[U-Boot] [PATCH 0/3] driver: ti_qspi: Use a SYSCON device to map the ctrl_mod_mmap register

[U-Boot] [PATCH 0/3] driver: ti_qspi: Use a SYSCON device to map the ctrl_mod_mmap register

[Jean-Jacques Hiblot]

This series allows the ti_qspi driver to use the same dts description as the
linux kernel. In Linux since 4.6 the ctrl_mod_mmap is described in the DTS as a
syscon. It used to be the 3rd memory range in "reg".

The first patch of this series is a generic one. It changes the regmap initialization
to take in account the required address translation.

The second patch adds the mechanism in the ti_qspi driver to get the address
of ctrl_mod_mmap from the syscon

The last patch simply enable the SYSCON feature in the default config for the
dra7xx and am57xx evms.

Jean-Jacques Hiblot (3):
  regmap: use fdt address translation
  drivers: ti_qspi: use syscon to get the address ctrl_mod_mmap register
  configs: dra7x/am57x: Enable the SYSCON and REGMAP features

 configs/am57xx_evm_defconfig    |  2 ++
 configs/am57xx_hs_evm_defconfig |  2 ++
 configs/dra7xx_evm_defconfig    |  2 ++
 configs/dra7xx_hs_evm_defconfig |  2 ++
 drivers/core/regmap.c           | 14 ++++++------
 drivers/spi/ti_qspi.c           | 47 ++++++++++++++++++++++++++++++++++++-----
 6 files changed, 58 insertions(+), 11 deletions(-)

-- 
1.9.1

[U-Boot] [PATCH 1/3] regmap: use fdt address translation

[Jean-Jacques Hiblot]

In the DTS, the addresses are defined relative to the parent bus. We need
to translate them to get the address as seen by the CPU core.

Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
---

to: sjg at chromium.org

 drivers/core/regmap.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/drivers/core/regmap.c b/drivers/core/regmap.c
index c68bcba..07a6614 100644
--- a/drivers/core/regmap.c
+++ b/drivers/core/regmap.c
@@ -70,6 +70,7 @@ int regmap_init_mem(struct udevice *dev, struct regmap **mapp)
 	int addr_len, size_len, both_len;
 	int parent;
 	int len;
+	int index;
 
 	parent = dev->parent->of_offset;
 	addr_len = fdt_address_cells(blob, parent);
@@ -86,13 +87,14 @@ int regmap_init_mem(struct udevice *dev, struct regmap **mapp)
 	if (!map)
 		return -ENOMEM;
 
-	map->base = fdtdec_get_number(cell, addr_len);
-
-	for (range = map->range; count > 0;
-	     count--, cell += both_len, range++) {
-		range->start = fdtdec_get_number(cell, addr_len);
-		range->size = fdtdec_get_number(cell + addr_len, size_len);
+	for (range = map->range, index = 0; count > 0;
+	     count--, cell += both_len, range++, index++) {
+		fdt_size_t sz;
+		range->start = fdtdec_get_addr_size_fixed(blob, dev->of_offset,
+				"reg", index, addr_len, size_len, &sz, true);
+		range->size = sz;
 	}
+	map->base = map->range[0].start;
 
 	*mapp = map;
 
-- 
1.9.1

[U-Boot] [PATCH 2/3] drivers: ti_qspi: use syscon to get the address ctrl_mod_mmap register

[Jean-Jacques Hiblot]

We used to get the address of the optionnal ctrl_mod_mmap register as the
third memory range of the "reg" property. the linux driver moved to use a
syscon instead. In order to keep the DTS as close as possible to that of
linux, we move to using a syscon as well.

