[PATCH] driver: mtd: update struct map_info's swap as per map requirement.

[PATCH] driver: mtd: update struct map_info's swap as per map requirement.

[Prabhakar Kushwaha]

It is not necessary for all device's maps to be CFI_HOST_ENDIAN. Maps device
can be Bigendian or little endian.

Currently it is being taken care using CONFIG_MTD_CFI_LE_BYTE_SWAP or
CONFIG_MTD_CFI_BE_BYTE_SWAP i.e. compile time.

Now update struct map_info's swap field based on device characteristics
defined in device tree.

Signed-off-by: Prabhakar Kushwaha <prabhakar.kushwaha@nxp.com>
---
 drivers/mtd/maps/physmap_of.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/mtd/maps/physmap_of.c b/drivers/mtd/maps/physmap_of.c
index 14e8909..f39607d 100644
--- a/drivers/mtd/maps/physmap_of.c
+++ b/drivers/mtd/maps/physmap_of.c
@@ -20,6 +20,7 @@
 #include <linux/mtd/map.h>
 #include <linux/mtd/partitions.h>
 #include <linux/mtd/concat.h>
+#include <linux/mtd/cfi_endian.h>
 #include <linux/of.h>
 #include <linux/of_address.h>
 #include <linux/of_platform.h>
@@ -243,6 +244,9 @@ static int of_flash_probe(struct platform_device *dev)
 		info->list[i].map.bankwidth = be32_to_cpup(width);
 		info->list[i].map.device_node = dp;
 
+		if (of_property_read_bool(dp->parent, "big-endian"))
+			info->list[i].map.swap = CFI_BIG_ENDIAN;
+
 		err = of_flash_probe_gemini(dev, dp, &info->list[i].map);
 		if (err)
 			return err;
-- 
2.7.4

Re: [PATCH] driver: mtd: update struct map_info's swap as per map requirement.

[Boris Brezillon]

Hi Prabhakar,

On Thu, 4 May 2017 12:50:37 +0530
Prabhakar Kushwaha <prabhakar.kushwaha@nxp.com> wrote:

> It is not necessary for all device's maps to be CFI_HOST_ENDIAN. Maps device
> can be Bigendian or little endian.
> 
> Currently it is being taken care using CONFIG_MTD_CFI_LE_BYTE_SWAP or
> CONFIG_MTD_CFI_BE_BYTE_SWAP i.e. compile time.
> 
> Now update struct map_info's swap field based on device characteristics
> defined in device tree.

You should update Documentation/devicetree/bindings/mtd/mtd-physmap.txt
and Cc DT maintainers.

> 
> Signed-off-by: Prabhakar Kushwaha <prabhakar.kushwaha@nxp.com>
> ---
>  drivers/mtd/maps/physmap_of.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/drivers/mtd/maps/physmap_of.c b/drivers/mtd/maps/physmap_of.c
> index 14e8909..f39607d 100644
> --- a/drivers/mtd/maps/physmap_of.c
> +++ b/drivers/mtd/maps/physmap_of.c
> @@ -20,6 +20,7 @@
>  #include <linux/mtd/map.h>
>  #include <linux/mtd/partitions.h>
>  #include <linux/mtd/concat.h>
> +#include <linux/mtd/cfi_endian.h>
>  #include <linux/of.h>
>  #include <linux/of_address.h>
>  #include <linux/of_platform.h>
> @@ -243,6 +244,9 @@ static int of_flash_probe(struct platform_device *dev)
>  		info->list[i].map.bankwidth = be32_to_cpup(width);
>  		info->list[i].map.device_node = dp;
>  
> +		if (of_property_read_bool(dp->parent, "big-endian"))
> +			info->list[i].map.swap = CFI_BIG_ENDIAN;
> +

Shouldn't we have

		else if (of_property_read_bool(dp->parent,
					       "little-endian"))
			info->list[i].map.swap = CFI_LITTE_ENDIAN;
		else if (of_property_read_bool(dp->parent,
					       "host-endian"))
			info->list[i].map.swap = CFI_HOST_ENDIAN;

Otherwise, you'll fallback to CFI_DEFAULT_ENDIAN (determined with the
CONFIG_MTD_CFI_BE/LE_BYTE_SWAP config options) if the endianness is
not big-endian, which is probably not what you want.

You can even make this property a string property:

		if (!of_property_read_string(dp->parent,
					     "cfi-endianness",
					     &endianness)) {
			if (!strcmp("big", endianness))
				info->list[i].map.swap = CFI_BIG_ENDIAN;
			else if (!strcmp("little", endianness))
				info->list[i].map.swap = CFI_LITTLE_ENDIAN;
			else if (!strcmp("host", endianness))
				info->list[i].map.swap = CFI_HOST_ENDIAN;
		}

>  		err = of_flash_probe_gemini(dev, dp, &info->list[i].map);
>  		if (err)
>  			return err;

Regards,

Boris

Re: [PATCH] driver: mtd: update struct map_info's swap as per map requirement.

[Scott Wood]

On Mon, 2017-05-15 at 10:48 +0200, Boris Brezillon wrote:
> 
> You can even make this property a string property:
> 
> 		if (!of_property_read_string(dp->parent,
> 					     "cfi-endianness",
> 					     &endianness)) {
> 			if (!strcmp("big", endianness))
> 				info->list[i].map.swap = CFI_BIG_ENDIAN;
> 			else if (!strcmp("little", endianness))
> 				info->list[i].map.swap = CFI_LITTLE_ENDIAN;
> 			else if (!strcmp("host", endianness))
> 				info->list[i].map.swap = CFI_HOST_ENDIAN;
> 		}

Boolean endianness properties are the standard elsewhere -- why should CFI be
different?

See Documentation/devicetree/bindings/common-properties.txt

-Scott

Re: [PATCH] driver: mtd: update struct map_info's swap as per map requirement.

[Boris Brezillon]

On Mon, 15 May 2017 12:28:33 -0500
Scott Wood <oss@buserror.net> wrote:

> On Mon, 2017-05-15 at 10:48 +0200, Boris Brezillon wrote:
> > 
> > You can even make this property a string property:
> > 
> > 		if (!of_property_read_string(dp->parent,
> > 					     "cfi-endianness",
> > 					     &endianness)) {
> > 			if (!strcmp("big", endianness))
> > 				info->list[i].map.swap = CFI_BIG_ENDIAN;
> > 			else if (!strcmp("little", endianness))
> > 				info->list[i].map.swap = CFI_LITTLE_ENDIAN;
> > 			else if (!strcmp("host", endianness))
> > 				info->list[i].map.swap = CFI_HOST_ENDIAN;
> > 		}  
> 
> Boolean endianness properties are the standard elsewhere -- why should CFI be
> different?
> 
> See Documentation/devicetree/bindings/common-properties.txt

Forget this suggestion, I didn't know xxx-endian were standard props. My
first comment still applies though.