[v4] aligned_alloc: conform to C17

Changes from V3:

* Simplified aligned_alloc symbol definition.

References:
https://patchwork.sourceware.org/project/glibc/patch/33ec9e0c1e587813b90e8aa771c2c8e6e379dd48.camel@posteo.net/
https://sourceware.org/bugzilla/show_bug.cgi?id=20137
https://sourceware.org/pipermail/libc-alpha/2023-February/145858.html

The memory.texi portion matches Martin's proposed patch.

man page portion, quoted to avoid CI/CD issues (I can send an official
patch separately after the glibc patch is applied):

> diff --git a/man3/posix_memalign.3 b/man3/posix_memalign.3
> index f5d6618b7..a73ff0421 100644
> --- a/man3/posix_memalign.3
> +++ b/man3/posix_memalign.3
> @@ -91,9 +91,8 @@ The function
>  is the same as
>  .BR memalign (),
>  except for the added restriction that
> -.I size
> -should be a multiple of
> -.IR alignment .
> +.I alignment
> +must be a power of two.
>  .PP
>  The obsolete function
>  .BR valloc ()

From 9cc7d558d9c06edb95102151212b319f21acf603 Mon Sep 17 00:00:00 2001
From: DJ Delorie <dj@redhat.com>
Date: Tue, 21 Mar 2023 00:46:43 -0400
Subject: aligned_alloc: conform to C17

This patch adds the strict checking for power-of-two alignments
in aligned_alloc(), and updates the manual accordingly.

Ping?

https://patchwork.sourceware.org/project/glibc/patch/xnh6uczcu3.fsf@greed.delorie.com/
https://sourceware.org/pipermail/libc-alpha/2023-March/146579.html

On 3/22/23 17:38, DJ Delorie via Libc-alpha wrote:
> 
> Changes from V3:
> 
> * Simplified aligned_alloc symbol definition.
> 

Looking forward to a v5.

> References:
> https://patchwork.sourceware.org/project/glibc/patch/33ec9e0c1e587813b90e8aa771c2c8e6e379dd48.camel@posteo.net/
> https://sourceware.org/bugzilla/show_bug.cgi?id=20137
> https://sourceware.org/pipermail/libc-alpha/2023-February/145858.html
> 
> The memory.texi portion matches Martin's proposed patch.
> 
> man page portion, quoted to avoid CI/CD issues (I can send an official
> patch separately after the glibc patch is applied):
> 
>> diff --git a/man3/posix_memalign.3 b/man3/posix_memalign.3
>> index f5d6618b7..a73ff0421 100644
>> --- a/man3/posix_memalign.3
>> +++ b/man3/posix_memalign.3
>> @@ -91,9 +91,8 @@ The function
>>  is the same as
>>  .BR memalign (),
>>  except for the added restriction that
>> -.I size
>> -should be a multiple of
>> -.IR alignment .
>> +.I alignment
>> +must be a power of two.
>>  .PP
>>  The obsolete function
>>  .BR valloc ()
> 
> From 9cc7d558d9c06edb95102151212b319f21acf603 Mon Sep 17 00:00:00 2001
> From: DJ Delorie <dj@redhat.com>
> Date: Tue, 21 Mar 2023 00:46:43 -0400
> Subject: aligned_alloc: conform to C17
> 
> This patch adds the strict checking for power-of-two alignments
> in aligned_alloc(), and updates the manual accordingly.
> 
> diff --git a/malloc/Makefile b/malloc/Makefile
> index dfb51d344c..691714fb52 100644
> --- a/malloc/Makefile
> +++ b/malloc/Makefile
> @@ -43,10 +43,12 @@ tests := mallocbug tst-malloc tst-valloc tst-calloc tst-obstack \
>  	 tst-tcfree1 tst-tcfree2 tst-tcfree3 \
>  	 tst-safe-linking \
>  	 tst-mallocalign1 \
> +	 tst-aligned-alloc \

OK. Adds a new test.

>  
>  tests-static := \
>  	 tst-interpose-static-nothread \
> -	 tst-interpose-static-thread
> +	 tst-interpose-static-thread \
> +	 tst-aligned-alloc-static

OK. Adds a new test (built static).

>  
>  # Test for the malloc_set_state symbol removed in glibc 2.25.
>  ifeq ($(have-GLIBC_2.23)$(build-shared),yesyes)
> diff --git a/malloc/malloc-debug.c b/malloc/malloc-debug.c
> index 3867d15698..da9d2340d3 100644
> --- a/malloc/malloc-debug.c
> +++ b/malloc/malloc-debug.c
> @@ -299,7 +299,14 @@ __debug_memalign (size_t alignment, size_t bytes)
>    return _debug_mid_memalign (alignment, bytes, RETURN_ADDRESS (0));
>  }
>  strong_alias (__debug_memalign, memalign)
> -strong_alias (__debug_memalign, aligned_alloc)
> +static void *
> +__debug_aligned_alloc (size_t alignment, size_t bytes)
> +{
> +  if (!powerof2 (alignment) || alignment == 0)
> +    return NULL;
> +  return _debug_mid_memalign (alignment, bytes, RETURN_ADDRESS (0));
> +}
> +strong_alias (__debug_aligned_alloc, aligned_alloc)

OK. Correctly provide a debug alias for aligned_alloc.

