Re: [PATCH v5 11/17] string: Improve generic memrchr

[Noah Goldstein <goldstein.w.n@gmail.com>]

On Mon, Sep 19, 2022 at 1:02 PM Adhemerval Zanella via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
> New algorithm have the following key differences:
>
>   - Use string-fz{b,i} functions.
>
> Checked on x86_64-linux-gnu, i686-linux-gnu, powerpc-linux-gnu,
> and powerpc64-linux-gnu by removing the arch-specific assembly
> implementation and disabling multi-arch (it covers both LE and BE
> for 64 and 32 bits).
>
> Co-authored-by: Richard Henderson  <rth@twiddle.net>
> ---
>  string/memrchr.c | 189 ++++++++---------------------------------------
>  1 file changed, 32 insertions(+), 157 deletions(-)
>
> diff --git a/string/memrchr.c b/string/memrchr.c
> index 8eb6829e45..5491689c66 100644
> --- a/string/memrchr.c
> +++ b/string/memrchr.c
> @@ -1,11 +1,6 @@
>  /* memrchr -- find the last occurrence of a byte in a memory block
>     Copyright (C) 1991-2022 Free Software Foundation, Inc.
>     This file is part of the GNU C Library.
> -   Based on strlen implementation by Torbjorn Granlund (tege@sics.se),
> -   with help from Dan Sahlin (dan@sics.se) and
> -   commentary by Jim Blandy (jimb@ai.mit.edu);
> -   adaptation to memchr suggested by Dick Karpinski (dick@cca.ucsf.edu),
> -   and implemented by Roland McGrath (roland@ai.mit.edu).
>
>     The GNU C Library is free software; you can redistribute it and/or
>     modify it under the terms of the GNU Lesser General Public
> @@ -21,177 +16,57 @@
>     License along with the GNU C Library; if not, see
>     <https://www.gnu.org/licenses/>.  */
>
> -#include <stdlib.h>
> -
> -#ifdef HAVE_CONFIG_H
> -# include <config.h>
> -#endif
> -
> -#if defined _LIBC
> -# include <string.h>
> -# include <memcopy.h>
> -#endif
> -
> -#if defined HAVE_LIMITS_H || defined _LIBC
> -# include <limits.h>
> -#endif
> -
> -#define LONG_MAX_32_BITS 2147483647
> -
> -#ifndef LONG_MAX
> -# define LONG_MAX LONG_MAX_32_BITS
> -#endif
> -
> -#include <sys/types.h>
> +#include <string-fzb.h>
> +#include <string-fzi.h>
> +#include <string-maskoff.h>
> +#include <string-opthr.h>
> +#include <string.h>
>
>  #undef __memrchr
>  #undef memrchr
>
> -#ifndef weak_alias
> -# define __memrchr memrchr
> +#ifdef MEMRCHR
> +# define __memrchr MEMRCHR
>  #endif
>
> -/* Search no more than N bytes of S for C.  */
>  void *
> -#ifndef MEMRCHR
> -__memrchr
> -#else
> -MEMRCHR
> -#endif
> -     (const void *s, int c_in, size_t n)
> +__memrchr (const void *s, int c_in, size_t n)
>  {
> -  const unsigned char *char_ptr;
> -  const unsigned long int *longword_ptr;
> -  unsigned long int longword, magic_bits, charmask;
> -  unsigned char c;
> -
> -  c = (unsigned char) c_in;
> -
>    /* Handle the last few characters by reading one character at a time.
> -     Do this until CHAR_PTR is aligned on a longword boundary.  */
> -  for (char_ptr = (const unsigned char *) s + n;
> -       n > 0 && ((unsigned long int) char_ptr
> -                & (sizeof (longword) - 1)) != 0;
> -       --n)
> -    if (*--char_ptr == c)
> +     Do this until CHAR_PTR is aligned on a word boundary, or
> +     the entirety of small inputs.  */
> +  const unsigned char *char_ptr = (const unsigned char *) (s + n);
> +  size_t align = (uintptr_t) char_ptr  % sizeof (op_t);
> +  if (n < OP_T_THRES || align > n)
> +    align = n;
> +  for (size_t i = 0; i < align; ++i)
> +    if (*--char_ptr == c_in)
>        return (void *) char_ptr;
>
> -  /* All these elucidatory comments refer to 4-byte longwords,
> -     but the theory applies equally well to 8-byte longwords.  */
> +  const op_t *word_ptr = (const op_t *) char_ptr;
> +  n -= align;
> +  if (__glibc_unlikely (n == 0))
> +    return NULL;
>
> -  longword_ptr = (const unsigned long int *) char_ptr;
> +  /* Compute the address of the word containing the initial byte. */
> +  const op_t *lword = word_containing (s);
>
> -  /* Bits 31, 24, 16, and 8 of this number are zero.  Call these bits
> -     the "holes."  Note that there is a hole just to the left of
> -     each byte, with an extra at the end:
> -
> -     bits:  01111110 11111110 11111110 11111111
> -     bytes: AAAAAAAA BBBBBBBB CCCCCCCC DDDDDDDD
> -
> -     The 1-bits make sure that carries propagate to the next 0-bit.
