Commit GSoC 2025 aio project by Ethan Miller:
https://blog.netbsd.org/tnf/entry/gsoc2025_asynchronous_i_o_framework

tests/lib/libc/sys/Makefile: Fix truncate_test.root_owned output.

Announce the creation of this file with ${_MKTARGET_CREATE}, not with
cryptic dd(1) output. And just use touch(1) instead of dd from
/dev/null, simpler and clearer what's going on.

No functional change intended, other than to the cosmetics of build
output.

t_poll: Expand tests of polling when other endpoint is closed.

This covers fifos, pipes, and sockets; reads and writes; and polling
when the endpoint is already closed as well as sleeping until the
endpoint is closed after a delay.

Unfortunately, there is disagreement right now between:

(a) POSIX,
(b) our poll(2) man page, and
(c) our implementation,

for many of these cases. And we recently changed the behaviour of
poll(2) on fifos between 9 and 10 (PR kern/56429: poll(2) should
yield POLLHUP when last writer to a fifo close()'s it) but didn't get
it quite right.

So I've annotated these test cases with xfails for all the cases I
think we're currently doing wrong -- under a justification from some
combination of POSIX, our own documentation, and/or a reasonable
interpretation of what the semantics _should_ be based on how read(2)
and write(2) will behave. These are mostly edge cases:

1. [pollclosed_pipe_*_write*] When the read side of a pipe is closed,
poll(2) on the write side currently returns POLLHUP|POLLOUT, which
POSIX forbids because it requires POLLHUP to be mutually exclusive
with POLLOUT. And our man page concurs with POSIX on this rule,
and has for a long time (going as far back as NetBSD 1.3 or
earlier, judging by https://man.NetBSD.org/NetBSD-1.3/poll.2).

=> Fixing this won't change whether anything wakes up earlier or
later -- it will only change which bits are set when the
wakeups happen.

2. [pollclosed_fifo*_immediate_readsome] When the write side of a
fifo was previously open with the read side, and has since been
closed before calling poll(2), poll(2) on the read side returns 0
instead of POLLHUP|POLLIN as it is expected to, even though
read(2) will return EOF without blocking.

=> Fixing this may lead to more wakeups than before, but only in
cases where read(2) would actually return immediately anyway.

3. [pollclosed_fifo*_delayed_*_read] When the write side of a fifo is
closed, poll(2) on the read side returns POLLHUP|POLLIN, as it is
expected to. But this state currently isn't persistent, even
though the behaviour of read(2) in returning EOF without blocking
is persistent.

=> Fixing this may lead to more wakeups than before, but only in
cases where read(2) would actually return immediately anyway.

That said, it turns out that we are correct, according to POSIX, in
not setting POLLHUP on a socket whose peer has been closed: POLLHUP
is only for devices, pipes, and FIFOs. So one of the issues I
reported in PR 59056 turns out to be bogus. (Also POLLHUP is only
for the reader side anyway, not for the writer side.)

https://pubs.opengroup.org/onlinepubs/9799919799/functions/poll.html

PR kern/59056: poll POLLHUP bugs

PR/58896: Martin Husemann: Demonstrate that a PT_STOPed process could not be
PT_KILLed.

branches: 1.76.2;
t_setrlimit: Narrow the scope of stack-protector warning suppression.

t_setrlimit uses alloca now

introduce new GCC 12 warning disables and use them in a few places

this introduces 4 new warning disable flags:

CC_WNO_MISSING_TEMPLATE_KEYWORD
CC_WNO_REGISTER
CC_WNO_STRINGOP_OVERREAD
CC_WNO_ARRAY_BOUNDS

and documents them in README.warnings. of these, the string op
and array bounds are both problematic (real bugs) and also spurious
(not real bugs), and the other 2 are mostly temporary for older
3rd party code.

add some new uses of CC_WNO_STRINGOP_OVERFLOW.

fix m68k build for gallium and GCC 12.

bsd.own.mk: rename to CC_WNO_ADDRESS_OF_PACKED_MEMBER

Provide a single variable
CC_WNO_ADDRESS_OF_PACKED_MEMBER
with options for both clang and gcc, to replace
CLANG_NO_ADDR_OF_PACKED_MEMBER
CC_NO_ADDR_OF_PACKED_MEMBER
GCC_NO_ADDR_OF_PACKED_MEMBER

Using the convention CC_compilerflag, where compilerflag
is based on the full compiler flag name.

branches: 1.72.2;

Provide _GNU_SOURCE for t_clone now that is required to make clone()
visible.

Add a regression test for PR kern/52239, "Changing protections of
already mmap'ed region can fail", based on the test program in the PR.

Test lib/libc/sys/t_timerfd often fails when run on QEMU because
QEMU misses clock interrupts.

Always check values against [ lower, upper ] bounds and use "4 * upper"
when run under QEMU.

Now becomes part of PR kern/43997 "Kernel timer discrepancies".

Add native implementations of eventfd(2) and timerfd(2), compatible with
the Linux interfaces of the same name.

branches: 1.68.2;
add support for new GCC 9 warnings that may be too much to fix
right now. new address-of-packed-member and format-overflow
warnings have new GCC_NO_ADDR_OF_PACKED_MEMBER amd
GCC_NO_FORMAT_OVERFLOW variables to remove these warnings.

apply to a bunch of the tree. mostly, these are real bugs that
should be fixed, but in many cases, only by removing the 'packed'
attribute from some structure that doesn't really need it. (i
looked at many different ones, and while perhaps 60-80% were
already properly aligned, it wasn't clear to me that the uses
were always coming from sane data vs network alignment, so it
doesn't seem safe to remove packed without careful research for
each affect struct.) clang already warned (and was not erroring)
for many of these cases, but gcc picked up dozens more.

New system call getrandom() compatible with Linux and others.

Three ways to call:

getrandom(p, n, 0) Blocks at boot until full entropy.
Returns up to n bytes at p; guarantees
up to 256 bytes even if interrupted
after blocking. getrandom(0,0,0)
serves as an entropy barrier: return
only after system has full entropy.

getrandom(p, n, GRND_INSECURE) Never blocks. Guarantees up to 256
bytes even if interrupted. Equivalent
to /dev/urandom. Safe only after
successful getrandom(...,0),
getrandom(...,GRND_RANDOM), or read
from /dev/random.

getrandom(p, n, GRND_RANDOM) May block at any time. Returns up to n
bytes at p, but no guarantees about how
many -- may return as short as 1 byte.
Equivalent to /dev/random. Legacy.
Provided only for source compatibility
with Linux.

Can also use flags|GRND_NONBLOCK to fail with EWOULDBLOCK/EAGAIN
without producing any output instead of blocking.

- The combination GRND_INSECURE|GRND_NONBLOCK is the same as
GRND_INSECURE, since GRND_INSECURE never blocks anyway.

- The combinations GRND_INSECURE|GRND_RANDOM and
GRND_INSECURE|GRND_RANDOM|GRND_NONBLOCK are nonsensical and fail
with EINVAL.

As proposed on tech-userlevel, tech-crypto, tech-security, and
tech-kern, and subsequently adopted by core (minus the getentropy part
of the proposal, because other operating systems and participants in
the discussion couldn't come to an agreement about getentropy and
blocking semantics):

https://mail-index.netbsd.org/tech-userlevel/2020/05/02/msg012333.html

Add ppoll() a compatibility wrapper around pollts(2)

Submitted by Apurva Nandan.

Fix build for vax; Compile t_ptrace_sigchld.c with -D__TEST_FENV.

Add a test for sigaltstack(2) and SA_ONSTACK

Add a NetBSD native futex implementation, mostly written by riastradh@.
Map the COMPAT_LINUX futex calls to the native ones.

PR/55177: Carlo Arenas: mremap(MAP_REMAPDUP) fails after fork()

Add new ptrace(2) test file

t_ptrace_sigchld - for SIGCHLD handler + ptrace(2).

Right now a single test is enabled (raise(SIGKILL)) and marked as failed
as it never finishes as the child is never collected before exiting the
parent uninterested about its child (SA_NOCLDWAIT).

fix variable name

Centralize the base rump libraries into a variable used by all the other
Makefiles so that we can make changes to it centrally as needed and have
less mess. Fixes the sun2 build that needs rumpvfs after librump after
the latest changes.

Disable t_ptrace_wait* tests for MKSANITIZER/MKLIBCSANITIZER

A subset of tests is not compatible with the sanitizers.

branches: 1.57.2;
Add a test for verifying procinfo note inside coredumps.

Add a first test for triggering a core dump in the debugged process
(via PT_DUMPCORE) and verifying it. The test finds procinfo note
and checks its contents.

The core dump is processed through libelf. However, it only provides
for finding all note segments (or sections?). I had to implement
finding and processing individual notes myself. I've added
a core_find_note() function that will be reused in future tests.

Reviewed by kamil.

Use LDADD, which apparently puts the libraries at the end of the command,
and makes static builds (aka sun2) happier.

Introduce check for the support of FPU exceptions

If FPU exceptions are unsupported, skip the SIGFPE crash tests.

Reuse code from tests/kernel/h_segv.c

Link t_ptrace_wait* tests with -pthread

While there, bump (c) for ATF ptrace(2) tests.

Add __used in infinite_thread() for consistency with other functions in
the file.

more tests

branches: 1.52.2;
Add a test to verify syscall() and/or __syscall() are working.
This used to trigger a panic (see PR kern/53261) on arm.

Add new ATF tests: t_fork and t_vfork

Test behavior of raise(signal) in either fork(2)ed or vfork(2)ed child.

Tests:
- raise1 SIGKILL
- raise2 SIGSTOP
- raise3 SIGTSTP
- raise4 SIGTTIN
- raise5 SIGTTOU
- raise6 SIGABRT
- raise7 SIGHUP
- raise8 SIGCONT

t_vfork:raise2 fails ignoring non-maskable SIGSTOP.

The remaining ones pass.

Sponsored by <The NetBSD Foundation>

branches: 1.50.2;
Add new tests in lib/libc/sys/t_ucontext

New tests:
- ucontext_sp
- ucontext_fp
- ucontext_pc
- ucontext_intrv

They test respectively:
- _UC_MACHINE_SP
- _UC_MACHINE_FP
- _UC_MACHINE_PC
- _UC_MACHINE_INTRV

These tests attempt to access and print the values from ucontext, without
interpreting the values.

This is a follow up of the _UC_MACHINE_FP() introduction.

These tests use PRIxREGISTER, and require to be built with -D_KERNTYPES.

Sponsored by <The NetBSD Foundation>

branches: 1.49.4;
Import ptrace(2) tests into appropriate directory tests/lib/libc/sys/

This is the correct directory documented in tests/README for such tests.

