
 
+	host->mmc_host_ops.execute_tuning = esdhc_execute_tuning;
+
 	if (IS_ERR(host))
 		return PTR_ERR(host);
 
-- 
2.1.0.27.g96db324

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[v2, 3/7] mmc: sdhci-of-esdhc: add tuning support

[Yangbo Lu]

eSDHC uses tuning block for tuning procedure. So the tuning
block control register must be configured properly before tuning.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
---
Changes for v3:
	- Replaced old function for mmc_host_ops.execute_tuning with
	  esdhc_execute_tuning to support eSDHC tuning.
---
 drivers/mmc/host/sdhci-esdhc.h    |  5 +++++
 drivers/mmc/host/sdhci-of-esdhc.c | 21 +++++++++++++++++++++
 2 files changed, 26 insertions(+)

diff --git a/drivers/mmc/host/sdhci-esdhc.h b/drivers/mmc/host/sdhci-esdhc.h
index 6869567..c4bbd74 100644
--- a/drivers/mmc/host/sdhci-esdhc.h
+++ b/drivers/mmc/host/sdhci-esdhc.h
@@ -53,9 +53,14 @@
 #define ESDHC_CLOCK_HCKEN		0x00000002
 #define ESDHC_CLOCK_IPGEN		0x00000001
 
+/* Tuning Block Control Register */
+#define ESDHC_TBCTL			0x120
+#define ESDHC_TB_EN			0x00000004
+
 /* Control Register for DMA transfer */
 #define ESDHC_DMA_SYSCTL		0x40c
 #define ESDHC_PERIPHERAL_CLK_SEL	0x00080000
+#define ESDHC_FLUSH_ASYNC_FIFO		0x00040000
 #define ESDHC_DMA_SNOOP			0x00000040
 
 #endif /* _DRIVERS_MMC_SDHCI_ESDHC_H */
diff --git a/drivers/mmc/host/sdhci-of-esdhc.c b/drivers/mmc/host/sdhci-of-esdhc.c
index 9ab6943..8007188 100644
--- a/drivers/mmc/host/sdhci-of-esdhc.c
+++ b/drivers/mmc/host/sdhci-of-esdhc.c
@@ -630,6 +630,25 @@ static int esdhc_signal_voltage_switch(struct mmc_host *mmc,
 	}
 }
 
+static int esdhc_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	u32 val;
+
+	/* Use tuning block for tuning procedure */
+	esdhc_clock_enable(host, false);
+	val = sdhci_readl(host, ESDHC_DMA_SYSCTL);
+	val |= ESDHC_FLUSH_ASYNC_FIFO;
+	sdhci_writel(host, val, ESDHC_DMA_SYSCTL);
+
+	val = sdhci_readl(host, ESDHC_TBCTL);
+	val |= ESDHC_TB_EN;
+	sdhci_writel(host, val, ESDHC_TBCTL);
+	esdhc_clock_enable(host, true);
+
+	return sdhci_execute_tuning(mmc, opcode);
+}
+
 #ifdef CONFIG_PM_SLEEP
 static u32 esdhc_proctl;
 static int esdhc_of_suspend(struct device *dev)
@@ -782,6 +801,8 @@ static int sdhci_esdhc_probe(struct platform_device *pdev)
 	host->mmc_host_ops.start_signal_voltage_switch =
 		esdhc_signal_voltage_switch;
 
+	host->mmc_host_ops.execute_tuning = esdhc_execute_tuning;
+
 	if (IS_ERR(host))
 		return PTR_ERR(host);
 
-- 
2.1.0.27.g96db324

[v2, 4/7] mmc: sdhci: add a quirk to restore delay in tuning

[Yangbo Lu]

There was a mdelay between tuning cycles which is removed since
eMMC spec didn't require that. However it's observed that eSDHC
needed this delay for HS200 successful tuning. Maybe it's also
needed for some other controllers. This patch is to add a quirk
to fix it.

Fixes: 197160d52e85 ("mmc: sdhci: remove mdelay in eMMC tuning")
Signed-off-by: Yangbo Lu <yangbo.lu-3arQi8VN3Tc@public.gmane.org>
---
Changes for v2:
	- None
---
 drivers/mmc/host/sdhci.c | 3 ++-
 drivers/mmc/host/sdhci.h | 2 ++
 2 files changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c
index 6fdd7a7..5d204d7 100644
--- a/drivers/mmc/host/sdhci.c
+++ b/drivers/mmc/host/sdhci.c
@@ -2094,7 +2094,8 @@ static void __sdhci_execute_tuning(struct sdhci_host *host, u32 opcode,
 		}
 
 		/* eMMC spec does not require a delay between tuning cycles */
-		if (opcode == MMC_SEND_TUNING_BLOCK)
+		if ((opcode == MMC_SEND_TUNING_BLOCK) ||
+		    (host->quirks2 & SDHCI_QUIRK2_DELAY_BETWEEN_TUNING_CYCLES))
 			mdelay(1);
 	}
 
diff --git a/drivers/mmc/host/sdhci.h b/drivers/mmc/host/sdhci.h
index edf3adf..b89589c 100644
--- a/drivers/mmc/host/sdhci.h
+++ b/drivers/mmc/host/sdhci.h
@@ -427,6 +427,8 @@ struct sdhci_host {
 #define SDHCI_QUIRK2_ACMD23_BROKEN			(1<<14)
 /* Broken Clock divider zero in controller */
 #define SDHCI_QUIRK2_CLOCK_DIV_ZERO_BROKEN		(1<<15)
+/* Delay is needed between tuning cycles */
+#define SDHCI_QUIRK2_DELAY_BETWEEN_TUNING_CYCLES	(1<<16)
 
 	int irq;		/* Device IRQ */
 	void __iomem *ioaddr;	/* Mapped address */
-- 
2.1.0.27.g96db324

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[v2, 4/7] mmc: sdhci: add a quirk to restore delay in tuning

[Yangbo Lu]

There was a mdelay between tuning cycles which is removed since
eMMC spec didn't require that. However it's observed that eSDHC
needed this delay for HS200 successful tuning. Maybe it's also
needed for some other controllers. This patch is to add a quirk
to fix it.

