2021-01-14 03:17:24

by Andrew Jeffery

[permalink] [raw]
Subject: [PATCH v7 0/6] mmc: sdhci-of-aspeed: Expose phase delay tuning

Hello,

This series implements support for the MMC core clk-phase-* devicetree bindings
in the Aspeed SD/eMMC driver. The relevant register was exposed on the AST2600
and is present for both the SD/MMC controller and the dedicated eMMC
controller.

v7 is just a small change to the the kunit testing in response to Adrian's
feedback.

I've just done a quick build test of v7 given the small change and more
extensive testing done with v5.

v6 can be found here:

https://lore.kernel.org/linux-mmc/[email protected]/

Please review!

Cheers,

Andrew

Andrew Jeffery (6):
mmc: core: Add helper for parsing clock phase properties
mmc: sdhci-of-aspeed: Expose clock phase controls
mmc: sdhci-of-aspeed: Add AST2600 bus clock support
mmc: sdhci-of-aspeed: Add KUnit tests for phase calculations
MAINTAINERS: Add entry for the ASPEED SD/MMC driver
ARM: dts: rainier: Add eMMC clock phase compensation

MAINTAINERS | 9 +
arch/arm/boot/dts/aspeed-bmc-ibm-rainier.dts | 1 +
drivers/mmc/core/host.c | 44 ++++
drivers/mmc/host/Kconfig | 14 +
drivers/mmc/host/sdhci-of-aspeed-test.c | 98 +++++++
drivers/mmc/host/sdhci-of-aspeed.c | 255 ++++++++++++++++++-
include/linux/mmc/host.h | 13 +
7 files changed, 423 insertions(+), 11 deletions(-)
create mode 100644 drivers/mmc/host/sdhci-of-aspeed-test.c

--
2.27.0


2021-01-14 03:17:30

by Andrew Jeffery

[permalink] [raw]
Subject: [PATCH v7 1/6] mmc: core: Add helper for parsing clock phase properties

Drivers for MMC hosts that accept phase corrections can take advantage
of the helper by embedding an instance of struct mmc_clk_phase_map in
their private data and invoking mmc_of_parse_clk_phase() to extract
phase parameters. It is the responsibility of the host driver to
translate and apply the extracted values to hardware as required.

Signed-off-by: Andrew Jeffery <[email protected]>
---
drivers/mmc/core/host.c | 44 ++++++++++++++++++++++++++++++++++++++++
include/linux/mmc/host.h | 13 ++++++++++++
2 files changed, 57 insertions(+)

diff --git a/drivers/mmc/core/host.c b/drivers/mmc/core/host.c
index 96b2ca1f1b06..74e853bb6948 100644
--- a/drivers/mmc/core/host.c
+++ b/drivers/mmc/core/host.c
@@ -163,6 +163,50 @@ static void mmc_retune_timer(struct timer_list *t)
mmc_retune_needed(host);
}

