public inbox for libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: Florian Weimer <fweimer@redhat.com>
To: Adhemerval Zanella via Libc-alpha <libc-alpha@sourceware.org>
Subject: Re: [PATCH 3/3] posix: Add posix_spawn_file_actions_closefrom
Date: Tue, 22 Dec 2020 12:32:45 +0100	[thread overview]
Message-ID: <87ft3yulcy.fsf@oldenburg2.str.redhat.com> (raw)
In-Reply-To: <20201221215027.2176966-3-adhemerval.zanella@linaro.org> (Adhemerval Zanella via Libc-alpha's message of "Mon, 21 Dec 2020 18:50:27 -0300")

* Adhemerval Zanella via Libc-alpha:

> +* The posix_spawn_file_actions_closefrom_np has been added, enabling
> +  posix_spawn and posix_spawnp to close all file descriptors greater than
> +  a giver interger.  This function is a GNU extension, although Solaris also
> +  provides a similar function.

Two typos: “giver interger”

> diff --git a/posix/spawn.h b/posix/spawn.h
> index be6bd591a3..21ea563425 100644
> --- a/posix/spawn.h
> +++ b/posix/spawn.h
> @@ -213,6 +213,13 @@ extern int posix_spawn_file_actions_addchdir_np (posix_spawn_file_actions_t *
>  extern int posix_spawn_file_actions_addfchdir_np (posix_spawn_file_actions_t *,
>  						  int __fd)
>       __THROW __nonnull ((1));
> +
> +/* Add an action to close all file descriptor greater than FROM during
> +   spawn.  This affects the subsequent file actions.  */
> +extern int posix_spawn_file_actions_addclosefrom_np (posix_spawn_file_actions_t *,
> +						     int __from)
> +     __THROW __nonnull ((1));
> +
>  #endif

Line length issue (but I'm not sure how to avoid that).

> diff --git a/posix/spawn_faction_addclosefrom.c b/posix/spawn_faction_addclosefrom.c
> new file mode 100644
> index 0000000000..3a295580bc
> --- /dev/null
> +++ b/posix/spawn_faction_addclosefrom.c

> +int
> +__posix_spawn_file_actions_addclosefrom (posix_spawn_file_actions_t
> +					 *file_actions, int from)
> +{
> +#if __SPAWN_SUPPORT_CLOSEFROM
> +  struct __spawn_action *rec;
> +
> +  if (!__spawn_valid_fd (from))
> +    return EBADF;
> +
> +  /* Allocate more memory if needed.  */
> +  if (file_actions->__used == file_actions->__allocated
> +      && __posix_spawn_file_actions_realloc (file_actions) != 0)
> +    /* This can only mean we ran out of memory.  */
> +    return ENOMEM;
> +
> +  /* Add the new value.  */
> +  rec = &file_actions->__actions[file_actions->__used];
> +  rec->tag = spawn_do_closefrom;
> +  rec->action.closefrom_action.from = from;
> +
> +  /* Account for the new entry.  */
> +  ++file_actions->__used;
> +
> +  return 0;
> +#else
> +  __set_errno (EINVAL);
> +  return -1;
> +#endif
> +}

Should this return EINVAL for the fallback case?

> diff --git a/posix/spawn_int_abi.h b/posix/spawn_int_abi.h
> new file mode 100644
> index 0000000000..5fda4d3442
> --- /dev/null
> +++ b/posix/spawn_int_abi.h
> @@ -0,0 +1,27 @@
> +/* Internal ABI specific for posix_spawn functionality.  Generic version.
> +   Copyright (C) 2020 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
> +   <http://www.gnu.org/licenses/>.  */
> +
> +#ifndef _SPAWN_INT_ABI_H
> +#define _SPAWN_INT_ABI_H
> +
> +/* The closefrom file actions requires either a syscall or an arch-specific
> +   way to interact over all file descriptors and act uppon them (such
> +   /proc/self/fd on Linux).  */
> +#define __SPAWN_SUPPOR_CLOSEFROM 0
> +
> +#endif /* _SPAWN_INT_H */

I think Hurd has support for this via __getdtablesize, so perhaps this
is not needed?

In any case, ABI is a bit of a misnomer here, and this file is
uncessary, given the sysdeps/generic file.

