Skip to contentSkip to conversation

mmc: fsl_esdhc: add esdhc_imx flag

PATCH New Y.b. Lu 3 participants 9 messages Feb 15, 2019 – Feb 26, 2019 MMC, SD and SDIO

lore
Share & tools
lore

Y.b. Lu

author
#0 lore raw

PATCH v2 1/3 mmc: fsl_esdhc: add esdhc_imx flag

The fsl_esdhc driver was for Freescale eSDHC on MPC83XX/MPC85XX initially. The later QoriQ series processors (which were evolutions of MPC83XX/MPC85XX) and i.MX series processors were using this driver for their eSDHCs too.

So there are two evolution directions for eSDHC now. For the two series processors, the eSDHCs are becoming more and more different. We should have split it into two drivers, like them (sdhci-of-esdhc.c/sdhci-esdhc-imx.c) in linux kernel. But it seemed to be a lot of work now. So let's keep as it is. Be very careful to change the driver if the changes are not common for all eSDHCs, and clarify i.MX eSDHC specific things to distingush them with QorIQ eSDHC.

This patch is to added an esdhc_imx flag for the development of i.MX eSDHC, to distinguish it with QoriQ eSDHC.

---
Changes for v2:
- Converted to use device_is_compatible().

---
 drivers/mmc/fsl_esdhc.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
drivers/mmc/fsl_esdhc.c +11 -0
index 21fa2ab1d4..a647bc7f1c 100644
@@ -147,6 +147,7 @@ struct fsl_esdhc_priv {
 	struct gpio_desc cd_gpio;
 	struct gpio_desc wp_gpio;
 #endif
+	bool esdhc_imx;
 };
 
 /* Return the XFERTYP flags for a given command and data packet */
@@ -1462,6 +1463,16 @@ static int fsl_esdhc_probe(struct udevice *dev)
 		priv->caps = data->caps;
 	}
 
+	/*
+	 * This is to specify whether current eSDHC is an i.MX eSDHC,
+	 * since the i.MX eSDHC has been becoming more and more different
+	 * with QorIQ eSDHC and initial MPC83XX/MPC85XX.
+	 */
+	if (!device_is_compatible(dev, "fsl,esdhc"))
+		priv->esdhc_imx = true;
+	else
+		priv->esdhc_imx = false;
+
 	val = dev_read_u32_default(dev, "bus-width", -1);
 	if (val == 8)
 		priv->bus_width = 8;
Signature · 1 line
2.17.1
author
#1 lore raw

PATCH v2 2/3 mmc: fsl_esdhc: clean up register definition

The fsl_esdhc driver was for Freescale eSDHC on MPC83XX/MPC85XX initially. The later QoriQ series processors (which are evolutions of MPC83XX/MPC85XX) and i.MX series processors were using this driver for their eSDHCs too.

So there are two evolution directions for eSDHC now. For the two series processors, the eSDHCs are becoming more and more different. We should have split it into two drivers, like them (sdhci-of-esdhc.c/sdhci-esdhc-imx.c) in linux kernel. But it seemed to be a lot of work now. So let's keep as it is. Be very careful to change the driver if the changes are not common for all eSDHCs, and clarify i.MX eSDHC specific things to distingush them with QorIQ eSDHC.

This patch is to clean up register definition. Besides the QorIQ eSDHC register definition, the i.MX eSDHC specific registers should be clarified.

---
Changes for v2:
- Added "Acked-by: Peng Fan".

---
 drivers/mmc/fsl_esdhc.c |  66 ++++++-------
 include/fsl_esdhc.h     | 208 ++++++++++++++++++++++------------------
 2 files changed, 149 insertions(+), 125 deletions(-)
drivers/mmc/fsl_esdhc.c +33 -33
index a647bc7f1c..e9622ce9a4 100644
@@ -54,23 +54,23 @@ struct fsl_esdhc {
 	uint    autoc12err;	/* Auto CMD error status register */
 	uint    hostcapblt;	/* Host controller capabilities register */
 	uint    wml;		/* Watermark level register */
-	uint    mixctrl;	/* For USDHC */
+	uint    mixctrl;	/* For i.MX USDHC */
 	char    reserved1[4];	/* reserved */
 	uint    fevt;		/* Force event register */
 	uint    admaes;		/* ADMA error status register */
 	uint    adsaddr;	/* ADMA system address register */
 	char    reserved2[4];
-	uint    dllctrl;
-	uint    dllstat;
-	uint    clktunectrlstatus;
+	uint    dllctrl;	/* For i.MX USDHC */
+	uint    dllstat;	/* For i.MX USDHC */
+	uint    clktunectrlstatus;	/* For i.MX USDHC */
 	char    reserved3[4];
-	uint	strobe_dllctrl;
-	uint	strobe_dllstat;
+	uint	strobe_dllctrl;	/* For i.MX USDHC */
+	uint	strobe_dllstat;	/* For i.MX USDHC */
 	char    reserved4[72];
-	uint    vendorspec;
-	uint    mmcboot;
-	uint    vendorspec2;
-	uint    tuning_ctrl;	/* on i.MX6/7/8 */
+	uint    vendorspec;	/* For i.MX USDHC */
+	uint    mmcboot;	/* For i.MX USDHC */
+	uint    vendorspec2;	/* For i.MX USDHC */
+	uint    tuning_ctrl;	/* For i.MX USDHC, on i.MX6/7/8 */
 	char	reserved5[44];
 	uint    hostver;	/* Host controller version register */
 	char    reserved6[4];	/* reserved */
@@ -80,11 +80,11 @@ struct fsl_esdhc {
 	char    reserved8[4];	/* reserved */
 	uint    hostcapblt2;	/* Host controller capabilities register 2 */
 	char    reserved9[8];	/* reserved */
-	uint    tcr;		/* Tuning control register */
+	uint    tbctl;		/* Tuning block control register */
 	char    reserved10[28];	/* reserved */
 	uint    sddirctl;	/* SD direction control register */
 	char    reserved11[712];/* reserved */
-	uint    scr;		/* eSDHC control register */
+	uint    esdhcctl;	/* eSDHC control register */
 };
 