Fixes: 197160d52e85 ("mmc: sdhci: remove mdelay in eMMC tuning")
Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
---
Changes for v2:
	- None
---
 drivers/mmc/host/sdhci.c | 3 ++-
 drivers/mmc/host/sdhci.h | 2 ++
 2 files changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c
index 6fdd7a7..5d204d7 100644
--- a/drivers/mmc/host/sdhci.c
+++ b/drivers/mmc/host/sdhci.c
@@ -2094,7 +2094,8 @@ static void __sdhci_execute_tuning(struct sdhci_host *host, u32 opcode,
 		}
 
 		/* eMMC spec does not require a delay between tuning cycles */
-		if (opcode == MMC_SEND_TUNING_BLOCK)
+		if ((opcode == MMC_SEND_TUNING_BLOCK) ||
+		    (host->quirks2 & SDHCI_QUIRK2_DELAY_BETWEEN_TUNING_CYCLES))
 			mdelay(1);
 	}
 
diff --git a/drivers/mmc/host/sdhci.h b/drivers/mmc/host/sdhci.h
index edf3adf..b89589c 100644
--- a/drivers/mmc/host/sdhci.h
+++ b/drivers/mmc/host/sdhci.h
@@ -427,6 +427,8 @@ struct sdhci_host {
 #define SDHCI_QUIRK2_ACMD23_BROKEN			(1<<14)
 /* Broken Clock divider zero in controller */
 #define SDHCI_QUIRK2_CLOCK_DIV_ZERO_BROKEN		(1<<15)
+/* Delay is needed between tuning cycles */
+#define SDHCI_QUIRK2_DELAY_BETWEEN_TUNING_CYCLES	(1<<16)
 
 	int irq;		/* Device IRQ */
 	void __iomem *ioaddr;	/* Mapped address */
-- 
2.1.0.27.g96db324

[v2, 5/7] mmc: sdhci-of-esdhc: add delay between tuning cycles

[Yangbo Lu]

It's observed that eSDHC needed delay between tuning cycles for
HS200 successful tuning. This patch is to add the quirk
SDHCI_QUIRK2_DELAY_BETWEEN_TUNING_CYCLES to fix it.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
---
Changes for v2:
	- None
---
 drivers/mmc/host/sdhci-of-esdhc.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/mmc/host/sdhci-of-esdhc.c b/drivers/mmc/host/sdhci-of-esdhc.c
index 8007188..7c4d8e8 100644
--- a/drivers/mmc/host/sdhci-of-esdhc.c
+++ b/drivers/mmc/host/sdhci-of-esdhc.c
@@ -836,6 +836,8 @@ static int sdhci_esdhc_probe(struct platform_device *pdev)
 		host->quirks2 |= SDHCI_QUIRK2_BROKEN_HOST_CONTROL;
 	}
 
+	host->quirks2 |= SDHCI_QUIRK2_DELAY_BETWEEN_TUNING_CYCLES;
+
 	/* call to generic mmc_of_parse to support additional capabilities */
 	ret = mmc_of_parse(host->mmc);
 	if (ret)
-- 
2.1.0.27.g96db324

[v2, 5/7] mmc: sdhci-of-esdhc: add delay between tuning cycles

[Yangbo Lu]

It's observed that eSDHC needed delay between tuning cycles for
HS200 successful tuning. This patch is to add the quirk
SDHCI_QUIRK2_DELAY_BETWEEN_TUNING_CYCLES to fix it.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
---
Changes for v2:
	- None
---
 drivers/mmc/host/sdhci-of-esdhc.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/mmc/host/sdhci-of-esdhc.c b/drivers/mmc/host/sdhci-of-esdhc.c
index 8007188..7c4d8e8 100644
--- a/drivers/mmc/host/sdhci-of-esdhc.c
+++ b/drivers/mmc/host/sdhci-of-esdhc.c
@@ -836,6 +836,8 @@ static int sdhci_esdhc_probe(struct platform_device *pdev)
 		host->quirks2 |= SDHCI_QUIRK2_BROKEN_HOST_CONTROL;
 	}
 
+	host->quirks2 |= SDHCI_QUIRK2_DELAY_BETWEEN_TUNING_CYCLES;
+
 	/* call to generic mmc_of_parse to support additional capabilities */
 	ret = mmc_of_parse(host->mmc);
 	if (ret)
-- 
2.1.0.27.g96db324

[v2, 6/7] arm64: dts: ls1046a: add clocks property and compatible for eSDHC node

[Yangbo Lu]

The eSDHC could select peripheral clock or platform clock as clock source.
In default, we use platform clock. This patch is to add clocks property
describing peripheral clock for eSDHC node. The driver could use common clk
APIs to get peripheral clock. Also add a compatible for ls1046a eSDHC node.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
---
Changes for v2:
	- None
---
 arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi b/arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi
index 4a164b8..141a513 100644
--- a/arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi
+++ b/arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi
@@ -209,10 +209,11 @@
 		};
 
 		esdhc: esdhc@1560000 {
-			compatible = "fsl,esdhc";
+			compatible = "fsl,ls1046a-esdhc", "fsl,esdhc";
 			reg = <0x0 0x1560000 0x0 0x10000>;
 			interrupts = <GIC_SPI 62 IRQ_TYPE_LEVEL_HIGH>;
 			clock-frequency = <0>;
+			clocks = <&clockgen 2 1>;
 			voltage-ranges = <1800 1800 3300 3300>;
 			sdhci,auto-cmd12;
 			big-endian;
-- 
2.1.0.27.g96db324

