linux-dev/tools/testing/selftests/pidfd, branch linus/master

selftests: fix an unused variable warning in pidfd selftest

2022-04-04T19:32:53Z

I fixed a few warnings like this in commit e2aa5e650b07 ("selftests: fixup build warnings in pidfd / clone3 tests"), but I missed this one by mistake. Since this variable is unused, remove it. Signed-off-by: Axel Rasmussen Reviewed-by: Christian Brauner Signed-off-by: Shuah Khan

selftests: fixup build warnings in pidfd / clone3 tests

2022-02-04T20:07:07Z

These are some trivial fixups, which were needed to build the tests with clang and -Werror. The following issues are fixed: - Remove various unused variables. - In child_poll_leader_exit_test, clang isn't smart enough to realize syscall(SYS_exit, 0) won't return, so it complains we never return from a non-void function. Add an extra exit(0) to appease it. - In test_pidfd_poll_leader_exit, ret may be branched on despite being uninitialized, if we have !use_waitpid. Initialize it to zero to get the right behavior in that case. Signed-off-by: Axel Rasmussen Acked-by: Christian Brauner Signed-off-by: Shuah Khan

pidfd: fix test failure due to stack overflow on some arches

2022-02-04T20:06:39Z

When running the pidfd_fdinfo_test on arm64, it fails for me. After some digging, the reason is that the child exits due to SIGBUS, because it overflows the 1024 byte stack we've reserved for it. To fix the issue, increase the stack size to 8192 bytes (this number is somewhat arbitrary, and was arrived at through experimentation -- I kept doubling until the failure no longer occurred). Also, let's make the issue easier to debug. wait_for_pid() returns an ambiguous value: it may return -1 in all of these cases: 1. waitpid() itself returned -1 2. waitpid() returned success, but we found !WIFEXITED(status). 3. The child process exited, but it did so with a -1 exit code. There's no way for the caller to tell the difference. So, at least log which occurred, so the test runner can debug things. While debugging this, I found that we had !WIFEXITED(), because the child exited due to a signal. This seems like a reasonably common case, so also print out whether or not we have WIFSIGNALED(), and the associated WTERMSIG() (if any). This lets us see the SIGBUS I'm fixing clearly when it occurs. Finally, I'm suspicious of allocating the child's stack on our stack. man clone(2) suggests that the correct way to do this is with mmap(), and in particular by setting MAP_STACK. So, switch to doing it that way instead. Signed-off-by: Axel Rasmussen Acked-by: Christian Brauner Signed-off-by: Shuah Khan

selftests: pidfd: drop needless linux/kcmp.h inclusion in pidfd_setns_test.c

2020-10-27T23:04:01Z

kcmp is not used in pidfd_setns_test.c, so do not include Signed-off-by: Tommi Rantala Acked-by: Christian Brauner Signed-off-by: Shuah Khan

selftests: pidfd: add CONFIG_CHECKPOINT_RESTORE=y to config

2020-10-27T23:02:03Z

kcmp syscall is used in pidfd_getfd_test.c, so add CONFIG_CHECKPOINT_RESTORE=y to config to ensure kcmp is available. Signed-off-by: Tommi Rantala Acked-by: Christian Brauner Signed-off-by: Shuah Khan

selftests: pidfd: skip test on kcmp() ENOSYS

2020-10-27T23:01:45Z

Skip test if kcmp() is not available, for example if kernel is compiled without CONFIG_CHECKPOINT_RESTORE=y. Signed-off-by: Tommi Rantala Acked-by: Christian Brauner Signed-off-by: Shuah Khan

selftests: pidfd: use ksft_test_result_skip() when skipping test

2020-10-27T23:01:24Z

There's planned tests != run tests in pidfd_test when some test is skipped: $ ./pidfd_test TAP version 13 1..8 [...] # pidfd_send_signal signal recycled pid test: Skipping test # Planned tests != run tests (8 != 7) # Totals: pass:7 fail:0 xfail:0 xpass:0 skip:0 error:0 Fix by using ksft_test_result_skip(): $ ./pidfd_test TAP version 13 1..8 [...] ok 8 # SKIP pidfd_send_signal signal recycled pid test: Unsharing pid namespace not permitted # Totals: pass:7 fail:0 xfail:0 xpass:0 skip:1 error:0 Signed-off-by: Tommi Rantala Acked-by: Christian Brauner Signed-off-by: Shuah Khan