Discussed with <martin>

Sponsored by <The NetBSD Foundation>

Restrict special mode/owner settings to truncate_test.root_owned;
otherwise the Atffile will be generated unreadable for group/others
as well.

branches: 1.47.2;
Set FILESBUILD=yes to actually run the creation script for the file.

Should fix the build by creating a file which install can then find.

PR/51844: Ngie Cooper: use root-owned file created during build instead of
/usr/bin/fpr

New clock_nanosleep(2) testcase.

Add new tests: tests/lib/libc/sys/t_wait_noproc and t_wait_noproc_wnohang

The t_wait_noproc test checks whether wait(2)-family of functions return
error and set ECHILD for lack of childs.

The t_wait_noproc_wnohang adds to options (except wait(2), wait3(2)) new
parameter WNOHANG and verifies that error is still signaled and errno set
to ECHILD.

Currently t_wait_noproc_wnohang reports failures, these have been marked as
expected and referenced to PR standards/51606.

The problem report is authored by Robert Elz, and the initial regression
has been notified by Nicolas Joly.

Remove redundant test in tests/lib/libc/sys/t_wait for wait6(2) with no
WNOHANG specified.

Sponsored by <The NetBSD Foundation>.

New getsockname(2) testcase for UNIX domain sockets.

branches: 1.42.2;
Add wait6(2) tests.

t_lwp_create needs kernel types.

Define _KERNTYPES for things that need it.

Don't build tests that depend on RUMP if BSD_MK_COMPAT_FILE is defined.

New test program for bind(2), with a single (currently eroneously failing)
test case from Tyler Retzlaff.

PR/49617: Kirk Russell: posix_fallocate() should be returning an error on
failure, without setting errno.

Add new minherit test.

Fix static linking for the tests: -lrump is also used by -lrumpuser,
so we also need -lrump after -lrumpuser. Fixes build for sun2.

branches: 1.34.2;
There wasn't a codegen bug - the test program itself was buggy.

Nothing to see hear... move along.

Compile tests/lib/libc/sys/t_msgrcv with -O0 on hppa to avoid compiler
bug.

branches: 1.32.4;
Rename testprogram and make it more general by adding other testcases.
One commented out, I didn't manage to get all signal handling correct
for now.

Add a test program for PR kern/47625, based on the sample code provided
by anthony mallet.

Move the bitmap(3) test to the "right" place. Note it in bitops(3). Xrefs.

Add a bitops test

branches: 1.28.2;
setcontext() used to be incompatible with -lpthread since it affected
the TLS pointer, therefore wrecking the pthread environement.

Some ports had _UC_TLSBASE flag or equivalent (_UC_UNIQUE on alpha)
that controlled whether setcontext() would change the TLS pointer.
This change let libpthread override setcontext() with its own version
that unsets _UC_TLSBASE, enabling safe usage of setcontext() with
-lpthread.

We also have the following required changes here:
- rename alpha's _UC_UNIQUE into _UC_TLSBASE
- add _UC_TLSBASE definition in header file for all ports
(powerpc, sh3, sparc and sparc64 lack the implementation for now)
- introduce a libc stub that can be overriden for setcontext()
- modify MD libcs swapcontext() implementations so that they use the
setcontext() libc stub instead of doing a plain system call.

While we are there:
- document various MD _UC_* flags in header file
- add libc and libpthread tests for swapcontext() behavior
(hopefully helpful to spot MD problems introduced with this change)

Future work:
- Deciding whether kernel support or _UC_TLSBASE should be added for
powerpc, sh3, sparc and sparc64 is left to portmasters
sparc64

Approved by core@

Exclude tests that use rump

tests for recvmmsg

Revert previous - instead of guessing the amount of needed memory locked
limits (often way too high) and skipping the test case if in doubt,
raise the limits as far as we can, and fix a few places in the test where
we could run into the limits and either skip or fail with a reasonable
message.

Try to estimate the number of locked pages the mincore() test will need and
check it against resource limits, skipping the tests if it probably is too
low.

Calling _lwp_create() with a bogus ucontext could trigger a kernel
assertion failure (and thus a crash in DIAGNOSTIC kernels). Independently
discovered by YAMAMOTO Takashi and Joel Sing.

To avoid this, introduce a cpu_mcontext_validate() function and move all
sanity checks from cpu_setmcontext() there. Also untangle the netbsd32
compat mess slightly and add a cpu_mcontext32_validate() cousin there.

Add an exhaustive atf test case, based partly on code from Joel Sing.

Should finally fix the remaining open part of PR kern/43903.

Add few unit tests for mlock(2), including a case for PR kern/44788.

Move the _lwp_ctl(2) preemption counter check to the right place.

branches: 1.20.2;
Add a test for PR/45618: Motoyuki OHMORI: kqueue EVFILT_TIMER with smaller
timeout value makes kernel busy or panic

Move connect(2), listen(2) and socketpair(2) tests to the right place.

Add 21 unit tests for the XSI message queues.

branches: 1.17.2;
Move all pipe2 tests to a separate file.

Move the sigaction(2) test to the right place.

Move the ucontext(2) test to the right place.

Move the writev(2) test to the right place.

Move the clock_gettime(2) timer test to the right place.

Move the mkdir(2) test to the right place.

Move the pipe(2) test to the right place.

Move the posix_fadvise(2) test to the right place.

Remove ptrace(2) tests that are not yet ready but were accidentally
committed.

Simplify.

Add a simple test case to check executable mapping rights for mprotect(2).

- provide an exec_prot_support() routine so $ARCH can indicate whether
it supports execution protection or not, and skip test accordingly.

- have a trivial 'return_one' shellcode to copy anywhere in a page, and
call it. The decision to keep the assembly solution is mine, reasons are
twofold:

- all pure-C implementations cannot be controlled easily:
bounds detection (beginning/end) of return_one is unpredictable,
or requires the use of overkill solutions like libelf. Using
dlsym(3) was a good proposal, however I can't use it to know the
end address of the payload. It makes copying of the shellcode a bit
more difficult: using a constant may be too small (code has not been
entirely copied, and can lead to errors that can be erroneously
detected as "test passed"), or too big (depending on where it is mapped
in memory, copying past the end of the function may trigger SIGSEGV).

- I have to ensure that the resulting assembly is the most compact
possible, especially as it will be reused to test other parts of
memory (stack, data, rodata, etc.).

Only i386 and amd64 are implemented so far. Others will come in due time.
FWIW, writing the exec_prot_support() callback and the return_one payload
should be enough. Writing callback requires good knowledge of the platform,
depending on chip revision, CPU, board, MMU... the protection level may vary.

Current files are put under lib/libc/arch/* and lib/libc/common/. Feel free to
move them around the tests/ tree. Keep in mind that the common/ and arch/ code
will be used outside of libc, so please do not hide them too deep in the tree.

I checked a few architectures via build.sh cross-compile, and will keep an
eye on buildbot for potential build breakage. Feel free to contact me in
case you see any, of course.

Fix previous.

Remove 't_cerror' (this is tested by numerous individual tests).
Rename 't_context' to 't_getcontext' for consistency.

Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

Make sure we build all of thests, not just the last one (use += vs =)

Continuing the (re)organization of the lib/libc atf regression tests

test for sigqueue

Move macros wrapping functions from msg.h into msg.h

This makes code reusing easier.

branches: 1.2.2;
ATF t_ptrace_wait*: Disable debug messages in msg.h

msg.h is a dummy IPC interface.

Disable additional debugging logging here, especially wanted in race*
tests.

Sponsored by <The NetBSD Foundation>

branches: 1.1.2; 1.1.4; 1.1.8; 1.1.14;
Import ptrace(2) tests into appropriate directory tests/lib/libc/sys/

This is the correct directory documented in tests/README for such tests.

Discussed with <martin>

Sponsored by <The NetBSD Foundation>

PR misc/54382: whenever open(2) is called with O_CREAT, make sure to
pass an open mode argument.

branches: 1.2.14;
more tests needing <sys/stat.h>

branches: 1.1.24;
Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

Commit GSoC 2025 aio project by Ethan Miller:
https://blog.netbsd.org/tnf/entry/gsoc2025_asynchronous_i_o_framework

Commit GSoC 2025 aio project by Ethan Miller:
https://blog.netbsd.org/tnf/entry/gsoc2025_asynchronous_i_o_framework

Commit GSoC 2025 aio project by Ethan Miller:
https://blog.netbsd.org/tnf/entry/gsoc2025_asynchronous_i_o_framework

Commit GSoC 2025 aio project by Ethan Miller:
https://blog.netbsd.org/tnf/entry/gsoc2025_asynchronous_i_o_framework

fix typo bindinging -> binding

- ATF_REQUIRE(EINVAL == errno);
+ ATF_REQUIRE(EAFNOSUPPORT == errno);

as per http://pubs.opengroup.org/onlinepubs/009695399/functions/bind.html

while here remove some unnecessary initialization of port and addr members
and replace with a comment explaining why.

New test program for bind(2), with a single (currently eroneously failing)
test case from Tyler Retzlaff.

Move the bitmap(3) test to the "right" place. Note it in bitops(3). Xrefs.

Add a bitops test

Remove 't_cerror' (this is tested by numerous individual tests).
Rename 't_context' to 't_getcontext' for consistency.

Continuing the (re)organization of the lib/libc atf regression tests

more tests needing <sys/stat.h>

branches: 1.1.24;
Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

t_clock_gettime: Break overlong lines.

No functional change intended.

t_clock_gettime.c: Fix my typo

tv_nsec prints with %ld not %jd (should fix build)

clock_getres(2): Support CLOCK_{PROCESS,THREAD}_CPUTIME_ID

The syscall previously returned EINVAL for these two clocks. It still has
no support for CLOCK_VIRTUAL and CLOCK_PROF but clock_gettime(2) doesn't
either.

Fixes PR kern/59127

branches: 1.6.2;
t_clock_gettime: Avoid signed/unsigned comparison.

clock_gettime(2): Fix CLOCK_PROCESS/THREAD_CPUTIME_ID.

Use same calculation as getrusage, not some ad-hoc arithmetic of
internal scheduler parameters that are periodically rewound.

PR kern/57512

XXX pullup-8
XXX pullup-9
XXX pullup-10

t_clock_gettime: Add test for PR kern/57512.

branches: 1.3.16; 1.3.28;
Don't play with "../.." in includes for h_macros.h; deal with it centrally.
Minor fixes.

PR/51849: Ngie Cooper: headers portability fixes

branches: 1.1.24;
Move the clock_gettime(2) timer test to the right place.

branches: 1.1.2;
New clock_nanosleep(2) testcase.

tests/lib/libc/sys/t_clone: Nix trailing whitespace.

No functional change intended.

centralize stack allocation/freeing. The test that broke with the new jemalloc
had broken stack allocation.

branches: 1.4.26;
not allowed to map RWX anymore.