[v2, 6/7] arm64: dts: ls1046a: add clocks property and compatible for eSDHC node

[Yangbo Lu]

The eSDHC could select peripheral clock or platform clock as clock source.
In default, we use platform clock. This patch is to add clocks property
describing peripheral clock for eSDHC node. The driver could use common clk
APIs to get peripheral clock. Also add a compatible for ls1046a eSDHC node.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
---
Changes for v2:
	- None
---
 arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi b/arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi
index 4a164b8..141a513 100644
--- a/arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi
+++ b/arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi
@@ -209,10 +209,11 @@
 		};
 
 		esdhc: esdhc at 1560000 {
-			compatible = "fsl,esdhc";
+			compatible = "fsl,ls1046a-esdhc", "fsl,esdhc";
 			reg = <0x0 0x1560000 0x0 0x10000>;
 			interrupts = <GIC_SPI 62 IRQ_TYPE_LEVEL_HIGH>;
 			clock-frequency = <0>;
+			clocks = <&clockgen 2 1>;
 			voltage-ranges = <1800 1800 3300 3300>;
 			sdhci,auto-cmd12;
 			big-endian;
-- 
2.1.0.27.g96db324

[v2, 7/7] arm64: dts: ls1046ardb: add MMC HS200/UHS-1 modes support

[Yangbo Lu]

Add HS200/UHS-1 properties in eSDHC node to support these
speed modes in driver.

Signed-off-by: Yangbo Lu <yangbo.lu-3arQi8VN3Tc@public.gmane.org>
---
Changes for v2:
	- None
---
 arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts b/arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts
index d1ccc00..08528c2 100644
--- a/arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts
+++ b/arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts
@@ -64,6 +64,14 @@
 	};
 };
 
+&esdhc {
+	mmc-hs200-1_8v;
+	sd-uhs-sdr104;
+	sd-uhs-sdr50;
+	sd-uhs-sdr25;
+	sd-uhs-sdr12;
+};
+
 &duart0 {
 	status = "okay";
 };
-- 
2.1.0.27.g96db324

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[v2, 7/7] arm64: dts: ls1046ardb: add MMC HS200/UHS-1 modes support

[Yangbo Lu]

Add HS200/UHS-1 properties in eSDHC node to support these
speed modes in driver.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
---
Changes for v2:
	- None
---
 arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts b/arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts
index d1ccc00..08528c2 100644
--- a/arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts
+++ b/arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts
@@ -64,6 +64,14 @@
 	};
 };
 
+&esdhc {
+	mmc-hs200-1_8v;
+	sd-uhs-sdr104;
+	sd-uhs-sdr50;
+	sd-uhs-sdr25;
+	sd-uhs-sdr12;
+};
+
 &duart0 {
 	status = "okay";
 };
-- 
2.1.0.27.g96db324

RE: [v2, 0/7] Add SD UHS-I and eMMC HS200 support for eSDHC

[Y.B. Lu]

Any comments on this patchset ?
Thanks :)


Best regards,
Yangbo Lu

> -----Original Message-----
> From: Yangbo Lu [mailto:yangbo.lu@nxp.com]
> Sent: Friday, March 03, 2017 4:19 PM
> To: linux-mmc@vger.kernel.org; ulf.hansson@linaro.org; Adrian Hunter; Rob
> Herring; Mark Rutland; Catalin Marinas; Will Deacon
> Cc: devicetree@vger.kernel.org; linux-arm-kernel@lists.infradead.org;
> Xiaobo Xie; Y.B. Lu
> Subject: [v2, 0/7] Add SD UHS-I and eMMC HS200 support for eSDHC
> 
> It's complicated to support SD UHS-I and eMMC HS200 for eSDHC because
> there're many differences between eSDHC and SD/eMMC spec. Several
> differences as below must be considered:
> 1. Peripheral clock must be used instead of platform clock.
>     - eSDHC could select peripheral clock or platform clock as its clock
>       source. According to RM, UHS-I/HS200 must use peripheral clock
> since
>       it supports higher frequency than platform clock.
>     - Patch 1 and patch 6 is to support this.
> 2. Signal voltage switching requires a control circuit out of eSDHC.
>     - eSDHC supports signal voltage switch from 3.3v to 1.8v by
>       eSDHC_PROCTL[VOLT_SEL] bit. This bit changes the value of output
>       signal SDHC_VS, and there must be a control circuit out of eSDHC
>       to change the signal voltage according to SDHC_VS output signal.
>     - Patch 2 is to support this.
> 3. eSDHC uses tuning block for tuning procedure.
>     - Tuning clock control register must be configured before tuning.
>     - Patch 3 is to support this.
> 4. Delay is needed between tuning cycles for HS200 tuning.
>     - Once a patch removed mdelay between tuning cycles.
>       But eSDHC needs it.
>     - Patch 4 and patch 5 is to support this.
> 5. UHS-I/HS200 modes could be enabled in dts node.
>     - Patch 7 is to support this.
> 
> Please review and merge these patches on mmc git tree if no changes are
> required.
> 
> Yangbo Lu (7):
>   mmc: sdhci-of-esdhc: add peripheral clock support
>   mmc: sdhci-of-esdhc: add support for signal voltage switch
>   mmc: sdhci-of-esdhc: add tuning support
>   mmc: sdhci: add a quirk to restore delay in tuning
>   mmc: sdhci-of-esdhc: add delay between tuning cycles
>   arm64: dts: ls1046a: add clocks property and compatible for eSDHC node
>   arm64: dts: ls1046ardb: add MMC HS200/UHS-1 modes support
> 
>  arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts |   8 ++
>  arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi    |   3 +-
>  drivers/mmc/host/sdhci-esdhc.h                    |   7 +
>  drivers/mmc/host/sdhci-of-esdhc.c                 | 167
> +++++++++++++++++++++-
>  drivers/mmc/host/sdhci.c                          |   3 +-
>  drivers/mmc/host/sdhci.h                          |   2 +
>  6 files changed, 186 insertions(+), 4 deletions(-)
> 
> --
> 2.1.0.27.g96db324

