unofficial mirror of libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: Carlos O'Donell via Libc-alpha <libc-alpha@sourceware.org>
To: Adhemerval Zanella <adhemerval.zanella@linaro.org>,
	libc-alpha@sourceware.org, Szabolcs Nagy <szabolcs.nagy@arm.com>,
	"H . J . Lu" <hjl.tools@gmail.com>
Cc: Christian Brauner <brauner@kernel.org>
Subject: Re: [PATCH v2 03/11] Linux: Do not align the stack for __clone3
Date: Wed, 11 Jan 2023 16:43:26 -0500	[thread overview]
Message-ID: <ca59571d-336b-f39a-4c03-d9cec9f1bd42@redhat.com> (raw)
In-Reply-To: <20221104190112.2566409-4-adhemerval.zanella@linaro.org>

On 11/4/22 15:01, Adhemerval Zanella via Libc-alpha wrote:
> Adjusting the stack on child thread is not async-signal safe, since
> kernel does not enforce stack alignmnet.  Although, all the internal
> usage of __clone_internal are done with all signal masked, removing
> stack alignment of __clone3 is still a net gain: is simplifies the
> function contract detail (mask/unmarsk signals) and simplifies the
> arch-specific code.

Please post v4.

Suggest:

All internal callers of __clone3 should provide an already aligned stack.
Removing the stack alignment in __clone3 is a net gain: it simplifies the
internal function contract (mask/unmask signals) along with the arch-specific
code.

> 
> The internal users are responsible for correct stack alignment.  If
> glibc ever exports a clone3 wrapper, it should adjust stack alignment.
> 
> Checked on x86_64-linux-gnu.
> ---
>  include/clone_internal.h                      |  5 ++
>  .../sysv/linux/tst-misalign-clone-internal.c  | 74 -------------------
>  sysdeps/unix/sysv/linux/x86_64/clone3.S       |  3 -
>  3 files changed, 5 insertions(+), 77 deletions(-)
>  delete mode 100644 sysdeps/unix/sysv/linux/tst-misalign-clone-internal.c
> 
> diff --git a/include/clone_internal.h b/include/clone_internal.h
> index 73b8114df4..5a48c7caa5 100644
> --- a/include/clone_internal.h
> +++ b/include/clone_internal.h
> @@ -11,6 +11,11 @@
>     Different than kernel, the implementation also returns EINVAL for an
>     invalid NULL __CL_ARGS or __FUNC (similar to __clone).
>  
> +   The stack is not aligned prior the syscall, different than exported
> +   __clone.  Adjusting the stack on child is not async-signal safe and since
> +   it is used only internally the caller is responsible for proper stack
> +   alignment.

Suggest:

All callers are responsible for correctly aligning the stack.  The stack is not aligned
prior to the syscall (this differs from the exported __clone).


> +
>     This function is only implemented if the ABI defines HAVE_CLONE3_WRAPPER.
>  */
>  extern int __clone3 (struct clone_args *__cl_args, size_t __size,
> diff --git a/sysdeps/unix/sysv/linux/tst-misalign-clone-internal.c b/sysdeps/unix/sysv/linux/tst-misalign-clone-internal.c
> deleted file mode 100644
> index e039ca5a3a..0000000000
> --- a/sysdeps/unix/sysv/linux/tst-misalign-clone-internal.c
> +++ /dev/null
> @@ -1,74 +0,0 @@
> -/* Verify that __clone_internal properly aligns the child stack.
> -   Copyright (C) 2021-2022 Free Software Foundation, Inc.
> -   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 <sched.h>
> -#include <stdbool.h>
> -#include <stdint.h>
> -#include <stdio.h>
> -#include <string.h>
> -#include <sys/wait.h>
> -#include <unistd.h>
> -#include <libc-pointer-arith.h>
> -#include <tst-stack-align.h>
> -#include <clone_internal.h>
> -#include <support/xunistd.h>
> -#include <support/check.h>
> -
> -static int
> -check_stack_alignment (void *arg)
> -{
> -  puts ("in f");
> -
> -  return TEST_STACK_ALIGN () ? 1 : 0;
> -}
> -
> -static int
> -do_test (void)
> -{
> -  puts ("in do_test");
> -
> -  if (TEST_STACK_ALIGN ())
> -    FAIL_EXIT1 ("stack isn't aligned\n");
> -
> -#ifdef __ia64__
> -# define STACK_SIZE (256 * 1024)
> -#else
> -# define STACK_SIZE (128 * 1024)
> -#endif
> -  char st[STACK_SIZE + 1];
> -  /* NB: Align child stack to 1 byte.  */
> -  char *stack = PTR_ALIGN_UP (&st[0], 2) + 1;
> -  struct clone_args clone_args =
> -    {
> -      .stack = (uintptr_t) stack,
> -      .stack_size = STACK_SIZE,
> -    };
> -  pid_t p = __clone_internal (&clone_args, check_stack_alignment, 0);
> -
> -  /* Clone must not fail.  */
> -  TEST_VERIFY_EXIT (p != -1);
> -
> -  int e;
> -  xwaitpid (p, &e, __WCLONE);
> -  TEST_VERIFY (WIFEXITED (e));
> -  TEST_COMPARE (WEXITSTATUS (e), 0);
> -
> -  return 0;
> -}
> -
> -#include <support/test-driver.c>

OK.

