Re: [PATCH 10/17] string: Improve generic memchr

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

On Fri, Sep 2, 2022 at 1:45 PM Adhemerval Zanella via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
> New algorithm have the following key differences:
>
>   - Reads first word unaligned and use string-maskoff function to
>     remove unwanted data.  This strategy follow arch-specific
>     optimization used on aarch64 and powerpc.
>
>   - Use string-fz{b,i} and string-opthr 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/memchr.c                               | 168 +++++-------------
>  .../powerpc32/power4/multiarch/memchr-ppc32.c |  14 +-
>  .../powerpc64/multiarch/memchr-ppc64.c        |   9 +-
>  3 files changed, 48 insertions(+), 143 deletions(-)
>
> diff --git a/string/memchr.c b/string/memchr.c
> index 422bcd0cd6..8fe0ac48ab 100644
> --- a/string/memchr.c
> +++ b/string/memchr.c
> @@ -1,10 +1,6 @@
> -/* Copyright (C) 1991-2022 Free Software Foundation, Inc.
> +/* Scan memory for a character.  Generic version
> +   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
> @@ -20,143 +16,65 @@
>     License along with the GNU C Library; if not, see
>     <https://www.gnu.org/licenses/>.  */
>
> -#ifndef _LIBC
> -# include <config.h>
> -#endif
> -
> +#include <intprops.h>
> +#include <string-fza.h>
> +#include <string-fzb.h>
> +#include <string-fzi.h>
> +#include <string-maskoff.h>
> +#include <string-opthr.h>
>  #include <string.h>
>
> -#include <stddef.h>
> +#undef memchr
>
> -#include <limits.h>
> -
> -#undef __memchr
> -#ifdef _LIBC
> -# undef memchr
> +#ifdef MEMCHR
> +# define __memchr MEMCHR
>  #endif
>
> -#ifndef weak_alias
> -# define __memchr memchr
> -#endif
> -
> -#ifndef MEMCHR
> -# define MEMCHR __memchr
> -#endif
> +static inline const char *
> +sadd (uintptr_t x, uintptr_t y)
> +{
> +  uintptr_t ret = INT_ADD_OVERFLOW (x, y) ? (uintptr_t)-1 : x + y;
> +  return (const char *)ret;
> +}
>
>  /* Search no more than N bytes of S for C.  */
>  void *
> -MEMCHR (void const *s, int c_in, size_t n)
> +__memchr (void const *s, int c_in, size_t n)
>  {
> -  /* On 32-bit hardware, choosing longword to be a 32-bit unsigned
> -     long instead of a 64-bit uintmax_t tends to give better
> -     performance.  On 64-bit hardware, unsigned long is generally 64
> -     bits already.  Change this typedef to experiment with
> -     performance.  */
> -  typedef unsigned long int longword;
> +  if (__glibc_unlikely (n == 0))
> +    return NULL;
>
> -  const unsigned char *char_ptr;
> -  const longword *longword_ptr;
> -  longword repeated_one;
> -  longword repeated_c;
> -  unsigned char c;
> +  uintptr_t s_int = (uintptr_t) s;
>
> -  c = (unsigned char) c_in;
> +  /* Set up a word, each of whose bytes is C.  */
> +  op_t repeated_c = repeat_bytes (c_in);
> +  op_t before_mask = create_mask (s_int);
>
> -  /* Handle the first few bytes by reading one byte at a time.
> -     Do this until CHAR_PTR is aligned on a longword boundary.  */
> -  for (char_ptr = (const unsigned char *) s;
> -       n > 0 && (size_t) char_ptr % sizeof (longword) != 0;
> -       --n, ++char_ptr)
> -    if (*char_ptr == c)
> -      return (void *) char_ptr;
> +  /* Compute the address of the last byte taking in consideration possible
> +     overflow.  */
> +  const char *lbyte = sadd (s_int, n - 1);
>
> -  longword_ptr = (const longword *) char_ptr;
> +  /* Compute the address of the word containing the last byte. */
> +  const op_t *lword = word_containing (lbyte);
>
> -  /* All these elucidatory comments refer to 4-byte longwords,
> -     but the theory applies equally well to any size longwords.  */
> +  /* Read the first word, but munge it so that bytes before the array
> +     will not match goal.  */
> +  const op_t * word_ptr = word_containing (s);
> +  op_t word = (*word_ptr | before_mask) ^ (repeated_c & before_mask);

Why do you xor with repeated_c & before_mask here?