[v2, 0/7] Add SD UHS-I and eMMC HS200 support for eSDHC

[Y.B. Lu]

Any comments on this patchset ?
Thanks :)


Best regards,
Yangbo Lu

> -----Original Message-----
> From: Yangbo Lu [mailto:yangbo.lu at nxp.com]
> Sent: Friday, March 03, 2017 4:19 PM
> To: linux-mmc at vger.kernel.org; ulf.hansson at linaro.org; Adrian Hunter; Rob
> Herring; Mark Rutland; Catalin Marinas; Will Deacon
> Cc: devicetree at vger.kernel.org; linux-arm-kernel at lists.infradead.org;
> Xiaobo Xie; Y.B. Lu
> Subject: [v2, 0/7] Add SD UHS-I and eMMC HS200 support for eSDHC
> 
> It's complicated to support SD UHS-I and eMMC HS200 for eSDHC because
> there're many differences between eSDHC and SD/eMMC spec. Several
> differences as below must be considered:
> 1. Peripheral clock must be used instead of platform clock.
>     - eSDHC could select peripheral clock or platform clock as its clock
>       source. According to RM, UHS-I/HS200 must use peripheral clock
> since
>       it supports higher frequency than platform clock.
>     - Patch 1 and patch 6 is to support this.
> 2. Signal voltage switching requires a control circuit out of eSDHC.
>     - eSDHC supports signal voltage switch from 3.3v to 1.8v by
>       eSDHC_PROCTL[VOLT_SEL] bit. This bit changes the value of output
>       signal SDHC_VS, and there must be a control circuit out of eSDHC
>       to change the signal voltage according to SDHC_VS output signal.
>     - Patch 2 is to support this.
> 3. eSDHC uses tuning block for tuning procedure.
>     - Tuning clock control register must be configured before tuning.
>     - Patch 3 is to support this.
> 4. Delay is needed between tuning cycles for HS200 tuning.
>     - Once a patch removed mdelay between tuning cycles.
>       But eSDHC needs it.
>     - Patch 4 and patch 5 is to support this.
> 5. UHS-I/HS200 modes could be enabled in dts node.
>     - Patch 7 is to support this.
> 
> Please review and merge these patches on mmc git tree if no changes are
> required.
> 
> Yangbo Lu (7):
>   mmc: sdhci-of-esdhc: add peripheral clock support
>   mmc: sdhci-of-esdhc: add support for signal voltage switch
>   mmc: sdhci-of-esdhc: add tuning support
>   mmc: sdhci: add a quirk to restore delay in tuning
>   mmc: sdhci-of-esdhc: add delay between tuning cycles
>   arm64: dts: ls1046a: add clocks property and compatible for eSDHC node
>   arm64: dts: ls1046ardb: add MMC HS200/UHS-1 modes support
> 
>  arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts |   8 ++
>  arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi    |   3 +-
>  drivers/mmc/host/sdhci-esdhc.h                    |   7 +
>  drivers/mmc/host/sdhci-of-esdhc.c                 | 167
> +++++++++++++++++++++-
>  drivers/mmc/host/sdhci.c                          |   3 +-
>  drivers/mmc/host/sdhci.h                          |   2 +
>  6 files changed, 186 insertions(+), 4 deletions(-)
> 
> --
> 2.1.0.27.g96db324

RE: [v2, 0/7] Add SD UHS-I and eMMC HS200 support for eSDHC

[Y.B. Lu]

Hi Adrian,

Could you help to review the new version patch-set?
Thanks a lot.