If CONFIG_SYSCON is no set, the driver reverts to the old way of getting
the address from the 3rd memory range

Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
---
 drivers/spi/ti_qspi.c | 47 ++++++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 42 insertions(+), 5 deletions(-)

diff --git a/drivers/spi/ti_qspi.c b/drivers/spi/ti_qspi.c
index 6f9f983..d964b50 100644
--- a/drivers/spi/ti_qspi.c
+++ b/drivers/spi/ti_qspi.c
@@ -17,6 +17,8 @@
 #include <asm/omap_common.h>
 #include <asm/ti-common/ti-edma3.h>
 #include <linux/kernel.h>
+#include <regmap.h>
+#include <syscon.h>
 
 DECLARE_GLOBAL_DATA_PTR;
 
@@ -549,21 +551,56 @@ static int ti_qspi_probe(struct udevice *bus)
 	return 0;
 }
 
+static void *map_syscon_chipselects(struct udevice *bus)
+{
+#if defined(CONFIG_SYSCON) && !defined(CONFIG_SPL_BUILD)
+	struct udevice *syscon;
+	struct regmap *regmap;
+	const fdt32_t *cell;
+	int len, err;
+
+	err = uclass_get_device_by_phandle(UCLASS_SYSCON, bus,
+					   "syscon-chipselects", &syscon);
+	if (err) {
+		debug("%s: unable to find syscon device (%d)\n", __func__,
+		      err);
+		return NULL;
+	}
+
+	regmap = syscon_get_regmap(syscon);
+	if (IS_ERR(regmap)) {
+		debug("%s: unable to find regmap (%ld)\n", __func__,
+		      PTR_ERR(regmap));
+		return NULL;
+	}
+
+	cell = fdt_getprop(gd->fdt_blob, bus->of_offset, "syscon-chipselects",
+			   &len);
+	if (len < 2*sizeof(fdt32_t)) {
+		debug("%s: offset not available\n", __func__);
+		return NULL;
+	}
+
+	return fdtdec_get_number(cell + 1, 1) + regmap_get_range(regmap, 0);
+#else
+	fdt_addr_t addr;
+	addr = dev_get_addr_index(bus, 2);
+	return (addr == FDT_ADDR_T_NONE) ? NULL :
+		map_physmem(addr, 0, MAP_NOCACHE);
+#endif
+}
+
 static int ti_qspi_ofdata_to_platdata(struct udevice *bus)
 {
 	struct ti_qspi_priv *priv = dev_get_priv(bus);
 	const void *blob = gd->fdt_blob;
 	int node = bus->of_offset;
-	fdt_addr_t addr;
-	void *mmap;
 
+	priv->ctrl_mod_mmap = map_syscon_chipselects(bus);
 	priv->base = map_physmem(dev_get_addr(bus), sizeof(struct ti_qspi_regs),
 				 MAP_NOCACHE);
 	priv->memory_map = map_physmem(dev_get_addr_index(bus, 1), 0,
 				       MAP_NOCACHE);
-	addr = dev_get_addr_index(bus, 2);
-	mmap = map_physmem(dev_get_addr_index(bus, 2), 0, MAP_NOCACHE);
-	priv->ctrl_mod_mmap = (addr == FDT_ADDR_T_NONE) ? NULL : mmap;
 
 	priv->max_hz = fdtdec_get_int(blob, node, "spi-max-frequency", -1);
 	if (priv->max_hz < 0) {
-- 
1.9.1

[U-Boot] [PATCH 3/3] configs: dra7x/am57x: Enable the SYSCON and REGMAP features

[Jean-Jacques Hiblot]

This is required by the ti_qspi driver to get from the DTS the address of
the ctrl_mod_mmap register.

Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
---
 configs/am57xx_evm_defconfig    | 2 ++
 configs/am57xx_hs_evm_defconfig | 2 ++
 configs/dra7xx_evm_defconfig    | 2 ++
 configs/dra7xx_hs_evm_defconfig | 2 ++
 4 files changed, 8 insertions(+)

