2013-01-07 21:41:20 +00:00
|
|
|
/****************************************************************************
|
2021-06-10 07:51:27 +02:00
|
|
|
* apps/testing/ostest/vfork.c
|
2013-01-07 21:41:20 +00:00
|
|
|
*
|
2024-12-29 16:26:09 +01:00
|
|
|
* SPDX-License-Identifier: Apache-2.0
|
|
|
|
|
*
|
2021-06-10 07:51:27 +02:00
|
|
|
* Licensed to the Apache Software Foundation (ASF) under one or more
|
|
|
|
|
* contributor license agreements. See the NOTICE file distributed with
|
|
|
|
|
* this work for additional information regarding copyright ownership. The
|
|
|
|
|
* ASF licenses this file to you under the Apache License, Version 2.0 (the
|
|
|
|
|
* "License"); you may not use this file except in compliance with the
|
|
|
|
|
* License. You may obtain a copy of the License at
|
2013-01-07 21:41:20 +00:00
|
|
|
*
|
2021-06-10 07:51:27 +02:00
|
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
2013-01-07 21:41:20 +00:00
|
|
|
*
|
2021-06-10 07:51:27 +02:00
|
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
|
|
|
* distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
|
|
|
|
|
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
|
|
|
|
|
* License for the specific language governing permissions and limitations
|
|
|
|
|
* under the License.
|
2013-01-07 21:41:20 +00:00
|
|
|
*
|
|
|
|
|
****************************************************************************/
|
|
|
|
|
|
|
|
|
|
/****************************************************************************
|
|
|
|
|
* Included Files
|
|
|
|
|
****************************************************************************/
|
|
|
|
|
|
|
|
|
|
#include <nuttx/config.h>
|
|
|
|
|
|
2023-03-16 10:35:15 +08:00
|
|
|
#include <assert.h>
|
|
|
|
|
#include <errno.h>
|
2013-01-07 21:41:20 +00:00
|
|
|
#include <stdio.h>
|
|
|
|
|
#include <stdlib.h>
|
testing/ostest: Split the fork test into vfork and fork.
ostest's "vfork" test was never testing vfork(). It has the child write a
global and the parent observe the write -- the defining property of *sharing*,
not of vfork(), whose defining property is that the parent is suspended and
whose contract forbids the child to write anything at all. It passed because
NuttX implemented fork() and vfork() as the same sharing primitive, which
apache/nuttx#19562 separates.
vfork.c is rewritten to test what vfork() promises. The child does only what
POSIX permits -- it calls _exit(42) and nothing else, not even exit(), which
would run atexit handlers and flush stdio in the parent's address space. Since
the child may not write memory and the parent cannot run while the child lives,
the observable is the child's exit status: had the parent not been suspended,
it would have reached waitpid() while the child was still alive. Where child
status is not retained -- ostest_main() sets SA_NOCLDWAIT for the whole run,
deliberately -- ECHILD is accepted as equally good evidence, since it says the
child was already gone when the parent asked.
fork.c is new and tests POSIX fork(): the child's writes to .data, .bss and
the heap are invisible to the parent and vice versa, a pointer to a stack local
taken before the fork names the same object in both, and the child does
everything a vfork() child may not -- calls malloc() and printf(), and returns
from the function that called fork().
Both run at the top of user_main(). They exercise the lowest-level machinery
in the suite -- address environments, stack setup, the architecture's register
context -- so a fault in one takes the process down instead of reporting a
failure. Learning that in seconds rather than after everything else has passed
matters when a port is being brought up.
Each test gates on the one primitive it tests, ARCH_HAVE_VFORK and
ARCH_HAVE_FORK respectively. There is no compatibility layer and no mapping
between symbols. vfork.c no longer requires SCHED_WAITPID: the suspension is
in the kernel primitive now, so the test's core assertion holds without it and
only the status check is conditional.
The simulator is the one exception. It selects ARCH_HAVE_VFORK, but ostest
takes the sim down as soon as the test runs there, so the call keeps the
!ARCH_SIM guard that apps ee7642793 put on the old test in 2024. The old gate
hid this: ARCH_HAVE_FORK is not set on the sim, so the test was not built
there at all.
The other in-tree callers are audited for which primitive they actually meant:
* interpreters/python's _posixsubprocess and netutils/libwebsockets'
LWS_HAVE_WORKING_VFORK want the fork-then-exec path -- ARCH_HAVE_VFORK.
* python's os.fork() and libwebsockets' LWS_HAVE_FORK mean real fork() and stay
on ARCH_HAVE_FORK, so they become *absent* rather than silently wrong.
* testing/fs/fdsantest's vfork case follows ARCH_HAVE_VFORK.
interpreters/bas is deliberately left alone. Its SHELL and EDIT statements
reach for vfork() under an ARCH_HAVE_FORK guard and want the same treatment,
but checkpatch.sh checks the whole of any file a patch touches and
bas_statement.c produces 1681 pre-existing findings against master, so a
one-line change there fails CI on its own. The consequence is small:
EXAMPLES_BAS_SHELL is EXPERIMENTAL and already depends on ARCH_HAVE_FORK, so it
becomes unselectable rather than misbehaving.
Assisted-by: Claude Code:claude-opus-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
2026-08-07 13:03:08 +02:00
|
|
|
#include <sys/wait.h>
|
2013-01-07 21:41:20 +00:00
|
|
|
#include <unistd.h>
|
|
|
|
|
|
|
|
|
|
#include "ostest.h"
|
|
|
|
|
|
|
|
|
|
/****************************************************************************
|
testing/ostest: Split the fork test into vfork and fork.
ostest's "vfork" test was never testing vfork(). It has the child write a
global and the parent observe the write -- the defining property of *sharing*,
not of vfork(), whose defining property is that the parent is suspended and
whose contract forbids the child to write anything at all. It passed because
NuttX implemented fork() and vfork() as the same sharing primitive, which
apache/nuttx#19562 separates.
vfork.c is rewritten to test what vfork() promises. The child does only what
POSIX permits -- it calls _exit(42) and nothing else, not even exit(), which
would run atexit handlers and flush stdio in the parent's address space. Since
the child may not write memory and the parent cannot run while the child lives,
the observable is the child's exit status: had the parent not been suspended,
it would have reached waitpid() while the child was still alive. Where child
status is not retained -- ostest_main() sets SA_NOCLDWAIT for the whole run,
deliberately -- ECHILD is accepted as equally good evidence, since it says the
child was already gone when the parent asked.
fork.c is new and tests POSIX fork(): the child's writes to .data, .bss and
the heap are invisible to the parent and vice versa, a pointer to a stack local
taken before the fork names the same object in both, and the child does
everything a vfork() child may not -- calls malloc() and printf(), and returns
from the function that called fork().
