mirror of
https://sourceware.org/git/glibc.git
synced 2024-11-22 21:10:07 +00:00
00cb84dde7
This patch adds generic string find and detection meant to be used in
generic vectorized string implementation. The idea is to decompose the
basic string operation so each architecture can reimplement if it
provides any specialized hardware instruction.
The 'string-misc.h' provides miscellaneous functions:
- extractbyte: extracts the byte from an specific index.
- repeat_bytes: setup an word by replicate the argument on each byte.
The 'string-fza.h' provides zero byte detection functions:
- find_zero_low, find_zero_all, find_eq_low, find_eq_all,
find_zero_eq_low, find_zero_eq_all, and find_zero_ne_all
The 'string-fzb.h' provides boolean zero byte detection functions:
- has_zero: determine if any byte within a word is zero.
- has_eq: determine byte equality between two words.
- has_zero_eq: determine if any byte within a word is zero along with
byte equality between two words.
The 'string-fzi.h' provides positions for string-fza.h results:
- index_first: return index of first zero byte within a word.
- index_last: return index of first byte different between two words.
The 'string-fzc.h' provides a combined version of fza and fzi:
- index_first_zero_eq: return index of first zero byte within a word or
first byte different between two words.
- index_first_zero_ne: return index of first zero byte within a word or
first byte equal between two words.
- index_last_zero: return index of last zero byte within a word.
- index_last_eq: return index of last byte different between two words.
The 'string-shift.h' provides a way to mask off parts of a work based on
some alignmnet (to handle unaligned arguments):
- shift_find, shift_find_last.
Co-authored-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Noah Goldstein <goldstein.w.n@gmail.com>
99 lines
3.2 KiB
C
99 lines
3.2 KiB
C
/* Basic zero byte detection. Generic C version.
|
|
Copyright (C) 2023 Free Software Foundation, Inc.
|
|
This file is part of the GNU C Library.
|
|
|
|
The GNU C Library is free software; you can redistribute it and/or
|
|
modify it under the terms of the GNU Lesser General Public
|
|
License as published by the Free Software Foundation; either
|
|
version 2.1 of the License, or (at your option) any later version.
|
|
|
|
The GNU C Library is distributed in the hope that it will be useful,
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
Lesser General Public License for more details.
|
|
|
|
You should have received a copy of the GNU Lesser General Public
|
|
License along with the GNU C Library; if not, see
|
|
<http://www.gnu.org/licenses/>. */
|
|
|
|
#ifndef _STRING_FZA_H
|
|
#define _STRING_FZA_H 1
|
|
|
|
#include <string-misc.h>
|
|
#include <string-optype.h>
|
|
|
|
/* The function return a byte mask. */
|
|
typedef op_t find_t;
|
|
|
|
/* This function returns non-zero if any byte in X is zero.
|
|
More specifically, at least one bit set within the least significant
|
|
byte that was zero; other bytes within the word are indeterminate. */
|
|
static __always_inline find_t
|
|
find_zero_low (op_t x)
|
|
{
|
|
/* This expression comes from
|
|
https://graphics.stanford.edu/~seander/bithacks.html#ZeroInWord
|
|
Subtracting 1 sets 0x80 in a byte that was 0; anding ~x clears
|
|
0x80 in a byte that was >= 128; anding 0x80 isolates that test bit. */
|
|
op_t lsb = repeat_bytes (0x01);
|
|
op_t msb = repeat_bytes (0x80);
|
|
return (x - lsb) & ~x & msb;
|
|
}
|
|
|
|
/* This function returns at least one bit set within every byte of X that
|
|
is zero. The result is exact in that, unlike find_zero_low, all bytes
|
|
are determinate. This is usually used for finding the index of the
|
|
most significant byte that was zero. */
|
|
static __always_inline find_t
|
|
find_zero_all (op_t x)
|
|
{
|
|
/* For each byte, find not-zero by
|
|
(0) And 0x7f so that we cannot carry between bytes,
|
|
(1) Add 0x7f so that non-zero carries into 0x80,
|
|
(2) Or in the original byte (which might have had 0x80 set).
|
|
Then invert and mask such that 0x80 is set iff that byte was zero. */
|
|
op_t m = repeat_bytes (0x7f);
|
|
return ~(((x & m) + m) | x | m);
|
|
}
|
|
|
|
/* With similar caveats, identify bytes that are equal between X1 and X2. */
|
|
static __always_inline find_t
|
|
find_eq_low (op_t x1, op_t x2)
|
|
{
|
|
return find_zero_low (x1 ^ x2);
|
|
}
|
|
|
|
static __always_inline find_t
|
|
find_eq_all (op_t x1, op_t x2)
|
|
{
|
|
return find_zero_all (x1 ^ x2);
|
|
}
|
|
|
|
/* With similar caveats, identify zero bytes in X1 and bytes that are
|
|
equal between in X1 and X2. */
|
|
static __always_inline find_t
|
|
find_zero_eq_low (op_t x1, op_t x2)
|
|
{
|
|
return find_zero_low (x1) | find_zero_low (x1 ^ x2);
|
|
}
|
|
|
|
static __always_inline find_t
|
|
find_zero_eq_all (op_t x1, op_t x2)
|
|
{
|
|
return find_zero_all (x1) | find_zero_all (x1 ^ x2);
|
|
}
|
|
|
|
/* With similar caveats, identify zero bytes in X1 and bytes that are
|
|
not equal between in X1 and X2. */
|
|
static __always_inline find_t
|
|
find_zero_ne_all (op_t x1, op_t x2)
|
|
{
|
|
op_t m = repeat_bytes (0x7f);
|
|
op_t eq = x1 ^ x2;
|
|
op_t nz1 = ((x1 & m) + m) | x1;
|
|
op_t ne2 = ((eq & m) + m) | eq;
|
|
return (ne2 | ~nz1) & ~m;
|
|
}
|
|
|
|
#endif /* _STRING_FZA_H */
|