 struct fsl_esdhc_plat {
@@ -496,7 +496,7 @@ static int esdhc_send_cmd_common(struct fsl_esdhc_priv *priv, struct mmc *mmc,
 
 	/* Switch voltage to 1.8V if CMD11 succeeded */
 	if (cmd->cmdidx == SD_CMD_SWITCH_UHS18V) {
-		esdhc_setbits32(&regs->vendorspec, ESDHC_VENDORSPEC_VSELECT);
+		esdhc_setbits32(&regs->vendorspec, VENDORSPEC_VSELECT);
 
 		printf("Run CMD11 1.8V switch\n");
 		/* Sleep for 5 ms - max time for card to switch to 1.8V */
@@ -728,23 +728,23 @@ static void esdhc_set_strobe_dll(struct mmc *mmc)
 	struct fsl_esdhc *regs = priv->esdhc_regs;
 	u32 val;
 
-	if (priv->clock > ESDHC_STROBE_DLL_CLK_FREQ) {
-		writel(ESDHC_STROBE_DLL_CTRL_RESET, &regs->strobe_dllctrl);
+	if (priv->clock > STROBE_DLL_CLK_FREQ) {
+		writel(STROBE_DLL_CTRL_RESET, &regs->strobe_dllctrl);
 
 		/*
 		 * enable strobe dll ctrl and adjust the delay target
 		 * for the uSDHC loopback read clock
 		 */
-		val = ESDHC_STROBE_DLL_CTRL_ENABLE |
+		val = STROBE_DLL_CTRL_ENABLE |
 			(priv->strobe_dll_delay_target <<
-			 ESDHC_STROBE_DLL_CTRL_SLV_DLY_TARGET_SHIFT);
+			 STROBE_DLL_CTRL_SLV_DLY_TARGET_SHIFT);
 		writel(val, &regs->strobe_dllctrl);
 		/* wait 1us to make sure strobe dll status register stable */
 		mdelay(1);
 		val = readl(&regs->strobe_dllstat);
-		if (!(val & ESDHC_STROBE_DLL_STS_REF_LOCK))
+		if (!(val & STROBE_DLL_STS_REF_LOCK))
 			pr_warn("HS400 strobe DLL status REF not lock!\n");
-		if (!(val & ESDHC_STROBE_DLL_STS_SLV_LOCK))
+		if (!(val & STROBE_DLL_STS_SLV_LOCK))
 			pr_warn("HS400 strobe DLL status SLV not lock!\n");
 	}
 }
@@ -817,9 +817,9 @@ static int esdhc_set_voltage(struct mmc *mmc)
 		}
 #endif
 
-		esdhc_clrbits32(&regs->vendorspec, ESDHC_VENDORSPEC_VSELECT);
+		esdhc_clrbits32(&regs->vendorspec, VENDORSPEC_VSELECT);
 		if (!(esdhc_read32(&regs->vendorspec) &
-		    ESDHC_VENDORSPEC_VSELECT))
+		    VENDORSPEC_VSELECT))
 			return 0;
 
 		return -EAGAIN;
@@ -833,8 +833,8 @@ static int esdhc_set_voltage(struct mmc *mmc)
 			}
 		}
 #endif
-		esdhc_setbits32(&regs->vendorspec, ESDHC_VENDORSPEC_VSELECT);
-		if (esdhc_read32(&regs->vendorspec) & ESDHC_VENDORSPEC_VSELECT)
+		esdhc_setbits32(&regs->vendorspec, VENDORSPEC_VSELECT);
+		if (esdhc_read32(&regs->vendorspec) & VENDORSPEC_VSELECT)
 			return 0;
 
 		return -EAGAIN;
@@ -952,7 +952,7 @@ static int esdhc_set_ios_common(struct fsl_esdhc_priv *priv, struct mmc *mmc)
 #ifdef CONFIG_FSL_ESDHC_USE_PERIPHERAL_CLK
 	/* Select to use peripheral clock */
 	esdhc_clock_control(priv, false);
-	esdhc_setbits32(&regs->scr, ESDHCCTL_PCS);
+	esdhc_setbits32(&regs->esdhcctl, ESDHCCTL_PCS);
 	esdhc_clock_control(priv, true);
 #endif
 	/* Set the clock speed */
@@ -1028,7 +1028,7 @@ static int esdhc_init_common(struct fsl_esdhc_priv *priv, struct mmc *mmc)
 	/* Put VEND_SPEC to default value */
 	if (priv->vs18_enable)
 		esdhc_write32(&regs->vendorspec, (VENDORSPEC_INIT |
-			      ESDHC_VENDORSPEC_VSELECT));
+			      VENDORSPEC_VSELECT));
 	else
 		esdhc_write32(&regs->vendorspec, VENDORSPEC_INIT);
 