Both run at the top of user_main(). They exercise the lowest-level machinery
in the suite -- address environments, stack setup, the architecture's register
context -- so a fault in one takes the process down instead of reporting a
failure. Learning that in seconds rather than after everything else has passed
matters when a port is being brought up.
Each test gates on the one primitive it tests, ARCH_HAVE_VFORK and
ARCH_HAVE_FORK respectively. There is no compatibility layer and no mapping
between symbols. vfork.c no longer requires SCHED_WAITPID: the suspension is
in the kernel primitive now, so the test's core assertion holds without it and
only the status check is conditional.
The simulator is the one exception. It selects ARCH_HAVE_VFORK, but ostest
takes the sim down as soon as the test runs there, so the call keeps the
!ARCH_SIM guard that apps ee7642793 put on the old test in 2024. The old gate
hid this: ARCH_HAVE_FORK is not set on the sim, so the test was not built
there at all.
The other in-tree callers are audited for which primitive they actually meant:
* interpreters/python's _posixsubprocess and netutils/libwebsockets'
LWS_HAVE_WORKING_VFORK want the fork-then-exec path -- ARCH_HAVE_VFORK.
* python's os.fork() and libwebsockets' LWS_HAVE_FORK mean real fork() and stay
on ARCH_HAVE_FORK, so they become *absent* rather than silently wrong.
* testing/fs/fdsantest's vfork case follows ARCH_HAVE_VFORK.
interpreters/bas is deliberately left alone. Its SHELL and EDIT statements
reach for vfork() under an ARCH_HAVE_FORK guard and want the same treatment,
but checkpatch.sh checks the whole of any file a patch touches and
bas_statement.c produces 1681 pre-existing findings against master, so a
one-line change there fails CI on its own. The consequence is small:
EXAMPLES_BAS_SHELL is EXPERIMENTAL and already depends on ARCH_HAVE_FORK, so it
becomes unselectable rather than misbehaving.
Assisted-by: Claude Code:claude-opus-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
2026-08-07 13:03:08 +02:00
|
|
|
* Public Functions
|
2013-01-07 21:41:20 +00:00
|
|
|
****************************************************************************/
|
|
|
|
|
|
|
|
|
|
/****************************************************************************
|
testing/ostest: Split the fork test into vfork and fork.
ostest's "vfork" test was never testing vfork(). It has the child write a
global and the parent observe the write -- the defining property of *sharing*,
not of vfork(), whose defining property is that the parent is suspended and
whose contract forbids the child to write anything at all. It passed because
NuttX implemented fork() and vfork() as the same sharing primitive, which
apache/nuttx#19562 separates.
vfork.c is rewritten to test what vfork() promises. The child does only what
POSIX permits -- it calls _exit(42) and nothing else, not even exit(), which
would run atexit handlers and flush stdio in the parent's address space. Since
the child may not write memory and the parent cannot run while the child lives,
the observable is the child's exit status: had the parent not been suspended,
it would have reached waitpid() while the child was still alive. Where child
status is not retained -- ostest_main() sets SA_NOCLDWAIT for the whole run,
deliberately -- ECHILD is accepted as equally good evidence, since it says the
child was already gone when the parent asked.
fork.c is new and tests POSIX fork(): the child's writes to .data, .bss and
the heap are invisible to the parent and vice versa, a pointer to a stack local
taken before the fork names the same object in both, and the child does
everything a vfork() child may not -- calls malloc() and printf(), and returns
from the function that called fork().
Both run at the top of user_main(). They exercise the lowest-level machinery
in the suite -- address environments, stack setup, the architecture's register
context -- so a fault in one takes the process down instead of reporting a
failure. Learning that in seconds rather than after everything else has passed
matters when a port is being brought up.
Each test gates on the one primitive it tests, ARCH_HAVE_VFORK and
ARCH_HAVE_FORK respectively. There is no compatibility layer and no mapping
between symbols. vfork.c no longer requires SCHED_WAITPID: the suspension is
in the kernel primitive now, so the test's core assertion holds without it and
only the status check is conditional.
The simulator is the one exception. It selects ARCH_HAVE_VFORK, but ostest
takes the sim down as soon as the test runs there, so the call keeps the
!ARCH_SIM guard that apps ee7642793 put on the old test in 2024. The old gate
hid this: ARCH_HAVE_FORK is not set on the sim, so the test was not built
there at all.
The other in-tree callers are audited for which primitive they actually meant:
* interpreters/python's _posixsubprocess and netutils/libwebsockets'
LWS_HAVE_WORKING_VFORK want the fork-then-exec path -- ARCH_HAVE_VFORK.
* python's os.fork() and libwebsockets' LWS_HAVE_FORK mean real fork() and stay
on ARCH_HAVE_FORK, so they become *absent* rather than silently wrong.
* testing/fs/fdsantest's vfork case follows ARCH_HAVE_VFORK.
interpreters/bas is deliberately left alone. Its SHELL and EDIT statements
reach for vfork() under an ARCH_HAVE_FORK guard and want the same treatment,
but checkpatch.sh checks the whole of any file a patch touches and
bas_statement.c produces 1681 pre-existing findings against master, so a
one-line change there fails CI on its own. The consequence is small:
EXAMPLES_BAS_SHELL is EXPERIMENTAL and already depends on ARCH_HAVE_FORK, so it
becomes unselectable rather than misbehaving.