Don't try to figure out how the stack grows at run time. It's not
portable and fails badly if the test function is removed...

branches: 1.2.2;
Rename test case names.

Continuing the (re)organization of the lib/libc atf regression tests

PR/51850: Ngie Cooper: add sys/socket.h for socket(2) and don't leak fds

branches: 1.2.2;
add another test program for connect(2) that checks that connect fails
with EAFNOSUPPORT (similar to the bind test) if the domain of the socket
does not match the address family of the supplied address.

test currently fails as it should

branches: 1.1.2;
Move connect(2), listen(2) and socketpair(2) tests to the right place.

Remove 't_cerror' (this is tested by numerous individual tests).
Rename 't_context' to 't_getcontext' for consistency.

Continuing the (re)organization of the lib/libc atf regression tests

Adjust to the new dup3 behavior for equal fds

PR/51851: Ngie Cooper: add stdbool.h #include for false/true

branches: 1.8.14;
Move the references to PRs from code comments to the test description. Once
ATF has the ability to output the metadata in the HTML reports, it should be
easy to traverse between releng and gnats -reports via links.

branches: 1.7.2;
Remove unneeded uneeded err.h include.

Do not try to open files up to the RLIMIT_NOFILE limit, which fails
when process limit is identical to the whole system limit. Just do it
for a small fixed value, and adjust the limit accordingly instead.

Verify that dup2(2) and dup3(2) fail with EBADF if the "new descriptor" is
specified to be larger than RLIMIT_NOFILE.

PR lib/45148 was fixed; remove xfail.

Add some cases for dup2(2) and dup3(2).

Rewrite the limit-test.

Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

Validate basic fcntl(2) behavior for eventfd and timerfd.

Add native implementations of eventfd(2) and timerfd(2), compatible with
the Linux interfaces of the same name.

branches: 1.1.2;
file t_eventfd.c was initially added on branch thorpej-futex.

s/estabilishing/establishing/ in copy pasted comment.

Add new tests in ATF t_fork/t_vfork

Verify whether nested fork(2)/vfork(2)/clone(2) calls are supported in a
fork(2)ed/vforked(2) child.

The interesting ones are non-forked parent and non-forked child scenarios,
in particular double vfork(2).

branches: 1.3.2; 1.3.4;
Reflect reality in the comment

The watcher process is emitting SIGKILL, not SIGTERM.

Sponsored by <The NetBSD Foundation>

Stop masking SIGSTOP in a vfork(2)ed child

Keep the traditional BSD behavior masking SIGTSTP, SIGTTIN and SIGTTOU in
a vfork(2)ed child before exec(3)/exit(3). This is useful in shells and
prevents deadlocking, when a parent cannot unstop the sleeping child.

Change the behavior for SIGSTOP. This signal is by design not maskable and
this property shall be obeyed without exceptions. The STOP behavior is
expected in the context of debuggers and useful in standalone programs.

It is still possible to stop a vfork(2)ed child, however it requires
proc.curproc.stopfork=1, but it is not a flexible solution.

FreeBSD and OpenBSD keep masking SIGSTOP in a vfork(2)ed child.
Linux does not mask stop signals in the same scenarios.

This fixes ATF test: t_vfork:raise2.
No known regressions reported in the existing ATF tests.

Discussed with <kre>

Sponsored by <The NetBSD Foundation>

Add new ATF tests: t_fork and t_vfork

Test behavior of raise(signal) in either fork(2)ed or vfork(2)ed child.

Tests:
- raise1 SIGKILL
- raise2 SIGSTOP
- raise3 SIGTSTP
- raise4 SIGTTIN
- raise5 SIGTTOU
- raise6 SIGABRT
- raise7 SIGHUP
- raise8 SIGCONT

t_vfork:raise2 fails ignoring non-maskable SIGSTOP.

The remaining ones pass.

Sponsored by <The NetBSD Foundation>

Move the references to PRs from code comments to the test description. Once
ATF has the ability to output the metadata in the HTML reports, it should be
easy to traverse between releng and gnats -reports via links.

branches: 1.1.2;
Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

futex(2): Sign-extend FUTEX_WAKE_OP oparg/cmparg as Linux does.

Also mask off bits in the FUTEX_OP macro as Linux does so that passing
negative arguments works like in Linux.

PR kern/59129: futex(3): missing sign extension in FUTEX_WAKE_OP

t_futex_ops: Test sign-extension of WAKE_OP oparg/cmparg.

PR kern/59129: futex(3): missing sign extension in FUTEX_WAKE_OP

t_futex_ops: Relax various ATF_REQUIRE_* to ATF_CHECK_*.

This makes the test output more usable for browsing diagnostics when
multi-part tests partially fail.

Preparation for:

PR kern/59129: futex(3): missing sign extension in FUTEX_WAKE_OP

t_futex_ops: Spruce up diagnostics.

- Use RL to report errno for any syscalls that set it.
- Sprinkle messages to show intermediate quantities.

Preparation for:

PR kern/59129: futex(3): missing sign extension in FUTEX_WAKE_OP

futex(2): Fix return value of FUTEX_CMP_REQUEUE.

The return value is the number of waiters woken _or requeued_, not
just the number of waiters woken:

FUTEX_CMP_REQUEUE
Returns the total number of waiters that were woken up or
requeued to the futex for the futex word at uaddr2. If
this value is greater than val, then the difference is the
number of waiters requeued to the futex for the futex word
at uaddr2.

https://man7.org/linux/man-pages/man2/futex.2.html

While here, clarify some of the arguments with comments so it's not
quite so cryptic with val/val2/val3 everywhere.

PR kern/56828: futex calls in Linux emulation sometimes hang

futex(2): Fix FUTEX_CMP_REQUEUE to always compare even if no waiters.

It must always compare the futex value and fail with EAGAIN on
mismatch, even if there are no waiters.

FUTEX_CMP_REQUEUE (since Linux 2.6.7)
This operation first checks whether the location uaddr
still contains the value val3. If not, the operation
fails with the error EAGAIN. Otherwise, the operation [...]

https://man7.org/linux/man-pages/man2/futex.2.html

PR kern/56828: futex calls in Linux emulation sometimes hang

tests/lib/libc/sys/t_futex_ops: Fix another FUTEX_CMP_REQUEUE case.

PR kern/56828: futex calls in Linux emulation sometimes hang

tests/lib/libc/sys/t_futex_ops: Fix FUTEX_CMP_REQUEUE return values.

The return value is the number of waiters woken _or requeued_, not
just the number of waiters woken:

FUTEX_CMP_REQUEUE
Returns the total number of waiters that were woken up or
requeued to the futex for the futex word at uaddr2. If
this value is greater than val, then the difference is the
number of waiters requeued to the futex for the futex word
at uaddr2.

https://man7.org/linux/man-pages/man2/futex.2.html

PR kern/56828: futex calls in Linux emulation sometimes hang

tests/lib/libc/sys/t_futex_ops: Test FUTEX_CMP_REQUEUE edge case.

It must always compare the futex value and fail with EAGAIN on
mismatch, even if there are no waiters.

PR kern/56828: futex calls in Linux emulation sometimes hang

branches: 1.5.2; 1.5.8; 1.5.12;
Fix a bug in the futex_wake_highest_pri test case. Still fails as
expected.

Add a test case for PR kern/55230. It is currently marked as expect-fail.

- In uvm_voaddr_acquire(), take an extra hold on the anon lock obj.
- In uvm_voaddr_release(), if the anon ref count drops to 0, call
uvm_anfree() rather than uvm_anon_release(). Unconditionally drop
the anon lock, and release the extra hold on the anon lock obj.

Fixes a panic that occurs if the backing store for a futex backed by
an anon memory location is unmapped while a thread is waiting in the
futex.

Add a test case that reproduced the panic to verify that it's fixed.

Make FUTEX_WAIT_BITSET(bitset=0) fail with EINVAL to match Linux.

Add a NetBSD native futex implementation, mostly written by riastradh@.
Map the COMPAT_LINUX futex calls to the native ones.

Oops, accidentally #if 0'd some tests, probably while debugging
something else. Correct this silly mistake.

Add a NetBSD native futex implementation, mostly written by riastradh@.
Map the COMPAT_LINUX futex calls to the native ones.

Adjust.

Add some errno-checks.

Remove 't_cerror' (this is tested by numerous individual tests).
Rename 't_context' to 't_getcontext' for consistency.

Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

t_getitimer: Test invalid itimerval arguments.

Use sig_atomic_t while here.

Ex-POSIX (POSIX.1-2008; the interface was removed in POSIX.1-2024)
prescribes EINVAL for all of these cases:

SYNOPSIS

#include <sys/time.h>

int getitimer(int which, struct itimerval *value);
int setitimer(int which, const struct itimerval *restrict value,
struct itimerval *restrict ovalue);
[...]
ERRORS

The setitimer() function shall fail if:

[EINVAL]
The value argument is not in canonical form. (In canonical
form, the number of microseconds is a non-negative integer
less than 1000000 and the number of seconds is a non-negative
integer.)

https://pubs.opengroup.org/onlinepubs/9699919799/functions/setitimer.html

branches: 1.4.4;
fix various typos, mainly in comments.

In the setitimer_old test case, allow for time passing between the two
setitimer() calls. Should fix PR kern/54370.

branches: 1.2.32;
dholland fixed PR/44927

branches: 1.1.2;
Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

t_getrandom: Fix false positives for short read(2)

This only affects really slooooow machines, 68030 for me :)

branches: 1.4.2;
getrandom(2): Fix return value checks in automatic tests.

The syscall only guarantees up to 256 bytes in a single go -- if
interrupted, it might return short, but if the caller requested at
least 256 bytes it will definitely return 256 bytes.

Fix getrandom() tests.

Use sigaction() without SA_RESTART -- signal() implies SA_RESTART so
we never got the EINTR.

While here, reduce the timeout to something more reasonable so we
don't waste 20min of testbed time if anything goes wrong and the
one-second alarm doesn't fire.

Split getrandom tests into several cases to find out which ones hang.

New system call getrandom() compatible with Linux and others.

Three ways to call:

getrandom(p, n, 0) Blocks at boot until full entropy.
Returns up to n bytes at p; guarantees
up to 256 bytes even if interrupted
after blocking. getrandom(0,0,0)
serves as an entropy barrier: return
only after system has full entropy.

getrandom(p, n, GRND_INSECURE) Never blocks. Guarantees up to 256
bytes even if interrupted. Equivalent
to /dev/urandom. Safe only after
successful getrandom(...,0),
getrandom(...,GRND_RANDOM), or read
from /dev/random.

getrandom(p, n, GRND_RANDOM) May block at any time. Returns up to n
bytes at p, but no guarantees about how
many -- may return as short as 1 byte.
Equivalent to /dev/random. Legacy.
Provided only for source compatibility
with Linux.