>  
>  static void *
>  __debug_pvalloc (size_t bytes)
> diff --git a/malloc/malloc.c b/malloc/malloc.c
> index 76c50e3f58..ece5b8a224 100644
> --- a/malloc/malloc.c
> +++ b/malloc/malloc.c
> @@ -3509,6 +3509,27 @@ __libc_memalign (size_t alignment, size_t bytes)
>    void *address = RETURN_ADDRESS (0);
>    return _mid_memalign (alignment, bytes, address);
>  }
> +libc_hidden_def (__libc_memalign)
> +
> +/* For ISO C11.  */

This is not true, it's for ISO C17 (starting with n2310).

> +void *
> +weak_function
> +aligned_alloc (size_t alignment, size_t bytes)
> +{
> +  if (!__malloc_initialized)
> +    ptmalloc_init ();
> +
> +  /* Similar to memalign, but ISO C17 requires an error for invalid
> +     alignments.  Valid alignments are non-negative powers of two.  */

This isn't quite correct.

Suggest:

/* Similar to memalign, but starting with ISO C17 the standard requires an error for
   alignments that are not supported by the implementation.  Valid alignments for
   the current implementation are non-negative powers of two.  */

The error is only in the case of alignments that are not supported by the implementation.

> +  if (!powerof2 (alignment) || alignment == 0)

OK. Assumes MALLOC_ALIGNMENT is a power of 2, but that's true generally.

> +    {
> +      __set_errno (EINVAL);
> +      return 0;
> +    }
> +
> +  void *address = RETURN_ADDRESS (0);
> +  return _mid_memalign (alignment, bytes, address);

OK. We'll likely see back a higher alignment due to MALLOC_ALIGNMENT, but that's still aligned.

> +}
>  
>  static void *
>  _mid_memalign (size_t alignment, size_t bytes, void *address)
> @@ -3567,9 +3588,6 @@ _mid_memalign (size_t alignment, size_t bytes, void *address)
>            ar_ptr == arena_for_chunk (mem2chunk (p)));
>    return tag_new_usable (p);
>  }
> -/* For ISO C11.  */
> -weak_alias (__libc_memalign, aligned_alloc)
> -libc_hidden_def (__libc_memalign)

OK.

>  
>  void *
>  __libc_valloc (size_t bytes)
> diff --git a/malloc/tst-aligned-alloc-static.c b/malloc/tst-aligned-alloc-static.c
> new file mode 100644
> index 0000000000..d504473094
> --- /dev/null
> +++ b/malloc/tst-aligned-alloc-static.c
> @@ -0,0 +1 @@
> +#include "tst-aligned-alloc.c"
> diff --git a/malloc/tst-aligned-alloc.c b/malloc/tst-aligned-alloc.c
> new file mode 100644
> index 0000000000..d6739376d4
> --- /dev/null
> +++ b/malloc/tst-aligned-alloc.c
> @@ -0,0 +1,56 @@
> +/* Copyright (C) 2023 Free Software Foundation, Inc.

Needs a first line explaining what the test does.

> +   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
> +   <https://www.gnu.org/licenses/>.  */
> +
> +#include <errno.h>
> +#include <malloc.h>
> +#include <stdint.h>
> +#include <stdio.h>
> +#include <stdlib.h>
> +#include <string.h>
> +#include <libc-diag.h>
> +#include <support/check.h>
> +
> +static int
> +do_test (void)
> +{
> +  void *p1;
> +  void *p2;
> +  void *p3;
> +
> +  errno = 0;
> +

Suggest:

/* The implementation supports alignments that are non-negative powers of 2.
   We test 5 distinct conditions here:
   - A non-negative power of 2 alignment e.g. 64.
   - A degenerate zero power of 2 alignment e.g. 1.
   - A non-power-of-2 alignment e.g. 65.
   - A zero alignment.
   - A corner case SIZE_MAX / 2 + 1 alignment. */

> +  p1 = aligned_alloc (64, 64);
> +
> +  if (p1 == NULL)
> +    FAIL_EXIT1 ("aligned_alloc(64, 64) failed");

Please add a degenerate test for alignment == 1 e.g. 2^(0)

> +
> +  p2 = aligned_alloc (65, 64);
> +
> +  if (p2 != NULL)
> +    FAIL_EXIT1 ("aligned_alloc(65, 64) did not fail");
> +
> +  p3 = aligned_alloc (0, 64);
> +
> +  if (p3 != NULL)
> +    FAIL_EXIT1 ("aligned_alloc(0, 64) did not fail");

Please add a SIZE_MAX / 2 + 1 alignment case that has to fail.

> +
> +  free (p1);
> +  return 0;
> +}
> +
> +#define TEST_FUNCTION do_test ()
> +#include "../test-skeleton.c"
> diff --git a/manual/memory.texi b/manual/memory.texi
> index 9d3398a326..8952ff2bfa 100644
> --- a/manual/memory.texi
> +++ b/manual/memory.texi
> @@ -995,7 +995,7 @@ power of two than that, use @code{aligned_alloc} or @code{posix_memalign}.
>  @c Alias to memalign.
>  The @code{aligned_alloc} function allocates a block of @var{size} bytes whose
>  address is a multiple of @var{alignment}.  The @var{alignment} must be a
> -power of two and @var{size} must be a multiple of @var{alignment}.
> +power of two.

OK.

>  
>  The @code{aligned_alloc} function returns a null pointer on error and sets
>  @code{errno} to one of the following values:
>