Re: [PATCH v5 08/17] string: Improve generic strchrnul

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

On Tue, Jan 10, 2023 at 6:18 AM Adhemerval Zanella Netto
<adhemerval.zanella@linaro.org> wrote:
>
>
>
> On 09/01/23 20:33, Richard Henderson wrote:
> > On 1/9/23 12:35, Adhemerval Zanella Netto wrote:
> >>
> >>
> >> On 05/01/23 20:17, Noah Goldstein wrote:
> >>> On Mon, Sep 19, 2022 at 1:04 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} functions.
> >>>>
> >>>> Checked on x86_64-linux-gnu, i686-linux-gnu, powerpc64-linux-gnu,
> >>>> and powerpc-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/strchrnul.c                            | 156 +++---------------
> >>>>   .../power4/multiarch/strchrnul-ppc32.c        |   4 -
> >>>>   sysdeps/s390/strchrnul-c.c                    |   2 -
> >>>>   3 files changed, 24 insertions(+), 138 deletions(-)
> >>>>
> >>>> diff --git a/string/strchrnul.c b/string/strchrnul.c
> >>>> index 0cc1fc6bb0..67defa3dab 100644
> >>>> --- a/string/strchrnul.c
> >>>> +++ b/string/strchrnul.c
> >>>> @@ -1,10 +1,5 @@
> >>>>   /* 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
> >>>> -   bug fix and commentary by Jim Blandy (jimb@ai.mit.edu);
> >>>> -   adaptation to strchr 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
> >>>> @@ -21,146 +16,43 @@
> >>>>      <https://www.gnu.org/licenses/>.  */
> >>>>
> >>>>   #include <string.h>
> >>>> -#include <memcopy.h>
> >>>>   #include <stdlib.h>
> >>>> +#include <stdint.h>
> >>>> +#include <string-fza.h>
> >>>> +#include <string-fzb.h>
> >>>> +#include <string-fzi.h>
> >>>> +#include <string-maskoff.h>
> >>>>
> >>>>   #undef __strchrnul
> >>>>   #undef strchrnul
> >>>>
> >>>> -#ifndef STRCHRNUL
> >>>> -# define STRCHRNUL __strchrnul
> >>>> +#ifdef STRCHRNUL
> >>>> +# define __strchrnul STRCHRNUL
> >>>>   #endif
> >>>>
> >>>>   /* Find the first occurrence of C in S or the final NUL byte.  */
> >>>>   char *
> >>>> -STRCHRNUL (const char *s, int c_in)
> >>>> +__strchrnul (const char *str, int c_in)
> >>>>   {
> >>>> -  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 first 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;
> >>>> -       ((unsigned long int) char_ptr & (sizeof (longword) - 1)) != 0;
> >>>> -       ++char_ptr)
> >>>> -    if (*char_ptr == c || *char_ptr == '\0')
> >>>> -      return (void *) char_ptr;
> >>>> -
> >>>> -  /* All these elucidatory comments refer to 4-byte longwords,
> >>>> -     but the theory applies equally well to 8-byte longwords.  */
> >>>> -
> >>>> -  longword_ptr = (unsigned long int *) char_ptr;
> >>>> -
> >>>> -  /* 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 (sizeof (longword) > 4)
> >>>> -    /* Do the shift in two steps to avoid a warning if long has 32 bits.  */
> >>>> -    charmask |= (charmask << 16) << 16;
> >>>> -  if (sizeof (longword) > 8)
> >>>> -    abort ();
> >>>> -
> >>>> -  /* 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.  */
> >>>> -  for (;;)
> >>>> -    {
> >>>> -      /* 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.
> >>>> +  /* Set up a word, each of whose bytes is C.  */
> >>>> +  op_t repeated_c = repeat_bytes (c_in);
> >>>>
> >>>> -        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.
> >>>> +  /* Align the input address to op_t.  */
> >>>> +  uintptr_t s_int = (uintptr_t) str;
> >>>> +  const op_t *word_ptr = word_containing (str);
> >>>>
> >>>> -        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!
> >>>> +  /* Read the first aligned word, but force bytes before the string to
> >>>> +     match neither zero nor goal (we make sure the high bit of each byte
> >>>> +     is 1, and the low 7 bits are all the opposite of the goal byte).  */
> >>>> +  op_t bmask = create_mask (s_int);
> >>>> +  op_t word = (*word_ptr | bmask) ^ (repeated_c & highbit_mask (bmask));
> >>>
> >>> Think much clearer (and probably better codegen) is:
> >>> find_zero_eq_low/all(word, repeated) >> (s_int * CHAR_BIT)
> >>
> >> It does not seem to work, at least not replacing the two lines with:
> >>
> >>    op_t word = find_zero_eq_all/low (*word_ptr, repeated_c) >> (s_int * CHAR_BIT);
> >
> > Oh, two fine points:
> >
> > (1) big-endian would want shifting left,
> > (2) alpha would want shifting by bits not bytes,
> >     because the cmpbge insn produces an 8-bit mask.
> >
> > so you'd need to hide this shift in the headers like create_mask().
>
> Alright, the following works:
>
>
> static __always_inline op_t
> check_mask (op_t word, uintptr_t s_int)
> {
>   if (__BYTE_ORDER == __LITTLE_ENDIAN)
>     return word >> (CHAR_BIT * (s_int % sizeof (s_int)));
>   else
>     return word << (CHAR_BIT * (s_int % sizeof (s_int)));
> }

Imo put this in with "[PATCH v5 03/17] Add string-maskoff.h generic header"
think may also be needed for memchr.
>
> char *
> __strchrnul (const char *str, int c_in)
> {
>   op_t repeated_c = repeat_bytes (c_in);
>
>   uintptr_t s_int = (uintptr_t) str;
>   const op_t *word_ptr = word_containing (str);
>
>   op_t word = *word_ptr;
>
>   op_t mask = check_mask (find_zero_eq_all (word, repeated_c), s_int);
>   if (mask != 0)
>     return (char *) str + index_first_(mask);
>
>   do
>     word = *++word_ptr;
>   while (! has_zero_eq (word, repeated_c));
>
>   op_t found = index_first_zero_eq (word, repeated_c);
>   return (char *) word_ptr + found;
> }
>
>
> I had to use find_zero_eq_all to avoid uninitialized bytes, that triggered
> some regression on tests that use strchr (for instance test-strpbrk).
>
> I will update the patch based on this version.