Assisted-by: Claude Code:claude-opus-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
2026-08-07 13:03:08 +02:00
|
|
|
* Name: vfork_test
|
|
|
|
|
*
|
|
|
|
|
* Description:
|
|
|
|
|
* Verify the defining property of vfork(): the parent is suspended until
|
|
|
|
|
* the child _exit()s or exec()s. The child does only what POSIX
|
|
|
|
|
* permits -- _exit(), not exit(), which would flush stdio in the parent's
|
|
|
|
|
* address space. Since the child may not write memory, the observable is
|
|
|
|
|
* its exit status: an unsuspended parent would reach waitpid() first.
|
|
|
|
|
*
|
2013-01-07 21:41:20 +00:00
|
|
|
****************************************************************************/
|
|
|
|
|
|
|
|
|
|
int vfork_test(void)
|
|
|
|
|
{
|
|
|
|
|
pid_t pid;
|
|
|
|
|
|
testing/ostest: Split the fork test into vfork and fork.
ostest's "vfork" test was never testing vfork(). It has the child write a
global and the parent observe the write -- the defining property of *sharing*,
not of vfork(), whose defining property is that the parent is suspended and
whose contract forbids the child to write anything at all. It passed because
NuttX implemented fork() and vfork() as the same sharing primitive, which
apache/nuttx#19562 separates.
vfork.c is rewritten to test what vfork() promises. The child does only what
POSIX permits -- it calls _exit(42) and nothing else, not even exit(), which
would run atexit handlers and flush stdio in the parent's address space. Since
the child may not write memory and the parent cannot run while the child lives,
the observable is the child's exit status: had the parent not been suspended,
it would have reached waitpid() while the child was still alive. Where child
status is not retained -- ostest_main() sets SA_NOCLDWAIT for the whole run,
deliberately -- ECHILD is accepted as equally good evidence, since it says the
child was already gone when the parent asked.
fork.c is new and tests POSIX fork(): the child's writes to .data, .bss and
the heap are invisible to the parent and vice versa, a pointer to a stack local
taken before the fork names the same object in both, and the child does
everything a vfork() child may not -- calls malloc() and printf(), and returns
from the function that called fork().
Both run at the top of user_main(). They exercise the lowest-level machinery
in the suite -- address environments, stack setup, the architecture's register
context -- so a fault in one takes the process down instead of reporting a
failure. Learning that in seconds rather than after everything else has passed
matters when a port is being brought up.
Each test gates on the one primitive it tests, ARCH_HAVE_VFORK and
ARCH_HAVE_FORK respectively. There is no compatibility layer and no mapping
between symbols. vfork.c no longer requires SCHED_WAITPID: the suspension is
in the kernel primitive now, so the test's core assertion holds without it and
only the status check is conditional.
The simulator is the one exception. It selects ARCH_HAVE_VFORK, but ostest
takes the sim down as soon as the test runs there, so the call keeps the
!ARCH_SIM guard that apps ee7642793 put on the old test in 2024. The old gate
hid this: ARCH_HAVE_FORK is not set on the sim, so the test was not built
there at all.
The other in-tree callers are audited for which primitive they actually meant:
* interpreters/python's _posixsubprocess and netutils/libwebsockets'
LWS_HAVE_WORKING_VFORK want the fork-then-exec path -- ARCH_HAVE_VFORK.
* python's os.fork() and libwebsockets' LWS_HAVE_FORK mean real fork() and stay
on ARCH_HAVE_FORK, so they become *absent* rather than silently wrong.
* testing/fs/fdsantest's vfork case follows ARCH_HAVE_VFORK.
interpreters/bas is deliberately left alone. Its SHELL and EDIT statements
reach for vfork() under an ARCH_HAVE_FORK guard and want the same treatment,
but checkpatch.sh checks the whole of any file a patch touches and
bas_statement.c produces 1681 pre-existing findings against master, so a
one-line change there fails CI on its own. The consequence is small:
EXAMPLES_BAS_SHELL is EXPERIMENTAL and already depends on ARCH_HAVE_FORK, so it
becomes unselectable rather than misbehaving.
Assisted-by: Claude Code:claude-opus-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
2026-08-07 13:03:08 +02:00
|
|
|
printf("vfork_test: Started\n");
|
|
|
|
|
|
2013-01-07 21:41:20 +00:00
|
|
|
pid = vfork();
|
|
|
|
|
if (pid == 0)
|
|
|
|
|
{
|
testing/ostest: Split the fork test into vfork and fork.
ostest's "vfork" test was never testing vfork(). It has the child write a
global and the parent observe the write -- the defining property of *sharing*,
not of vfork(), whose defining property is that the parent is suspended and
whose contract forbids the child to write anything at all. It passed because
NuttX implemented fork() and vfork() as the same sharing primitive, which
apache/nuttx#19562 separates.
vfork.c is rewritten to test what vfork() promises. The child does only what
POSIX permits -- it calls _exit(42) and nothing else, not even exit(), which
would run atexit handlers and flush stdio in the parent's address space. Since
the child may not write memory and the parent cannot run while the child lives,
the observable is the child's exit status: had the parent not been suspended,
it would have reached waitpid() while the child was still alive. Where child
status is not retained -- ostest_main() sets SA_NOCLDWAIT for the whole run,
deliberately -- ECHILD is accepted as equally good evidence, since it says the
child was already gone when the parent asked.
fork.c is new and tests POSIX fork(): the child's writes to .data, .bss and
the heap are invisible to the parent and vice versa, a pointer to a stack local
taken before the fork names the same object in both, and the child does
everything a vfork() child may not -- calls malloc() and printf(), and returns
from the function that called fork().
Both run at the top of user_main(). They exercise the lowest-level machinery
in the suite -- address environments, stack setup, the architecture's register
context -- so a fault in one takes the process down instead of reporting a
failure. Learning that in seconds rather than after everything else has passed
matters when a port is being brought up.
Each test gates on the one primitive it tests, ARCH_HAVE_VFORK and
ARCH_HAVE_FORK respectively. There is no compatibility layer and no mapping
between symbols. vfork.c no longer requires SCHED_WAITPID: the suspension is
in the kernel primitive now, so the test's core assertion holds without it and
only the status check is conditional.