@@ -1038,7 +1038,7 @@ static int esdhc_init_common(struct fsl_esdhc_priv *priv, struct mmc *mmc)
 
 #ifndef ARCH_MXC
 	/* Enable cache snooping */
-	esdhc_write32(&regs->scr, 0x00000040);
+	esdhc_write32(&regs->esdhcctl, 0x00000040);
 #endif
 
 #ifndef CONFIG_FSL_USDHC
@@ -1184,7 +1184,7 @@ static int fsl_esdhc_init(struct fsl_esdhc_priv *priv,
 #endif
 
 	if (priv->vs18_enable)
-		esdhc_setbits32(&regs->vendorspec, ESDHC_VENDORSPEC_VSELECT);
+		esdhc_setbits32(&regs->vendorspec, VENDORSPEC_VSELECT);
 
 	writel(SDHCI_IRQ_EN_BITS, &regs->irqstaten);
 	cfg = &plat->cfg;
@@ -1272,11 +1272,11 @@ static int fsl_esdhc_init(struct fsl_esdhc_priv *priv,
 		if (priv->flags & ESDHC_FLAG_STD_TUNING) {
 			u32 val = readl(&regs->tuning_ctrl);
 
-			val |= ESDHC_STD_TUNING_EN;
-			val &= ~ESDHC_TUNING_START_TAP_MASK;
+			val |= STD_TUNING_EN;
+			val &= ~TUNING_START_TAP_MASK;
 			val |= priv->tuning_start_tap;
-			val &= ~ESDHC_TUNING_STEP_MASK;
-			val |= (priv->tuning_step) << ESDHC_TUNING_STEP_SHIFT;
+			val &= ~TUNING_STEP_MASK;
+			val |= (priv->tuning_step) << TUNING_STEP_SHIFT;
 			writel(val, &regs->tuning_ctrl);
 		}
 	}
@@ -1484,10 +1484,10 @@ static int fsl_esdhc_probe(struct udevice *dev)
 	val = fdtdec_get_int(fdt, node, "fsl,tuning-step", 1);
 	priv->tuning_step = val;
 	val = fdtdec_get_int(fdt, node, "fsl,tuning-start-tap",
-			     ESDHC_TUNING_START_TAP_DEFAULT);
+			     TUNING_START_TAP_DEFAULT);
 	priv->tuning_start_tap = val;
 	val = fdtdec_get_int(fdt, node, "fsl,strobe-dll-delay-target",
-			     ESDHC_STROBE_DLL_CTRL_SLV_DLY_TARGET_DEFAULT);
+			     STROBE_DLL_CTRL_SLV_DLY_TARGET_DEFAULT);
 	priv->strobe_dll_delay_target = val;
 
 	if (dev_read_bool(dev, "non-removable")) {
include/fsl_esdhc.h +116 -92

265 lines not on this page · raw message

Signature · 1 line
2.17.1
author
#2 lore raw

PATCH v2 3/3 mmc: fsl_esdhc: clarify i.MX eSDHC specific functions

In reply to Y.b. Lu

A previous patch had added SDR104/HS200 support for fsl_esdhc. However this was only for i.MX eSDHC, and QorIQ eSDHC used different registers and method.

This patch is to clarify i.MX eSDHC specific functions defined in that patch, and to use them only for i.MX eSDHC. The QorIQ eSDHC SDR104/HS200 support will be added in the future.

Fixes 51313b49f2b8 (mmc: fsl_esdhc: support SDR104 and HS200)

---
Changes for v2:
- Added "Reviewed-by: Peng Fan".

---
 drivers/mmc/fsl_esdhc.c | 79 +++++++++++++++++++++--------------------
 1 file changed, 41 insertions(+), 38 deletions(-)
drivers/mmc/fsl_esdhc.c +41 -38
index e9622ce9a4..92924b63ac 100644
@@ -682,7 +682,7 @@ static void esdhc_clock_control(struct fsl_esdhc_priv *priv, bool enable)
 #endif
 
 #ifdef MMC_SUPPORTS_TUNING
-static int esdhc_change_pinstate(struct udevice *dev)
+static int esdhc_imx_change_pinstate(struct udevice *dev)
 {
 	struct fsl_esdhc_priv *priv = dev_get_priv(dev);
 	int ret;
@@ -708,7 +708,7 @@ static int esdhc_change_pinstate(struct udevice *dev)
 	return ret;
 }
 
-static void esdhc_reset_tuning(struct mmc *mmc)
+static void esdhc_imx_reset_tuning(struct mmc *mmc)
 {
 	struct fsl_esdhc_priv *priv = dev_get_priv(mmc->dev);
 	struct fsl_esdhc *regs = priv->esdhc_regs;
@@ -722,7 +722,7 @@ static void esdhc_reset_tuning(struct mmc *mmc)
 	}
 }
 
-static void esdhc_set_strobe_dll(struct mmc *mmc)
+static void esdhc_imx_set_strobe_dll(struct mmc *mmc)
 {
 	struct fsl_esdhc_priv *priv = dev_get_priv(mmc->dev);
 	struct fsl_esdhc *regs = priv->esdhc_regs;
@@ -749,7 +749,7 @@ static void esdhc_set_strobe_dll(struct mmc *mmc)
 	}
 }
 