> diff --git a/posix/tst-spawn5.c b/posix/tst-spawn5.c
> new file mode 100644
> index 0000000000..1f901d112b
> --- /dev/null
> +++ b/posix/tst-spawn5.c

> +enum
> +{
> +  maximum_fd = 100,
> +  half_fd = maximum_fd / 2,
> +};
> +static int fds[maximum_fd + 1];

The usual comments about file descriptor layout. 8-)

> +static void
> +do_test_closefrom (void)
> +{
> +  for (int i = 0; i < array_length (fds); i++)
> +    fds[i] = xopen ("/dev/null", O_WRONLY, 0);

This should perhaps check that the descriptors 57, …, 90 have actually
been opened, perhaps implicitly by checking that the resulting
descriptors end up as one block.

> +  /* Close the remmaining but the last one.  */

Typo: “remmaining”

> +  if (restart)
> +    handle_restart (argc, argv);
> +
> +  initial_argv[0] = argv[1]; /* path for ld.so  */
> +  initial_argv[1] = argv[2]; /* "--library-path"  */
> +  initial_argv[2] = argv[3]; /* the library path  */
> +  initial_argv[3] = argv[4]; /* the application name  */
> +  initial_argv[4] = (char *) "--direct";
> +  initial_argv[5] = (char *) "--restart";
> +
> +  do_test_closefrom ();
> +
> +  return 0;
> +}
> +
> +#define TEST_FUNCTION_ARGV do_test
> +#include <support/test-driver.c>

You could simplify the reinvocation logic by making this a static or a
container test.

> diff --git a/sysdeps/unix/sysv/linux/spawni.c b/sysdeps/unix/sysv/linux/spawni.c
> index f157bfffd2..f496578d19 100644
> --- a/sysdeps/unix/sysv/linux/spawni.c
> +++ b/sysdeps/unix/sysv/linux/spawni.c
> @@ -16,22 +16,17 @@
>     License along with the GNU C Library; if not, see
>     <https://www.gnu.org/licenses/>.  */
>  
> -#include <spawn.h>
> -#include <fcntl.h>
> -#include <paths.h>
> -#include <string.h>
> -#include <sys/resource.h>
> -#include <sys/wait.h>
> -#include <sys/param.h>
> -#include <sys/mman.h>
> -#include <not-cancel.h>
> +#include <arch-fd_to_filename.h>
> +#include <internal-signals.h>
> +#include <ldsodefs.h>
>  #include <local-setxid.h>
> +#include <not-cancel.h>
> +#include <paths.h>
>  #include <shlib-compat.h>
> -#include <nptl/pthreadP.h>
> -#include <dl-sysdep.h>
> -#include <libc-pointer-arith.h>
> -#include <ldsodefs.h>
> -#include "spawn_int.h"
> +#include <spawn.h>
> +#include <spawn_int.h>
> +#include <sysdep.h>
> +#include <sys/resource.h>

Suprious changes?

Thanks,
Florian
-- 
Red Hat GmbH, https://de.redhat.com/ , Registered seat: Grasbrunn,
Commercial register: Amtsgericht Muenchen, HRB 153243,
Managing Directors: Charles Cachera, Brian Klemm, Laurie Krebs, Michael O'Neill


  reply	other threads:[~2020-12-22 11:32 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-12-21 21:50 [PATCH 1/3] Linux: Add close_range Adhemerval Zanella
2020-12-21 21:50 ` [PATCH 2/3] io: Add closefrom [BZ #10353] Adhemerval Zanella
2020-12-22 10:49   ` Florian Weimer
2020-12-22 11:50     ` Adhemerval Zanella
2020-12-22 11:57       ` Florian Weimer
2020-12-22 12:41         ` Adhemerval Zanella
2020-12-22 13:17           ` Florian Weimer
2020-12-22 13:28             ` Adhemerval Zanella
2020-12-22 13:34               ` Florian Weimer
2020-12-21 21:50 ` [PATCH 3/3] posix: Add posix_spawn_file_actions_closefrom Adhemerval Zanella
2020-12-22 11:32   ` Florian Weimer [this message]
2020-12-22 11:56     ` Adhemerval Zanella
2020-12-22 12:00       ` Florian Weimer
2020-12-22 12:46         ` Adhemerval Zanella
2020-12-22  9:05 ` [PATCH 1/3] Linux: Add close_range Florian Weimer
2020-12-22 11:35   ` Adhemerval Zanella
2020-12-22 11:41     ` Florian Weimer
2020-12-22 11:47       ` Adhemerval Zanella
2020-12-23 12:33       ` Christian Brauner

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

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

  git send-email \
    --in-reply-to=87ft3yulcy.fsf@oldenburg2.str.redhat.com \
    --to=fweimer@redhat.com \
    --cc=libc-alpha@sourceware.org \
    /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).