The simulator is the one exception. It selects ARCH_HAVE_VFORK, but ostest
takes the sim down as soon as the test runs there, so the call keeps the
!ARCH_SIM guard that apps ee7642793 put on the old test in 2024. The old gate
hid this: ARCH_HAVE_FORK is not set on the sim, so the test was not built
there at all.
The other in-tree callers are audited for which primitive they actually meant:
* interpreters/python's _posixsubprocess and netutils/libwebsockets'
LWS_HAVE_WORKING_VFORK want the fork-then-exec path -- ARCH_HAVE_VFORK.
* python's os.fork() and libwebsockets' LWS_HAVE_FORK mean real fork() and stay
on ARCH_HAVE_FORK, so they become *absent* rather than silently wrong.
* testing/fs/fdsantest's vfork case follows ARCH_HAVE_VFORK.
interpreters/bas is deliberately left alone. Its SHELL and EDIT statements
reach for vfork() under an ARCH_HAVE_FORK guard and want the same treatment,
but checkpatch.sh checks the whole of any file a patch touches and
bas_statement.c produces 1681 pre-existing findings against master, so a
one-line change there fails CI on its own. The consequence is small:
EXAMPLES_BAS_SHELL is EXPERIMENTAL and already depends on ARCH_HAVE_FORK, so it
becomes unselectable rather than misbehaving.
Assisted-by: Claude Code:claude-opus-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
2026-08-07 13:03:08 +02:00
|
|
|
/* The only thing a vfork() child may do is leave. */
|
2013-01-07 21:41:20 +00:00
|
|
|
|
testing/ostest: Split the fork test into vfork and fork.
ostest's "vfork" test was never testing vfork(). It has the child write a
global and the parent observe the write -- the defining property of *sharing*,
not of vfork(), whose defining property is that the parent is suspended and
whose contract forbids the child to write anything at all. It passed because
NuttX implemented fork() and vfork() as the same sharing primitive, which
apache/nuttx#19562 separates.
vfork.c is rewritten to test what vfork() promises. The child does only what
POSIX permits -- it calls _exit(42) and nothing else, not even exit(), which
would run atexit handlers and flush stdio in the parent's address space. Since
the child may not write memory and the parent cannot run while the child lives,
the observable is the child's exit status: had the parent not been suspended,
it would have reached waitpid() while the child was still alive. Where child
status is not retained -- ostest_main() sets SA_NOCLDWAIT for the whole run,
deliberately -- ECHILD is accepted as equally good evidence, since it says the
child was already gone when the parent asked.
fork.c is new and tests POSIX fork(): the child's writes to .data, .bss and
the heap are invisible to the parent and vice versa, a pointer to a stack local
taken before the fork names the same object in both, and the child does
everything a vfork() child may not -- calls malloc() and printf(), and returns
from the function that called fork().
Both run at the top of user_main(). They exercise the lowest-level machinery
in the suite -- address environments, stack setup, the architecture's register
context -- so a fault in one takes the process down instead of reporting a
failure. Learning that in seconds rather than after everything else has passed
matters when a port is being brought up.
Each test gates on the one primitive it tests, ARCH_HAVE_VFORK and
ARCH_HAVE_FORK respectively. There is no compatibility layer and no mapping
between symbols. vfork.c no longer requires SCHED_WAITPID: the suspension is
in the kernel primitive now, so the test's core assertion holds without it and
only the status check is conditional.
The simulator is the one exception. It selects ARCH_HAVE_VFORK, but ostest
takes the sim down as soon as the test runs there, so the call keeps the
!ARCH_SIM guard that apps ee7642793 put on the old test in 2024. The old gate
hid this: ARCH_HAVE_FORK is not set on the sim, so the test was not built
there at all.
The other in-tree callers are audited for which primitive they actually meant:
* interpreters/python's _posixsubprocess and netutils/libwebsockets'
LWS_HAVE_WORKING_VFORK want the fork-then-exec path -- ARCH_HAVE_VFORK.
* python's os.fork() and libwebsockets' LWS_HAVE_FORK mean real fork() and stay
on ARCH_HAVE_FORK, so they become *absent* rather than silently wrong.
* testing/fs/fdsantest's vfork case follows ARCH_HAVE_VFORK.
interpreters/bas is deliberately left alone. Its SHELL and EDIT statements
reach for vfork() under an ARCH_HAVE_FORK guard and want the same treatment,
but checkpatch.sh checks the whole of any file a patch touches and
bas_statement.c produces 1681 pre-existing findings against master, so a
one-line change there fails CI on its own. The consequence is small:
EXAMPLES_BAS_SHELL is EXPERIMENTAL and already depends on ARCH_HAVE_FORK, so it
becomes unselectable rather than misbehaving.
Assisted-by: Claude Code:claude-opus-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
2026-08-07 13:03:08 +02:00
|
|
|
_exit(42);
|
2013-01-07 21:41:20 +00:00
|
|
|
}
|
|
|
|
|
else if (pid < 0)
|
|
|
|
|
{
|
2022-10-15 23:36:29 +09:00
|
|
|
printf("vfork_test: ERROR vfork() failed: %d\n", errno);
|
|
|
|
|
ASSERT(false);
|
2013-01-07 21:41:20 +00:00
|
|
|
return -1;
|
|
|
|
|
}
|
testing/ostest: Split the fork test into vfork and fork.
ostest's "vfork" test was never testing vfork(). It has the child write a
global and the parent observe the write -- the defining property of *sharing*,
not of vfork(), whose defining property is that the parent is suspended and
whose contract forbids the child to write anything at all. It passed because
NuttX implemented fork() and vfork() as the same sharing primitive, which
apache/nuttx#19562 separates.
vfork.c is rewritten to test what vfork() promises. The child does only what
POSIX permits -- it calls _exit(42) and nothing else, not even exit(), which
would run atexit handlers and flush stdio in the parent's address space. Since
the child may not write memory and the parent cannot run while the child lives,
the observable is the child's exit status: had the parent not been suspended,
it would have reached waitpid() while the child was still alive. Where child
status is not retained -- ostest_main() sets SA_NOCLDWAIT for the whole run,
deliberately -- ECHILD is accepted as equally good evidence, since it says the
child was already gone when the parent asked.
fork.c is new and tests POSIX fork(): the child's writes to .data, .bss and
the heap are invisible to the parent and vice versa, a pointer to a stack local
taken before the fork names the same object in both, and the child does
everything a vfork() child may not -- calls malloc() and printf(), and returns
from the function that called fork().
