diff options
author | Ahmad Fatoum <a.fatoum@pengutronix.de> | 2023-11-22 18:29:39 +0100 |
---|---|---|
committer | Sascha Hauer <s.hauer@pengutronix.de> | 2023-11-23 15:50:35 +0100 |
commit | dde1c3e59b88a74a2eb85e9874384936ae20cb57 (patch) | |
tree | 06fc76c2535ddae56085ef42b8888d7adb14eef7 /arch/powerpc | |
parent | 2376924e24c003def53712d91524f8ffaa6903e7 (diff) | |
download | barebox-dde1c3e59b88a74a2eb85e9874384936ae20cb57.tar.gz barebox-dde1c3e59b88a74a2eb85e9874384936ae20cb57.tar.xz |
include: import Linux word-at-a-time.h
The Linux <linux/word-at-a-time.h> interface is used to optimize
searching for bytes in strings by doing word-size comparisons. This will
be used in the implementation of strscpy in a follow-up commit, so
import the generic version here.
A good overview on the interface is available at LWN[1]. Note that
it discuss Linux v3.5. The asm-generic version imported here works
also on little-endian and not only big-endian[2].
[1]: https://lwn.net/Articles/501492/
[2]: Linux kernel commit a6e2f029ae34 ("Make asm/word-at-a-time.h
available on all architectures").
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Link: https://lore.barebox.org/20231122172951.376531-9-a.fatoum@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Diffstat (limited to 'arch/powerpc')
-rw-r--r-- | arch/powerpc/include/asm/bitops.h | 1 | ||||
-rw-r--r-- | arch/powerpc/include/asm/word-at-a-time.h | 162 |
2 files changed, 163 insertions, 0 deletions
diff --git a/arch/powerpc/include/asm/bitops.h b/arch/powerpc/include/asm/bitops.h index 543612e6ed..62b09c7da4 100644 --- a/arch/powerpc/include/asm/bitops.h +++ b/arch/powerpc/include/asm/bitops.h @@ -184,6 +184,7 @@ static inline int ffs(int x) return __ilog2(x & -x) + 1; } +#include <asm-generic/bitops/__fls.h> #include <asm-generic/bitops/fls64.h> #include <asm-generic/bitops/hweight.h> diff --git a/arch/powerpc/include/asm/word-at-a-time.h b/arch/powerpc/include/asm/word-at-a-time.h new file mode 100644 index 0000000000..029740c6de --- /dev/null +++ b/arch/powerpc/include/asm/word-at-a-time.h @@ -0,0 +1,162 @@ +#ifndef _ASM_WORD_AT_A_TIME_H +#define _ASM_WORD_AT_A_TIME_H + +/* + * Word-at-a-time interfaces for PowerPC. + */ + +#include <linux/kernel.h> +#include <linux/bitops.h> + +#ifdef __BIG_ENDIAN__ + +struct word_at_a_time { + const unsigned long high_bits, low_bits; +}; + +#define WORD_AT_A_TIME_CONSTANTS { REPEAT_BYTE(0xfe) + 1, REPEAT_BYTE(0x7f) } + +/* Bit set in the bytes that have a zero */ +static inline long prep_zero_mask(unsigned long val, unsigned long rhs, const struct word_at_a_time *c) +{ + unsigned long mask = (val & c->low_bits) + c->low_bits; + return ~(mask | rhs); +} + +#define create_zero_mask(mask) (mask) + +static inline long find_zero(unsigned long mask) +{ + long leading_zero_bits; + +#ifdef __powerpc64__ + asm ("cntlzd %0,%1" : "=r" (leading_zero_bits) : "r" (mask)); +#else + asm ("cntlzw %0,%1" : "=r" (leading_zero_bits) : "r" (mask)); +#endif + + return leading_zero_bits >> 3; +} + +static inline unsigned long has_zero(unsigned long val, unsigned long *data, const struct word_at_a_time *c) +{ + unsigned long rhs = val | c->low_bits; + *data = rhs; + return (val + c->high_bits) & ~rhs; +} + +static inline unsigned long zero_bytemask(unsigned long mask) +{ + return ~1ul << __fls(mask); +} + +#else + +#ifdef CONFIG_64BIT + +/* unused */ +struct word_at_a_time { +}; + +#define WORD_AT_A_TIME_CONSTANTS { } + +/* This will give us 0xff for a NULL char and 0x00 elsewhere */ +static inline unsigned long has_zero(unsigned long a, unsigned long *bits, const struct word_at_a_time *c) +{ + unsigned long ret; + unsigned long zero = 0; + + asm("cmpb %0,%1,%2" : "=r" (ret) : "r" (a), "r" (zero)); + *bits = ret; + + return ret; +} + +static inline unsigned long prep_zero_mask(unsigned long a, unsigned long bits, const struct word_at_a_time *c) +{ + return bits; +} + +/* Alan Modra's little-endian strlen tail for 64-bit */ +static inline unsigned long create_zero_mask(unsigned long bits) +{ + unsigned long leading_zero_bits; + long trailing_zero_bit_mask; + + asm("addi %1,%2,-1\n\t" + "andc %1,%1,%2\n\t" + "popcntd %0,%1" + : "=r" (leading_zero_bits), "=&r" (trailing_zero_bit_mask) + : "b" (bits)); + + return leading_zero_bits; +} + +static inline unsigned long find_zero(unsigned long mask) +{ + return mask >> 3; +} + +/* This assumes that we never ask for an all 1s bitmask */ +static inline unsigned long zero_bytemask(unsigned long mask) +{ + return (1UL << mask) - 1; +} + +#else /* 32-bit case */ + +struct word_at_a_time { + const unsigned long one_bits, high_bits; +}; + +#define WORD_AT_A_TIME_CONSTANTS { REPEAT_BYTE(0x01), REPEAT_BYTE(0x80) } + +/* + * This is largely generic for little-endian machines, but the + * optimal byte mask counting is probably going to be something + * that is architecture-specific. If you have a reliably fast + * bit count instruction, that might be better than the multiply + * and shift, for example. + */ + +/* Carl Chatfield / Jan Achrenius G+ version for 32-bit */ +static inline long count_masked_bytes(long mask) +{ + /* (000000 0000ff 00ffff ffffff) -> ( 1 1 2 3 ) */ + long a = (0x0ff0001+mask) >> 23; + /* Fix the 1 for 00 case */ + return a & mask; +} + +static inline unsigned long create_zero_mask(unsigned long bits) +{ + bits = (bits - 1) & ~bits; + return bits >> 7; +} + +static inline unsigned long find_zero(unsigned long mask) +{ + return count_masked_bytes(mask); +} + +/* Return nonzero if it has a zero */ +static inline unsigned long has_zero(unsigned long a, unsigned long *bits, const struct word_at_a_time *c) +{ + unsigned long mask = ((a - c->one_bits) & ~a) & c->high_bits; + *bits = mask; + return mask; +} + +static inline unsigned long prep_zero_mask(unsigned long a, unsigned long bits, const struct word_at_a_time *c) +{ + return bits; +} + +/* The mask we created is directly usable as a bytemask */ +#define zero_bytemask(mask) (mask) + +#endif /* CONFIG_64BIT */ + +#endif /* __BIG_ENDIAN__ */ + +#endif /* _ASM_WORD_AT_A_TIME_H */ |