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

[Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>]

On 10/01/23 14:17, Noah Goldstein wrote:
> 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.

Yeap, this is what I have done.