> diff --git a/sysdeps/unix/sysv/linux/x86_64/clone3.S b/sysdeps/unix/sysv/linux/x86_64/clone3.S
> index 18fc2ee966..989b62bdf6 100644
> --- a/sysdeps/unix/sysv/linux/x86_64/clone3.S
> +++ b/sysdeps/unix/sysv/linux/x86_64/clone3.S
> @@ -73,9 +73,6 @@ L(thread_start):
>  	   the outermost frame obviously.  */
>  	xorl	%ebp, %ebp
>  
> -	/* Align stack to 16 bytes per the x86-64 psABI.  */
> -	and	$-16, %RSP_LP

OK.

> -
>  	/* Set up arguments for the function call.  */
>  	mov	%R8_LP, %RDI_LP	/* Argument.  */
>  	call	*%rdx		/* Call function.  */

-- 
Cheers,
Carlos.


  reply	other threads:[~2023-01-11 21:43 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-11-04 19:01 [PATCH v2 00/11] Optimize posix_spawn signal setup with clone3 Adhemerval Zanella via Libc-alpha
2022-11-04 19:01 ` [PATCH v2 01/11] linux: Do not reset signal handler in posix_spawn if it is already SIG_DFL Adhemerval Zanella via Libc-alpha
2023-01-11 21:27   ` Carlos O'Donell via Libc-alpha
2022-11-04 19:01 ` [PATCH v2 02/11] linux: Extend internal clone3 documentation Adhemerval Zanella via Libc-alpha
2022-11-07 16:40   ` Szabolcs Nagy via Libc-alpha
2023-01-11 21:35   ` Carlos O'Donell via Libc-alpha
2022-11-04 19:01 ` [PATCH v2 03/11] Linux: Do not align the stack for __clone3 Adhemerval Zanella via Libc-alpha
2023-01-11 21:43   ` Carlos O'Donell via Libc-alpha [this message]
2022-11-04 19:01 ` [PATCH v2 04/11] linux: Add clone3 CLONE_CLEAR_SIGHAND optimization to posix_spawn Adhemerval Zanella via Libc-alpha
2023-01-11 21:44   ` Carlos O'Donell via Libc-alpha
2022-11-04 19:01 ` [PATCH v2 05/11] powerpc64: Add the clone3 wrapper Adhemerval Zanella via Libc-alpha
2022-11-28 16:38   ` Paul E Murphy via Libc-alpha
2022-11-29 20:43     ` Adhemerval Zanella Netto via Libc-alpha
2022-11-04 19:01 ` [PATCH v2 06/11] aarch64: " Adhemerval Zanella via Libc-alpha
2022-11-07 16:14   ` Szabolcs Nagy via Libc-alpha
2022-11-04 19:01 ` [PATCH v2 07/11] s390x: " Adhemerval Zanella via Libc-alpha
2022-11-04 19:01 ` [PATCH v2 08/11] riscv: " Adhemerval Zanella via Libc-alpha
2022-11-04 19:01 ` [PATCH v2 09/11] arm: " Adhemerval Zanella via Libc-alpha
2022-11-04 19:01 ` [PATCH v2 10/11] mips: " Adhemerval Zanella via Libc-alpha
2022-11-04 19:01 ` [PATCH v2 11/11] Linux: optimize clone3 internal usage Adhemerval Zanella via Libc-alpha
2023-01-11 21:45   ` Carlos O'Donell via Libc-alpha
2023-01-11 21:24 ` [PATCH v2 00/11] Optimize posix_spawn signal setup with clone3 Carlos O'Donell via Libc-alpha

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: https://www.gnu.org/software/libc/involved.html

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=ca59571d-336b-f39a-4c03-d9cec9f1bd42@redhat.com \
    --to=libc-alpha@sourceware.org \
    --cc=adhemerval.zanella@linaro.org \
    --cc=brauner@kernel.org \
    --cc=carlos@redhat.com \
    --cc=hjl.tools@gmail.com \
    --cc=szabolcs.nagy@arm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).