about summary refs log tree commit diff
path: root/sysdeps/unix/sysv/linux/spawni.c
diff options
context:
space:
mode:
authorAdhemerval Zanella <adhemerval.zanella@linaro.org>2016-01-19 17:33:32 -0200
committerAdhemerval Zanella <adhemerval.zanella@linaro.com>2016-03-07 11:53:47 +0700
commit9ff72da471a509a8c19791efe469f47fa6977410 (patch)
tree8c5f0563c5141f1401c5dd43d4255cde78085194 /sysdeps/unix/sysv/linux/spawni.c
parent1eb8930608705702d5746e5491bab4e4429fcb83 (diff)
downloadglibc-9ff72da471a509a8c19791efe469f47fa6977410.tar.gz
glibc-9ff72da471a509a8c19791efe469f47fa6977410.tar.xz
glibc-9ff72da471a509a8c19791efe469f47fa6977410.zip
posix: New Linux posix_spawn{p} implementation
This patch implements a new posix_spawn{p} implementation for Linux.  The main
difference is it uses the clone syscall directly with CLONE_VM and CLONE_VFORK
flags and a direct allocated stack.  The new stack and start function solves
most the vfork limitation (possible parent clobber due stack spilling).  The
remaning issue are related to signal handling:

  1. That no signal handlers must run in child context, to avoid corrupt
     parent's state.
  2. Child must synchronize with parent to enforce stack deallocation and
     to possible return execv issues.

The first one is solved by blocking all signals in child, even NPTL-internal
ones (SIGCANCEL and SIGSETXID).  The second issue is done by a stack allocation
in parent and a synchronization with using a pipe or waitpid (in case or error).
The pipe has the advantage of allowing the child signal an exec error (checked
with new tst-spawn2 test).

There is an inherent race condition in pipe2 usage for architectures that do not
support the syscall directly.  In such cases the a pipe plus fctnl is used
instead and it may lead to file descriptor leak in parent (as decribed by fcntl
documentation).

The child process stack is allocate with a mmap with MAP_STACK flag using
default architecture stack size.  Although it is slower than use a stack buffer
from parent, it allows some slack for the compatibility code to run scripts
with no shebang (which may use a buffer with size depending of argument list
count).

Performance should be similar to the vfork default posix implementation and
way faster than fork path (vfork on mostly linux ports are basically
clone with CLONE_VM plus CLONE_VFORK).  The only difference is the syscalls
required for the stack allocation/deallocation.

It fixes BZ#10354, BZ#14750, and BZ#18433.

Tested on i386, x86_64, powerpc64le, and aarch64.