diff --git a/configs/am57xx_evm_defconfig b/configs/am57xx_evm_defconfig
index a582055..15bb474 100644
--- a/configs/am57xx_evm_defconfig
+++ b/configs/am57xx_evm_defconfig
@@ -50,6 +50,8 @@ CONFIG_CMD_FS_GENERIC=y
 CONFIG_OF_CONTROL=y
 CONFIG_OF_LIST="am57xx-beagle-x15 am57xx-beagle-x15-revb1 am572x-idk am571x-idk"
 CONFIG_DM=y
+CONFIG_REGMAP=y
+CONFIG_SYSCON=y
 # CONFIG_BLK is not set
 CONFIG_DFU_MMC=y
 CONFIG_DFU_RAM=y
diff --git a/configs/am57xx_hs_evm_defconfig b/configs/am57xx_hs_evm_defconfig
index 4dcfce4..fc37b58 100644
--- a/configs/am57xx_hs_evm_defconfig
+++ b/configs/am57xx_hs_evm_defconfig
@@ -52,6 +52,8 @@ CONFIG_CMD_FAT=y
 CONFIG_CMD_FS_GENERIC=y
 CONFIG_OF_CONTROL=y
 CONFIG_DM=y
+CONFIG_REGMAP=y
+CONFIG_SYSCON=y
 # CONFIG_BLK is not set
 CONFIG_DFU_MMC=y
 CONFIG_DFU_RAM=y
diff --git a/configs/dra7xx_evm_defconfig b/configs/dra7xx_evm_defconfig
index 64fe038..624cfe4 100644
--- a/configs/dra7xx_evm_defconfig
+++ b/configs/dra7xx_evm_defconfig
@@ -50,6 +50,8 @@ CONFIG_CMD_FS_GENERIC=y
 CONFIG_OF_CONTROL=y
 CONFIG_OF_LIST="dra7-evm dra72-evm dra72-evm-revc dra71-evm"
 CONFIG_DM=y
+CONFIG_REGMAP=y
+CONFIG_SYSCON=y
 # CONFIG_BLK is not set
 CONFIG_DFU_MMC=y
 CONFIG_DFU_RAM=y
diff --git a/configs/dra7xx_hs_evm_defconfig b/configs/dra7xx_hs_evm_defconfig
index b890a06..028b212 100644
--- a/configs/dra7xx_hs_evm_defconfig
+++ b/configs/dra7xx_hs_evm_defconfig
@@ -55,6 +55,8 @@ CONFIG_CMD_FS_GENERIC=y
 CONFIG_OF_CONTROL=y
 CONFIG_OF_LIST="dra7-evm dra72-evm dra72-evm-revc dra71-evm"
 CONFIG_DM=y
+CONFIG_REGMAP=y
+CONFIG_SYSCON=y
 # CONFIG_BLK is not set
 CONFIG_DFU_MMC=y
 CONFIG_DFU_RAM=y
-- 
1.9.1

[U-Boot] [PATCH 1/3] regmap: use fdt address translation

[Simon Glass]

+Stephen

On 7 February 2017 at 08:45, Jean-Jacques Hiblot <jjhiblot@ti.com> wrote:
> In the DTS, the addresses are defined relative to the parent bus. We need
> to translate them to get the address as seen by the CPU core.
>
> Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
> ---
>
> to: sjg at chromium.org
>
>  drivers/core/regmap.c | 14 ++++++++------
>  1 file changed, 8 insertions(+), 6 deletions(-)

Reviewed-by: Simon Glass <sjg@chromium.org>

>
> diff --git a/drivers/core/regmap.c b/drivers/core/regmap.c
> index c68bcba..07a6614 100644
> --- a/drivers/core/regmap.c
> +++ b/drivers/core/regmap.c
> @@ -70,6 +70,7 @@ int regmap_init_mem(struct udevice *dev, struct regmap **mapp)
>         int addr_len, size_len, both_len;
>         int parent;
>         int len;
> +       int index;
>
>         parent = dev->parent->of_offset;
>         addr_len = fdt_address_cells(blob, parent);
> @@ -86,13 +87,14 @@ int regmap_init_mem(struct udevice *dev, struct regmap **mapp)
>         if (!map)
>                 return -ENOMEM;
>
> -       map->base = fdtdec_get_number(cell, addr_len);
> -
> -       for (range = map->range; count > 0;
> -            count--, cell += both_len, range++) {
> -               range->start = fdtdec_get_number(cell, addr_len);
> -               range->size = fdtdec_get_number(cell + addr_len, size_len);
> +       for (range = map->range, index = 0; count > 0;
> +            count--, cell += both_len, range++, index++) {
> +               fdt_size_t sz;
> +               range->start = fdtdec_get_addr_size_fixed(blob, dev->of_offset,
> +                               "reg", index, addr_len, size_len, &sz, true);
> +               range->size = sz;
>         }
> +       map->base = map->range[0].start;
>
>         *mapp = map;
>
> --
> 1.9.1
>

