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.
* sysdeps/unix/sysv/linux/ia64/clone2.S: Check for NULL stakc
pointers to match other architectures.
* sysdeps/unix/sysv/linux/Makefile [subdirs=misc] (tests): Add
tst-clone.
* sysdeps/unix/sysv/linux/tst-clone.c: New file.
malloc'ed wbuffer which isn't freed in error conditions.
2004-12-07 Jakub Jelinek <jakub@redhat.com>
* sysdeps/unix/sysv/linux/ia64/clone2.S (__clone2): Add support for
NPTL where the PID is stored at userlevel and needs to be reset when
CLONE_THREAD is not used. Restore gp before calling _exit.
2004-04-21 Ulrich Drepper <drepper@redhat.com>
* sysdeps/unix/sysv/linux/ia64/setjmp.S: Define _GI___sigsetjmp
and use it internally instead of __sigsetjmp.
* sysdeps/ia64/fpu/s_frexpf.c (frexpf): Use _GI___libm_frexp_4f.
* sysdeps/ia64/fpu/s_frexpl.c (frexpl): Use _GI___libm_frexp_4l.
* sysdeps/ia64/fpu/libm_frexp4.S: Define _GI___libm_frexp_4.
* sysdeps/ia64/fpu/libm_frexp4f.S: Define _GI___libm_frexp_4f.
* sysdeps/ia64/fpu/libm_frexp4l.S: Define _GI___libm_frexp_4l.
* sysdeps/ia64/fpu/s_frexp.c (frexp): Use _GI___libm_frexp_4.
* sysdeps/ia64/fpu/libm_support.h: Declare _GI___libm_frexp_4,
_GI___libm_frexp_4f, _GI___libm_frexp_4l.
* sysdeps/ia64/fpu/bits/mathinline.h: New file.
* sysdeps/unix/sysv/linux/ia64/__start_context.S: Use
HIDDEN_JUMPTARGET for exit call.
* sysdeps/unix/sysv/linux/ia64/clone2.S: Use HIDDEN_JUMPTARGET for
_exit call.
* sysdeps/ia64/bcopy.S: Use HIDDEN_JUMPTARGET for memmove call.
* sysdeps/ia64/strcat.S: Use HIDDEN_JUMPTARGET for strlen and
strcpy calls.
2003-12-02 David Mosberger <davidm@hpl.hp.com>
* sysdeps/ia64/elf/initfini.c: Add unwind info.
* sysdeps/ia64/dl-machine.h (elf_machine_matches_host): Mark with
attribute "unused".
(elf_machine_dynamic): Mark with attributes "unused" and "const".
(elf_machine_runtime_setup): Likewise.
* sysdeps/generic/dl-fptr.c (make_fptr_table): Mark with
attribute "always_inline".
* sysdeps/ia64/dl-machine.h (__ia64_init_bootstrap_fdesc_table):
Likewise.
* configure.in: Check whether compiler has libunwind support.
* config.make.in (have-cc-with-libunwind): New variable.
* config.h.in (HAVE_CC_WITH_LIBUNWIND): New macro.
* Makeconfig (gnulib): If have-cc-withh-libunwind is "yes", also
mention -lunwind.
003-11-12 David Mosberger <davidm@hpl.hp.com>
* sysdeps/unix/sysv/linux/ia64/sysdep.h: Define DO_CALL_VIA_BREAK.
Redefine DO_CALL to use vdso if supported, otherwise DO_CALL_VIA_BREAK.
Likewise for DO_INLINE_SYSCALL. Make INTERNAL_SYSCALL use
DO_INLINE_SYSCALL.
* sysdeps/unix/sysv/linux/ia64/vfork.S: Use DO_CALL_VIA_BREAK()
instead of DO_CALL().
* sysdeps/unix/sysv/linux/ia64/clone2.S: Use break directly instead
of DO_CALL().
* sysdeps/unix/sysv/linux/ia64/brk.S (__curbrk): Restructure it
to take advantage of DO_CALL() macro.
* sysdeps/unix/sysv/linux/ia64/setcontext.S: Likewise.
* sysdeps/unix/sysv/linux/ia64/getcontext.S: Likewise.
* elf/rtld.c (dl_main): Restrict dl_sysinfo_dso check to first
program header. On ia64, the check failed previously because
there are two program headers.
* sysdeps/generic/s_nexttowardf.c: Likewise.
* math/bug-nexttoward.c: New file.
2003-03-13 Jakub Jelinek <jakub@redhat.com>
* sysdeps/unix/sysv/linux/ia64/sysdep.h (BREAK_INSN_1, BREAK_INSN,
ASM_OUTARGS_0, ASM_OUTARGS_1, ASM_OUTARGS_2, ASM_OUTARGS_3,
ASM_OUTARGS_4, ASM_OUTARGS_5, ASM_OUTARGS_6): Define.
(INTERNAL_SYSCALL, INLINE_SYSCALL): Use it. Make syscall arguments
clobbered by the syscall.
(ASM_ARGS_1, ASM_ARGS_2, ASM_ARGS_3, ASM_ARGS_4, ASM_ARGS_5,
ASM_ARGS_6): Change constraints from r to index of corresponding
output register.
* sysdeps/unix/sysv/linux/ia64/clone2.S (__clone2): Swap
ptid and ctid to match kernel.
* sysdeps/unix/sysv/linux/ia64/system.c (FORK): Likewise.
* sysdeps/powerpc/powerpc64/elf/configure.in: Remove
* sysdeps/powerpc/powerpc64/elf/configure: Regenerated.
allocate space even for the trailing '/'.
Reported by John Reiser <jreiser@BitWagon.com>.
* sysdeps/unix/sysv/linux/ia64/sysdep.h (LOAD_ARGS_6, ASM_ARGS_6,
ASM_CLOBBERS_6): Define.
(ASM_CLOBBERS_5): Use ASM_CLOBBERS_6.
* sysdeps/unix/sysv/linux/ia64/clone2.S (__clone2): Reorder arguments
to match IA-32 order.
* sysdeps/unix/sysv/linux/i386/clone.S: Fix comment.
2003-03-03 Ian Wienand <ianw@gelato.unsw.edu.au>
* sysdeps/unix/sysv/linux/ia64/clone2.S: Update to take extra clone
flags.
* include/sched.h: Update clone2 prototype.
2001-07-06 Paul Eggert <eggert@twinsun.com>
* manual/argp.texi: Remove ignored LGPL copyright notice; it's
not appropriate for documentation anyway.
* manual/libc-texinfo.sh: "Library General Public License" ->
"Lesser General Public License".
2001-07-06 Andreas Jaeger <aj@suse.de>
* All files under GPL/LGPL version 2: Place under LGPL version
2.1.
2001-02-18 Mark Kettenis <kettenis@gnu.org>
* malloc/mtrace.c [USE_IN_LIBIO]: Define fopen as _IO_fopen64.
(mtrace): Revert 2001-02-13 patch: use fopen instead of fopen64.
2001-02-17 H.J. Lu <hjl@gnu.org>
* sysdeps/unix/sysv/linux/ia64/clone2.S: Use clone2 and don't use
scratch registers across the system call.