+static void mmc_of_parse_timing_phase(struct device *dev, const char *prop,
+ struct mmc_clk_phase *phase)
+{
+ int degrees[2] = {0};
+ int rc;
+
+ rc = device_property_read_u32_array(dev, prop, degrees, 2);
+ phase->valid = !rc;
+ if (phase->valid) {
+ phase->in_deg = degrees[0];
+ phase->out_deg = degrees[1];
+ }
+}
+
+void
+mmc_of_parse_clk_phase(struct mmc_host *host, struct mmc_clk_phase_map *map)
+{
+ struct device *dev = host->parent;
+
+ mmc_of_parse_timing_phase(dev, "clk-phase-legacy",
+ &map->phase[MMC_TIMING_LEGACY]);
+ mmc_of_parse_timing_phase(dev, "clk-phase-mmc-hs",
+ &map->phase[MMC_TIMING_MMC_HS]);
+ mmc_of_parse_timing_phase(dev, "clk-phase-sd-hs",
+ &map->phase[MMC_TIMING_SD_HS]);
+ mmc_of_parse_timing_phase(dev, "clk-phase-uhs-sdr12",
+ &map->phase[MMC_TIMING_UHS_SDR12]);
+ mmc_of_parse_timing_phase(dev, "clk-phase-uhs-sdr25",
+ &map->phase[MMC_TIMING_UHS_SDR25]);
+ mmc_of_parse_timing_phase(dev, "clk-phase-uhs-sdr50",
+ &map->phase[MMC_TIMING_UHS_SDR50]);
+ mmc_of_parse_timing_phase(dev, "clk-phase-uhs-sdr104",
+ &map->phase[MMC_TIMING_UHS_SDR104]);
+ mmc_of_parse_timing_phase(dev, "clk-phase-uhs-ddr50",
+ &map->phase[MMC_TIMING_UHS_DDR50]);
+ mmc_of_parse_timing_phase(dev, "clk-phase-mmc-ddr52",
+ &map->phase[MMC_TIMING_MMC_DDR52]);
+ mmc_of_parse_timing_phase(dev, "clk-phase-mmc-hs200",
+ &map->phase[MMC_TIMING_MMC_HS200]);
+ mmc_of_parse_timing_phase(dev, "clk-phase-mmc-hs400",
+ &map->phase[MMC_TIMING_MMC_HS400]);
+}
+EXPORT_SYMBOL(mmc_of_parse_clk_phase);
+
/**
* mmc_of_parse() - parse host's device-tree node
* @host: host whose node should be parsed.
diff --git a/include/linux/mmc/host.h b/include/linux/mmc/host.h
index 01bba36545c5..8cf686d98a68 100644
--- a/include/linux/mmc/host.h
+++ b/include/linux/mmc/host.h
@@ -79,6 +79,17 @@ struct mmc_ios {
bool enhanced_strobe; /* hs400es selection */
};

+struct mmc_clk_phase {
+ bool valid;
+ u16 in_deg;
+ u16 out_deg;
+};
+
+#define MMC_NUM_CLK_PHASES (MMC_TIMING_MMC_HS400 + 1)
+struct mmc_clk_phase_map {
+ struct mmc_clk_phase phase[MMC_NUM_CLK_PHASES];
+};
+
struct mmc_host;

struct mmc_host_ops {
@@ -490,6 +501,8 @@ struct mmc_host *mmc_alloc_host(int extra, struct device *);
int mmc_add_host(struct mmc_host *);
void mmc_remove_host(struct mmc_host *);
void mmc_free_host(struct mmc_host *);
+void mmc_of_parse_clk_phase(struct mmc_host *host,
+ struct mmc_clk_phase_map *map);
int mmc_of_parse(struct mmc_host *host);
int mmc_of_parse_voltage(struct device_node *np, u32 *mask);

--
2.27.0

2021-01-14 03:18:31

by Andrew Jeffery

[permalink] [raw]
Subject: [PATCH v7 2/6] mmc: sdhci-of-aspeed: Expose clock phase controls

The Aspeed SD/eMMC controllers expose configurable clock phase
correction by inserting delays of up to 15 logic elements in length into
the bus clock path. The hardware supports independent configuration for
both bus directions on a per-slot basis.

The timing delay per element encoded in the driver was experimentally
determined by scope measurements.

The phase controls for both slots are grouped together in a single
register of the global register block of the SD/MMC controller(s), which
drives the use of a locking scheme between the SDHCIs and the global
register set.

Signed-off-by: Andrew Jeffery <[email protected]>
---
drivers/mmc/host/sdhci-of-aspeed.c | 216 +++++++++++++++++++++++++++--
1 file changed, 208 insertions(+), 8 deletions(-)

diff --git a/drivers/mmc/host/sdhci-of-aspeed.c b/drivers/mmc/host/sdhci-of-aspeed.c
index 4f008ba3280e..b1a14e7dda82 100644
--- a/drivers/mmc/host/sdhci-of-aspeed.c
+++ b/drivers/mmc/host/sdhci-of-aspeed.c
@@ -6,6 +6,7 @@
#include <linux/delay.h>
#include <linux/device.h>
#include <linux/io.h>
+#include <linux/math64.h>
#include <linux/mmc/host.h>
#include <linux/module.h>
#include <linux/of_address.h>
@@ -16,9 +17,19 @@