[U-Boot] [PATCH 2/3] drivers: ti_qspi: use syscon to get the address ctrl_mod_mmap register

[Simon Glass]

On 7 February 2017 at 08:45, Jean-Jacques Hiblot <jjhiblot@ti.com> wrote:
> We used to get the address of the optionnal ctrl_mod_mmap register as the
> third memory range of the "reg" property. the linux driver moved to use a
> syscon instead. In order to keep the DTS as close as possible to that of
> linux, we move to using a syscon as well.
>
> If CONFIG_SYSCON is no set, the driver reverts to the old way of getting
> the address from the 3rd memory range
>
> Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
> ---
>  drivers/spi/ti_qspi.c | 47 ++++++++++++++++++++++++++++++++++++++++++-----
>  1 file changed, 42 insertions(+), 5 deletions(-)

Reviewed-by: Simon Glass <sjg@chromium.org>

[U-Boot] [PATCH 3/3] configs: dra7x/am57x: Enable the SYSCON and REGMAP features

[Simon Glass]

On 7 February 2017 at 08:45, Jean-Jacques Hiblot <jjhiblot@ti.com> wrote:
> This is required by the ti_qspi driver to get from the DTS the address of
> the ctrl_mod_mmap register.
>
> Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
> ---
>  configs/am57xx_evm_defconfig    | 2 ++
>  configs/am57xx_hs_evm_defconfig | 2 ++
>  configs/dra7xx_evm_defconfig    | 2 ++
>  configs/dra7xx_hs_evm_defconfig | 2 ++
>  4 files changed, 8 insertions(+)

Reviewed-by: Simon Glass <sjg@chromium.org>

[U-Boot] [PATCH 1/3] regmap: use fdt address translation

[Stephen Warren]

On 02/10/2017 09:22 AM, Simon Glass wrote:
> +Stephen
>
> On 7 February 2017 at 08:45, Jean-Jacques Hiblot <jjhiblot@ti.com> wrote:
>> In the DTS, the addresses are defined relative to the parent bus. We need
>> to translate them to get the address as seen by the CPU core.
>>
>> Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
>> ---
>>
>> to: sjg at chromium.org
>>
>>  drivers/core/regmap.c | 14 ++++++++------
>>  1 file changed, 8 insertions(+), 6 deletions(-)
>
> Reviewed-by: Simon Glass <sjg@chromium.org>

Reviewed-by: Stephen Warren <swarren@nvidia.com>

[U-Boot] [PATCH 2/3] drivers: ti_qspi: use syscon to get the address ctrl_mod_mmap register

[Vignesh R]

On Tuesday 07 February 2017 09:15 PM, Jean-Jacques Hiblot wrote:
> We used to get the address of the optionnal ctrl_mod_mmap register as the
> third memory range of the "reg" property. the linux driver moved to use a
> syscon instead. In order to keep the DTS as close as possible to that of
> linux, we move to using a syscon as well.
> 
> If CONFIG_SYSCON is no set, the driver reverts to the old way of getting
> the address from the 3rd memory range
> 
> Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
> ---
>  drivers/spi/ti_qspi.c | 47 ++++++++++++++++++++++++++++++++++++++++++-----
>  1 file changed, 42 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/spi/ti_qspi.c b/drivers/spi/ti_qspi.c
> index 6f9f983..d964b50 100644
> --- a/drivers/spi/ti_qspi.c
> +++ b/drivers/spi/ti_qspi.c
> @@ -17,6 +17,8 @@
>  #include <asm/omap_common.h>
>  #include <asm/ti-common/ti-edma3.h>
>  #include <linux/kernel.h>
> +#include <regmap.h>
> +#include <syscon.h>
>  
>  DECLARE_GLOBAL_DATA_PTR;
>  
> @@ -549,21 +551,56 @@ static int ti_qspi_probe(struct udevice *bus)
>  	return 0;
>  }
>  
> +static void *map_syscon_chipselects(struct udevice *bus)
> +{
> +#if defined(CONFIG_SYSCON) && !defined(CONFIG_SPL_BUILD)