Doesn't the has_eq(word, repeated_c) do that?
>
> -  /* Compute auxiliary longword values:
> -     repeated_one is a value which has a 1 in every byte.
> -     repeated_c has c in every byte.  */
> -  repeated_one = 0x01010101;
> -  repeated_c = c | (c << 8);
> -  repeated_c |= repeated_c << 16;
> -  if (0xffffffffU < (longword) -1)
> +  while (has_eq (word, repeated_c) == 0)
>      {
> -      repeated_one |= repeated_one << 31 << 1;
> -      repeated_c |= repeated_c << 31 << 1;
> -      if (8 < sizeof (longword))
> -       {
> -         size_t i;
> -
> -         for (i = 64; i < sizeof (longword) * 8; i *= 2)
> -           {
> -             repeated_one |= repeated_one << i;
> -             repeated_c |= repeated_c << i;
> -           }
> -       }
> +      if (word_ptr == lword)
> +       return NULL;
> +      word = *++word_ptr;
>      }
>
> -  /* Instead of the traditional loop which tests each byte, 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 equal to c.  We first use an xor
> -     with repeated_c.  This reduces the task to testing whether *any of the
> -     four* bytes in longword1 is zero.
> -
> -     We compute tmp =
> -       ((longword1 - repeated_one) & ~longword1) & (repeated_one << 7).
> -     That is, we perform the following operations:
> -       1. Subtract repeated_one.
> -       2. & ~longword1.
> -       3. & a mask consisting of 0x80 in every byte.
> -     Consider what happens in each byte:
> -       - If a byte of longword1 is zero, step 1 and 2 transform it into 0xff,
> -        and step 3 transforms it into 0x80.  A carry can also be propagated
> -        to more significant bytes.
> -       - If a byte of longword1 is nonzero, let its lowest 1 bit be at
> -        position k (0 <= k <= 7); so the lowest k bits are 0.  After step 1,
> -        the byte ends in a single bit of value 0 and k bits of value 1.
> -        After step 2, the result is just k bits of value 1: 2^k - 1.  After
> -        step 3, the result is 0.  And no carry is produced.
> -     So, if longword1 has only non-zero bytes, tmp is zero.
> -     Whereas if longword1 has a zero byte, call j the position of the least
> -     significant zero byte.  Then the result has a zero at positions 0, ...,
> -     j-1 and a 0x80 at position j.  We cannot predict the result at the more
> -     significant bytes (positions j+1..3), but it does not matter since we
> -     already have a non-zero bit at position 8*j+7.
> -
> -     So, the test whether any byte in longword1 is zero is equivalent to
> -     testing whether tmp is nonzero.  */
> -
> -  while (n >= sizeof (longword))
> -    {
> -      longword longword1 = *longword_ptr ^ repeated_c;
> -
> -      if ((((longword1 - repeated_one) & ~longword1)
> -          & (repeated_one << 7)) != 0)
> -       break;
> -      longword_ptr++;
> -      n -= sizeof (longword);
> -    }
> -
> -  char_ptr = (const unsigned char *) longword_ptr;
> -
> -  /* At this point, we know that either n < sizeof (longword), or one of the
> -     sizeof (longword) bytes starting at char_ptr is == c.  On little-endian
> -     machines, we could determine the first such byte without any further
> -     memory accesses, just by looking at the tmp result from the last loop
> -     iteration.  But this does not work on big-endian machines.  Choose code
> -     that works in both cases.  */
> -
> -  for (; n > 0; --n, ++char_ptr)
> -    {
> -      if (*char_ptr == c)
> -       return (void *) char_ptr;
> -    }
> -
> -  return NULL;
> +  /* We found a match, but it might be in a byte past the end
> +     of the array.  */
> +  char *ret = (char *) word_ptr + index_first_eq (word, repeated_c);
> +  return (ret <= lbyte) ? ret : NULL;
>  }
> -#ifdef weak_alias
> +#ifndef MEMCHR
>  weak_alias (__memchr, memchr)
> -#endif
>  libc_hidden_builtin_def (memchr)
> +#endif
> diff --git a/sysdeps/powerpc/powerpc32/power4/multiarch/memchr-ppc32.c b/sysdeps/powerpc/powerpc32/power4/multiarch/memchr-ppc32.c
> index fc69df54b3..02877d3c98 100644
> --- a/sysdeps/powerpc/powerpc32/power4/multiarch/memchr-ppc32.c
> +++ b/sysdeps/powerpc/powerpc32/power4/multiarch/memchr-ppc32.c
> @@ -18,17 +18,11 @@
>
>  #include <string.h>
>
> -#define MEMCHR  __memchr_ppc
> +extern __typeof (memchr) __memchr_ppc attribute_hidden;
>
> -#undef weak_alias
> -#define weak_alias(a, b)
> +#define MEMCHR  __memchr_ppc
> +#include <string/memchr.c>
>
>  #ifdef SHARED
> -# undef libc_hidden_builtin_def
> -# define libc_hidden_builtin_def(name) \
> -  __hidden_ver1(__memchr_ppc, __GI_memchr, __memchr_ppc);
> +__hidden_ver1(__memchr_ppc, __GI_memchr, __memchr_ppc);
>  #endif
> -
> -extern __typeof (memchr) __memchr_ppc attribute_hidden;
> -
> -#include <string/memchr.c>
> diff --git a/sysdeps/powerpc/powerpc64/multiarch/memchr-ppc64.c b/sysdeps/powerpc/powerpc64/multiarch/memchr-ppc64.c
> index 3c966f4403..15beca787b 100644
> --- a/sysdeps/powerpc/powerpc64/multiarch/memchr-ppc64.c
> +++ b/sysdeps/powerpc/powerpc64/multiarch/memchr-ppc64.c
> @@ -18,14 +18,7 @@
>
>  #include <string.h>
>
> -#define MEMCHR  __memchr_ppc
> -
> -#undef weak_alias
> -#define weak_alias(a, b)
> -
> -# undef libc_hidden_builtin_def
> -# define libc_hidden_builtin_def(name)
> -
>  extern __typeof (memchr) __memchr_ppc attribute_hidden;
>
> +#define MEMCHR  __memchr_ppc
>  #include <string/memchr.c>
> --
> 2.34.1
>