#include "sdhci-pltfm.h"

-#define ASPEED_SDC_INFO 0x00
-#define ASPEED_SDC_S1MMC8 BIT(25)
-#define ASPEED_SDC_S0MMC8 BIT(24)
+#define ASPEED_SDC_INFO 0x00
+#define ASPEED_SDC_S1_MMC8 BIT(25)
+#define ASPEED_SDC_S0_MMC8 BIT(24)
+#define ASPEED_SDC_PHASE 0xf4
+#define ASPEED_SDC_S1_PHASE_IN GENMASK(25, 21)
+#define ASPEED_SDC_S0_PHASE_IN GENMASK(20, 16)
+#define ASPEED_SDC_S1_PHASE_OUT GENMASK(15, 11)
+#define ASPEED_SDC_S1_PHASE_IN_EN BIT(10)
+#define ASPEED_SDC_S1_PHASE_OUT_EN GENMASK(9, 8)
+#define ASPEED_SDC_S0_PHASE_OUT GENMASK(7, 3)
+#define ASPEED_SDC_S0_PHASE_IN_EN BIT(2)
+#define ASPEED_SDC_S0_PHASE_OUT_EN GENMASK(1, 0)
+#define ASPEED_SDC_PHASE_MAX 31

struct aspeed_sdc {
struct clk *clk;
@@ -28,9 +39,36 @@ struct aspeed_sdc {
void __iomem *regs;
};

+struct aspeed_sdhci_tap_param {
+ bool valid;
+
+#define ASPEED_SDHCI_TAP_PARAM_INVERT_CLK BIT(4)
+ u8 in;
+ u8 out;
+};
+
+struct aspeed_sdhci_tap_desc {
+ u32 tap_mask;
+ u32 enable_mask;
+ u8 enable_value;
+};
+
+struct aspeed_sdhci_phase_desc {
+ struct aspeed_sdhci_tap_desc in;
+ struct aspeed_sdhci_tap_desc out;
+};
+
+struct aspeed_sdhci_pdata {
+ const struct aspeed_sdhci_phase_desc *phase_desc;
+ size_t nr_phase_descs;
+};
+
struct aspeed_sdhci {
+ const struct aspeed_sdhci_pdata *pdata;
struct aspeed_sdc *parent;
u32 width_mask;
+ struct mmc_clk_phase_map phase_map;
+ const struct aspeed_sdhci_phase_desc *phase_desc;
};

static void aspeed_sdc_configure_8bit_mode(struct aspeed_sdc *sdc,
@@ -50,10 +88,118 @@ static void aspeed_sdc_configure_8bit_mode(struct aspeed_sdc *sdc,
spin_unlock(&sdc->lock);
}