Best regards,
Yangbo Lu

> -----Original Message-----
> From: Y.B. Lu
> Sent: Thursday, March 09, 2017 10:23 AM
> To: linux-mmc@vger.kernel.org; ulf.hansson@linaro.org; Adrian Hunter; Rob
> Herring; Mark Rutland; Catalin Marinas; Will Deacon
> Cc: devicetree@vger.kernel.org; linux-arm-kernel@lists.infradead.org;
> Xiaobo Xie
> Subject: RE: [v2, 0/7] Add SD UHS-I and eMMC HS200 support for eSDHC
> 
> Any comments on this patchset ?
> Thanks :)
> 
> 
> Best regards,
> Yangbo Lu
> 
> > -----Original Message-----
> > From: Yangbo Lu [mailto:yangbo.lu@nxp.com]
> > Sent: Friday, March 03, 2017 4:19 PM
> > To: linux-mmc@vger.kernel.org; ulf.hansson@linaro.org; Adrian Hunter;
> > Rob Herring; Mark Rutland; Catalin Marinas; Will Deacon
> > Cc: devicetree@vger.kernel.org; linux-arm-kernel@lists.infradead.org;
> > Xiaobo Xie; Y.B. Lu
> > Subject: [v2, 0/7] Add SD UHS-I and eMMC HS200 support for eSDHC
> >
> > It's complicated to support SD UHS-I and eMMC HS200 for eSDHC because
> > there're many differences between eSDHC and SD/eMMC spec. Several
> > differences as below must be considered:
> > 1. Peripheral clock must be used instead of platform clock.
> >     - eSDHC could select peripheral clock or platform clock as its
> clock
> >       source. According to RM, UHS-I/HS200 must use peripheral clock
> > since
> >       it supports higher frequency than platform clock.
> >     - Patch 1 and patch 6 is to support this.
> > 2. Signal voltage switching requires a control circuit out of eSDHC.
> >     - eSDHC supports signal voltage switch from 3.3v to 1.8v by
> >       eSDHC_PROCTL[VOLT_SEL] bit. This bit changes the value of output
> >       signal SDHC_VS, and there must be a control circuit out of eSDHC
> >       to change the signal voltage according to SDHC_VS output signal.
> >     - Patch 2 is to support this.
> > 3. eSDHC uses tuning block for tuning procedure.
> >     - Tuning clock control register must be configured before tuning.
> >     - Patch 3 is to support this.
> > 4. Delay is needed between tuning cycles for HS200 tuning.
> >     - Once a patch removed mdelay between tuning cycles.
> >       But eSDHC needs it.
> >     - Patch 4 and patch 5 is to support this.
> > 5. UHS-I/HS200 modes could be enabled in dts node.
> >     - Patch 7 is to support this.
> >
> > Please review and merge these patches on mmc git tree if no changes
> > are required.
> >
> > Yangbo Lu (7):
> >   mmc: sdhci-of-esdhc: add peripheral clock support
> >   mmc: sdhci-of-esdhc: add support for signal voltage switch
> >   mmc: sdhci-of-esdhc: add tuning support
> >   mmc: sdhci: add a quirk to restore delay in tuning
> >   mmc: sdhci-of-esdhc: add delay between tuning cycles
> >   arm64: dts: ls1046a: add clocks property and compatible for eSDHC
> node
> >   arm64: dts: ls1046ardb: add MMC HS200/UHS-1 modes support
> >
> >  arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts |   8 ++
> >  arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi    |   3 +-
> >  drivers/mmc/host/sdhci-esdhc.h                    |   7 +
> >  drivers/mmc/host/sdhci-of-esdhc.c                 | 167
> > +++++++++++++++++++++-
> >  drivers/mmc/host/sdhci.c                          |   3 +-
> >  drivers/mmc/host/sdhci.h                          |   2 +
> >  6 files changed, 186 insertions(+), 4 deletions(-)
> >
> > --
> > 2.1.0.27.g96db324

[v2, 0/7] Add SD UHS-I and eMMC HS200 support for eSDHC

[Y.B. Lu]

Hi Adrian,

Could you help to review the new version patch-set?
Thanks a lot.


Best regards,
Yangbo Lu

> -----Original Message-----
> From: Y.B. Lu
> Sent: Thursday, March 09, 2017 10:23 AM
> To: linux-mmc at vger.kernel.org; ulf.hansson at linaro.org; Adrian Hunter; Rob
> Herring; Mark Rutland; Catalin Marinas; Will Deacon
> Cc: devicetree at vger.kernel.org; linux-arm-kernel at lists.infradead.org;
> Xiaobo Xie
> Subject: RE: [v2, 0/7] Add SD UHS-I and eMMC HS200 support for eSDHC
> 
> Any comments on this patchset ?
> Thanks :)
> 
> 
> Best regards,
> Yangbo Lu
> 
> > -----Original Message-----
> > From: Yangbo Lu [mailto:yangbo.lu at nxp.com]
> > Sent: Friday, March 03, 2017 4:19 PM
> > To: linux-mmc at vger.kernel.org; ulf.hansson at linaro.org; Adrian Hunter;
> > Rob Herring; Mark Rutland; Catalin Marinas; Will Deacon
> > Cc: devicetree at vger.kernel.org; linux-arm-kernel at lists.infradead.org;
> > Xiaobo Xie; Y.B. Lu
> > Subject: [v2, 0/7] Add SD UHS-I and eMMC HS200 support for eSDHC
> >
> > It's complicated to support SD UHS-I and eMMC HS200 for eSDHC because
> > there're many differences between eSDHC and SD/eMMC spec. Several
> > differences as below must be considered:
> > 1. Peripheral clock must be used instead of platform clock.
> >     - eSDHC could select peripheral clock or platform clock as its
> clock
> >       source. According to RM, UHS-I/HS200 must use peripheral clock
> > since
> >       it supports higher frequency than platform clock.
> >     - Patch 1 and patch 6 is to support this.
> > 2. Signal voltage switching requires a control circuit out of eSDHC.
> >     - eSDHC supports signal voltage switch from 3.3v to 1.8v by
> >       eSDHC_PROCTL[VOLT_SEL] bit. This bit changes the value of output
> >       signal SDHC_VS, and there must be a control circuit out of eSDHC
> >       to change the signal voltage according to SDHC_VS output signal.
> >     - Patch 2 is to support this.
> > 3. eSDHC uses tuning block for tuning procedure.
> >     - Tuning clock control register must be configured before tuning.
> >     - Patch 3 is to support this.
> > 4. Delay is needed between tuning cycles for HS200 tuning.
> >     - Once a patch removed mdelay between tuning cycles.
> >       But eSDHC needs it.
> >     - Patch 4 and patch 5 is to support this.
> > 5. UHS-I/HS200 modes could be enabled in dts node.
> >     - Patch 7 is to support this.
> >
> > Please review and merge these patches on mmc git tree if no changes
> > are required.
> >
> > Yangbo Lu (7):
> >   mmc: sdhci-of-esdhc: add peripheral clock support
> >   mmc: sdhci-of-esdhc: add support for signal voltage switch
> >   mmc: sdhci-of-esdhc: add tuning support
> >   mmc: sdhci: add a quirk to restore delay in tuning
> >   mmc: sdhci-of-esdhc: add delay between tuning cycles
> >   arm64: dts: ls1046a: add clocks property and compatible for eSDHC
> node
> >   arm64: dts: ls1046ardb: add MMC HS200/UHS-1 modes support
> >
> >  arch/arm64/boot/dts/freescale/fsl-ls1046a-rdb.dts |   8 ++
> >  arch/arm64/boot/dts/freescale/fsl-ls1046a.dtsi    |   3 +-
> >  drivers/mmc/host/sdhci-esdhc.h                    |   7 +
> >  drivers/mmc/host/sdhci-of-esdhc.c                 | 167
> > +++++++++++++++++++++-
> >  drivers/mmc/host/sdhci.c                          |   3 +-
> >  drivers/mmc/host/sdhci.h                          |   2 +
> >  6 files changed, 186 insertions(+), 4 deletions(-)
> >
> > --
> > 2.1.0.27.g96db324

