Re: [PATCH v6 13/26] bitfield: Add non-constant field_{prep,get}() helpers

[Yury Norov <yury.norov@gmail.com>]

On Thu, Nov 06, 2025 at 02:34:01PM +0100, Geert Uytterhoeven wrote:
> The existing FIELD_{GET,PREP}() macros are limited to compile-time
> constants.  However, it is very common to prepare or extract bitfield
> elements where the bitfield mask is not a compile-time constant.
> 
> To avoid this limitation, the AT91 clock driver and several other
> drivers already have their own non-const field_{prep,get}() macros.
> Make them available for general use by adding them to
> <linux/bitfield.h>, and improve them slightly:
>   1. Avoid evaluating macro parameters more than once,
>   2. Replace "ffs() - 1" by "__ffs()",
>   3. Support 64-bit use on 32-bit architectures,
>   4. Wire field_{get,prep}() to FIELD_{GET,PREP}() when mask is
>      actually constant.
> 
> This is deliberately not merged into the existing FIELD_{GET,PREP}()
> macros, as people expressed the desire to keep stricter variants for
> increased safety, or for performance critical paths.
> 
> Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
> Acked-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
> Acked-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
> Acked-by: Crt Mori <cmo@melexis.com>
> Acked-by: Nuno Sá <nuno.sa@analog.com>
> Acked-by: Richard Genoud <richard.genoud@bootlin.com>
> Reviewed-by: Andy Shevchenko <andriy.shevchenko@intel.com>
> ---
> v6:
>   - Align \,
>   - Add Return sections to kerneldoc,
>   - Add Reviewed-by,
>   - Document field_{get,prep} in top comment block,
>   - Use less-checking __FIELD_{GET,PREP}() to avoid build issues with
>     clang and W=1:
>     https://lore.kernel.org/all/202510281335.UpSLYJG9-lkp@intel.com
>     https://lore.kernel.org/all/202510281414.DnejZh4n-lkp@intel.com
>     https://lore.kernel.org/all/202510281304.RK3J3c3t-lkp@intel.com
> 
> v5:
>   - Add Acked-by,
>   - Split off changes outside <linux/bitfield.h>,
>   - Document that mask must be non-zero,
>   - Document typical usage pattern,
>   - Recommend using FIELD_{PREP,GET}() directly to ensure compile-time
>     constant masks,
>   - Check BITS_PER_TYPE(mask) instead of sizeof(mask),
>   - Wire field_{get,prep}() to FIELD_{GET,PREP}() when mask is
>     constant, to improve type checking.
> 
> v4:
>   - Add Acked-by,
>   - Rebase on top of commit 7c68005a46108ffa ("crypto: qat - relocate
>     power management debugfs helper APIs") in v6.17-rc1,
>   - Convert more recently introduced upstream copies:
>       - drivers/edac/ie31200_edac.c
>       - drivers/iio/dac/ad3530r.c
> 
> v3:
>   - Add Acked-by,
>   - Drop underscores from macro parameters,
>   - Use __auto_type where possible,
>   - Correctly cast reg to the mask type,
>   - Introduces __val and __reg intermediates to simplify the actual
>     operation,
>   - Drop unneeded parentheses,
>   - Clarify having both FIELD_{GET,PREP}() and field_{get,prep}(),
> 
> v2:
>   - Cast val resp. reg to the mask type,
>   - Fix 64-bit use on 32-bit architectures,
>   - Convert new upstream users:
>       - drivers/crypto/intel/qat/qat_common/adf_gen4_pm_debugfs.c
>       - drivers/gpio/gpio-aspeed.c
>       - drivers/iio/temperature/mlx90614.c
>       - drivers/pinctrl/nuvoton/pinctrl-ma35.c
>       - sound/usb/mixer_quirks.c
>   - Convert new user queued in renesas-devel for v6.15:
>       - drivers/soc/renesas/rz-sysc.c
> ---
>  include/linux/bitfield.h | 59 ++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 59 insertions(+)
> 
> diff --git a/include/linux/bitfield.h b/include/linux/bitfield.h
> index bf8e0ae4b5b41038..f92e18c9629a59c6 100644
> --- a/include/linux/bitfield.h
> +++ b/include/linux/bitfield.h
> @@ -17,6 +17,7 @@
>   * FIELD_{GET,PREP} macros take as first parameter shifted mask
>   * from which they extract the base mask and shift amount.
>   * Mask must be a compilation time constant.
> + * field_{get,prep} are variants that take a non-const mask.
>   *
>   * Example:
>   *
> @@ -240,4 +241,62 @@ __MAKE_OP(64)
>  #undef __MAKE_OP
>  #undef ____MAKE_OP
>  
> +#define __field_prep(mask, val)						\
> +	({								\
> +		__auto_type __mask = (mask);				\
> +		typeof(mask) __val = (val);				\

typeof(__mask), I guess.

> +		unsigned int __shift = BITS_PER_TYPE(mask) <= 32 ?	\

Same here: BITS_PER_TYPE(__mask). BITS_PER_TYPE() doesn't evaluate the
expression because it is based on sizeof(). But we definitely don't
want people check it every time they read the code.

No need to send another version for this. I can fix it inplace.

Reviewed-by: Yury Norov (NVIDIA) <yury.norov@gmail.com>

> +				       __ffs(__mask) : __ffs64(__mask);	\
> +		(__val << __shift) & __mask;				\
> +	})
> +
> +#define __field_get(mask, reg)						\
> +	({								\
> +		__auto_type __mask = (mask);				\
> +		typeof(mask) __reg =  (reg);				\
> +		unsigned int __shift = BITS_PER_TYPE(mask) <= 32 ?	\
> +				       __ffs(__mask) : __ffs64(__mask);	\
> +		(__reg & __mask) >> __shift;				\
> +	})
> +
> +/**
> + * field_prep() - prepare a bitfield element
> + * @mask: shifted mask defining the field's length and position, must be
> + *        non-zero
> + * @val:  value to put in the field
> + *
> + * Return: field value masked and shifted to its final destination
> + *
> + * field_prep() masks and shifts up the value.  The result should be
> + * combined with other fields of the bitfield using logical OR.
> + * Unlike FIELD_PREP(), @mask is not limited to a compile-time constant.
> + * Typical usage patterns are a value stored in a table, or calculated by
> + * shifting a constant by a variable number of bits.
> + * If you want to ensure that @mask is a compile-time constant, please use
> + * FIELD_PREP() directly instead.
> + */
> +#define field_prep(mask, val)						\
> +	(__builtin_constant_p(mask) ? __FIELD_PREP(mask, val, "field_prep: ") \
> +				    : __field_prep(mask, val))
> +
> +/**
> + * field_get() - extract a bitfield element
> + * @mask: shifted mask defining the field's length and position, must be
> + *        non-zero
> + * @reg:  value of entire bitfield
> + *
> + * Return: extracted field value
> + *
> + * field_get() extracts the field specified by @mask from the
> + * bitfield passed in as @reg by masking and shifting it down.
> + * Unlike FIELD_GET(), @mask is not limited to a compile-time constant.
> + * Typical usage patterns are a value stored in a table, or calculated by
> + * shifting a constant by a variable number of bits.
> + * If you want to ensure that @mask is a compile-time constant, please use
> + * FIELD_GET() directly instead.
> + */
> +#define field_get(mask, reg)						\
> +	(__builtin_constant_p(mask) ? __FIELD_GET(mask, reg, "field_get: ") \
> +				    : __field_get(mask, reg))
> +
>  #endif
> -- 
> 2.43.0