Can also use flags|GRND_NONBLOCK to fail with EWOULDBLOCK/EAGAIN
without producing any output instead of blocking.

- The combination GRND_INSECURE|GRND_NONBLOCK is the same as
GRND_INSECURE, since GRND_INSECURE never blocks anyway.

- The combinations GRND_INSECURE|GRND_RANDOM and
GRND_INSECURE|GRND_RANDOM|GRND_NONBLOCK are nonsensical and fail
with EINVAL.

As proposed on tech-userlevel, tech-crypto, tech-security, and
tech-kern, and subsequently adopted by core (minus the getentropy part
of the proposal, because other operating systems and participants in
the discussion couldn't come to an agreement about getentropy and
blocking semantics):

https://mail-index.netbsd.org/tech-userlevel/2020/05/02/msg012333.html

Require at least 64 MB RAM to run the getrusage_maxrss test case as it
allocates 40 MB and we should leave some for the system, too.

don't print "long" with "%zu".

Make the getrusage_maxrss test more stable by preventing the compiler to
optimize out a dummy loop. While there print more details when failing.

add tests for maxrss, msgsnd

branches: 1.5.12;
PR/51852: Ngie Cooper: mark signo __unused

PR port-ia64/51261

Make this compile for ia64

branches: 1.3.2;
The nop instruction on or1k is l.nop

Requires stdint.h.

Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

branches: 1.1.2;
New getsockname(2) testcase for UNIX domain sockets.

Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

PR/55663: Ruslan Nikolaev: Add support for EVFILT_USER in kqueue(2)

Reference PRs consistently.

Skip kqueue_unsupported_fd test if /dev/drvctl is not configured
in the kernel. PR bin/49644.

XXX Ideally, it's better to modify to use more generic device file.

Skip kqueue_unsupported_fd if the drvctl device is not present.

Add a test for adding an event to an unsupported fd.

branches: 1.4.2;
adapt to new reality

Add a (skipped for now) test case for PR 46463

branches: 1.2.2;
Move the references to PRs from code comments to the test description. Once
ATF has the ability to output the metadata in the HTML reports, it should be
easy to traverse between releng and gnats -reports via links.

branches: 1.1.2; 1.1.6;
Add a test for PR/45618: Motoyuki OHMORI: kqueue EVFILT_TIMER with smaller
timeout value makes kernel busy or panic

tests/lib/libc/sys/t_kill: Test kill(INT_MIN, ...) fails with ESRCH.

PR kern/58425

branches: 1.1.44; 1.1.56; 1.1.58;
Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

PR/51853: Ngie Cooper: add limits.h for MAXPATHLEN in FreeBSD

branches: 1.2.6;
When a hard link to a directory fails, accept both EPERM and EACCESS as
valid error codes.

branches: 1.1.2; 1.1.8; 1.1.18;
Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

return return atf_no_error() instead of 0 for consistency.

suggested by moritzbuhl in https://github.com/NetBSD/src/pull/11/

branches: 1.5.14;
PR/51854: Ngie Cooper: sys/socket.h include and check result from socket(2)

branches: 1.4.14;
Move the references to PRs from code comments to the test description. Once
ATF has the ability to output the metadata in the HTML reports, it should be
easy to traverse between releng and gnats -reports via links.

This is no longer expected to fail.

Check the error values from listen(2), incl. case for standards/46150.

branches: 1.1.2;
Move connect(2), listen(2) and socketpair(2) tests to the right place.

s/priviledge/privilege/ in comments/log messages and fix one typo in descriptor.

Add a test case to ensure that _lwp_create() fails with the
expected error code when a bad new-lwp-id pointer is passed.

branches: 1.2.2;
Typo in comment

branches: 1.1.2;
Calling _lwp_create() with a bogus ucontext could trigger a kernel
assertion failure (and thus a crash in DIAGNOSTIC kernels). Independently
discovered by YAMAMOTO Takashi and Joel Sing.

To avoid this, introduce a cpu_mcontext_validate() function and move all
sanity checks from cpu_setmcontext() there. Also untangle the netbsd32
compat mess slightly and add a cpu_mcontext32_validate() cousin there.

Add an exhaustive atf test case, based partly on code from Joel Sing.

Should finally fix the remaining open part of PR kern/43903.

branches: 1.2.2;
Get rid of the weird macros.

Move the _lwp_ctl(2) preemption counter check to the right place.

Remove unit tests for the never-exposed _lwp_gettid(2).

branches: 1.1.2;
Add support for lazily generating a "global thread ID" for a LWP. This
identifier uniquely identifies an LWP across the entire system, and will
be used in future improvements in user-space synchronization primitives.

(Test disabled and libc stub not included intentionally so as to avoid
multiple libc version bumps.)

0x%p --> %p for non-external codes.

branches: 1.14.4;

Skip the mincore_shmseg test case if the SYSVSHM option is not configured
in the kernel.

Allow the test to build on systems where sys/mman.h does not provide
PROT_MPROTECT(), or if the test is built without _NETBSD_SOURCE defined,
by providing a dummy #define if the real one does not exist. The test
might then fail, if the PROT_MPROTECT() is really needed, but that's
better than not building.

Use PROT_MPROTECT() (which would have been better had it been called
PROT_MAXPROTECT or PROT_ALLOWPROTECT or something) on the mmap() call
which specifies PROT_NONE, and which we later want to change to PROT_READ,
otherwise when PAX is enabled, the mprotect() will fail.

Add the error in syscall failure.

Pr/51866: Ngie Cooper: fix leak and release limit on memory to prevent
false positives with mlock/mmap

more tests needing <sys/stat.h>

branches: 1.8.14;
Degrade all intermediate failures due to memory/resource shortage to
"skip" instead of "fail" - only if we get through to the real meat, we
can tell wether mlockall/mincore work or not.

Revert previous - instead of guessing the amount of needed memory locked
limits (often way too high) and skipping the test case if in doubt,
raise the limits as far as we can, and fix a few places in the test where
we could run into the limits and either skip or fail with a reasonable
message.

Try to estimate the number of locked pages the mincore() test will need and
check it against resource limits, skipping the tests if it probably is too
low.

Simplify creation of a temporary file slightly, fix sizeof(buf) confusion
when buf is a pointer.

Skip the ENOMEM/RLIMIT_MEMLOCK case when doing mlockall(2).

branches: 1.3.2;
Remove the 't_mincore' test case, as it is essentially a less elegant
version of the recently added page residency checks.

Add the test cases from the old ../regress/sys/uvm/mmap/mmap.c.

Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

branches: 1.1.2; 1.1.6;
Add new minherit test.

Add few basic checks.

Move the mkdir(2) test to the right place.

Add mkfifo{,at}(2) mode in mknod{,at}(2) as requested by POSIX

mknod with mode & S_IFIFO and dev=0 shall behave like mkfifo.

Update the documentation to reflect this state.

Add ATF tests.

This is an in-kernel implementation as typically user-space programs use
mkfifo(2) directly, however whenever there is need to bypass libc (like in
valgrind) then portable POSIX software calls the mknod syscall.

Noted on tech-kern@ by Greg Troxel.

branches: 1.2.42;
Fix a test case to test what the comment says.

Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

Move the references to PRs from code comments to the test description. Once
ATF has the ability to output the metadata in the HTML reports, it should be
easy to traverse between releng and gnats -reports via links.

branches: 1.1.2;
Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

Don't print the pagesize it messes up atf-report

Add a bunch of fprintf(stderr,...) diagnostics to try and help determine
what is going wrong which is causing this test to crash the kernel.

branches: 1.6.14;

This test should be testing how the kernel should behave, not how
it actually (used to) behave when it was incorrect... Aside from
a possible EPERM (which is not tested) the only error possible from
mlock() is ENOMEM. POSIX also allows EINVAL if the address is not
page aligned and the implementation does not round down to the previous
page boundary, but NetBSD does.

The kernel was recently fixed to return the correct errors for mlock()
so now we really need the test to be checking them, and not expecting
the incorrect errno values that the kernel used to return.

Same for munlock() - there ENOMEM is the only possible error, again,
EINVAL cannot happen as the kernel rounds to page boundaries.
For munlock() the kernel has not yet been corrected (that is coming
real soon...) and one of the munlock() tests will currently fail
(as of the time this commit is made, hopefully not for much longer)
as it should - it is indicating a kernel bug.

Note that NetBSD mlock(2) talks about EINVAL for cases where the length
parameter is negative ... but that is a size_t - good luck having that
ever occur (the man page will soon be corrected as well.)

Remove hardcoded VM_MIN_ADDRESS and instead query current value via
sysctl vm.minaddress.

branches: 1.4.2;
Make this compile on archs where VM_MIN_ADDRESS is not defined (or not
exported to userland)

Deal with architectures defining VM_MIN_ADDRESS > 0.
Add additional error code tests using a pointer one page past current brk.

branches: 1.2.2;
Avoid harmless compiler (integer) warnings.

Add few unit tests for mlock(2), including a case for PR kern/44788.

tests/lib/libc: Test mmap(2) with bad hints.

Add a regression test for PR kern/52239, "Changing protections of
already mmap'ed region can fail", based on the test program in the PR.

In the mmap_err test case, mmap the address, not the address of the address.
Should fix PR kern/56780.

When the t_mmap:mmap_err test fails due to errno not having the expected
value of EINVAL, print the actual value.

Reference PRs consistently.

Handle dk disks and print errors.

PR/51891: Ngie Cooper: use sys/disklabel.h instead of machine/disklabel.h

branches: 1.11.2;
PR/51855: Ngie Cooper: fix leaks

more tests needing <sys/stat.h>

branches: 1.9.2;
Make sure to use the values calculated in the loop that is supposed to crash
the client process - so the compiler does not optimize it all away.
Also improve diagnostic in case of common bug: SIGSEGV instead of SIGBUS.

add truncate causing signal test

Disable the mmap_block test again, it doesn't panic when mmaping /dev/wd0d
only by accident. PR kern/46592.

Do not skip the block device mmap test, as it does not crash
the kernel any more. Mark it as expected failure instead.

mmap_block:
do not use a hardcoded block device list, but query the kernel for attached
disks instead, then try to mmap the raw partition.

Use atf_tc_skip().

A test case for serious PR kern/38889: crash on open/mmap/close of block
device. The test case is skipped for the time being as it replicates the
panic described in the PR (tested on NetBSD/amd64 6.0 BETA).

branches: 1.2.2; 1.2.4;
Move the uvm "loan" test to tests(7).

Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

PR/55177: Carlo Arenas: mremap(MAP_REMAPDUP) fails after fork()

PR misc/54382: whenever open(2) is called with O_CREAT, make sure to
pass an open mode argument.

branches: 1.7.10;
Extend the mmap(2) interface to allow requesting protections for later
use with mprotect(2), but without enabling them immediately.