Both run at the top of user_main(). They exercise the lowest-level machinery
in the suite -- address environments, stack setup, the architecture's register
context -- so a fault in one takes the process down instead of reporting a
failure. Learning that in seconds rather than after everything else has passed
matters when a port is being brought up.
Each test gates on the one primitive it tests, ARCH_HAVE_VFORK and
ARCH_HAVE_FORK respectively. There is no compatibility layer and no mapping
between symbols. vfork.c no longer requires SCHED_WAITPID: the suspension is
in the kernel primitive now, so the test's core assertion holds without it and
only the status check is conditional.
The simulator is the one exception. It selects ARCH_HAVE_VFORK, but ostest
takes the sim down as soon as the test runs there, so the call keeps the
!ARCH_SIM guard that apps ee7642793 put on the old test in 2024. The old gate
hid this: ARCH_HAVE_FORK is not set on the sim, so the test was not built
there at all.
The other in-tree callers are audited for which primitive they actually meant:
* interpreters/python's _posixsubprocess and netutils/libwebsockets'
LWS_HAVE_WORKING_VFORK want the fork-then-exec path -- ARCH_HAVE_VFORK.
* python's os.fork() and libwebsockets' LWS_HAVE_FORK mean real fork() and stay
on ARCH_HAVE_FORK, so they become *absent* rather than silently wrong.
* testing/fs/fdsantest's vfork case follows ARCH_HAVE_VFORK.
interpreters/bas is deliberately left alone. Its SHELL and EDIT statements
reach for vfork() under an ARCH_HAVE_FORK guard and want the same treatment,
but checkpatch.sh checks the whole of any file a patch touches and
bas_statement.c produces 1681 pre-existing findings against master, so a
one-line change there fails CI on its own. The consequence is small:
EXAMPLES_BAS_SHELL is EXPERIMENTAL and already depends on ARCH_HAVE_FORK, so it
becomes unselectable rather than misbehaving.
Assisted-by: Claude Code:claude-opus-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
2026-08-07 13:03:08 +02:00
|
|
|
|
|
|
|
|
/* Reached only once the child has exited or exec'ed. */
|
|
|
|
|
|
|
|
|
|
#ifdef CONFIG_SCHED_WAITPID
|
2013-01-07 21:41:20 +00:00
|
|
|
{
|
testing/ostest: Split the fork test into vfork and fork.
ostest's "vfork" test was never testing vfork(). It has the child write a
global and the parent observe the write -- the defining property of *sharing*,
not of vfork(), whose defining property is that the parent is suspended and
whose contract forbids the child to write anything at all. It passed because
NuttX implemented fork() and vfork() as the same sharing primitive, which
apache/nuttx#19562 separates.
vfork.c is rewritten to test what vfork() promises. The child does only what
POSIX permits -- it calls _exit(42) and nothing else, not even exit(), which
would run atexit handlers and flush stdio in the parent's address space. Since
the child may not write memory and the parent cannot run while the child lives,
the observable is the child's exit status: had the parent not been suspended,
it would have reached waitpid() while the child was still alive. Where child
status is not retained -- ostest_main() sets SA_NOCLDWAIT for the whole run,
deliberately -- ECHILD is accepted as equally good evidence, since it says the
child was already gone when the parent asked.
fork.c is new and tests POSIX fork(): the child's writes to .data, .bss and
the heap are invisible to the parent and vice versa, a pointer to a stack local
taken before the fork names the same object in both, and the child does
everything a vfork() child may not -- calls malloc() and printf(), and returns
from the function that called fork().
Both run at the top of user_main(). They exercise the lowest-level machinery
in the suite -- address environments, stack setup, the architecture's register
context -- so a fault in one takes the process down instead of reporting a
failure. Learning that in seconds rather than after everything else has passed
matters when a port is being brought up.
Each test gates on the one primitive it tests, ARCH_HAVE_VFORK and
ARCH_HAVE_FORK respectively. There is no compatibility layer and no mapping
between symbols. vfork.c no longer requires SCHED_WAITPID: the suspension is
in the kernel primitive now, so the test's core assertion holds without it and
only the status check is conditional.
The simulator is the one exception. It selects ARCH_HAVE_VFORK, but ostest
takes the sim down as soon as the test runs there, so the call keeps the
!ARCH_SIM guard that apps ee7642793 put on the old test in 2024. The old gate
hid this: ARCH_HAVE_FORK is not set on the sim, so the test was not built
there at all.
The other in-tree callers are audited for which primitive they actually meant:
* interpreters/python's _posixsubprocess and netutils/libwebsockets'
LWS_HAVE_WORKING_VFORK want the fork-then-exec path -- ARCH_HAVE_VFORK.
* python's os.fork() and libwebsockets' LWS_HAVE_FORK mean real fork() and stay
on ARCH_HAVE_FORK, so they become *absent* rather than silently wrong.
* testing/fs/fdsantest's vfork case follows ARCH_HAVE_VFORK.
interpreters/bas is deliberately left alone. Its SHELL and EDIT statements
reach for vfork() under an ARCH_HAVE_FORK guard and want the same treatment,
but checkpatch.sh checks the whole of any file a patch touches and
bas_statement.c produces 1681 pre-existing findings against master, so a
one-line change there fails CI on its own. The consequence is small:
EXAMPLES_BAS_SHELL is EXPERIMENTAL and already depends on ARCH_HAVE_FORK, so it
becomes unselectable rather than misbehaving.