+static u32
+aspeed_sdc_set_phase_tap(const struct aspeed_sdhci_tap_desc *desc,
+ u8 tap, bool enable, u32 reg)
+{
+ reg &= ~(desc->enable_mask | desc->tap_mask);
+ if (enable) {
+ reg |= tap << __ffs(desc->tap_mask);
+ reg |= desc->enable_value << __ffs(desc->enable_mask);
+ }
+
+ return reg;
+}
+
+static void
+aspeed_sdc_set_phase_taps(struct aspeed_sdc *sdc,
+ const struct aspeed_sdhci_phase_desc *desc,
+ const struct aspeed_sdhci_tap_param *taps)
+{
+ u32 reg;
+
+ spin_lock(&sdc->lock);
+ reg = readl(sdc->regs + ASPEED_SDC_PHASE);
+
+ reg = aspeed_sdc_set_phase_tap(&desc->in, taps->in, taps->valid, reg);
+ reg = aspeed_sdc_set_phase_tap(&desc->out, taps->out, taps->valid, reg);
+
+ writel(reg, sdc->regs + ASPEED_SDC_PHASE);
+ spin_unlock(&sdc->lock);
+}
+
+#define PICOSECONDS_PER_SECOND 1000000000000ULL
+#define ASPEED_SDHCI_NR_TAPS 15
+/* Measured value with *handwave* environmentals and static loading */
+#define ASPEED_SDHCI_MAX_TAP_DELAY_PS 1253
+static int aspeed_sdhci_phase_to_tap(struct device *dev, unsigned long rate_hz,
+ int phase_deg)
+{
+ u64 phase_period_ps;
+ u64 prop_delay_ps;
+ u64 clk_period_ps;
+ unsigned int tap;
+ u8 inverted;
+
+ phase_deg %= 360;
+
+ if (phase_deg >= 180) {
+ inverted = ASPEED_SDHCI_TAP_PARAM_INVERT_CLK;
+ phase_deg -= 180;
+ dev_dbg(dev,
+ "Inverting clock to reduce phase correction from %d to %d degrees\n",
+ phase_deg + 180, phase_deg);
+ } else {
+ inverted = 0;
+ }
+
+ prop_delay_ps = ASPEED_SDHCI_MAX_TAP_DELAY_PS / ASPEED_SDHCI_NR_TAPS;
+ clk_period_ps = div_u64(PICOSECONDS_PER_SECOND, (u64)rate_hz);
+ phase_period_ps = div_u64((u64)phase_deg * clk_period_ps, 360ULL);
+
+ tap = div_u64(phase_period_ps, prop_delay_ps);
+ if (tap > ASPEED_SDHCI_NR_TAPS) {
+ dev_warn(dev,
+ "Requested out of range phase tap %d for %d degrees of phase compensation at %luHz, clamping to tap %d\n",
+ tap, phase_deg, rate_hz, ASPEED_SDHCI_NR_TAPS);
+ tap = ASPEED_SDHCI_NR_TAPS;
+ }
+
+ return inverted | tap;
+}
+
+static void
+aspeed_sdhci_phases_to_taps(struct device *dev, unsigned long rate,
+ const struct mmc_clk_phase *phases,
+ struct aspeed_sdhci_tap_param *taps)
+{
+ taps->valid = phases->valid;
+
+ if (!phases->valid)
+ return;
+
+ taps->in = aspeed_sdhci_phase_to_tap(dev, rate, phases->in_deg);
+ taps->out = aspeed_sdhci_phase_to_tap(dev, rate, phases->out_deg);
+}
+
+static void
+aspeed_sdhci_configure_phase(struct sdhci_host *host, unsigned long rate)
+{
+ struct aspeed_sdhci_tap_param _taps = {0}, *taps = &_taps;
+ struct mmc_clk_phase *params;
+ struct aspeed_sdhci *sdhci;
+ struct device *dev;
+
+ dev = host->mmc->parent;
+ sdhci = sdhci_pltfm_priv(sdhci_priv(host));
+
+ if (!sdhci->phase_desc)
+ return;
+
+ params = &sdhci->phase_map.phase[host->timing];
+ aspeed_sdhci_phases_to_taps(dev, rate, params, taps);
+ aspeed_sdc_set_phase_taps(sdhci->parent, sdhci->phase_desc, taps);
+ dev_dbg(dev,
+ "Using taps [%d, %d] for [%d, %d] degrees of phase correction at %luHz (%d)\n",
+ taps->in & ASPEED_SDHCI_NR_TAPS,
+ taps->out & ASPEED_SDHCI_NR_TAPS,
+ params->in_deg, params->out_deg, rate, host->timing);
+}
+
static void aspeed_sdhci_set_clock(struct sdhci_host *host, unsigned int clock)
{
struct sdhci_pltfm_host *pltfm_host;
- unsigned long parent;
+ unsigned long parent, bus;
int div;
u16 clk;

@@ -69,13 +215,17 @@ static void aspeed_sdhci_set_clock(struct sdhci_host *host, unsigned int clock)
clock = host->max_clk;

for (div = 2; div < 256; div *= 2) {
- if ((parent / div) <= clock)
+ bus = parent / div;
+ if (bus <= clock)
break;
}
+
div >>= 1;

clk = div << SDHCI_DIVIDER_SHIFT;