> -     The 0-bits provide holes for carries to fall into.  */
> -  magic_bits = -1;
> -  magic_bits = magic_bits / 0xff * 0xfe << 1 >> 1 | 1;
> -
> -  /* Set up a longword, each of whose bytes is C.  */
> -  charmask = c | (c << 8);
> -  charmask |= charmask << 16;
> -#if LONG_MAX > LONG_MAX_32_BITS
> -  charmask |= charmask << 32;
> -#endif
> +  /* Set up a word, each of whose bytes is C.  */
> +  op_t repeated_c = repeat_bytes (c_in);
>
> -  /* Instead of the traditional loop which tests each character,
> -     we will test a longword at a time.  The tricky part is testing
> -     if *any of the four* bytes in the longword in question are zero.  */
> -  while (n >= sizeof (longword))
> +  while (word_ptr != lword)
again I would make lword, lword - 1, and move the out-of-bounds return check to
the return outside of the loop.
>      {
> -      /* We tentatively exit the loop if adding MAGIC_BITS to
> -        LONGWORD fails to change any of the hole bits of LONGWORD.
> -
> -        1) Is this safe?  Will it catch all the zero bytes?
> -        Suppose there is a byte with all zeros.  Any carry bits
> -        propagating from its left will fall into the hole at its
> -        least significant bit and stop.  Since there will be no
> -        carry from its most significant bit, the LSB of the
> -        byte to the left will be unchanged, and the zero will be
> -        detected.
> -
> -        2) Is this worthwhile?  Will it ignore everything except
> -        zero bytes?  Suppose every byte of LONGWORD has a bit set
> -        somewhere.  There will be a carry into bit 8.  If bit 8
> -        is set, this will carry into bit 16.  If bit 8 is clear,
> -        one of bits 9-15 must be set, so there will be a carry
> -        into bit 16.  Similarly, there will be a carry into bit
> -        24.  If one of bits 24-30 is set, there will be a carry
> -        into bit 31, so all of the hole bits will be changed.
> -
> -        The one misfire occurs when bits 24-30 are clear and bit
> -        31 is set; in this case, the hole at bit 31 is not
> -        changed.  If we had access to the processor carry flag,
> -        we could close this loophole by putting the fourth hole
> -        at bit 32!
> -
> -        So it ignores everything except 128's, when they're aligned
> -        properly.
> -
> -        3) But wait!  Aren't we looking for C, not zero?
> -        Good point.  So what we do is XOR LONGWORD with a longword,
> -        each of whose bytes is C.  This turns each byte that is C
> -        into a zero.  */
> -
> -      longword = *--longword_ptr ^ charmask;
> -
> -      /* Add MAGIC_BITS to LONGWORD.  */
> -      if ((((longword + magic_bits)
> -
> -           /* Set those bits that were unchanged by the addition.  */
> -           ^ ~longword)
> -
> -          /* Look at only the hole bits.  If any of the hole bits
> -             are unchanged, most likely one of the bytes was a
> -             zero.  */
> -          & ~magic_bits) != 0)
> +      op_t word = *--word_ptr;
> +      if (has_eq (word, repeated_c))
>         {
> -         /* Which of the bytes was C?  If none of them were, it was
> -            a misfire; continue the search.  */
> -
> -         const unsigned char *cp = (const unsigned char *) longword_ptr;
> -
> -#if LONG_MAX > 2147483647
> -         if (cp[7] == c)
> -           return (void *) &cp[7];
> -         if (cp[6] == c)
> -           return (void *) &cp[6];
> -         if (cp[5] == c)
> -           return (void *) &cp[5];
> -         if (cp[4] == c)
> -           return (void *) &cp[4];
> -#endif
> -         if (cp[3] == c)
> -           return (void *) &cp[3];
> -         if (cp[2] == c)
> -           return (void *) &cp[2];
> -         if (cp[1] == c)
> -           return (void *) &cp[1];
> -         if (cp[0] == c)
> -           return (void *) cp;
> +         /* We found a match, but it might be in a byte past the start
> +            of the array.  */
> +         char *ret = (char *) word_ptr + index_last_eq (word, repeated_c);
> +         return ret >= (char *) s ? ret : NULL;
>         }
> -
> -      n -= sizeof (longword);
>      }
> -
> -  char_ptr = (const unsigned char *) longword_ptr;
> -
> -  while (n-- > 0)
> -    {
> -      if (*--char_ptr == c)
> -       return (void *) char_ptr;
> -    }
> -
> -  return 0;
> +  return NULL;
>  }
>  #ifndef MEMRCHR
> -# ifdef weak_alias
>  weak_alias (__memrchr, memrchr)
> -# endif
>  #endif
> --
> 2.34.1
>