Re: [v2, 4/7] mmc: sdhci: add a quirk to restore delay in tuning

[Adrian Hunter]

On 03/03/17 10:18, Yangbo Lu wrote:
> There was a mdelay between tuning cycles which is removed since
> eMMC spec didn't require that. However it's observed that eSDHC
> needed this delay for HS200 successful tuning. Maybe it's also
> needed for some other controllers. This patch is to add a quirk
> to fix it.
> 
> Fixes: 197160d52e85 ("mmc: sdhci: remove mdelay in eMMC tuning")
> Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
> ---

I would prefer to do something like this:

---
From: Adrian Hunter <adrian.hunter@intel.com>
Date: Thu, 23 Mar 2017 08:54:06 +0200
Subject: [PATCH] mmc: sdhci: Control the delay between tuning commands

The delay between tuning commands for SD cards is not part of the
specification. A driver that needs it probably needs it for eMMC too,
whereas most drivers would probably like to set it to 0. Make it a
host member (host->tuning_delay) that defaults to the existing behaviour.
Drivers can set it to zero to eliminate the delay, or set it to a positive
value to always have a delay.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
---
 drivers/mmc/host/sdhci.c | 11 ++++++++---
 drivers/mmc/host/sdhci.h |  1 +
 2 files changed, 9 insertions(+), 3 deletions(-)

diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c
index a33102fc800b..2fc33485df1d 100644
--- a/drivers/mmc/host/sdhci.c
+++ b/drivers/mmc/host/sdhci.c
@@ -2100,9 +2100,9 @@ static void __sdhci_execute_tuning(struct sdhci_host *host, u32 opcode)
 			break;
 		}
 
-		/* eMMC spec does not require a delay between tuning cycles */
-		if (opcode == MMC_SEND_TUNING_BLOCK)
-			mdelay(1);
+		/* Spec does not require a delay between tuning cycles */
+		if (host->tuning_delay > 0)
+			mdelay(host->tuning_delay);
 	}
 
 	pr_info("%s: Tuning failed, falling back to fixed sampling clock\n",
@@ -2164,6 +2164,9 @@ int sdhci_execute_tuning(struct mmc_host *mmc, u32 opcode)
 
 	host->mmc->retune_period = tuning_count;
 
+	if (host->tuning_delay < 0)
+		host->tuning_delay = opcode == MMC_SEND_TUNING_BLOCK;
+
 	sdhci_start_tuning(host);
 
 	__sdhci_execute_tuning(host, opcode);
@@ -3108,6 +3111,8 @@ struct sdhci_host *sdhci_alloc_host(struct device *dev,
 	host->cqe_ier     = SDHCI_CQE_INT_MASK;
 	host->cqe_err_ier = SDHCI_CQE_INT_ERR_MASK;
 
+	host->tuning_delay = -1;
+
 	return host;
 }
 
diff --git a/drivers/mmc/host/sdhci.h b/drivers/mmc/host/sdhci.h
index 35b41da0a636..b0f416fc9548 100644
--- a/drivers/mmc/host/sdhci.h
+++ b/drivers/mmc/host/sdhci.h
@@ -538,6 +538,7 @@ struct sdhci_host {
 #define SDHCI_TUNING_MODE_1	0
 #define SDHCI_TUNING_MODE_2	1
 #define SDHCI_TUNING_MODE_3	2
+	int			tuning_delay;	/* Delay (ms) between tuning commands */
 
 	unsigned long private[0] ____cacheline_aligned;
 };
-- 
1.9.1

[v2, 4/7] mmc: sdhci: add a quirk to restore delay in tuning

[Adrian Hunter]

On 03/03/17 10:18, Yangbo Lu wrote:
> There was a mdelay between tuning cycles which is removed since
> eMMC spec didn't require that. However it's observed that eSDHC
> needed this delay for HS200 successful tuning. Maybe it's also
> needed for some other controllers. This patch is to add a quirk
> to fix it.
> 
> Fixes: 197160d52e85 ("mmc: sdhci: remove mdelay in eMMC tuning")
> Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
> ---

I would prefer to do something like this:

---
From: Adrian Hunter <adrian.hunter@intel.com>
Date: Thu, 23 Mar 2017 08:54:06 +0200
Subject: [PATCH] mmc: sdhci: Control the delay between tuning commands