	[BZ #14750]
	[BZ #10354]
	[BZ #18433]
	* include/sched.h (__clone): Add hidden prototype.
	(__clone2): Likewise.
	* include/unistd.h (__dup): Likewise.
	* posix/Makefile (tests): Add tst-spawn2.
	* posix/tst-spawn2.c: New file.
	* sysdeps/posix/dup.c (__dup): Add hidden definition.
	* sysdeps/unix/sysv/linux/aarch64/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/alpha/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/arm/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/hppa/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/i386/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/ia64/clone2.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/m68k/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/microblaze/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/mips/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/nios2/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/powerpc/powerpc32/clone.S (__clone):
	Likewise.
	* sysdeps/unix/sysv/linux/powerpc/powerpc64/clone.S (__clone):
	Likewise.
	* sysdeps/unix/sysv/linux/s390/s390-32/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/s390/s390-64/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/sh/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/sparc/sparc32/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/sparc/sparc64/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/tile/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/x86_64/clone.S (__clone): Likewise.
	* sysdeps/unix/sysv/linux/nptl-signals.h
	(____nptl_is_internal_signal): New function.
	* sysdeps/unix/sysv/linux/spawni.c: New file.
Diffstat (limited to 'sysdeps/unix/sysv/linux/spawni.c')
-rw-r--r--sysdeps/unix/sysv/linux/spawni.c404
1 files changed, 404 insertions, 0 deletions
diff --git a/sysdeps/unix/sysv/linux/spawni.c b/sysdeps/unix/sysv/linux/spawni.c
new file mode 100644
index 0000000000..454462be47
--- /dev/null
+++ b/sysdeps/unix/sysv/linux/spawni.c
@@ -0,0 +1,404 @@
+/* POSIX spawn interface.  Linux version.
+   Copyright (C) 2016 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/>.  */
+
+#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 <local-setxid.h>
+#include <shlib-compat.h>
+#include <nptl/pthreadP.h>
+#include <dl-sysdep.h>
+#include <ldsodefs.h>
+#include <libc-internal.h>
+#include "spawn_int.h"
+
+/* The Linux implementation of posix_spawn{p} uses the clone syscall directly
+   with CLONE_VM and CLONE_VFORK flags and an allocated stack.  The new stack
+   and start function solves most the vfork limitation (possible parent
+   clobber due stack spilling). The remaining issue are:
+
+   1. That no signal handlers must run in child context, to avoid corrupting
+      parent's state.
+   2. The parent must ensure child's stack freeing.
+   3. Child must synchronize with parent to enforce 2. and to possible
+      return execv issues.
+
+   The first issue is solved by blocking all signals in child, even the
+   NPTL-internal ones (SIGCANCEL and SIGSETXID).  The second and third issue
+   is done by a stack allocation in parent and a synchronization with using
+   a pipe or waitpid (in case or error).  The pipe has the advantage of
+   allowing the child the communicate an exec error.  */
+
+
+/* The Unix standard contains a long explanation of the way to signal
+   an error after the fork() was successful.  Since no new wait status
+   was wanted there is no way to signal an error using one of the
+   available methods.  The committee chose to signal an error by a
+   normal program exit with the exit code 127.  */
+#define SPAWN_ERROR	127
+
+/* We need to block both SIGCANCEL and SIGSETXID.  */
+#define SIGALL_SET \
+  ((__sigset_t) { .__val = {[0 ...  _SIGSET_NWORDS-1 ] =  -1 } })
+
+#ifdef __ia64__
+# define CLONE(__fn, __stack, __stacksize, __flags, __args) \
+  __clone2 (__fn, __stack, __stacksize, __flags, __args, 0, 0, 0)
+#else
+# define CLONE(__fn, __stack, __stacksize, __flags, __args) \
+  __clone (__fn, __stack, __flags, __args)
+#endif
+
+#if _STACK_GROWS_DOWN
+# define STACK(__stack, __stack_size) (__stack + __stack_size)
+#elif _STACK_GROWS_UP
+# define STACK(__stack, __stack_size) (__stack)
+#endif
+
+
+struct posix_spawn_args
+{
+  int pipe[2];
+  sigset_t oldmask;
+  const char *file;
+  int (*exec) (const char *, char *const *, char *const *);
+  const posix_spawn_file_actions_t *fa;
+  const posix_spawnattr_t *restrict attr;
+  char *const *argv;
+  ptrdiff_t argc;
+  char *const *envp;
+  int xflags;
+};
+
+/* Older version requires that shell script without shebang definition
+   to be called explicitly using /bin/sh (_PATH_BSHELL).  */
+static void
+maybe_script_execute (struct posix_spawn_args *args)