Extend the mremap(2) interface to allow duplicating mappings, i.e.
create a second range of virtual addresses references the same physical
pages. Duplicated mappings can have different effective protections.

Adjust PAX mprotect logic to disallow effective protections of W&X, but
allow one mapping W and another X protections. This obsoletes using
temporary files for purposes like JIT.

Adjust PAX logic for mmap(2) and mprotect(2) to fail if W&X is requested
and not silently drop the X protection.

Improve test cases to ensure correct operation of the changed
interfaces.

branches: 1.6.2;
Rather than ghecking system global state to see if mprotect might be
enabled, use the newly-committed sysctl proc.curproc.paxflags variable

Do not toggle global security.pax.mprotect state in an attempt to
activate it for the current process. It does not work and tests should
not change global system state anyway.
Instead: skip the test is pax.mprotect is not globally enabled. We could
use a better check for this (querying the current processes pax flags),
but unfortunately we don't have one.

branches: 1.4.2; 1.4.4;
Skip the mprotect_exec test if PaX MPROTECT restrictions are enabled.

Modify exec test to be more resilient regarding partial exec protection:
skip test instead of returning a failure that could be misinterpreted.

Suggested by Jukka in private mail.

Add a simple test case to check executable mapping rights for mprotect(2).

- provide an exec_prot_support() routine so $ARCH can indicate whether
it supports execution protection or not, and skip test accordingly.

- have a trivial 'return_one' shellcode to copy anywhere in a page, and
call it. The decision to keep the assembly solution is mine, reasons are
twofold:

- all pure-C implementations cannot be controlled easily:
bounds detection (beginning/end) of return_one is unpredictable,
or requires the use of overkill solutions like libelf. Using
dlsym(3) was a good proposal, however I can't use it to know the
end address of the payload. It makes copying of the shellcode a bit
more difficult: using a constant may be too small (code has not been
entirely copied, and can lead to errors that can be erroneously
detected as "test passed"), or too big (depending on where it is mapped
in memory, copying past the end of the function may trigger SIGSEGV).

- I have to ensure that the resulting assembly is the most compact
possible, especially as it will be reused to test other parts of
memory (stack, data, rodata, etc.).

Only i386 and amd64 are implemented so far. Others will come in due time.
FWIW, writing the exec_prot_support() callback and the return_one payload
should be enough. Writing callback requires good knowledge of the platform,
depending on chip revision, CPU, board, MMU... the protection level may vary.

Current files are put under lib/libc/arch/* and lib/libc/common/. Feel free to
move them around the tests/ tree. Keep in mind that the common/ and arch/ code
will be used outside of libc, so please do not hide them too deep in the tree.

I checked a few architectures via build.sh cross-compile, and will keep an
eye on buildbot for potential build breakage. Feel free to contact me in
case you see any, of course.

Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

Avoid plabels on hppa to make some tests not crash

branches: 1.1.2;
PR/55177: Carlo Arenas: mremap(MAP_REMAPDUP) fails after fork()

Avoid plabels on hppa to make some tests not crash

branches: 1.1.2;
Oops, need the header too.

If we have SYSVMSG in kernel (test finds it) restore the signal handler
to its previous state, so any signals that occur during the rest of the
test gat handled just as they would (core dump) as if the handler was never
executed. No need to bother if the SYSVMSG test fails, as in that case,
nothing else will be done anyway.

Make this test skip if running on a kernel without the SYSVMSG option.
(Assuming this behaves as intended the other tests which should behave
the same way will get upgrades as well.)

PR/51856: Ngie Cooper: add limits.h for INT_MAX

branches: 1.4.8;
Use an abs wide enough to cover the types involved.

branches: 1.3.6;
Fix typo.

branches: 1.2.2;
Fix some typos.

Add 21 unit tests for the XSI message queues.

Don't bother attempting any of the (real) test cases if the SYSVMSG
option isn't configured in the kernel.

Use an abs wide enough to cover the types involved.

branches: 1.1.2; 1.1.8;
Add 21 unit tests for the XSI message queues.

Don't bother attempting any of the (real) test cases if the SYSVMSG
option isn't configured in the kernel.

PR/51856: Ngie Cooper: add limits.h for INT_MAX

branches: 1.3.8;
Fix msgsz confusion.

branches: 1.2.2; 1.2.6;
Fix typo.

branches: 1.1.2;
Add 21 unit tests for the XSI message queues.

Don't bother attempting any of the (real) test cases if the SYSVMSG
option isn't configured in the kernel.

PR/51856: Ngie Cooper: add limits.h for INT_MAX

branches: 1.2.2; 1.2.24;
Fix some typos.

Add 21 unit tests for the XSI message queues.

PR/51867: Ngie Cooper: misc cleanups

branches: 1.2.14;
Don't rely on INT_MAX being unmapped. Use mmap to get a page and then unmap
it. Then use that address for msync.

branches: 1.1.2;
Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

PR/51857: Ngie Cooper: mark signo __unused for -Wunused

branches: 1.3.10;
add more test cases for nanosleep

branches: 1.2.2;
Note PR bin/14558.

branches: 1.1.2;
Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

Reference PRs consistently.

s/inital/initial/

branches: 1.5.14;
Don't play with "../.." in includes for h_macros.h; deal with it centrally.
Minor fixes.

PR/51858: Ngie Cooper: don't leak buf

branches: 1.3.24;
Move all pipe2 tests to a separate file.

New testcase that check for close-on-exec flag on pipe2 syscall.

Move the pipe(2) test to the right place.

ATF test additions for O_CLOFORK.

These are Ricardo Branco's ATF test modifications for O_CLOFORK
& SOCK_CLOFORK (with some FD_CLOFORK included).

These have all been made able to be built & run in an environment
where O_CLOFORK (etc) are not yet visible to userland.

branches: 1.9.30;
PR/51859: Ngie Cooper: use closefrom/restore RLIMIT_NOFILE when done

branches: 1.8.14;
Remove also redundant comment (the reference to the PR is already in the
metadata).

Simplify the test for PR kern/4645 and make it independend of resource
settings.

Fix typo in comment.

Enable the test for PR kern/46457 now that it does not crash the
kernel any more.

Add a case for PR kern/46457. This is skipped for the time being, as it
reproduces the panic described in the PR.

branches: 1.3.2;
add a signopipe test

branches: 1.2.2;
make this work

Move all pipe2 tests to a separate file.

t_poll: Extend the hangup tests to ptys too.

Of course, ptys work slightly differently from pipes and fifos and
sockets, presumably for hysterical raisins...

While here: add a more detailed comment to each xfail for the reason
why I think it is broken.

PR kern/59056: poll POLLHUP bugs

t_poll: Expand tests of polling when other endpoint is closed.

This covers fifos, pipes, and sockets; reads and writes; and polling
when the endpoint is already closed as well as sleeping until the
endpoint is closed after a delay.

Unfortunately, there is disagreement right now between:

(a) POSIX,
(b) our poll(2) man page, and
(c) our implementation,

for many of these cases. And we recently changed the behaviour of
poll(2) on fifos between 9 and 10 (PR kern/56429: poll(2) should
yield POLLHUP when last writer to a fifo close()'s it) but didn't get
it quite right.

So I've annotated these test cases with xfails for all the cases I
think we're currently doing wrong -- under a justification from some
combination of POSIX, our own documentation, and/or a reasonable
interpretation of what the semantics _should_ be based on how read(2)
and write(2) will behave. These are mostly edge cases:

1. [pollclosed_pipe_*_write*] When the read side of a pipe is closed,
poll(2) on the write side currently returns POLLHUP|POLLOUT, which
POSIX forbids because it requires POLLHUP to be mutually exclusive
with POLLOUT. And our man page concurs with POSIX on this rule,
and has for a long time (going as far back as NetBSD 1.3 or
earlier, judging by https://man.NetBSD.org/NetBSD-1.3/poll.2).

=> Fixing this won't change whether anything wakes up earlier or
later -- it will only change which bits are set when the
wakeups happen.

2. [pollclosed_fifo*_immediate_readsome] When the write side of a
fifo was previously open with the read side, and has since been
closed before calling poll(2), poll(2) on the read side returns 0
instead of POLLHUP|POLLIN as it is expected to, even though
read(2) will return EOF without blocking.

=> Fixing this may lead to more wakeups than before, but only in
cases where read(2) would actually return immediately anyway.

3. [pollclosed_fifo*_delayed_*_read] When the write side of a fifo is
closed, poll(2) on the read side returns POLLHUP|POLLIN, as it is
expected to. But this state currently isn't persistent, even
though the behaviour of read(2) in returning EOF without blocking
is persistent.

=> Fixing this may lead to more wakeups than before, but only in
cases where read(2) would actually return immediately anyway.

That said, it turns out that we are correct, according to POSIX, in
not setting POLLHUP on a socket whose peer has been closed: POLLHUP
is only for devices, pipes, and FIFOs. So one of the issues I
reported in PR 59056 turns out to be bogus. (Also POLLHUP is only
for the reader side anyway, not for the writer side.)

https://pubs.opengroup.org/onlinepubs/9799919799/functions/poll.html

PR kern/59056: poll POLLHUP bugs

t_poll: Tweak to improve diagnostics.

PR kern/59056: poll POLLHUP bugs

t_poll: Make sure POLLOUT is clear when testing POLLHUP.

POLLOUT and POLLHUP are mutually exclusive in POSIX. Also, we're
only polling for POLLIN anyway.

PR kern/59056: poll POLLHUP bugs

branches: 1.8.4;
- Strenghen the poll(2) fifo_inout test to ensure that once the reader
has read enough that exactly PIPE_BUF space is available that the FIFO
becomes writable again.
- When creating a FIFO, ensure that the receive low water mark is 1
(a FIFO must be readable when at least 1 byte is available); this
was already the case implicitly, but this makes it explicit.
- Similarly, set the send low water mark to PIPE_BUF to ensure that
the pipe is writable when at least PIPE_BUF bytes of space are available
in the send buffer. Without this change, the strengthened test case
above does not pass (the default send low water mark is larger than
PIPE_BUF; see soreserve()).
- Make the same low water mark changes to the PIPE_SOCKETPAIR case.

Add a fifo_inout test case that validates the expected POLLIN / POLLOUT
behavior for FIFOs:
- A FIFO is readable so long as at least 1 byte is available.
- A FIFO is writable so long as at least PIPE_BUF (obtained with _PC_PIPE_BUF)
space is avaiable.

This will be cloned for a forthcoming kevent test case.

In the fifo_hup1 test, also ensure that POLLHUP is de-asserted when a
new writer appears.