+ aspeed_sdhci_configure_phase(host, bus);
+
sdhci_enable_clk(host, clk);
}

@@ -157,6 +307,7 @@ static inline int aspeed_sdhci_calculate_slot(struct aspeed_sdhci *dev,

static int aspeed_sdhci_probe(struct platform_device *pdev)
{
+ const struct aspeed_sdhci_pdata *aspeed_pdata;
struct sdhci_pltfm_host *pltfm_host;
struct aspeed_sdhci *dev;
struct sdhci_host *host;
@@ -164,12 +315,15 @@ static int aspeed_sdhci_probe(struct platform_device *pdev)
int slot;
int ret;

+ aspeed_pdata = of_device_get_match_data(&pdev->dev);
+
host = sdhci_pltfm_init(pdev, &aspeed_sdhci_pdata, sizeof(*dev));
if (IS_ERR(host))
return PTR_ERR(host);

pltfm_host = sdhci_priv(host);
dev = sdhci_pltfm_priv(pltfm_host);
+ dev->pdata = aspeed_pdata;
dev->parent = dev_get_drvdata(pdev->dev.parent);

res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
@@ -180,8 +334,17 @@ static int aspeed_sdhci_probe(struct platform_device *pdev)
else if (slot >= 2)
return -EINVAL;

- dev_info(&pdev->dev, "Configuring for slot %d\n", slot);
- dev->width_mask = !slot ? ASPEED_SDC_S0MMC8 : ASPEED_SDC_S1MMC8;
+ if (dev->pdata && slot < dev->pdata->nr_phase_descs) {
+ dev->phase_desc = &dev->pdata->phase_desc[slot];
+ } else {
+ dev_info(&pdev->dev,
+ "Phase control not supported for slot %d\n", slot);
+ dev->phase_desc = NULL;
+ }
+
+ dev->width_mask = !slot ? ASPEED_SDC_S0_MMC8 : ASPEED_SDC_S1_MMC8;
+
+ dev_info(&pdev->dev, "Configured for slot %d\n", slot);

sdhci_get_of_property(pdev);

@@ -199,6 +362,9 @@ static int aspeed_sdhci_probe(struct platform_device *pdev)
if (ret)
goto err_sdhci_add;

+ if (dev->phase_desc)
+ mmc_of_parse_clk_phase(host->mmc, &dev->phase_map);
+
ret = sdhci_add_host(host);
if (ret)
goto err_sdhci_add;
@@ -230,10 +396,44 @@ static int aspeed_sdhci_remove(struct platform_device *pdev)
return 0;
}

+static const struct aspeed_sdhci_phase_desc ast2600_sdhci_phase[] = {
+ /* SDHCI/Slot 0 */
+ [0] = {
+ .in = {
+ .tap_mask = ASPEED_SDC_S0_PHASE_IN,
+ .enable_mask = ASPEED_SDC_S0_PHASE_IN_EN,
+ .enable_value = 1,
+ },
+ .out = {
+ .tap_mask = ASPEED_SDC_S0_PHASE_OUT,
+ .enable_mask = ASPEED_SDC_S0_PHASE_OUT_EN,
+ .enable_value = 3,
+ },
+ },
+ /* SDHCI/Slot 1 */
+ [1] = {
+ .in = {
+ .tap_mask = ASPEED_SDC_S1_PHASE_IN,
+ .enable_mask = ASPEED_SDC_S1_PHASE_IN_EN,
+ .enable_value = 1,
+ },
+ .out = {
+ .tap_mask = ASPEED_SDC_S1_PHASE_OUT,
+ .enable_mask = ASPEED_SDC_S1_PHASE_OUT_EN,
+ .enable_value = 3,
+ },
+ },
+};
+
+static const struct aspeed_sdhci_pdata ast2600_sdhci_pdata = {
+ .phase_desc = ast2600_sdhci_phase,
+ .nr_phase_descs = ARRAY_SIZE(ast2600_sdhci_phase),
+};
+
static const struct of_device_id aspeed_sdhci_of_match[] = {
{ .compatible = "aspeed,ast2400-sdhci", },
{ .compatible = "aspeed,ast2500-sdhci", },
- { .compatible = "aspeed,ast2600-sdhci", },
+ { .compatible = "aspeed,ast2600-sdhci", .data = &ast2600_sdhci_pdata, },
{ }
};

--
2.27.0

2021-01-14 03:19:13

by Andrew Jeffery

[permalink] [raw]
Subject: [PATCH v7 5/6] MAINTAINERS: Add entry for the ASPEED SD/MMC driver

Add myself as the maintainer.