-static int esdhc_set_timing(struct mmc *mmc)
+static int esdhc_imx_set_timing(struct mmc *mmc)
 {
 	struct fsl_esdhc_priv *priv = dev_get_priv(mmc->dev);
 	struct fsl_esdhc *regs = priv->esdhc_regs;
@@ -761,13 +761,13 @@ static int esdhc_set_timing(struct mmc *mmc)
 	switch (mmc->selected_mode) {
 	case MMC_LEGACY:
 	case SD_LEGACY:
-		esdhc_reset_tuning(mmc);
+		esdhc_imx_reset_tuning(mmc);
 		writel(mixctrl, &regs->mixctrl);
 		break;
 	case MMC_HS_400:
 		mixctrl |= MIX_CTRL_DDREN | MIX_CTRL_HS400_EN;
 		writel(mixctrl, &regs->mixctrl);
-		esdhc_set_strobe_dll(mmc);
+		esdhc_imx_set_strobe_dll(mmc);
 		break;
 	case MMC_HS:
 	case MMC_HS_52:
@@ -791,14 +791,13 @@ static int esdhc_set_timing(struct mmc *mmc)
 
 	priv->mode = mmc->selected_mode;
 
-	return esdhc_change_pinstate(mmc->dev);
+	return esdhc_imx_change_pinstate(mmc->dev);
 }
 
-static int esdhc_set_voltage(struct mmc *mmc)
+static int esdhc_imx_set_voltage(struct mmc *mmc)
 {
 	struct fsl_esdhc_priv *priv = dev_get_priv(mmc->dev);
 	struct fsl_esdhc *regs = priv->esdhc_regs;
-	int ret;
 
 	priv->signal_voltage = mmc->signal_voltage;
 	switch (mmc->signal_voltage) {
@@ -975,18 +974,18 @@ static int esdhc_set_ios_common(struct fsl_esdhc_priv *priv, struct mmc *mmc)
 #endif
 	}
 
-	if (priv->mode != mmc->selected_mode) {
-		ret = esdhc_set_timing(mmc);
+	if (priv->esdhc_imx && priv->mode != mmc->selected_mode) {
+		ret = esdhc_imx_set_timing(mmc);
 		if (ret) {
-			printf("esdhc_set_timing error %d\n", ret);
+			printf("esdhc_imx_set_timing error %d\n", ret);
 			return ret;
 		}
 	}
 
-	if (priv->signal_voltage != mmc->signal_voltage) {
-		ret = esdhc_set_voltage(mmc);
+	if (priv->esdhc_imx && priv->signal_voltage != mmc->signal_voltage) {
+		ret = esdhc_imx_set_voltage(mmc);
 		if (ret) {
-			printf("esdhc_set_voltage error %d\n", ret);
+			printf("esdhc_imx_set_voltage error %d\n", ret);
 			return ret;
 		}
 	}
@@ -1267,20 +1266,21 @@ static int fsl_esdhc_init(struct fsl_esdhc_priv *priv,
 
 	cfg->b_max = CONFIG_SYS_MMC_MAX_BLK_COUNT;
 
-	writel(0, &regs->dllctrl);
-	if (priv->flags & ESDHC_FLAG_USDHC) {
-		if (priv->flags & ESDHC_FLAG_STD_TUNING) {
-			u32 val = readl(&regs->tuning_ctrl);
-
-			val |= STD_TUNING_EN;
-			val &= ~TUNING_START_TAP_MASK;
-			val |= priv->tuning_start_tap;
-			val &= ~TUNING_STEP_MASK;
-			val |= (priv->tuning_step) << TUNING_STEP_SHIFT;
-			writel(val, &regs->tuning_ctrl);
+	if (priv->esdhc_imx) {
+		writel(0, &regs->dllctrl);
+		if (priv->flags & ESDHC_FLAG_USDHC) {
+			if (priv->flags & ESDHC_FLAG_STD_TUNING) {
+				u32 val = readl(&regs->tuning_ctrl);
+
+				val |= STD_TUNING_EN;
+				val &= ~TUNING_START_TAP_MASK;
+				val |= priv->tuning_start_tap;
+				val &= ~TUNING_STEP_MASK;
+				val |= priv->tuning_step << TUNING_STEP_SHIFT;
+				writel(val, &regs->tuning_ctrl);
+			}
 		}
 	}
-
 	return 0;
 }
 
@@ -1481,14 +1481,17 @@ static int fsl_esdhc_probe(struct udevice *dev)
 	else
 		priv->bus_width = 1;
 
-	val = fdtdec_get_int(fdt, node, "fsl,tuning-step", 1);
-	priv->tuning_step = val;
-	val = fdtdec_get_int(fdt, node, "fsl,tuning-start-tap",
-			     TUNING_START_TAP_DEFAULT);
-	priv->tuning_start_tap = val;
-	val = fdtdec_get_int(fdt, node, "fsl,strobe-dll-delay-target",
-			     STROBE_DLL_CTRL_SLV_DLY_TARGET_DEFAULT);
-	priv->strobe_dll_delay_target = val;
+	if (priv->esdhc_imx) {
+		priv->tuning_step =
+			fdtdec_get_int(fdt, node, "fsl,tuning-step", 1);
+		priv->tuning_start_tap =
+			fdtdec_get_int(fdt, node, "fsl,tuning-start-tap",
+				       TUNING_START_TAP_DEFAULT);
+		priv->strobe_dll_delay_target =
+			fdtdec_get_int(fdt, node,
+				       "fsl,strobe-dll-delay-target",
+				       STROBE_DLL_CTRL_SLV_DLY_TARGET_DEFAULT);
+	}
 