The delay between tuning commands for SD cards is not part of the
specification. A driver that needs it probably needs it for eMMC too,
whereas most drivers would probably like to set it to 0. Make it a
host member (host->tuning_delay) that defaults to the existing behaviour.
Drivers can set it to zero to eliminate the delay, or set it to a positive
value to always have a delay.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
---
 drivers/mmc/host/sdhci.c | 11 ++++++++---
 drivers/mmc/host/sdhci.h |  1 +
 2 files changed, 9 insertions(+), 3 deletions(-)

diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c
index a33102fc800b..2fc33485df1d 100644
--- a/drivers/mmc/host/sdhci.c
+++ b/drivers/mmc/host/sdhci.c
@@ -2100,9 +2100,9 @@ static void __sdhci_execute_tuning(struct sdhci_host *host, u32 opcode)
 			break;
 		}
 
-		/* eMMC spec does not require a delay between tuning cycles */
-		if (opcode == MMC_SEND_TUNING_BLOCK)
-			mdelay(1);
+		/* Spec does not require a delay between tuning cycles */
+		if (host->tuning_delay > 0)
+			mdelay(host->tuning_delay);
 	}
 
 	pr_info("%s: Tuning failed, falling back to fixed sampling clock\n",
@@ -2164,6 +2164,9 @@ int sdhci_execute_tuning(struct mmc_host *mmc, u32 opcode)
 
 	host->mmc->retune_period = tuning_count;
 
+	if (host->tuning_delay < 0)
+		host->tuning_delay = opcode == MMC_SEND_TUNING_BLOCK;
+
 	sdhci_start_tuning(host);
 
 	__sdhci_execute_tuning(host, opcode);
@@ -3108,6 +3111,8 @@ struct sdhci_host *sdhci_alloc_host(struct device *dev,
 	host->cqe_ier     = SDHCI_CQE_INT_MASK;
 	host->cqe_err_ier = SDHCI_CQE_INT_ERR_MASK;
 
+	host->tuning_delay = -1;
+
 	return host;
 }
 
diff --git a/drivers/mmc/host/sdhci.h b/drivers/mmc/host/sdhci.h
index 35b41da0a636..b0f416fc9548 100644
--- a/drivers/mmc/host/sdhci.h
+++ b/drivers/mmc/host/sdhci.h
@@ -538,6 +538,7 @@ struct sdhci_host {
 #define SDHCI_TUNING_MODE_1	0
 #define SDHCI_TUNING_MODE_2	1
 #define SDHCI_TUNING_MODE_3	2
+	int			tuning_delay;	/* Delay (ms) between tuning commands */
 
 	unsigned long private[0] ____cacheline_aligned;
 };
-- 
1.9.1

RE: [v2, 4/7] mmc: sdhci: add a quirk to restore delay in tuning

[Y.B. Lu]

Hi Adrian,

Thanks a lot for your suggestion.
I have sent out the new version patch-set.

Please help to review :)


Best regards,
Yangbo Lu

> -----Original Message-----
> From: Adrian Hunter [mailto:adrian.hunter@intel.com]
> Sent: Thursday, March 23, 2017 3:04 PM
> To: Y.B. Lu; linux-mmc@vger.kernel.org; ulf.hansson@linaro.org; Rob
> Herring; Mark Rutland; Catalin Marinas; Will Deacon
> Cc: devicetree@vger.kernel.org; linux-arm-kernel@lists.infradead.org;
> Xiaobo Xie
> Subject: Re: [v2, 4/7] mmc: sdhci: add a quirk to restore delay in tuning
> 
> On 03/03/17 10:18, Yangbo Lu wrote:
> > There was a mdelay between tuning cycles which is removed since eMMC
> > spec didn't require that. However it's observed that eSDHC needed this
> > delay for HS200 successful tuning. Maybe it's also needed for some
> > other controllers. This patch is to add a quirk to fix it.
> >
> > Fixes: 197160d52e85 ("mmc: sdhci: remove mdelay in eMMC tuning")
> > Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
> > ---
> 
> I would prefer to do something like this:
> 
> ---
> From: Adrian Hunter <adrian.hunter@intel.com>
> Date: Thu, 23 Mar 2017 08:54:06 +0200
> Subject: [PATCH] mmc: sdhci: Control the delay between tuning commands
> 
> The delay between tuning commands for SD cards is not part of the
> specification. A driver that needs it probably needs it for eMMC too,
> whereas most drivers would probably like to set it to 0. Make it a host
> member (host->tuning_delay) that defaults to the existing behaviour.
> Drivers can set it to zero to eliminate the delay, or set it to a
> positive value to always have a delay.
> 
> Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
> ---
>  drivers/mmc/host/sdhci.c | 11 ++++++++---  drivers/mmc/host/sdhci.h |  1
> +
>  2 files changed, 9 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c index
> a33102fc800b..2fc33485df1d 100644
> --- a/drivers/mmc/host/sdhci.c
> +++ b/drivers/mmc/host/sdhci.c
> @@ -2100,9 +2100,9 @@ static void __sdhci_execute_tuning(struct
> sdhci_host *host, u32 opcode)
>  			break;
>  		}
> 
> -		/* eMMC spec does not require a delay between tuning cycles
> */
> -		if (opcode == MMC_SEND_TUNING_BLOCK)
> -			mdelay(1);
> +		/* Spec does not require a delay between tuning cycles */
> +		if (host->tuning_delay > 0)
> +			mdelay(host->tuning_delay);
>  	}
> 
>  	pr_info("%s: Tuning failed, falling back to fixed sampling clock\n",
> @@ -2164,6 +2164,9 @@ int sdhci_execute_tuning(struct mmc_host *mmc, u32
> opcode)
> 
>  	host->mmc->retune_period = tuning_count;
> 
> +	if (host->tuning_delay < 0)
> +		host->tuning_delay = opcode == MMC_SEND_TUNING_BLOCK;
> +
>  	sdhci_start_tuning(host);
> 
>  	__sdhci_execute_tuning(host, opcode);
> @@ -3108,6 +3111,8 @@ struct sdhci_host *sdhci_alloc_host(struct device
> *dev,
>  	host->cqe_ier     = SDHCI_CQE_INT_MASK;
>  	host->cqe_err_ier = SDHCI_CQE_INT_ERR_MASK;
> 
> +	host->tuning_delay = -1;
> +
>  	return host;
>  }
> 
> diff --git a/drivers/mmc/host/sdhci.h b/drivers/mmc/host/sdhci.h index
> 35b41da0a636..b0f416fc9548 100644
> --- a/drivers/mmc/host/sdhci.h
> +++ b/drivers/mmc/host/sdhci.h
> @@ -538,6 +538,7 @@ struct sdhci_host {
>  #define SDHCI_TUNING_MODE_1	0
>  #define SDHCI_TUNING_MODE_2	1
>  #define SDHCI_TUNING_MODE_3	2
> +	int			tuning_delay;	/* Delay (ms) between tuning
> commands */
> 
>  	unsigned long private[0] ____cacheline_aligned;  };
> --
> 1.9.1
> 
> 