Signed-off-by: Andrew Jeffery <[email protected]>
---
MAINTAINERS | 9 +++++++++
1 file changed, 9 insertions(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index e451dcce054f..eae4322aae67 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -2826,6 +2826,15 @@ F: Documentation/devicetree/bindings/interrupt-controller/aspeed,ast2xxx-scu-ic.
F: drivers/irqchip/irq-aspeed-scu-ic.c
F: include/dt-bindings/interrupt-controller/aspeed-scu-ic.h

+ASPEED SD/MMC DRIVER
+M: Andrew Jeffery <[email protected]>
+L: [email protected] (moderated for non-subscribers)
+L: [email protected] (moderated for non-subscribers)
+L: [email protected]
+S: Maintained
+F: Documentation/devicetree/bindings/mmc/aspeed,sdhci.yaml
+F: drivers/mmc/host/sdhci-of-aspeed*
+
ASPEED VIDEO ENGINE DRIVER
M: Eddie James <[email protected]>
L: [email protected]
--
2.27.0

2021-01-14 03:19:16

by Andrew Jeffery

[permalink] [raw]
Subject: [PATCH v7 6/6] ARM: dts: rainier: Add eMMC clock phase compensation

Determined by scope measurements at speed.

Signed-off-by: Andrew Jeffery <[email protected]>
---
arch/arm/boot/dts/aspeed-bmc-ibm-rainier.dts | 1 +
1 file changed, 1 insertion(+)

diff --git a/arch/arm/boot/dts/aspeed-bmc-ibm-rainier.dts b/arch/arm/boot/dts/aspeed-bmc-ibm-rainier.dts
index 21ae880c7530..ab8d37d49f30 100644
--- a/arch/arm/boot/dts/aspeed-bmc-ibm-rainier.dts
+++ b/arch/arm/boot/dts/aspeed-bmc-ibm-rainier.dts
@@ -186,6 +186,7 @@ &pinctrl_emmc_default {

&emmc {
status = "okay";
+ clk-phase-mmc-hs200 = <180>, <180>;
};

&fsim0 {
--
2.27.0

2021-01-14 03:20:15

by Andrew Jeffery

[permalink] [raw]
Subject: [PATCH v7 3/6] mmc: sdhci-of-aspeed: Add AST2600 bus clock support

The AST2600 can achieve HS200 speeds with a change to the bus clock
divisor behaviour. The divisor can also be more accurate with respect
to the requested clock rate, but keep the one-hot behaviour for
backwards compatibility with the AST2400 and AST2500.

Signed-off-by: Andrew Jeffery <[email protected]>
---
drivers/mmc/host/sdhci-of-aspeed.c | 37 ++++++++++++++++++++++++++----
1 file changed, 33 insertions(+), 4 deletions(-)

diff --git a/drivers/mmc/host/sdhci-of-aspeed.c b/drivers/mmc/host/sdhci-of-aspeed.c
index b1a14e7dda82..4b30f6a1178e 100644
--- a/drivers/mmc/host/sdhci-of-aspeed.c
+++ b/drivers/mmc/host/sdhci-of-aspeed.c
@@ -59,6 +59,7 @@ struct aspeed_sdhci_phase_desc {
};

struct aspeed_sdhci_pdata {
+ unsigned int clk_div_start;
const struct aspeed_sdhci_phase_desc *phase_desc;
size_t nr_phase_descs;
};
@@ -200,10 +201,13 @@ static void aspeed_sdhci_set_clock(struct sdhci_host *host, unsigned int clock)
{
struct sdhci_pltfm_host *pltfm_host;
unsigned long parent, bus;
+ struct aspeed_sdhci *sdhci;
int div;
u16 clk;

pltfm_host = sdhci_priv(host);
+ sdhci = sdhci_pltfm_priv(pltfm_host);
+
parent = clk_get_rate(pltfm_host->clk);

sdhci_writew(host, 0, SDHCI_CLOCK_CONTROL);
@@ -214,7 +218,23 @@ static void aspeed_sdhci_set_clock(struct sdhci_host *host, unsigned int clock)
if (WARN_ON(clock > host->max_clk))
clock = host->max_clk;