- fifo_poll(): If the last writer has disappeared, detect this and return
POLLHUP, per POSIX.
- fifo_close(): Use the new fifo_socantrcvmore(), which is like the
garden-variety socantrcvmore(), except it specifies POLL_HUP rather
than POLL_IN (so the correct code for SIGIO is sent).
- sowakeup(): Allow POLL_HUP as a code (notifies poll'ers with POLLHUP).
- Add test cases for correct POLLHUP behavior with FIFOs.

Fixes PR kern/56429.

Add ppoll() a compatibility wrapper around pollts(2)

Submitted by Apurva Nandan.

branches: 1.3.34;
Move the references to PRs from code comments to the test description. Once
ATF has the ability to output the metadata in the HTML reports, it should be
easy to traverse between releng and gnats -reports via links.

branches: 1.2.2;
Follow the design principles of tests(7) by merging 't_poll3w' to 't_poll'
as a test case instead of separate file.

Add the syscall tests back. Couple of files and tests were also renamed for
consistency. All of these now mimic the libc structure.

t_pollts.c: fix typo in a description string (NFC)

Add ppoll() a compatibility wrapper around pollts(2)

Submitted by Apurva Nandan.

t_posix_fadvise: Don't check whether errno is preserved.

I can find no guarantee in POSIX about posix_fadvise preserving
errno; until such language is found I'm going to assume there is no
such guarantee.

What is happening is that, sometimes, rump_sys_posix_fadvise waits on
a mutex or condvar, which uses _lwp_park internally, which sometimes
wakes up early with EALREADY because a wakeup was already pending for
the thread by the time it entered _lwp_park. And that EALREADY is
delivered by _lwp_park via errno.

PR kern/53931: posix_fadvise_reg test case fails randomly on real
hardware

branches: 1.3.18;
No semicolon after macro do ... while (0) wrapper.

branches: 1.2.12;
Don't play with "../.." in includes for h_macros.h; deal with it centrally.
Minor fixes.

branches: 1.1.24;
Move the posix_fadvise(2) test to the right place.

PR/49617: Kirk Russell: posix_fallocate() should be returning an error on
failure, without setting errno.

Add ppoll() a compatibility wrapper around pollts(2)

Submitted by Apurva Nandan.

tests/lib/libc/sys/t_ptrace*: Sprinkle improved diagnostics.

This should make it easier to look at the test failure output and
figure out what's going on.

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

branches: 1.5.4;
s/estabilishing/establishing/ in copy pasted comment.

Revert previous change in t_ptrace.c

By a mistake this file started to include <sys/mman.h>
This is not needed.

The include was intended to be add just in t_ptrace_wait.c.

Sponsored by <The NetBSD Foundation>

Add a new ptrace(2) ATF exploit for: CVE-2018-8897 (POP SS debug exception)

The original expolit and mitigation have been developed by Maxime Villard.

I've reworked the shared code and adapted for the ATF context.
I've included PIE-aware version (RIP/EIP relative) and introduced
additional comments to the explanation how to trigger the flaw.

This test passes on NetBSD/8.99.17.

Sponsored by <The NetBSD Foundation>

Add two new ptrace(2) ATF tests

Added:

- traceme_pid1_parent
Assert that a process cannot mark its parent a debugger twice

- traceme_twice
Verify that PT_TRACE_ME is not allowed when our parent is PID1

All tests pass.

Sponsored by <The NetBSD Foundation>

branches: 1.1.2; 1.1.4; 1.1.14;
Import ptrace(2) tests into appropriate directory tests/lib/libc/sys/

This is the correct directory documented in tests/README for such tests.

Discussed with <martin>

Sponsored by <The NetBSD Foundation>

tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

branches: 1.12.14;
Combine x86 register tests into unified test function

Reduce the code duplication and improve maintainability of x86 register
tests by combining all of them to a single base function.

Add more debug to register writing tests

Add tests for PT_SETREGS

fix unused variable warnings.

Introduce tests for reading r8..r15 registers.

Add tests for reading pre-set general purpose register values via PT_GETREGS.

Add new tests verifying that PT_GETREGS provides correct register values
from the underlying program. The test uses inline assembly in the child program
to set registers to well-known constants, then compares the results obtained
via PT_GETREGS against them.

Reviewed by kamil.

Link t_ptrace_wait* tests with -pthread

While there, bump (c) for ATF ptrace(2) tests.

Add __used in infinite_thread() for consistency with other functions in
the file.

branches: 1.5.2;
Port the CVE 2018-8897 mitigation to i386 ATF ptrace(2) tests

On i386 there is no need to switch execution mode.
Use 0x23 SS selector for i386, amd64 used 0x4f.

Based on pointers from <maxv>.

Sponsored by <The NetBSD Foundation>

Simplify the x86_64_cve_2018_8897 ATF ptrace(2) test

Do not call _exit() from the child, ad this code shall not be reached.
Put there assert().

No functional change. The test still passes.

Sponsored by <The NetBSD Foundation>

Add a new ptrace(2) ATF exploit for: CVE-2018-8897 (POP SS debug exception)

The original expolit and mitigation have been developed by Maxime Villard.

I've reworked the shared code and adapted for the ATF context.
I've included PIE-aware version (RIP/EIP relative) and introduced
additional comments to the explanation how to trigger the flaw.

This test passes on NetBSD/8.99.17.

Sponsored by <The NetBSD Foundation>

branches: 1.2.2;
sync a bit more with reality; some things still fail, some new failures.
reduce spewage, be more explanatory about syscall errors.

branches: 1.1.2; 1.1.4; 1.1.8;
Import ptrace(2) tests into appropriate directory tests/lib/libc/sys/

This is the correct directory documented in tests/README for such tests.

Discussed with <martin>

Sponsored by <The NetBSD Foundation>

tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

branches: 1.1.12;
Move byte transfer tests out of t_ptrace_wait.c to t_ptrace_bytetransfer_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

branches: 1.3.12;
Fix potential race in ptrace(2) clone(2) tests

Instead of comparing old and new signal mask, just after the cloning
operation, check whether the expected signal is still masked and in
another test whether it is still ignored.

Catch up after t_ptrace_fork_wait.h r. 1.2.

Bump (c) year

Move clone() tests out of t_ptrace_wait.c to t_ptrace_clone_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

branches: 1.9.2;
core_dump_procinfo: Fix false positive for vax

Increment PC after trapped by `bpt`, as done for many other platforms.

t_ptrace_core_wait.h: Increment PC after trap, also for riscv

Fix *:core_dump_procinfo tests for riscv64.

There are two hard problems in computer science cache invalidation, naming
and off-by-one errors.

Fix a naming and off-by-one bug... Thanks to Tom Lane for spotting my typo.

PR/56865: hppa: t_ptrace_core_wait test doesn't know it must advance PC after breakpoint

Add __hppa__ to the list of architectures that need to advance to the instruction
after the breakpoint

Thanks to Tom Lane for the fix.

Add sparc* to the list of architectures that need an explicit address
with PT_CONTINUE in this test.

For sh3, increment PC when PT_CONTINUE from trigger_trap(), as already
done for aarch64, arm, and powerpc. Otherwise, child is trapped to the
PTRACE_BREAKPOINT_ASM (== trapa) instruction indefinitely.

Fix tests/lib/libc/sys/t_ptrace_wait*:core_dump_procinfo.

Apply fix in rev 1.2 for core_dump_procinfo to aarch64 and arm:

http://cvsweb.netbsd.org/bsdweb.cgi/src/tests/lib/libc/sys/t_ptrace_core_wait.h#rev1.2

> For powerpc, program counter is not automatically incremented by trap
> instruction. We cannot increment PC in the trap handler, which breaks
> applications depending on this behavior, e.g., GDB.

This statement is true for aarch64 and arm.

Also, use PTRACE_BREAKPOINT_SIZE instead of hard-coded 4 to address
instruction next to PC.

OK ryo

Fix core_dump_procinfo tests for powerpc, for which child process was
stalled indefinitely in trap instruction even after PT_CONTINUE.

For powerpc, program counter is not automatically incremented by trap
instruction. We cannot increment PC in the trap handler, which breaks
applications depending on this behavior, e.g., GDB.

Therefore, we need to pass (PC + 4) instead of (void *)1 (== PC) to
PT_CONTINUE when child process traps itself.

Move core tests out of t_ptrace_wait.c to t_ptrace_core_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

branches: 1.1.12;
Move eventmask tests out of t_ptrace_wait.c to t_ptrace_eventmask_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move exec() tests out of t_ptrace_wait.c to t_ptrace_exec_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

branches: 1.7.12;
Add fork/vfork/posix_spawn tests for processes within pgrp

New tests:
- fork_setpgid
- vfork_setpgid
- posix_spawn_setpgid
- unrelated_tracer_fork_setpgid
- unrelated_tracer_vfork_setpgid
- unrelated_tracer_posix_spawn_setpgid

These tests trigger a kernel assert for pg_jobc going negative.

The tests are temporarily skipped.

Ignore interception of SIGCHLD signals in the debugger

Set SIGPASS for SIGCHLD for the traced child in the following tests:
- unrelated_tracer_fork*
- unrelated_tracer_vfork*
- unrelated_tracer_posix_spawn*

There is a race that SIGCHLD might be blocked during forking and dropped.

PR/55241 by Andreas Gustafsson

Fix typo in test names

Noted by <joerg>

Ignore interception of SIGCHLD signals in the debugger

Set SIGPASS for SIGCHLD for the traced child in the following tests:

- posix_spawn_singalmasked
- posix_spawn_singalignored
- fork_singalmasked
- fork_singalignored
- vfork_singalmasked
- vfork_singalignored
- vforkdone_singalmasked
- vforkdone_singalignored

There is a race that SIGCHLD might be blocked during forking and dropped.

PR/55241 by Andreas Gustafsson

Ignore interception of the SIGCHLD signals.

SIGCHLD once blocked is discarded by the kernel as it has the
SA_IGNORE property. During the fork(2) operation all signals can be
shortly blocked and missed (unless there is a registered signal
handler in the traced child). This leads to a race in this test if
there would be an intention to catch SIGCHLD.

Fixes PR lib/55241 by Andreas Gustafsson

Fix race in fork_singalmasked

ELF RTLD after rtld.c r. 1.204 introduced locking that wraps the fork
syscall. This locking changes signal mask of the calling process during
the forking process.

Instead of comparing old and new signal mask, just after the forking
operation, check whether the expected signal is still masked and in
another test whether it is still ignored.

Move fork/vfork/posix_spawn tests out of t_ptrace_wait.c to t_ptrace_fork_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Combine x86 register tests into unified test function

Reduce the code duplication and improve maintainability of x86 register
tests by combining all of them to a single base function.

Add more debug to register writing tests

Add tests for PT_SETREGS

fix unused variable warnings.