Assisted-by: Claude Code:claude-opus-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
2026-08-07 13:03:08 +02:00
|
|
|
int status = 0;
|
|
|
|
|
pid_t ret;
|
|
|
|
|
|
|
|
|
|
ret = waitpid(pid, &status, 0);
|
|
|
|
|
|
|
|
|
|
/* Two answers are correct, and which one comes back is a property of
|
|
|
|
|
* the configuration: a retained status must be exit(42), and ECHILD
|
|
|
|
|
* is equally good evidence -- it says the child was already gone when
|
|
|
|
|
* we asked. ostest_main() sets SA_NOCLDWAIT for the whole run, so
|
|
|
|
|
* testing CONFIG_SCHED_CHILD_STATUS alone is not enough.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
if (ret == pid)
|
2013-01-07 21:41:20 +00:00
|
|
|
{
|
testing/ostest: Split the fork test into vfork and fork.
ostest's "vfork" test was never testing vfork(). It has the child write a
global and the parent observe the write -- the defining property of *sharing*,
not of vfork(), whose defining property is that the parent is suspended and
whose contract forbids the child to write anything at all. It passed because
NuttX implemented fork() and vfork() as the same sharing primitive, which
apache/nuttx#19562 separates.
vfork.c is rewritten to test what vfork() promises. The child does only what
POSIX permits -- it calls _exit(42) and nothing else, not even exit(), which
would run atexit handlers and flush stdio in the parent's address space. Since
the child may not write memory and the parent cannot run while the child lives,
the observable is the child's exit status: had the parent not been suspended,
it would have reached waitpid() while the child was still alive. Where child
status is not retained -- ostest_main() sets SA_NOCLDWAIT for the whole run,
deliberately -- ECHILD is accepted as equally good evidence, since it says the
child was already gone when the parent asked.
fork.c is new and tests POSIX fork(): the child's writes to .data, .bss and
the heap are invisible to the parent and vice versa, a pointer to a stack local
taken before the fork names the same object in both, and the child does
everything a vfork() child may not -- calls malloc() and printf(), and returns
from the function that called fork().
Both run at the top of user_main(). They exercise the lowest-level machinery
in the suite -- address environments, stack setup, the architecture's register
context -- so a fault in one takes the process down instead of reporting a
failure. Learning that in seconds rather than after everything else has passed
matters when a port is being brought up.
Each test gates on the one primitive it tests, ARCH_HAVE_VFORK and
ARCH_HAVE_FORK respectively. There is no compatibility layer and no mapping
between symbols. vfork.c no longer requires SCHED_WAITPID: the suspension is
in the kernel primitive now, so the test's core assertion holds without it and
only the status check is conditional.
The simulator is the one exception. It selects ARCH_HAVE_VFORK, but ostest
takes the sim down as soon as the test runs there, so the call keeps the
!ARCH_SIM guard that apps ee7642793 put on the old test in 2024. The old gate
hid this: ARCH_HAVE_FORK is not set on the sim, so the test was not built
there at all.
The other in-tree callers are audited for which primitive they actually meant:
* interpreters/python's _posixsubprocess and netutils/libwebsockets'
LWS_HAVE_WORKING_VFORK want the fork-then-exec path -- ARCH_HAVE_VFORK.
* python's os.fork() and libwebsockets' LWS_HAVE_FORK mean real fork() and stay
on ARCH_HAVE_FORK, so they become *absent* rather than silently wrong.
* testing/fs/fdsantest's vfork case follows ARCH_HAVE_VFORK.
interpreters/bas is deliberately left alone. Its SHELL and EDIT statements
reach for vfork() under an ARCH_HAVE_FORK guard and want the same treatment,
but checkpatch.sh checks the whole of any file a patch touches and
bas_statement.c produces 1681 pre-existing findings against master, so a
one-line change there fails CI on its own. The consequence is small:
EXAMPLES_BAS_SHELL is EXPERIMENTAL and already depends on ARCH_HAVE_FORK, so it
becomes unselectable rather than misbehaving.
Assisted-by: Claude Code:claude-opus-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
2026-08-07 13:03:08 +02:00
|
|
|
if (!WIFEXITED(status) || WEXITSTATUS(status) != 42)
|
|
|
|
|
{
|
|
|
|
|
printf("vfork_test: ERROR Child %d status 0x%04x, expected "
|
|
|
|
|
"exit(42)\n", pid, status);
|
|
|
|
|
ASSERT(false);
|
|
|
|
|
return -1;
|
|
|
|
|
}
|
2013-01-07 21:41:20 +00:00
|
|
|
}
|
testing/ostest: Split the fork test into vfork and fork.
ostest's "vfork" test was never testing vfork(). It has the child write a
global and the parent observe the write -- the defining property of *sharing*,
not of vfork(), whose defining property is that the parent is suspended and
whose contract forbids the child to write anything at all. It passed because
NuttX implemented fork() and vfork() as the same sharing primitive, which
apache/nuttx#19562 separates.
vfork.c is rewritten to test what vfork() promises. The child does only what
POSIX permits -- it calls _exit(42) and nothing else, not even exit(), which
would run atexit handlers and flush stdio in the parent's address space. Since
the child may not write memory and the parent cannot run while the child lives,
the observable is the child's exit status: had the parent not been suspended,
it would have reached waitpid() while the child was still alive. Where child
status is not retained -- ostest_main() sets SA_NOCLDWAIT for the whole run,
deliberately -- ECHILD is accepted as equally good evidence, since it says the
child was already gone when the parent asked.
fork.c is new and tests POSIX fork(): the child's writes to .data, .bss and
the heap are invisible to the parent and vice versa, a pointer to a stack local
taken before the fork names the same object in both, and the child does
everything a vfork() child may not -- calls malloc() and printf(), and returns
from the function that called fork().
Both run at the top of user_main(). They exercise the lowest-level machinery
in the suite -- address environments, stack setup, the architecture's register
context -- so a fault in one takes the process down instead of reporting a
failure. Learning that in seconds rather than after everything else has passed
matters when a port is being brought up.
Each test gates on the one primitive it tests, ARCH_HAVE_VFORK and
ARCH_HAVE_FORK respectively. There is no compatibility layer and no mapping
between symbols. vfork.c no longer requires SCHED_WAITPID: the suspension is
in the kernel primitive now, so the test's core assertion holds without it and
only the status check is conditional.