 	if (dev_read_bool(dev, "non-removable")) {
 		priv->non_removable = 1;
@@ -1628,9 +1631,9 @@ static const struct dm_mmc_ops fsl_esdhc_ops = {
 #endif
 
 static struct esdhc_soc_data usdhc_imx7d_data = {
-	.flags = ESDHC_FLAG_USDHC | ESDHC_FLAG_STD_TUNING
-			| ESDHC_FLAG_HAVE_CAP1 | ESDHC_FLAG_HS200
-			| ESDHC_FLAG_HS400,
+	.flags = ESDHC_FLAG_USDHC | ESDHC_FLAG_STD_TUNING |
+		 ESDHC_FLAG_HAVE_CAP1 | ESDHC_FLAG_HS200 |
+		 ESDHC_FLAG_HS400,
 	.caps = UHS_CAPS | MMC_MODE_HS200 | MMC_MODE_DDR_52MHz |
 		MMC_MODE_HS_52MHz | MMC_MODE_HS,
 };
Signature · 1 line
2.17.1
#3 lore raw

PATCH v2 1/3 mmc: fsl_esdhc: add esdhc_imx flag

In reply to Y.b. Lu
Quoted 54 lines

-----Original Message-----
From: Y.b. Lu
Sent: 2019年2月15日 10:26
To: u-boot at lists.denx.de
Cc: Jaehoon Chung <[email protected]>; Prabhakar Kushwaha <[email protected]>; Peng Fan <[email protected]>; Y.b. Lu <[email protected]>
Subject: [v2, 1/3] mmc: fsl_esdhc: add esdhc_imx flag

The fsl_esdhc driver was for Freescale eSDHC on MPC83XX/MPC85XX initially. The later QoriQ series processors (which were evolutions of MPC83XX/MPC85XX) and i.MX series processors were using this driver for their eSDHCs too.

So there are two evolution directions for eSDHC now. For the two series processors, the eSDHCs are becoming more and more different.
We should have split it into two drivers, like them (sdhci-of-esdhc.c/sdhci-esdhc-imx.c) in linux kernel. But it seemed to be a lot of work now. So let's keep as it is. Be very careful to change the driver if the changes are not common for all eSDHCs, and clarify i.MX eSDHC specific things to distingush them with QorIQ eSDHC.

This patch is to added an esdhc_imx flag for the development of i.MX eSDHC, to distinguish it with QoriQ eSDHC.

---
Changes for v2:
- Converted to use device_is_compatible().

---
 drivers/mmc/fsl_esdhc.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
drivers/mmc/fsl_esdhc.c index renamed from drivers/mmc/fsl_esdhc.c +0 -0

21fa2ab1d4..a647bc7f1c 100644

drivers/mmc/fsl_esdhc.c +4 -1
@@ -147,6 +147,7 @@ struct fsl_esdhc_priv {
 	struct gpio_desc cd_gpio;
 	struct gpio_desc wp_gpio;
 #endif
+	bool esdhc_imx;
 };
 
 /* Return the XFERTYP flags for a given command and data packet */ @@
-1462,6 +1463,16 @@ static int fsl_esdhc_probe(struct udevice *dev)
 		priv->caps = data->caps;
 	}
 
+	/*
+	 * This is to specify whether current eSDHC is an i.MX eSDHC,
+	 * since the i.MX eSDHC has been becoming more and more different

drivers/mmc/fsl_esdhc.c struct fsl_esdhc_priv { · line 1478

Quoted 10 lines
drivers/mmc/fsl_esdhc.c +7 -0
+	 * with QorIQ eSDHC and initial MPC83XX/MPC85XX.
+	 */
+	if (!device_is_compatible(dev, "fsl,esdhc"))
+		priv->esdhc_imx = true;
+	else
+		priv->esdhc_imx = false;
+
 	val = dev_read_u32_default(dev, "bus-width", -1);
 	if (val == 8)
 		priv->bus_width = 8;
Quoted 2 lines

--
2.17.1

#4 lore raw

PATCH v2 1/3 mmc: fsl_esdhc: add esdhc_imx flag

In reply to Y.b. Lu

Dear "Y.b. Lu",

Quoted 9 lines

The fsl_esdhc driver was for Freescale eSDHC on MPC83XX/MPC85XX initially. The later QoriQ series processors (which were evolutions of MPC83XX/MPC85XX) and i.MX series processors were using this driver for their eSDHCs too.

So there are two evolution directions for eSDHC now. For the two series processors, the eSDHCs are becoming more and more different. We should have split it into two drivers, like them (sdhci-of-esdhc.c/sdhci-esdhc-imx.c) in linux kernel.

Why we cannot do it right just from the very beginning?

In the end of the day we will try to mimic Linux kernel anyway, IMHO it is better to start the split now and save some effort on a half-step solution.

Quoted 37 lines

But it seemed
to be a lot of work now. So let's keep as it is. Be very careful to change the driver if the changes are not common for all eSDHCs, and clarify i.MX eSDHC specific things to distingush them with QorIQ eSDHC.

This patch is to added an esdhc_imx flag for the development of i.MX eSDHC, to distinguish it with QoriQ eSDHC.

---
Changes for v2:
- Converted to use device_is_compatible().