- for (div = 2; div < 256; div *= 2) {
+ /*
+ * Regarding the AST2600:
+ *
+ * If (EMMC12C[7:6], EMMC12C[15:8] == 0) then
+ * period of SDCLK = period of SDMCLK.
+ *
+ * If (EMMC12C[7:6], EMMC12C[15:8] != 0) then
+ * period of SDCLK = period of SDMCLK * 2 * (EMMC12C[7:6], EMMC[15:8])
+ *
+ * If you keep EMMC12C[7:6] = 0 and EMMC12C[15:8] as one-hot,
+ * 0x1/0x2/0x4/etc, you will find it is compatible to AST2400 or AST2500
+ *
+ * Keep the one-hot behaviour for backwards compatibility except for
+ * supporting the value 0 in (EMMC12C[7:6], EMMC12C[15:8]), and capture
+ * the 0-value capability in clk_div_start.
+ */
+ for (div = sdhci->pdata->clk_div_start; div < 256; div *= 2) {
bus = parent / div;
if (bus <= clock)
break;
@@ -316,6 +336,10 @@ static int aspeed_sdhci_probe(struct platform_device *pdev)
int ret;

aspeed_pdata = of_device_get_match_data(&pdev->dev);
+ if (!aspeed_pdata) {
+ dev_err(&pdev->dev, "Missing platform configuration data\n");
+ return -EINVAL;
+ }

host = sdhci_pltfm_init(pdev, &aspeed_sdhci_pdata, sizeof(*dev));
if (IS_ERR(host))
@@ -334,7 +358,7 @@ static int aspeed_sdhci_probe(struct platform_device *pdev)
else if (slot >= 2)
return -EINVAL;

- if (dev->pdata && slot < dev->pdata->nr_phase_descs) {
+ if (slot < dev->pdata->nr_phase_descs) {
dev->phase_desc = &dev->pdata->phase_desc[slot];
} else {
dev_info(&pdev->dev,
@@ -396,6 +420,10 @@ static int aspeed_sdhci_remove(struct platform_device *pdev)
return 0;
}

+static const struct aspeed_sdhci_pdata ast2400_sdhci_pdata = {
+ .clk_div_start = 2,
+};
+
static const struct aspeed_sdhci_phase_desc ast2600_sdhci_phase[] = {
/* SDHCI/Slot 0 */
[0] = {
@@ -426,13 +454,14 @@ static const struct aspeed_sdhci_phase_desc ast2600_sdhci_phase[] = {
};

static const struct aspeed_sdhci_pdata ast2600_sdhci_pdata = {
+ .clk_div_start = 1,
.phase_desc = ast2600_sdhci_phase,
.nr_phase_descs = ARRAY_SIZE(ast2600_sdhci_phase),
};

static const struct of_device_id aspeed_sdhci_of_match[] = {
- { .compatible = "aspeed,ast2400-sdhci", },
- { .compatible = "aspeed,ast2500-sdhci", },
+ { .compatible = "aspeed,ast2400-sdhci", .data = &ast2400_sdhci_pdata, },
+ { .compatible = "aspeed,ast2500-sdhci", .data = &ast2400_sdhci_pdata, },
{ .compatible = "aspeed,ast2600-sdhci", .data = &ast2600_sdhci_pdata, },
{ }
};
--
2.27.0

2021-01-20 20:26:03

by Ulf Hansson

[permalink] [raw]
Subject: Re: [PATCH v7 0/6] mmc: sdhci-of-aspeed: Expose phase delay tuning