The simulator is the one exception. It selects ARCH_HAVE_VFORK, but ostest
takes the sim down as soon as the test runs there, so the call keeps the
!ARCH_SIM guard that apps ee7642793 put on the old test in 2024. The old gate
hid this: ARCH_HAVE_FORK is not set on the sim, so the test was not built
there at all.
The other in-tree callers are audited for which primitive they actually meant:
* interpreters/python's _posixsubprocess and netutils/libwebsockets'
LWS_HAVE_WORKING_VFORK want the fork-then-exec path -- ARCH_HAVE_VFORK.
* python's os.fork() and libwebsockets' LWS_HAVE_FORK mean real fork() and stay
on ARCH_HAVE_FORK, so they become *absent* rather than silently wrong.
* testing/fs/fdsantest's vfork case follows ARCH_HAVE_VFORK.
interpreters/bas is deliberately left alone. Its SHELL and EDIT statements
reach for vfork() under an ARCH_HAVE_FORK guard and want the same treatment,
but checkpatch.sh checks the whole of any file a patch touches and
bas_statement.c produces 1681 pre-existing findings against master, so a
one-line change there fails CI on its own. The consequence is small:
EXAMPLES_BAS_SHELL is EXPERIMENTAL and already depends on ARCH_HAVE_FORK, so it
becomes unselectable rather than misbehaving.
Assisted-by: Claude Code:claude-opus-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
2026-08-07 13:03:08 +02:00
|
|
|
else if (ret >= 0 || errno != ECHILD)
|
2013-01-07 21:41:20 +00:00
|
|
|
{
|
testing/ostest: Split the fork test into vfork and fork.
ostest's "vfork" test was never testing vfork(). It has the child write a
global and the parent observe the write -- the defining property of *sharing*,
not of vfork(), whose defining property is that the parent is suspended and
whose contract forbids the child to write anything at all. It passed because
NuttX implemented fork() and vfork() as the same sharing primitive, which
apache/nuttx#19562 separates.
vfork.c is rewritten to test what vfork() promises. The child does only what
POSIX permits -- it calls _exit(42) and nothing else, not even exit(), which
would run atexit handlers and flush stdio in the parent's address space. Since
the child may not write memory and the parent cannot run while the child lives,
the observable is the child's exit status: had the parent not been suspended,
it would have reached waitpid() while the child was still alive. Where child
status is not retained -- ostest_main() sets SA_NOCLDWAIT for the whole run,
deliberately -- ECHILD is accepted as equally good evidence, since it says the
child was already gone when the parent asked.
fork.c is new and tests POSIX fork(): the child's writes to .data, .bss and
the heap are invisible to the parent and vice versa, a pointer to a stack local
taken before the fork names the same object in both, and the child does
everything a vfork() child may not -- calls malloc() and printf(), and returns
from the function that called fork().
Both run at the top of user_main(). They exercise the lowest-level machinery
in the suite -- address environments, stack setup, the architecture's register
context -- so a fault in one takes the process down instead of reporting a
failure. Learning that in seconds rather than after everything else has passed
matters when a port is being brought up.
Each test gates on the one primitive it tests, ARCH_HAVE_VFORK and
ARCH_HAVE_FORK respectively. There is no compatibility layer and no mapping
between symbols. vfork.c no longer requires SCHED_WAITPID: the suspension is
in the kernel primitive now, so the test's core assertion holds without it and
only the status check is conditional.
The simulator is the one exception. It selects ARCH_HAVE_VFORK, but ostest
takes the sim down as soon as the test runs there, so the call keeps the
!ARCH_SIM guard that apps ee7642793 put on the old test in 2024. The old gate
hid this: ARCH_HAVE_FORK is not set on the sim, so the test was not built
there at all.
The other in-tree callers are audited for which primitive they actually meant:
* interpreters/python's _posixsubprocess and netutils/libwebsockets'
LWS_HAVE_WORKING_VFORK want the fork-then-exec path -- ARCH_HAVE_VFORK.
* python's os.fork() and libwebsockets' LWS_HAVE_FORK mean real fork() and stay
on ARCH_HAVE_FORK, so they become *absent* rather than silently wrong.
* testing/fs/fdsantest's vfork case follows ARCH_HAVE_VFORK.
interpreters/bas is deliberately left alone. Its SHELL and EDIT statements
reach for vfork() under an ARCH_HAVE_FORK guard and want the same treatment,
but checkpatch.sh checks the whole of any file a patch touches and
bas_statement.c produces 1681 pre-existing findings against master, so a
one-line change there fails CI on its own. The consequence is small:
EXAMPLES_BAS_SHELL is EXPERIMENTAL and already depends on ARCH_HAVE_FORK, so it
becomes unselectable rather than misbehaving.
Assisted-by: Claude Code:claude-opus-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
2026-08-07 13:03:08 +02:00
|
|
|
printf("vfork_test: ERROR waitpid() returned %d (%d), expected "
|
|
|
|
|
"the child's status or ECHILD\n", ret, errno);
|
2022-10-15 23:36:29 +09:00
|
|
|
ASSERT(false);
|
2013-01-07 21:41:20 +00:00
|
|
|
return -1;
|
|
|
|
|
}
|
|
|
|
|
}
|
testing/ostest: Split the fork test into vfork and fork.
ostest's "vfork" test was never testing vfork(). It has the child write a
global and the parent observe the write -- the defining property of *sharing*,
not of vfork(), whose defining property is that the parent is suspended and
whose contract forbids the child to write anything at all. It passed because
NuttX implemented fork() and vfork() as the same sharing primitive, which
apache/nuttx#19562 separates.
vfork.c is rewritten to test what vfork() promises. The child does only what
POSIX permits -- it calls _exit(42) and nothing else, not even exit(), which
would run atexit handlers and flush stdio in the parent's address space. Since
the child may not write memory and the parent cannot run while the child lives,
the observable is the child's exit status: had the parent not been suspended,
it would have reached waitpid() while the child was still alive. Where child
status is not retained -- ostest_main() sets SA_NOCLDWAIT for the whole run,
deliberately -- ECHILD is accepted as equally good evidence, since it says the
child was already gone when the parent asked.
fork.c is new and tests POSIX fork(): the child's writes to .data, .bss and
the heap are invisible to the parent and vice versa, a pointer to a stack local
taken before the fork names the same object in both, and the child does
everything a vfork() child may not -- calls malloc() and printf(), and returns
from the function that called fork().