---
 drivers/mmc/fsl_esdhc.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
drivers/mmc/fsl_esdhc.c +4 -0
index 21fa2ab1d4..a647bc7f1c 100644
@@ -147,6 +147,7 @@ struct fsl_esdhc_priv {
 	struct gpio_desc cd_gpio;
 	struct gpio_desc wp_gpio;
 #endif
+	bool esdhc_imx;
 };
 
 /* Return the XFERTYP flags for a given command and data packet */
@@ -1462,6 +1463,16 @@ static int fsl_esdhc_probe(struct udevice *dev)
 		priv->caps = data->caps;
 	}
 
+	/*
+	 * This is to specify whether current eSDHC is an i.MX eSDHC,
+	 * since the i.MX eSDHC has been becoming more and more

different

drivers/mmc/fsl_esdhc.c static int fsl_esdhc_probe(struct udevice *dev) · line 1478

Quoted 10 lines
drivers/mmc/fsl_esdhc.c +7 -0
+	 * with QorIQ eSDHC and initial MPC83XX/MPC85XX.
+	 */
+	if (!device_is_compatible(dev, "fsl,esdhc"))
+		priv->esdhc_imx = true;
+	else
+		priv->esdhc_imx = false;
+
 	val = dev_read_u32_default(dev, "bus-width", -1);
 	if (val == 8)
 		priv->bus_width = 8;

Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH, Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email: lukma at denx.de -------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 488 bytes
Desc: OpenPGP digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20190218/9305e354/attachment.sig>

author
#5 lore raw

PATCH v2 1/3 mmc: fsl_esdhc: add esdhc_imx flag

Hi Lukasz,

Quoted 19 lines

-----Original Message-----
From: Lukasz Majewski <[email protected]>
Sent: Monday, February 18, 2019 7:12 AM
To: Y.b. Lu <[email protected]>
Cc: u-boot at lists.denx.de
Subject: Re: [U-Boot] [v2, 1/3] mmc: fsl_esdhc: add esdhc_imx flag

Dear "Y.b. Lu",

Quoted 9 lines

The fsl_esdhc driver was for Freescale eSDHC on MPC83XX/MPC85XX initially. The later QoriQ series processors (which were evolutions of MPC83XX/MPC85XX) and i.MX series processors were using this driver for their eSDHCs too.

So there are two evolution directions for eSDHC now. For the two series processors, the eSDHCs are becoming more and more different. We should have split it into two drivers, like them (sdhci-of-esdhc.c/sdhci-esdhc-imx.c) in linux kernel.

Quoted 5 lines

Why we cannot do it right just from the very beginning?

In the end of the day we will try to mimic Linux kernel anyway, IMHO it is better to start the split now and save some effort on a half-step solution.

[Y.b. Lu] I understand. The perfect way is to split them.
However, if you grep 'CONFIG_FSL_ESDHC' in the whole u-boot source, you would find there are 607 results. There will be so many companies boards affected. I just don't know who could make all of these boards tested.

My current patch-set is also to cleaning up the driver waiting for splitting them someday on the other hand.
After you check CONFIG_FSL_ESDHC in u-boot source, if you think it's better we should split them right now, I could work on the driver splitting.

Thanks a lot.

Quoted 61 lines
Quoted 37 lines

But it seemed
to be a lot of work now. So let's keep as it is. Be very careful to change the driver if the changes are not common for all eSDHCs, and clarify i.MX eSDHC specific things to distingush them with QorIQ eSDHC.

This patch is to added an esdhc_imx flag for the development of i.MX eSDHC, to distinguish it with QoriQ eSDHC.

---
Changes for v2:
- Converted to use device_is_compatible().

---
 drivers/mmc/fsl_esdhc.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
drivers/mmc/fsl_esdhc.c index renamed from drivers/mmc/fsl_esdhc.c +0 -0

21fa2ab1d4..a647bc7f1c 100644

drivers/mmc/fsl_esdhc.c +4 -1
@@ -147,6 +147,7 @@ struct fsl_esdhc_priv {
 	struct gpio_desc cd_gpio;
 	struct gpio_desc wp_gpio;
 #endif
+	bool esdhc_imx;
 };
 
 /* Return the XFERTYP flags for a given command and data packet */ @@
-1462,6 +1463,16 @@ static int fsl_esdhc_probe(struct udevice *dev)
 		priv->caps = data->caps;
 	}
 
+	/*
+	 * This is to specify whether current eSDHC is an i.MX eSDHC,
+	 * since the i.MX eSDHC has been becoming more and more

different

drivers/mmc/fsl_esdhc.c static int fsl_esdhc_probe(struct udevice *dev) · line 1478

Quoted 10 lines
drivers/mmc/fsl_esdhc.c +7 -0
+	 * with QorIQ eSDHC and initial MPC83XX/MPC85XX.
+	 */
+	if (!device_is_compatible(dev, "fsl,esdhc"))
+		priv->esdhc_imx = true;
+	else
+		priv->esdhc_imx = false;
+
 	val = dev_read_u32_default(dev, "bus-width", -1);
 	if (val == 8)
 		priv->bus_width = 8;

Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH, Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email:
lukma at denx.de

#6 lore raw

PATCH v2 1/3 mmc: fsl_esdhc: add esdhc_imx flag

Hi "Y.b. Lu",

Quoted 23 lines

Hi Lukasz,

Quoted 21 lines

-----Original Message-----
From: Lukasz Majewski <[email protected]>
Sent: Monday, February 18, 2019 7:12 AM
To: Y.b. Lu <[email protected]>
Cc: u-boot at lists.denx.de
Subject: Re: [U-Boot] [v2, 1/3] mmc: fsl_esdhc: add esdhc_imx flag

Dear "Y.b. Lu",

Quoted 9 lines

The fsl_esdhc driver was for Freescale eSDHC on MPC83XX/MPC85XX initially. The later QoriQ series processors (which were evolutions of MPC83XX/MPC85XX) and i.MX series processors were using this driver for their eSDHCs too.