Reintroduce i386 tests for general-purpose registers, without MMX req

Reintroduce the test for reading i386 general-purpose register values
from the debugged program. This time it's split into one test for the six
registers not normally used by the compiler (eax..edx, esi, edi), and another
for ebp, esp. This leaves enough free registers for the latter to work
without MMX.

Revert i386_regs_gp_read as it fails on non-MMX systems.

Add tests for reading pre-set general purpose register values via PT_GETREGS.

Add new tests verifying that PT_GETREGS provides correct register values
from the underlying program. The test uses inline assembly in the child program
to set registers to well-known constants, then compares the results obtained
via PT_GETREGS against them.

Reviewed by kamil.

Link t_ptrace_wait* tests with -pthread

While there, bump (c) for ATF ptrace(2) tests.

Add __used in infinite_thread() for consistency with other functions in
the file.

branches: 1.2.4;
sync a bit more with reality; some things still fail, some new failures.
reduce spewage, be more explanatory about syscall errors.

branches: 1.1.2; 1.1.4; 1.1.8;
Import ptrace(2) tests into appropriate directory tests/lib/libc/sys/

This is the correct directory documented in tests/README for such tests.

Discussed with <martin>

Sponsored by <The NetBSD Foundation>

branches: 1.2.4;
tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

PR/58896: Martin Husemann: Demonstrate that a PT_STOPed process could not be
PT_KILLed.

Move kill()-like tests out of t_ptrace_wait.c to t_ptrace_kill_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

branches: 1.1.12;
Move LWP tests out of t_ptrace_wait.c to t_ptrace_lwp_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move misc tests out of t_ptrace_wait.c to t_ptrace_misc_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

branches: 1.4.4;
Remove trailing whitespace

Simplify previous

Emit SIGKILL without a loop and without checking for exact process status
once it is collected.

Do not fail when trying to kill a dying process

A dying process can disappear for a while. Rather than aborting, retry
sending SIGKILL to it.

Move register tests out of t_ptrace_wait.c to t_ptrace_register_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

t_ptrace_*: Turn debug output back off.

Even with the output reduction I recently implemented, this seems to
be filling up the releng testbed disk with excess output. Not
completely obviated by the recent diagnostic improvements -- e.g.,
the traceme_signalmasked_crash_fpe test still uses DPRINTF for some
output that isn't available any other way. But let's get the tests
running to completion again before improving that.

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

t_ptrace_*.c: Make debug output a little less verbose.

Maybe it will stop overflowing the releng testbed file systems this
way.

t_ptrace_wait, t_ptrace_sigchld: Enable debug output.

Otherwise failures lead only to nearly useless output like:

FAILED: /work/src/tests/lib/libc/sys/t_ptrace_signal_wait.h:852: info.psi_siginfo.si_signo != sig

What _was_ si_signo, you ask? Sorry, you have to guess or run the
test yourself!

branches: 1.3.12;
Skip traceme_raise* tests rather than failing upfront

This avoid reporting the tests as unexpected failure.

branches: 1.2.2;
Remove duplicated lines

Add new ptrace(2) test file

t_ptrace_sigchld - for SIGCHLD handler + ptrace(2).

Right now a single test is enabled (raise(SIGKILL)) and marked as failed
as it never finishes as the child is never collected before exiting the
parent uninterested about its child (SA_NOCLDWAIT).

tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

branches: 1.1.12;
Move siginfo tests out of t_ptrace_wait.c to t_ptrace_siginfo_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

t_ptrace_*: Fix softfloat conditionalization.

Evidently __SOFTFP__ is an armism.

Should improve the situation on sh3 and others.

PR misc/56820: Many FPE related tests fail on softfloat machines

t_ptrace_signal_wait.h: Add some more debug output for softfloat.

PR misc/56820: Many FPE related tests fail on softfloat machines

t_ptrace_*: More diagnostic output for assertion failures.

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

t_ptrace_wait*: Deal with quirk of softfloat trap SIGFPE delivery.

PR misc/56820: Many FPE related tests fail on softfloat machines

branches: 1.5.10;
Sprinkle a few more \n's and the end of some debug printfs.

branches: 1.4.2;
Work around QEMU bug #1668041 differently, by which floating-point
division by zero is not correctly trapped for i386 and amd64:

https://bugs.launchpad.net/qemu/+bug/1668041

Make trigger_fpe() cause integer division by zero for x86, and
floating-point one for other architectures. Also, assertions for
si_code in *_crash_fpe tests are commented out for now. They
should be cleaned up after the bug is fixed.

Now, *_crash_fpe tests are working also on powerpc.

Suggested by kamil.

Turn trigger_fpe() back to integer division by zero for a while
until QEMU bug #1668041 is fixed:

https://bugs.launchpad.net/qemu/+bug/1668041

by which floating-point division by zero is not trapped correctly
both on amd64 and i386.

Skip *_crash_fpe tests on powerpc, where integer division by zero
is never trapped.

Let trigger_fpe() cause floating-point divide by zero exception, instead of
integer one, which is not trapped for powerpc and aarch64.

Note that it is checked in lib/libc/gen/t_siginfo whether integer divide by
zero is interpreted as SIGFPE or not.

Now, all *_crash_fpe tests pass for powerpc, and nothing changes for amd64
at least.

Move signal tests out of t_ptrace_wait.c to t_ptrace_signal_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move PT_STEP tests out of t_ptrace_wait.c to t_ptrace_step_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

branches: 1.3.2;
Adjust pc/npc before syscall allowing EJUSTRETURN to return
to the next instruction. Only ERESTART should return to
the same instruction. Differences to sparc64 reduced.

Test t_ptrace_wait:syscallemu1 now passes on sparc.

Fixes PR kern/52166 "syscallemu does not work on sparc (32-bit)"

Ok: Martin Husemann

branches: 1.2.2;
Skip the lib/libc/sys/t_ptrace_waitid:syscall_signal_on_sce test case
with a reference to PR lib/55087. Marking it as an expected failure
would be more appropriate, but that doesn't work for test cases that
fail by timing out.

Move syscall tests out of t_ptrace_wait.c to t_ptrace_syscall_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

tests/lib/libc/sys/t_ptrace*: Sprinkle improved diagnostics.

This should make it easier to look at the test failure output and
figure out what's going on.

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

branches: 1.1.12;
Move threads tests out of t_ptrace_wait.c to t_ptrace_threads_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

tests/lib/libc/sys/t_ptrace*: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

branches: 1.1.12;
Move topology tests out of t_ptrace_wait.c to t_ptrace_topology_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Allow code to include either (or both) of <sys/elfdefinitions.h> and
<sys/exec_elf.h>, in any order.

This is a work-around to address PR lib/59564 (build breakage when
both of these files are #included).

With this changeset, whichever file gets included first 'wins'.
Prior changes in NetBSD and upstream at Elftoolchain have aligned the
'generic' ELF definitions in NetBSD's <sys/exec_elf.h> with those
in Elftoolchain's <sys/elfdefinitions.h>, so that either file can
be used interchangeably when building NetBSD.

This changeset also removes prior workarounds in the source tree
for this issue.

PR lib/59564

t_ptrace_*: Turn debug output back off.

Even with the output reduction I recently implemented, this seems to
be filling up the releng testbed disk with excess output. Not
completely obviated by the recent diagnostic improvements -- e.g.,
the traceme_signalmasked_crash_fpe test still uses DPRINTF for some
output that isn't available any other way. But let's get the tests
running to completion again before improving that.

PR misc/59381: t_ptrace_wait* and t_ptrace_sigchld tests are flaky

t_ptrace_*.c: Make debug output a little less verbose.

Maybe it will stop overflowing the releng testbed file systems this
way.

t_ptrace_wait, t_ptrace_sigchld: Enable debug output.

Otherwise failures lead only to nearly useless output like:

FAILED: /work/src/tests/lib/libc/sys/t_ptrace_signal_wait.h:852: info.psi_siginfo.si_signo != sig

What _was_ si_signo, you ask? Sorry, you have to guess or run the
test yourself!

branches: 1.192.2;
elftoolchain: Be consistent about which ELF header files we use.

1. For tools that use elftoolchain: always use elftoolchain's
elfdefinitions.h. Don't even think about looking at the host's
sys/exec_elf.h, which makes no sense and should never happen.

(ELF tools that don't use elftoolchain, like m68k-elf2coff,
continue to use nbincludes/sys/exec_elf.h. But no more nbincludes
hacks in elftoolchain.)

2. For kernel components (solaris, zfs, dtrace): always use
sys/exec_elf.h, even in Solaris components via sys/elf.h.
elfdefinitions.h is not wired up in the kernel build at all.

3. For most userland components that involve libelf: use
elfdefinitions.h via libelf header files (libelf.h, gelf.h).

libdtrace in particular requires _all_ R_* reloc type definitions,
but sys/exec_elf.h brings in only the _current machine's_ R_*
reloc type definitions. (While here: Use uintptr_t instead of
Elf_Addr for pointer-to-integer cast, since Elf_Addr is MD and
provided only by sys/exec_elf.h, not by elfdefinitions.h.)

And most userland components using libelf don't rely on any
properties of the current machine from sys/exec_elf.h, so they can
use libelf's elfdefinition.h.

Exceptions:

- dtrace drti.c relies on link.h -> link_elf.h -> sys/exec_elf.h,
but it also relies on sys/dtrace.h -> sys/elf.h ->
elfdefinitions.h like other userland components using sys/elf.h.

- kdump-ioctl.c uses sys/exec_elf.h directly and sys/dtrace.h ->
sys/elf.h -> elfdefinitions like other userland components using
sys/elf.h.

- t_ptrace_wait.c (via t_ptrace_core_wait.h) uses libelf to parse
core files, but relies on sys/exec_elf.h for struct
netbsd_elfcore_procinfo.

None of these exceptions needs all R_* reloc type definitions, so
as a workaround, we can just suppress libelf's elfdefinitions.h by
defining _SYS_ELFDEFINITIONS_H_ and use sys/exec_elf.h in these
exceptions.

And undo the whole BUILTIN_ELF_HEADERS mistake. This was:

- half bogus workarounds for missing build_install dependencies in
tools/Makefile, which are no longer missing now, and
- half futile attempt to use src/sys/sys/exec_elf.h via nbincludes in
tools involving libelf instead of libelf's elfdefinitions.h, which
collides.

Longer-term, we may wish to unify sys/exec_elf.h and libelf's
elfdefinitions.h, so we don't have to play these games.

But at least now the games are limited to three .c files (one of
which is generated by Makefile.ioctl-c), rather than haphazardly
applied tree-wide by monstrous kludges in widely used .h files with
broken hackarounds to get the tools build lurching to completion.

