of: add support for reading s32 property value
Subsystems (2)
PATCH 1/2 of: add support for reading s32 property value
Add of_property_read_s32() to read the signed 32bit number from dt property value. This supports to pass the -ve numbers from dt. Use 2's complement method for represnting negative number and passed as u32 from dts. When reading back the value, again converted to 2's complement if msb shows as 1.
---
We required to pass the -ve value from DT and after looking the discussion
http://www.mail-archive.com/[email protected]/msg05108.html
I created this patch.
I CCed to David and Stephen on this.
include/linux/of.h | 16 ++++++++++++++++ 1 files changed, 16 insertions(+), 0 deletions(-)
include/linux/of.h +16 -0
@@ -601,6 +601,22 @@ static inline int of_property_read_u32(const struct device_node *np,return of_property_read_u32_array (np, propname, out_value,1 ); } +static inline int of_property_read_s32 (const struct device_node *np, +const char *propname, +s32 *out_value) +{ +u32 val; +int ret; + + ret =of_property_read_u32 (np, propname, &val); +if (ret <0 ) +return ret; + +/* 2's complement if MSB is 1 */ + *out_value = (val &0x80000000U ) ? -((val ^0xFFFFFFFFU ) +1 ) : val; +return ret; +} +#define of_property_for_each_u32 (np, propname, prop, p, u) \ for (prop = of_find_property(np, propname, NULL), \ p = of_prop_next_u32(prop, NULL, &u); \
Signature · 1 line
1.7.0.4
PATCH 2/2 dt-binding: add header to define types and conversion
Add a header file to define the macros for type conversion to be used by DTS file.
Add macro for the S32 to U32 conversion by using 2's complement.
--- include/dt-bindings/types.h | 18 ++++++++++++++++++ 1 files changed, 18 insertions(+), 0 deletions(-) create mode 100644 include/dt-bindings/types.h
include/dt-bindings/types.h new file +18 -0
@@ -0,0 +1,18 @@ +/* +* This header provides macros for different types and conversions +*/ + +#ifndef _DT_BINDINGS_TYPES_H_ +#define _DT_BINDINGS_TYPES_H_ + +/* +* S32_TO_U32: This macro converts the signed number to 2's complement +* unisgned number. E.g. S32_TO_U32(-3) will be 0xfffffffd and +* S32_TO_U32(3) will be 0x3; +* Use of_property_read_s32() for getting back the correct signed value +* in driver. +*/ +#define S32_TO_U32 (x) (((x) < 0) ? (((-(x)) ^ 0xFFFFFFFFU) + 1) : (x)) + +#endif +
Signature · 1 line
1.7.0.4
Quoted 11 lines · Laxman Dewangan
Add
of_property_read_s32()to read the signed 32bit number from dt property value. This supports to pass the -ve numbers from dt. Use 2's complement method for represnting negative number and passed as u32 from dts. When reading back the value, again converted to 2's complement if msb shows as 1.include/linux/of.h +3
-0+static inline int of_property_read_s32 (const struct device_node *np, +const char *propname, +s32 *out_value)
include/linux/of.h static inline int of_property_read_u32(const struct device_node *np, · line 616
Quoted 10 lines
include/linux/of.h +10
-0+{ +u32 val; +int ret; + + ret =of_property_read_u32 (np, propname, &val); +if (ret <0 ) +return ret; + +/* 2's complement if MSB is 1 */ + *out_value = (val &0x80000000U ) ? -((val ^0xFFFFFFFFU ) +1 ) : val;
I may not be thinking straight today since I have a cold, but doesn't patch 2/2 encode negative values as 2's complement, and an s32 variable in the kernel is also encoded as 2's complement, so all you need here is a cast:
*out_value = (s32)val;
... since the cast doesn't change the binary representation?
Quoted 25 lines · Stephen Warren
Quoted 9 lines · Laxman Dewangan
Add
of_property_read_s32()to read the signed 32bit number from dt property value. This supports to pass the -ve numbers from dt. Use 2's complement method for represnting negative number and passed as u32 from dts. When reading back the value, again converted to 2's complement if msb shows as 1.include/linux/of.h +3
-0+static inline int of_property_read_s32 (const struct device_node *np, +const char *propname, +s32 *out_value)include/linux/of.h static inline int of_property_read_u32(const struct device_node *np, · line 616
Quoted 10 lines
include/linux/of.h +10
-0+{ +u32 val; +int ret; + + ret =of_property_read_u32 (np, propname, &val); +if (ret <0 ) +return ret; + +/* 2's complement if MSB is 1 */ + *out_value = (val &0x80000000U ) ? -((val ^0xFFFFFFFFU ) +1 ) : val;I may not be thinking straight today since I have a cold, but doesn't patch 2/2 encode negative values as 2's complement, and an s32 variable in the kernel is also encoded as 2's complement, so all you need here is a cast:
Quoted 3 lines
*out_value = (s32)val;... since the cast doesn't change the binary representation?
This is correct if all architecture follows the 2's complement for -Ve number representation. If this is true then the patch will become more easy.
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html