So there are two evolution directions for eSDHC now. For the two series processors, the eSDHCs are becoming more and more different. We should have split it into two drivers, like them (sdhci-of-esdhc.c/sdhci-esdhc-imx.c) in linux kernel.

Why we cannot do it right just from the very beginning?

Quoted 7 lines
Quoted 3 lines

In the end of the day we will try to mimic Linux kernel anyway, IMHO it is better to start the split now and save some effort on a half-step solution.

[Y.b. Lu] I understand. The perfect way is to split them. However, if you grep 'CONFIG_FSL_ESDHC' in the whole u-boot source, you would find there are 607 results.

Those are boards, which use the driver. The modification would be done in one or two files (fsl_esdhc.c|h).

Quoted 2 lines

There will be so many companies
boards affected.

I guess that the task of your patch set is to separate imx6q and ppc specific code for those IP blocks.

Quoted 2 lines

I just don't know who could make all of these boards tested.

Your patch set also makes some changes in the generic driver depending on the priv->esdhc_imx flag. Those changes also need to be tested on all before mentioned boards.

In the end you logically split the driver,
having it in a single file, which IMHO is not proper long-term solution.

Quoted 5 lines

My current patch-set is also to cleaning up the driver waiting for splitting them someday on the other hand. After you check CONFIG_FSL_ESDHC in u-boot source, if you think it's better we should split them right now, I could work on the driver splitting.

You can think about having common part (in one file - fsl_esdhc.c) and then separate files with code specific to imx and ppc. This would reduce the number of changes needed.

Quoted 65 lines

Thanks a lot.

Quoted 62 lines
Quoted 38 lines

But it seemed
to be a lot of work now. So let's keep as it is. Be very careful to change the driver if the changes are not common for all eSDHCs, and clarify i.MX eSDHC specific things to distingush them with QorIQ eSDHC.

This patch is to added an esdhc_imx flag for the development of i.MX eSDHC, to distinguish it with QoriQ eSDHC.

---
Changes for v2:
- Converted to use device_is_compatible().

---
 drivers/mmc/fsl_esdhc.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
drivers/mmc/fsl_esdhc.c +1 -0
index 21fa2ab1d4..a647bc7f1c 100644
@@ -147,6 +147,7 @@ struct fsl_esdhc_priv {
 	struct gpio_desc cd_gpio;
 	struct gpio_desc wp_gpio;
 #endif
+	bool esdhc_imx;
 };
 
 /* Return the XFERTYP flags for a given command and data packet

*/ @@ -1462,6 +1463,16 @@ static int fsl_esdhc_probe(struct udevice *dev) priv->caps = data->caps;

drivers/mmc/fsl_esdhc.c +2 -0
 	}
 
+	/*
+	 * This is to specify whether current eSDHC is an i.MX

eSDHC,

drivers/mmc/fsl_esdhc.c +1 -0
+	 * since the i.MX eSDHC has been becoming more and more

different

drivers/mmc/fsl_esdhc.c static int fsl_esdhc_probe(struct udevice *dev) · line 1478

Quoted 10 lines
drivers/mmc/fsl_esdhc.c +7 -0
+	 * with QorIQ eSDHC and initial MPC83XX/MPC85XX.
+	 */
+	if (!device_is_compatible(dev, "fsl,esdhc"))
+		priv->esdhc_imx = true;
+	else
+		priv->esdhc_imx = false;
+
 	val = dev_read_u32_default(dev, "bus-width", -1);
 	if (val == 8)
 		priv->bus_width = 8;  

Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH, Managing Director: Wolfgang
Denk HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email:
lukma at denx.de

Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH, Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email: lukma at denx.de -------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 488 bytes
Desc: OpenPGP digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20190219/4cb801b8/attachment.sig>

author
#7 lore raw

PATCH v2 1/3 mmc: fsl_esdhc: add esdhc_imx flag

Hi Lukasz,

Quoted 36 lines

-----Original Message-----
From: Lukasz Majewski <[email protected]>
Sent: Tuesday, February 19, 2019 2:39 PM
To: Y.b. Lu <[email protected]>
Cc: u-boot at lists.denx.de
Subject: Re: [U-Boot] [v2, 1/3] mmc: fsl_esdhc: add esdhc_imx flag

Hi "Y.b. Lu",

Quoted 27 lines

Hi Lukasz,

Quoted 24 lines

-----Original Message-----
From: Lukasz Majewski <[email protected]>
Sent: Monday, February 18, 2019 7:12 AM
To: Y.b. Lu <[email protected]>
Cc: u-boot at lists.denx.de
Subject: Re: [U-Boot] [v2, 1/3] mmc: fsl_esdhc: add esdhc_imx flag

Dear "Y.b. Lu",

Quoted 9 lines

The fsl_esdhc driver was for Freescale eSDHC on MPC83XX/MPC85XX initially. The later QoriQ series processors (which were evolutions of MPC83XX/MPC85XX) and i.MX series processors were using this driver for their eSDHCs too.

So there are two evolution directions for eSDHC now. For the two series processors, the eSDHCs are becoming more and more different. We should have split it into two drivers, like them (sdhci-of-esdhc.c/sdhci-esdhc-imx.c) in linux kernel.