On Thu, 14 Jan 2021 at 04:14, Andrew Jeffery <[email protected]> wrote:
>
> Hello,
>
> This series implements support for the MMC core clk-phase-* devicetree bindings
> in the Aspeed SD/eMMC driver. The relevant register was exposed on the AST2600
> and is present for both the SD/MMC controller and the dedicated eMMC
> controller.
>
> v7 is just a small change to the the kunit testing in response to Adrian's
> feedback.
>
> I've just done a quick build test of v7 given the small change and more
> extensive testing done with v5.
>
> v6 can be found here:
>
> https://lore.kernel.org/linux-mmc/[email protected]/
>
> Please review!
>
> Cheers,
>
> Andrew
>
> Andrew Jeffery (6):
> mmc: core: Add helper for parsing clock phase properties
> mmc: sdhci-of-aspeed: Expose clock phase controls
> mmc: sdhci-of-aspeed: Add AST2600 bus clock support
> mmc: sdhci-of-aspeed: Add KUnit tests for phase calculations
> MAINTAINERS: Add entry for the ASPEED SD/MMC driver
> ARM: dts: rainier: Add eMMC clock phase compensation
>
> MAINTAINERS | 9 +
> arch/arm/boot/dts/aspeed-bmc-ibm-rainier.dts | 1 +
> drivers/mmc/core/host.c | 44 ++++
> drivers/mmc/host/Kconfig | 14 +
> drivers/mmc/host/sdhci-of-aspeed-test.c | 98 +++++++
> drivers/mmc/host/sdhci-of-aspeed.c | 255 ++++++++++++++++++-
> include/linux/mmc/host.h | 13 +
> 7 files changed, 423 insertions(+), 11 deletions(-)
> create mode 100644 drivers/mmc/host/sdhci-of-aspeed-test.c
>

Applied patch 1 to patch 5 applied for next (patch 6 should go via arm
soc), thanks!

Thanks for stepping and helping with maintenance as well!

Kind regards
Uffe

2021-01-21 03:37:28

by Andrew Jeffery

[permalink] [raw]
Subject: Re: [PATCH v7 0/6] mmc: sdhci-of-aspeed: Expose phase delay tuning



On Thu, 21 Jan 2021, at 00:26, Ulf Hansson wrote:
> On Thu, 14 Jan 2021 at 04:14, Andrew Jeffery <[email protected]> wrote:
> >
> > Hello,
> >
> > This series implements support for the MMC core clk-phase-* devicetree bindings
> > in the Aspeed SD/eMMC driver. The relevant register was exposed on the AST2600
> > and is present for both the SD/MMC controller and the dedicated eMMC
> > controller.
> >
> > v7 is just a small change to the the kunit testing in response to Adrian's
> > feedback.
> >
> > I've just done a quick build test of v7 given the small change and more
> > extensive testing done with v5.
> >
> > v6 can be found here:
> >
> > https://lore.kernel.org/linux-mmc/[email protected]/
> >
> > Please review!
> >
> > Cheers,
> >
> > Andrew
> >
> > Andrew Jeffery (6):
> > mmc: core: Add helper for parsing clock phase properties
> > mmc: sdhci-of-aspeed: Expose clock phase controls
> > mmc: sdhci-of-aspeed: Add AST2600 bus clock support
> > mmc: sdhci-of-aspeed: Add KUnit tests for phase calculations
> > MAINTAINERS: Add entry for the ASPEED SD/MMC driver
> > ARM: dts: rainier: Add eMMC clock phase compensation
> >
> > MAINTAINERS | 9 +
> > arch/arm/boot/dts/aspeed-bmc-ibm-rainier.dts | 1 +
> > drivers/mmc/core/host.c | 44 ++++
> > drivers/mmc/host/Kconfig | 14 +
> > drivers/mmc/host/sdhci-of-aspeed-test.c | 98 +++++++
> > drivers/mmc/host/sdhci-of-aspeed.c | 255 ++++++++++++++++++-
> > include/linux/mmc/host.h | 13 +
> > 7 files changed, 423 insertions(+), 11 deletions(-)
> > create mode 100644 drivers/mmc/host/sdhci-of-aspeed-test.c
> >
>
> Applied patch 1 to patch 5 applied for next (patch 6 should go via arm
> soc), thanks!

Yep. Joel, can you pick it up?

>
> Thanks for stepping and helping with maintenance as well!

Happy to help :)

Andrew