+{
+  if (SHLIB_COMPAT (libc, GLIBC_2_2, GLIBC_2_15)
+      && (args->xflags & SPAWN_XFLAGS_TRY_SHELL) && errno == ENOEXEC)
+    {
+      char *const *argv = args->argv;
+      ptrdiff_t argc = args->argc;
+
+      /* Construct an argument list for the shell.  */
+      char *new_argv[argc + 1];
+      new_argv[0] = (char *) _PATH_BSHELL;
+      new_argv[1] = (char *) args->file;
+      if (argc > 1)
+	memcpy (new_argv + 2, argv + 1, argc * sizeof(char *));
+      else
+	new_argv[2] = NULL;
+
+      /* Execute the shell.  */
+      args->exec (new_argv[0], new_argv, args->envp);
+    }
+}
+
+/* Function used in the clone call to setup the signals mask, posix_spawn
+   attributes, and file actions.  It run on its own stack (provided by the
+   posix_spawn call).  */
+static int
+__spawni_child (void *arguments)
+{
+  struct posix_spawn_args *args = arguments;
+  const posix_spawnattr_t *restrict attr = args->attr;
+  const posix_spawn_file_actions_t *file_actions = args->fa;
+  int p = args->pipe[1];
+  int ret;
+
+  close_not_cancel (args->pipe[0]);
+
+  /* The child must ensure that no signal handler are enabled because it shared
+     memory with parent, so the signal disposition must be either SIG_DFL or
+     SIG_IGN.  It does by iterating over all signals and although it could
+     possibly be more optimized (by tracking which signal potentially have a
+     signal handler), it might requires system specific solutions (since the
+     sigset_t data type can be very different on different architectures).  */
+  struct sigaction sa;
+  memset (&sa, '\0', sizeof (sa));
+
+  sigset_t hset;
+  __sigprocmask (SIG_BLOCK, 0, &hset);
+  for (int sig = 1; sig < _NSIG; ++sig)
+    {
+      if ((attr->__flags & POSIX_SPAWN_SETSIGDEF)
+	  && sigismember (&attr->__sd, sig))
+	{
+	  sa.sa_handler = SIG_DFL;
+	}
+      else if (sigismember (&hset, sig))
+	{
+	  if (__nptl_is_internal_signal (sig))
+	    sa.sa_handler = SIG_IGN;
+	  else
+	    {
+	      __libc_sigaction (sig, 0, &sa);
+	      if (sa.sa_handler == SIG_IGN)
+		continue;
+	      sa.sa_handler = SIG_DFL;
+	    }
+	}
+      else
+	continue;
+
+      __libc_sigaction (sig, &sa, 0);
+    }
+
+#ifdef _POSIX_PRIORITY_SCHEDULING
+  /* Set the scheduling algorithm and parameters.  */
+  if ((attr->__flags & (POSIX_SPAWN_SETSCHEDPARAM | POSIX_SPAWN_SETSCHEDULER))
+      == POSIX_SPAWN_SETSCHEDPARAM)
+    {
+      if ((ret = __sched_setparam (0, &attr->__sp)) == -1)
+	goto fail;
+    }
+  else if ((attr->__flags & POSIX_SPAWN_SETSCHEDULER) != 0)
+    {
+      if ((ret = __sched_setscheduler (0, attr->__policy, &attr->__sp)) == -1)
+	goto fail;
+    }
+#endif
+
+  /* Set the process group ID.  */
+  if ((attr->__flags & POSIX_SPAWN_SETPGROUP) != 0
+      && (ret = __setpgid (0, attr->__pgrp)) != 0)
+    goto fail;
+
+  /* Set the effective user and group IDs.  */
+  if ((attr->__flags & POSIX_SPAWN_RESETIDS) != 0
+      && ((ret = local_seteuid (__getuid ())) != 0
+	  || (ret = local_setegid (__getgid ())) != 0))
+    goto fail;
+
+  /* Execute the file actions.  */
+  if (file_actions != 0)
+    {
+      int cnt;
+      struct rlimit64 fdlimit;
+      bool have_fdlimit = false;
+
+      for (cnt = 0; cnt < file_actions->__used; ++cnt)
+	{
+	  struct __spawn_action *action = &file_actions->__actions[cnt];
+
+	  /* Dup the pipe fd onto an unoccupied one to avoid any file
+	     operation to clobber it.  */
+	  if ((action->action.close_action.fd == p)
+	      || (action->action.open_action.fd == p)
+	      || (action->action.dup2_action.fd == p))
+	    {
+	      if ((ret = __dup (p)) < 0)
+		goto fail;
+	      p = ret;
+	    }
+
+	  switch (action->tag)
+	    {
+	    case spawn_do_close:
+	      if ((ret =
+		   close_not_cancel (action->action.close_action.fd)) != 0)
+		{
+		  if (!have_fdlimit)
+		    {
+		      __getrlimit64 (RLIMIT_NOFILE, &fdlimit);
+		      have_fdlimit = true;
+		    }
+
+		  /* Signal errors only for file descriptors out of range.  */
+		  if (action->action.close_action.fd < 0
+		      || action->action.close_action.fd >= fdlimit.rlim_cur)
+		    goto fail;
+		}
+	      break;
+
+	    case spawn_do_open:
+	      {
+		ret = open_not_cancel (action->action.open_action.path,
+				       action->action.
+				       open_action.oflag | O_LARGEFILE,
+				       action->action.open_action.mode);
+
+		if (ret == -1)
+		  goto fail;
+
+		int new_fd = ret;
+
+		/* Make sure the desired file descriptor is used.  */
+		if (ret != action->action.open_action.fd)
+		  {
+		    if ((ret = __dup2 (new_fd, action->action.open_action.fd))
+			!= action->action.open_action.fd)
+		      goto fail;
+
+		    if ((ret = close_not_cancel (new_fd)) != 0)