Why we cannot do it right just from the very beginning?

In the end of the day we will try to mimic Linux kernel anyway, IMHO it is better to start the split now and save some effort on a half-step solution.

Quoted 6 lines
Quoted 3 lines

[Y.b. Lu] I understand. The perfect way is to split them. However, if you grep 'CONFIG_FSL_ESDHC' in the whole u-boot source, you would find there are 607 results.

Those are boards, which use the driver. The modification would be done in one or two files (fsl_esdhc.c|h).

[Y.b. Lu] Sorry, I can't get your point.
Since we split fsl_esdhc into fsl_esdhc and fsl_esdhc_imx driver, all i.mx boards files should convert to use new CONFIG_FSL_ESDHC_IMX instead as I understand.

Quoted 4 lines
Quoted 3 lines

There will be so many companies
boards affected.

Quoted 8 lines

I guess that the task of your patch set is to separate imx6q and ppc specific code for those IP blocks.

Quoted 1 line

I just don't know who could make all of these boards tested.

Your patch set also makes some changes in the generic driver depending on the priv->esdhc_imx flag. Those changes also need to be tested on all before mentioned boards.

[Y.b. Lu] Right...

Quoted 5 lines

In the end you logically split the driver, having it in a single file, which IMHO is not proper long-term solution.

Quoted 1 line
Quoted 8 lines
Quoted 4 lines

My current patch-set is also to cleaning up the driver waiting for splitting them someday on the other hand. After you check CONFIG_FSL_ESDHC in u-boot source, if you think it's better we should split them right now, I could work on the driver splitting.

You can think about having common part (in one file - fsl_esdhc.c) and then separate files with code specific to imx and ppc. This would reduce the number of changes needed.

[Y.b. Lu] Actually, it's not only imx and ppc.
In linux, esdhc driver is for MPC83XX/MPC85XX, QorIQ PowerPC processors, QoriQ ARM processors (Layerscape). Esdhc-imx driver is for i.MX processors.

Anyway there would be a lot files changed if we split the driver. Let me send out a patch-set to split them for your reviewing. Thanks a lot for your comments.

Quoted 80 lines
Quoted 66 lines

Thanks a lot.

Quoted 62 lines
Quoted 38 lines

But it seemed
to be a lot of work now. So let's keep as it is. Be very careful to change the driver if the changes are not common for all eSDHCs, and clarify i.MX eSDHC specific things to distingush them with QorIQ eSDHC.

This patch is to added an esdhc_imx flag for the development of i.MX eSDHC, to distinguish it with QoriQ eSDHC.

---
Changes for v2:
- Converted to use device_is_compatible().

---
 drivers/mmc/fsl_esdhc.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
drivers/mmc/fsl_esdhc.c +1 -0
index 21fa2ab1d4..a647bc7f1c 100644
@@ -147,6 +147,7 @@ struct fsl_esdhc_priv {
 	struct gpio_desc cd_gpio;
 	struct gpio_desc wp_gpio;
 #endif
+	bool esdhc_imx;
 };
 
 /* Return the XFERTYP flags for a given command and data packet

*/ @@ -1462,6 +1463,16 @@ static int fsl_esdhc_probe(struct udevice *dev) priv->caps = data->caps;

drivers/mmc/fsl_esdhc.c +2 -0
 	}
 
+	/*
+	 * This is to specify whether current eSDHC is an i.MX

eSDHC,

drivers/mmc/fsl_esdhc.c +1 -0
+	 * since the i.MX eSDHC has been becoming more and more

different

drivers/mmc/fsl_esdhc.c static int fsl_esdhc_probe(struct udevice *dev) · line 1478

Quoted 10 lines
drivers/mmc/fsl_esdhc.c +7 -0
+	 * with QorIQ eSDHC and initial MPC83XX/MPC85XX.
+	 */
+	if (!device_is_compatible(dev, "fsl,esdhc"))
+		priv->esdhc_imx = true;
+	else
+		priv->esdhc_imx = false;
+
 	val = dev_read_u32_default(dev, "bus-width", -1);
 	if (val == 8)
 		priv->bus_width = 8;

Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH, Managing Director: Wolfgang
Denk HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email:
lukma at denx.de

Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH, Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email:
lukma at denx.de

author
#8 lore raw

PATCH v2 1/3 mmc: fsl_esdhc: add esdhc_imx flag

[...]

Quoted 10 lines
Quoted 9 lines
Quoted 5 lines

My current patch-set is also to cleaning up the driver waiting for splitting them someday on the other hand. After you check CONFIG_FSL_ESDHC in u-boot source, if you think it's better we should split them right now, I could work on the driver splitting.

You can think about having common part (in one file - fsl_esdhc.c) and then separate files with code specific to imx and ppc. This would reduce the number of changes needed.

Quoted 9 lines

[Y.b. Lu] Actually, it's not only imx and ppc.
In linux, esdhc driver is for MPC83XX/MPC85XX, QorIQ PowerPC processors, QoriQ ARM processors (Layerscape).
Esdhc-imx driver is for i.MX processors.

Anyway there would be a lot files changed if we split the driver. Let me send out a patch-set to split them for your reviewing.
Thanks a lot for your comments.

[Y.b. Lu] Any comments or suggestions on the patch-set. Thanks a lot :)
http://patchwork.ozlabs.org/project/uboot/list/?series=93317&state=*

[...]