Both run at the top of user_main(). They exercise the lowest-level machinery
in the suite -- address environments, stack setup, the architecture's register
context -- so a fault in one takes the process down instead of reporting a
failure. Learning that in seconds rather than after everything else has passed
matters when a port is being brought up.
Each test gates on the one primitive it tests, ARCH_HAVE_VFORK and
ARCH_HAVE_FORK respectively. There is no compatibility layer and no mapping
between symbols. vfork.c no longer requires SCHED_WAITPID: the suspension is
in the kernel primitive now, so the test's core assertion holds without it and
only the status check is conditional.
The simulator is the one exception. It selects ARCH_HAVE_VFORK, but ostest
takes the sim down as soon as the test runs there, so the call keeps the
!ARCH_SIM guard that apps ee7642793 put on the old test in 2024. The old gate
hid this: ARCH_HAVE_FORK is not set on the sim, so the test was not built
there at all.
The other in-tree callers are audited for which primitive they actually meant:
* interpreters/python's _posixsubprocess and netutils/libwebsockets'
LWS_HAVE_WORKING_VFORK want the fork-then-exec path -- ARCH_HAVE_VFORK.
* python's os.fork() and libwebsockets' LWS_HAVE_FORK mean real fork() and stay
on ARCH_HAVE_FORK, so they become *absent* rather than silently wrong.
* testing/fs/fdsantest's vfork case follows ARCH_HAVE_VFORK.
interpreters/bas is deliberately left alone. Its SHELL and EDIT statements
reach for vfork() under an ARCH_HAVE_FORK guard and want the same treatment,
but checkpatch.sh checks the whole of any file a patch touches and
bas_statement.c produces 1681 pre-existing findings against master, so a
one-line change there fails CI on its own. The consequence is small:
EXAMPLES_BAS_SHELL is EXPERIMENTAL and already depends on ARCH_HAVE_FORK, so it
becomes unselectable rather than misbehaving.
Assisted-by: Claude Code:claude-opus-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
2026-08-07 13:03:08 +02:00
|
|
|
#endif
|
2013-01-07 21:41:20 +00:00
|
|
|
|
testing/ostest: Split the fork test into vfork and fork.
ostest's "vfork" test was never testing vfork(). It has the child write a
global and the parent observe the write -- the defining property of *sharing*,
not of vfork(), whose defining property is that the parent is suspended and
whose contract forbids the child to write anything at all. It passed because
NuttX implemented fork() and vfork() as the same sharing primitive, which
apache/nuttx#19562 separates.
vfork.c is rewritten to test what vfork() promises. The child does only what
POSIX permits -- it calls _exit(42) and nothing else, not even exit(), which
would run atexit handlers and flush stdio in the parent's address space. Since
the child may not write memory and the parent cannot run while the child lives,
the observable is the child's exit status: had the parent not been suspended,
it would have reached waitpid() while the child was still alive. Where child
status is not retained -- ostest_main() sets SA_NOCLDWAIT for the whole run,
deliberately -- ECHILD is accepted as equally good evidence, since it says the
child was already gone when the parent asked.
fork.c is new and tests POSIX fork(): the child's writes to .data, .bss and
the heap are invisible to the parent and vice versa, a pointer to a stack local
taken before the fork names the same object in both, and the child does
everything a vfork() child may not -- calls malloc() and printf(), and returns
from the function that called fork().
Both run at the top of user_main(). They exercise the lowest-level machinery
in the suite -- address environments, stack setup, the architecture's register
context -- so a fault in one takes the process down instead of reporting a
failure. Learning that in seconds rather than after everything else has passed
matters when a port is being brought up.
Each test gates on the one primitive it tests, ARCH_HAVE_VFORK and
ARCH_HAVE_FORK respectively. There is no compatibility layer and no mapping
between symbols. vfork.c no longer requires SCHED_WAITPID: the suspension is
in the kernel primitive now, so the test's core assertion holds without it and
only the status check is conditional.
The simulator is the one exception. It selects ARCH_HAVE_VFORK, but ostest
takes the sim down as soon as the test runs there, so the call keeps the
!ARCH_SIM guard that apps ee7642793 put on the old test in 2024. The old gate
hid this: ARCH_HAVE_FORK is not set on the sim, so the test was not built
there at all.
The other in-tree callers are audited for which primitive they actually meant:
* interpreters/python's _posixsubprocess and netutils/libwebsockets'
LWS_HAVE_WORKING_VFORK want the fork-then-exec path -- ARCH_HAVE_VFORK.
* python's os.fork() and libwebsockets' LWS_HAVE_FORK mean real fork() and stay
on ARCH_HAVE_FORK, so they become *absent* rather than silently wrong.
* testing/fs/fdsantest's vfork case follows ARCH_HAVE_VFORK.
interpreters/bas is deliberately left alone. Its SHELL and EDIT statements
reach for vfork() under an ARCH_HAVE_FORK guard and want the same treatment,
but checkpatch.sh checks the whole of any file a patch touches and
bas_statement.c produces 1681 pre-existing findings against master, so a
one-line change there fails CI on its own. The consequence is small:
EXAMPLES_BAS_SHELL is EXPERIMENTAL and already depends on ARCH_HAVE_FORK, so it
becomes unselectable rather than misbehaving.
Assisted-by: Claude Code:claude-opus-5
Signed-off-by: Marco Casaroli <marco.casaroli@gmail.com>
2026-08-07 13:03:08 +02:00
|
|
|
printf("vfork_test: Child %d ran and exited before the parent resumed\n",
|
|
|
|
|
pid);
|
2013-01-07 21:41:20 +00:00
|
|
|
return 0;
|
|
|
|
|
}
|