Please change this to #if CONFIG_IS_ENABLED(SYSCON)



-- 
Regards
Vignesh

[U-Boot] [PATCH 0/3] driver: ti_qspi: Use a SYSCON device to map the ctrl_mod_mmap register

[Vignesh R]

On Tuesday 07 February 2017 09:15 PM, Jean-Jacques Hiblot wrote:
> This series allows the ti_qspi driver to use the same dts description as the
> linux kernel. In Linux since 4.6 the ctrl_mod_mmap is described in the DTS as a
> syscon. It used to be the 3rd memory range in "reg".
> 
> The first patch of this series is a generic one. It changes the regmap initialization
> to take in account the required address translation.
> 
> The second patch adds the mechanism in the ti_qspi driver to get the address
> of ctrl_mod_mmap from the syscon
> 
> The last patch simply enable the SYSCON feature in the default config for the
> dra7xx and am57xx evms.
> 
> Jean-Jacques Hiblot (3):
>   regmap: use fdt address translation
>   drivers: ti_qspi: use syscon to get the address ctrl_mod_mmap register
>   configs: dra7x/am57x: Enable the SYSCON and REGMAP features
>

Could you rebase this series on top of Lokesh's series enabling support
for SPL_DM][1][2]. The changes that you need to do to this patch series
is minimum and restricted to defconfig files:


diff --git a/configs/dra7xx_evm_defconfig b/configs/dra7xx_evm_defconfig
index a52acc51d365..591c64dc7453 100644
--- a/configs/dra7xx_evm_defconfig
+++ b/configs/dra7xx_evm_defconfig
@@ -60,7 +60,9 @@ CONFIG_OF_LIST="dra7-evm dra72-evm dra72-evm-revc
dra71-evm"
 CONFIG_DM=y
 CONFIG_SPL_DM=y
 CONFIG_REGMAP=y
+CONFIG_SPL_REGMAP=y
 CONFIG_SYSCON=y
+CONFIG_SPL_SYSCON=y
 # CONFIG_BLK is not set
 CONFIG_DFU_MMC=y
 CONFIG_DFU_RAM=y
@@ -103,3 +105,5 @@ CONFIG_USB_GADGET_DOWNLOAD=y
 CONFIG_G_DNL_MANUFACTURER="Texas Instruments"
 CONFIG_G_DNL_VENDOR_NUM=0x0451
 CONFIG_G_DNL_PRODUCT_NUM=0xd022
+CONFIG_SPL_DM_SEQ_ALIAS=y
+CONFIG_SPL_OF_TRANSLATE=y


[1] https://www.mail-archive.com/u-boot at lists.denx.de/msg238751.html
[2] https://www.mail-archive.com/u-boot at lists.denx.de/msg238964.html



 --
Regards
Vignesh

[U-Boot] [PATCH 3/3] configs: dra7x/am57x: Enable the SYSCON and REGMAP features

[Simon Glass]

Hi,

On 10 February 2017 at 09:22, Simon Glass <sjg@chromium.org> wrote:
> On 7 February 2017 at 08:45, Jean-Jacques Hiblot <jjhiblot@ti.com> wrote:
>> This is required by the ti_qspi driver to get from the DTS the address of
>> the ctrl_mod_mmap register.
>>
>> Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
>> ---
>>  configs/am57xx_evm_defconfig    | 2 ++
>>  configs/am57xx_hs_evm_defconfig | 2 ++
>>  configs/dra7xx_evm_defconfig    | 2 ++
>>  configs/dra7xx_hs_evm_defconfig | 2 ++
>>  4 files changed, 8 insertions(+)
>
> Reviewed-by: Simon Glass <sjg@chromium.org>

This doesn't apply, and given that patch 1 has been dropped in favour
of something else, would you mind checking this and resending what you
need to on top of dm/testing?

Regards,
Simon