[v2, 4/7] mmc: sdhci: add a quirk to restore delay in tuning

[Y.B. Lu]

Hi Adrian,

Thanks a lot for your suggestion.
I have sent out the new version patch-set.

Please help to review :)


Best regards,
Yangbo Lu

> -----Original Message-----
> From: Adrian Hunter [mailto:adrian.hunter at intel.com]
> Sent: Thursday, March 23, 2017 3:04 PM
> To: Y.B. Lu; linux-mmc at vger.kernel.org; ulf.hansson at linaro.org; Rob
> Herring; Mark Rutland; Catalin Marinas; Will Deacon
> Cc: devicetree at vger.kernel.org; linux-arm-kernel at lists.infradead.org;
> Xiaobo Xie
> Subject: Re: [v2, 4/7] mmc: sdhci: add a quirk to restore delay in tuning
> 
> On 03/03/17 10:18, Yangbo Lu wrote:
> > There was a mdelay between tuning cycles which is removed since eMMC
> > spec didn't require that. However it's observed that eSDHC needed this
> > delay for HS200 successful tuning. Maybe it's also needed for some
> > other controllers. This patch is to add a quirk to fix it.
> >
> > Fixes: 197160d52e85 ("mmc: sdhci: remove mdelay in eMMC tuning")
> > Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
> > ---
> 
> I would prefer to do something like this:
> 
> ---
> From: Adrian Hunter <adrian.hunter@intel.com>
> Date: Thu, 23 Mar 2017 08:54:06 +0200
> Subject: [PATCH] mmc: sdhci: Control the delay between tuning commands
> 
> The delay between tuning commands for SD cards is not part of the
> specification. A driver that needs it probably needs it for eMMC too,
> whereas most drivers would probably like to set it to 0. Make it a host
> member (host->tuning_delay) that defaults to the existing behaviour.
> Drivers can set it to zero to eliminate the delay, or set it to a
> positive value to always have a delay.
> 
> Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
> ---
>  drivers/mmc/host/sdhci.c | 11 ++++++++---  drivers/mmc/host/sdhci.h |  1
> +
>  2 files changed, 9 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c index
> a33102fc800b..2fc33485df1d 100644
> --- a/drivers/mmc/host/sdhci.c
> +++ b/drivers/mmc/host/sdhci.c
> @@ -2100,9 +2100,9 @@ static void __sdhci_execute_tuning(struct
> sdhci_host *host, u32 opcode)
>  			break;
>  		}
> 
> -		/* eMMC spec does not require a delay between tuning cycles
> */
> -		if (opcode == MMC_SEND_TUNING_BLOCK)
> -			mdelay(1);
> +		/* Spec does not require a delay between tuning cycles */
> +		if (host->tuning_delay > 0)
> +			mdelay(host->tuning_delay);
>  	}
> 
>  	pr_info("%s: Tuning failed, falling back to fixed sampling clock\n",
> @@ -2164,6 +2164,9 @@ int sdhci_execute_tuning(struct mmc_host *mmc, u32
> opcode)
> 
>  	host->mmc->retune_period = tuning_count;
> 
> +	if (host->tuning_delay < 0)
> +		host->tuning_delay = opcode == MMC_SEND_TUNING_BLOCK;
> +
>  	sdhci_start_tuning(host);
> 
>  	__sdhci_execute_tuning(host, opcode);
> @@ -3108,6 +3111,8 @@ struct sdhci_host *sdhci_alloc_host(struct device
> *dev,
>  	host->cqe_ier     = SDHCI_CQE_INT_MASK;
>  	host->cqe_err_ier = SDHCI_CQE_INT_ERR_MASK;
> 
> +	host->tuning_delay = -1;
> +
>  	return host;
>  }
> 
> diff --git a/drivers/mmc/host/sdhci.h b/drivers/mmc/host/sdhci.h index
> 35b41da0a636..b0f416fc9548 100644
> --- a/drivers/mmc/host/sdhci.h
> +++ b/drivers/mmc/host/sdhci.h
> @@ -538,6 +538,7 @@ struct sdhci_host {
>  #define SDHCI_TUNING_MODE_1	0
>  #define SDHCI_TUNING_MODE_2	1
>  #define SDHCI_TUNING_MODE_3	2
> +	int			tuning_delay;	/* Delay (ms) between tuning
> commands */
> 
>  	unsigned long private[0] ____cacheline_aligned;  };
> --
> 1.9.1
> 
>