Move misc tests out of t_ptrace_wait.c to t_ptrace_misc_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move core tests out of t_ptrace_wait.c to t_ptrace_core_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move siginfo tests out of t_ptrace_wait.c to t_ptrace_siginfo_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move threads tests out of t_ptrace_wait.c to t_ptrace_threads_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move topology tests out of t_ptrace_wait.c to t_ptrace_topology_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move exec() tests out of t_ptrace_wait.c to t_ptrace_exec_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move LWP tests out of t_ptrace_wait.c to t_ptrace_lwp_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Remove the duplicate ATF_TP_ADD_TC() entries

Event mask tests are already defined in
ATF_TP_ADD_TCS_PTRACE_WAIT_EVENTMASK().

Move eventmask tests out of t_ptrace_wait.c to t_ptrace_eventmask_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Remove the duplicate ATF_TP_ADD_TC() entries

Byte transfer tests are already defined in
ATF_TP_ADD_TCS_PTRACE_WAIT_BYTETRANSFER().

Move signal tests out of t_ptrace_wait.c to t_ptrace_signal_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move fork/vfork/posix_spawn tests out of t_ptrace_wait.c to t_ptrace_fork_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move clone() tests out of t_ptrace_wait.c to t_ptrace_clone_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move byte transfer tests out of t_ptrace_wait.c to t_ptrace_bytetransfer_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move kill()-like tests out of t_ptrace_wait.c to t_ptrace_kill_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move PT_STEP tests out of t_ptrace_wait.c to t_ptrace_step_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move syscall tests out of t_ptrace_wait.c to t_ptrace_syscall_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Move register tests out of t_ptrace_wait.c to t_ptrace_register_wait.h

The same tests are now included with the preprocessor in t_ptrace_wait.c.

No functional change intended.

Reduce assumptions about LWP numbers

Update for new LWP behavior -- as of 9.99.59, the LWP ID of a single-LWP
process is the PID, not 1.

Switch from C11 specific static_assert() to __CTASSERT()

Add timeout to syscall_signal_on_sce that hangs from time to time

Try to fix the build. This is why all those inlines should really be in a
separate file as regular function. The code is too large and hard to manage
this way, and only increases in complexity as time goes by.

Move macros wrapping functions from msg.h into msg.h

This makes code reusing easier.

Add new signal trap entry tests in t_ptrace_wait*

New tests:
- syscall_signal_on_sce
- syscall_detach_on_sce

The first one fails as a signal emitted to the child is caught by the
debugger.

The second one passes.

Add new test in t_ptrace_wait*

New test: syscall_killed_on_sce

Test passes correctly.

Disable t_ptrace_wait* tests for MKSANITIZER/MKLIBCSANITIZER

A subset of tests is not compatible with the sanitizers.

Minor changes to t_ptrace_wait*

Add check for security.models.extensions.user_set_dbregs in the
concurrent event tests.

Use unconditionally atf_tc_skip(); as the test is racy.

Fix clang(1) build

Namespace a symbol that can be unused.

Enable combined breakpoint, watchpoint and signal tests

Extend concurrent events test to watchpoints

Extend concurrent events test to breakpoints

Add testing for concurrent breakpoint hits. The code is currently
x86-specific but since it reuses most of the generic concurrent event
test code, it's put in t_ptrace_wait.c with arch-specific hooks.

Include LWP events in concurrent event test

LWP events will be necessary when breakpoint/watchpoint support is added
to the test. However, they also benefit us by testing whether TRAP_LWP
events do not break signal reporting.

Test whether signal handler is called in concurrent bombarding

Extend concurrent signal tests to pass signal back to the thread

Generalize thread_concurrent_signals to a generic factory

Build thread_concurrent_signals on a generic parametrized
thread_concurrent_test() function. At first, this permits adjusting
the number of running signal threads but in the followup commits
support for different concurrent events (breakpoints, watchpoints)
and different signal handling will be added.

Rewrite the t_ptrace_wait resume1 test

Switch to pthread functions.
Rename to 'resume'.
Synchronize the threads with pthread barriers.
Avoid race in the test.

Test passes correctly.

Fixes PR bin/54893 by Andreas Gustafsson

Rename suspend2 to suspend_no_deadlock

Refactor signal9 and signal10 tests in t_ptrace_wait*

Remove signal9 and signal10. Reimplement the idea of these test in new
trace_thread_lwpexit_masked_sigtrap, trace_thread_lwpcreate_masked_sigtrap
and trace_thread_lwpcreate_and_exit_masked_sigtrap.

This removes 2 tests directly using low-level lwp APIs.

Disable debug code

Add new ATF ptrace(2) test in t_ptrace_wait*

threads_and_exec - verify that the expected LWP events are reported for a
multithreaded process that calls execve(2).

Test passes.

Add new ATF tests for unrelated tracer in t_ptrace_wait*

New tests:
- unrelated_tracer_posix_spawn_detach_spawner
- unrelated_tracer_fork_detach_forker
- unrelated_tracer_vfork_detach_vforker
- unrelated_tracer_vfork_detach_vforkerdone
- unrelated_tracer_posix_spawn_kill_spawner
- unrelated_tracer_fork_kill_forker
- unrelated_tracer_vfork_kill_vforker
- unrelated_tracer_vfork_kill_vforkerdone

All tests pass.

Add new fork/vfork/posix_spawn ATF tests in t_ptrace_wait*

Add unrelated tracer variation of tests: fork1-16, vfork1-16,
posix_spawn1-16.

All tests pass.

Fix the build: access_regs_set_unaligned_pc_* tests depend on availability
of gpregs

Add new ATF tests in t_ptrace_wait*

New tests:
- access_regs_set_unaligned_pc_0x1
- access_regs_set_unaligned_pc_0x3
- access_regs_set_unaligned_pc_0x7

The purpose of these tests is check whether unaligned PC can panic the
kernel. If there is a panic it is fine to return EINVAL on PT_SETREGS.
New tests check respectively pc |= 0x1, pc |= 0x3 and pc |= 0x7.

Inspired by https://marc.info/?l=openbsd-bugs&m=107558043319084&w=2

While there fix access_regs6 to really test PT_SETREGS.

Fix alignment when reading core notes

Both desc and note header needs to be aligned. Therefore, we need
to realign after skipping past desc as well.

While at it, fix the other alignment fix to use roundup() macro.

Read TLS base pointer with _lwp_getprivate()

This avoid usage of fast versions that are mostly internal to RTLD.

Define _RTLD_SOURCE in t_ptrace_wait.c

The __lwp_gettcb_fast() and __lwp_getprivate_fast() functions are
namespaced on a selection of ports.

Introduce PT_LWPSTATUS + PT_LWPNEXT, obsolete PT_LWPINFO

PT_LWPINFO is a legacy ptrace(2) operation that was originally intended
to retrieve the thread (LWP) information inside a traced process.

It has a number of flaws and is confused with PT_LWPINFO from FreeBSD.

PT_LWPSTATUS and PT_LWPNEXT address the problems (shortly by: rename,
removal of pl_event) and introduces new features: signal context
(pl_sigpend, pl_sigmask), LWP name (pl_name), LWP TLS base address
(pl_private). The private pointer was so far missing information for
a debugger.

PT_LWPSTATUS@nnn is now shipped with core(5) files and contain LWP specific
information, so far missed in the core(5) files.

PT_LWPSTATUS retrieves LWP information for the prompted thread.
PT_LWPNEXT retrieves LWP information for the next thread, borrowing the
semantics from NetBSD specific PT_LWPINFO.

PT_LWPINFO is namespaced with __LEGACY_PT_LWPINFO and still available for
the foreseeable future, without plans of removing it.

Add ATF tests for PT_LWPSTATUS + PT_LWPNEXT.

Keep ATF tests for PT_LWPINFO.

Switch GDB to new API.

Proposed on tech-kern@.

Log PID.LWP in t_ptrace_wait* tests for enabled debug

This is useful for multithreaded test scenarios.

Rework thread_concurrent_signals and trace_thread_lwpcreate_and_exit

Change the code to remove the LWP id assumptions that broke after
src/sys/kern/kern_lwp.c r. 1.206.

Original code by <mgorny>, tested and tweaked by myself.

Remove preprocessor switch TEST_LWP_ENABLED in t_ptrace_wait*

LWP tests are now enabled always and confirmed to be stable.

Remove preprocessor switch TEST_VFORK_ENABLED in t_ptrace_wait*

vfork(2) tests are now enabled always and confirmed to be stable.

Fix a race condition when handling concurrent LWP signals and add a test

Fix a race condition that caused PT_GET_SIGINFO to return incorrect
information when multiple signals were delivered concurrently
to different LWPs. Add a regression test that verifies that when 50
threads concurrently use pthread_kill() on themselves, the debugger
receives all signals with correct information.

The kernel uses separate signal queues for each LWP. However,
the signal context used to implement PT_GET_SIGINFO is stored in 'struct
proc' and therefore common to all LWPs in the process. Previously,
this member was filled in kpsignal2(), i.e. when the signal was sent.
This meant that if another LWP managed to send another signal
concurrently, the data was overwritten before the process was stopped.

As a result, PT_GET_SIGINFO did not report the correct LWP and signal
(it could even report a different signal than wait()). This can be
quite reliably reproduced with the number of 20 LWPs, however it can
also occur with 10.

This patch moves setting of signal context to issignal(), just before
the process is actually stopped. The data is taken from per-LWP
or per-process signal queue. The added test confirms that the debugger
correctly receives all signals, and PT_GET_SIGINFO reports both correct
LWP and signal number.

Reviewed by kamil.

Fix race in t_ptrace_wait* LWP tests

Increment the done variable under a mutex. This variable was updated
non-atomically and sometimes not reaching the expected treshold.

Enable TEST_LWP_ENABLED in t_ptrace_wait*

The LWP events (created, exited) are now reliable in my local tests.

PR kern/51420
PR kern/51995

Bump the number of threads in LWP t_ptrace_wait*() tests from to 100

This is a torture, but on purpose to catch LWP_CREATED and LWP_EXITED bugs.
The threads do nothing other than calling sched_yield() in a loop.

MAXLWP is now defined to 2048 on all ports.

This change effects the following tests:

- trace_thread_nolwpevents
- trace_thread_lwpexit
- trace_thread_lwpcreate
- trace_thread_lwpcreate_and_exit

Remove ATF t_ptrace_wait test suspend1

The tests verifies that a thread can be suspended by a debugger and later
resumed by a tracee.

This is an unwanted behavior and will be fixed soon.

Merge PT_GET_PROCESS_STATE checks into traceme_raise ATF ptrace tests

Assert that PT_GET_PROCESS_STATE for !child and !lwp events returns
non-error and zeroed struct ptrace_state.

These checks are not really special to traceme_raise, it's just an
opportunity to reuse them in an existing tests without writing a dedicated
one.

This behavior is needed to maintain in 3rd party software (GDB).