+		      goto fail;
+		  }
+	      }
+	      break;
+
+	    case spawn_do_dup2:
+	      if ((ret = __dup2 (action->action.dup2_action.fd,
+				 action->action.dup2_action.newfd))
+		  != action->action.dup2_action.newfd)
+		goto fail;
+	      break;
+	    }
+	}
+    }
+
+  /* Set the initial signal mask of the child if POSIX_SPAWN_SETSIGMASK
+     is set, otherwise restore the previous one.  */
+  __sigprocmask (SIG_SETMASK, (attr->__flags & POSIX_SPAWN_SETSIGMASK)
+		 ? &attr->__ss : &args->oldmask, 0);
+
+  args->exec (args->file, args->argv, args->envp);
+
+  /* This is compatibility function required to enable posix_spawn run
+     script without shebang definition for older posix_spawn versions
+     (2.15).  */
+  maybe_script_execute (args);
+
+  ret = -errno;
+
+fail:
+  /* Since sizeof errno < PIPE_BUF, the write is atomic. */
+  ret = -ret;
+  if (ret)
+    while (write_not_cancel (p, &ret, sizeof ret) < 0)
+      continue;
+  exit (SPAWN_ERROR);
+}
+
+/* Spawn a new process executing PATH with the attributes describes in *ATTRP.
+   Before running the process perform the actions described in FILE-ACTIONS. */
+static int
+__spawnix (pid_t * pid, const char *file,
+	   const posix_spawn_file_actions_t * file_actions,
+	   const posix_spawnattr_t * attrp, char *const argv[],
+	   char *const envp[], int xflags,
+	   int (*exec) (const char *, char *const *, char *const *))
+{
+  pid_t new_pid;
+  struct posix_spawn_args args;
+  int ec;
+
+  if (__pipe2 (args.pipe, O_CLOEXEC))
+    return errno;
+
+  /* To avoid imposing hard limits on posix_spawn{p} the total number of
+     arguments is first calculated to allocate a mmap to hold all possible
+     values.  */
+  ptrdiff_t argc = 0;
+  /* Linux allows at most max (0x7FFFFFFF, 1/4 stack size) arguments
+     to be used in a execve call.  We limit to INT_MAX minus one due the
+     compatiblity code that may execute a shell script (maybe_script_execute)
+     where it will construct another argument list with an additional
+     argument.  */
+  ptrdiff_t limit = INT_MAX - 1;
+  while (argv[argc++] != NULL)
+    if (argc == limit)
+      {
+	errno = E2BIG;
+	return errno;
+      }
+
+  int prot = (PROT_READ | PROT_WRITE
+	     | ((GL (dl_stack_flags) & PF_X) ? PROT_EXEC : 0));
+
+  /* Add a slack area for child's stack.  */
+  size_t argv_size = (argc * sizeof (void *)) + 512;
+  size_t stack_size = ALIGN_UP (argv_size, GLRO(dl_pagesize));
+  void *stack = __mmap (NULL, stack_size, prot,
+			MAP_PRIVATE | MAP_ANONYMOUS | MAP_STACK, -1, 0);
+  if (__glibc_unlikely (stack == MAP_FAILED))
+    {
+      close_not_cancel (args.pipe[0]);
+      close_not_cancel (args.pipe[1]);
+      return errno;
+    }
+
+  /* Disable asynchronous cancellation.  */
+  int cs = LIBC_CANCEL_ASYNC ();
+
+  args.file = file;
+  args.exec = exec;
+  args.fa = file_actions;
+  args.attr = attrp ? attrp : &(const posix_spawnattr_t) { 0 };
+  args.argv = argv;
+  args.argc = argc;
+  args.envp = envp;
+  args.xflags = xflags;
+
+  __sigprocmask (SIG_BLOCK, &SIGALL_SET, &args.oldmask);
+
+  /* The clone flags used will create a new child that will run in the same
+     memory space (CLONE_VM) and the execution of calling thread will be
+     suspend until the child calls execve or _exit.  These condition as
+     signal below either by pipe write (_exit with SPAWN_ERROR) or
+     a successful execve.
+     Also since the calling thread execution will be suspend, there is not
+     need for CLONE_SETTLS.  Although parent and child share the same TLS
+     namespace, there will be no concurrent access for TLS variables (errno
+     for instance).  */
+  new_pid = CLONE (__spawni_child, STACK (stack, stack_size), stack_size,
+		   CLONE_VM | CLONE_VFORK | SIGCHLD, &args);
+
+  close_not_cancel (args.pipe[1]);
+
+  if (new_pid > 0)
+    {
+      if (__read (args.pipe[0], &ec, sizeof ec) != sizeof ec)
+	ec = 0;
+      else
+	__waitpid (new_pid, NULL, 0);
+    }
+  else
+    ec = -new_pid;
+
+  __munmap (stack, stack_size);
+
+  close_not_cancel (args.pipe[0]);
+
+  if (!ec && new_pid)
+    *pid = new_pid;
+
+  __sigprocmask (SIG_SETMASK, &args.oldmask, 0);
+
+  LIBC_CANCEL_RESET (cs);
+
+  return ec;
+}
+
+/* Spawn a new process executing PATH with the attributes describes in *ATTRP.
+   Before running the process perform the actions described in FILE-ACTIONS. */
+int
+__spawni (pid_t * pid, const char *file,
+	  const posix_spawn_file_actions_t * acts,
+	  const posix_spawnattr_t * attrp, char *const argv[],
+	  char *const envp[], int xflags)
+{
+  return __spawnix (pid, file, acts, attrp, argv, envp, xflags,
+		    xflags & SPAWN_XFLAGS_USE_PATH ? __execvpe : __execve);
+}