selftests: pidfd: fix compilation errors due to wait.h

2020-10-27T22:51:36Z

Drop unneeded header inclusion to fix pidfd compilation errors seen in Fedora 32: In file included from pidfd_open_test.c:9: ../../../../usr/include/linux/wait.h:17:16: error: expected identifier before numeric constant 17 | #define P_ALL 0 | ^ Signed-off-by: Tommi Rantala Reviewed-by: Kees Cook Acked-by: Christian Brauner Signed-off-by: Shuah Khan

Merge tag 'threads-v5.10' of git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux

2020-10-14T21:39:20Z

Pull pidfd updates from Christian Brauner: "This introduces a new extension to the pidfd_open() syscall. Users can now raise the new PIDFD_NONBLOCK flag to support non-blocking pidfd file descriptors. This has been requested for uses in async process management libraries such as async-pidfd in Rust. Ever since the introduction of pidfds and more advanced async io various programming languages such as Rust have grown support for async event libraries. These libraries are created to help build epoll-based event loops around file descriptors. A common pattern is to automatically make all file descriptors they manage to O_NONBLOCK. For such libraries the EAGAIN error code is treated specially. When a function is called that returns EAGAIN the function isn't called again until the event loop indicates the the file descriptor is ready. Supporting EAGAIN when waiting on pidfds makes such libraries just work with little effort. This introduces a new flag PIDFD_NONBLOCK that is equivalent to O_NONBLOCK. This follows the same patterns we have for other (anon inode) file descriptors such as EFD_NONBLOCK, IN_NONBLOCK, SFD_NONBLOCK, TFD_NONBLOCK and the same for close-on-exec flags. Passing a non-blocking pidfd to waitid() currently has no effect, i.e. is not supported. There are users which would like to use waitid() on pidfds that are O_NONBLOCK and mix it with pidfds that are blocking and both pass them to waitid(). The expected behavior is to have waitid() return -EAGAIN for non-blocking pidfds and to block for blocking pidfds without needing to perform any additional checks for flags set on the pidfd before passing it to waitid(). Non-blocking pidfds will return EAGAIN from waitid() when no child process is ready yet. Returning -EAGAIN for non-blocking pidfds makes it easier for event loops that handle EAGAIN specially. It also makes the API more consistent and uniform. In essence, waitid() is treated like a read on a non-blocking pidfd or a recvmsg() on a non-blocking socket. With the addition of support for non-blocking pidfds we support the same functionality that sockets do. For sockets() recvmsg() supports MSG_DONTWAIT for pidfds waitid() supports WNOHANG. Both flags are per-call options. In contrast non-blocking pidfds and non-blocking sockets are a setting on an open file description affecting all threads in the calling process as well as other processes that hold file descriptors referring to the same open file description. Both behaviors, per call and per open file description, have genuine use-cases. The interaction with the WNOHANG flag is documented as follows: - If a non-blocking pidfd is passed and WNOHANG is not raised we simply raise the WNOHANG flag internally. When do_wait() returns indicating that there are eligible child processes but none have exited yet we set EAGAIN. If no child process exists we continue returning ECHILD. - If a non-blocking pidfd is passed and WNOHANG is raised waitid() will continue returning 0, i.e. it will not set EAGAIN. This ensure backwards compatibility with applications passing WNOHANG explicitly with pidfds" * tag 'threads-v5.10' of git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux: tests: remove O_NONBLOCK before waiting for WSTOPPED tests: add waitid() tests for non-blocking pidfds tests: port pidfd_wait to kselftest harness pidfd: support PIDFD_NONBLOCK in pidfd_open() exit: support non-blocking pidfds

tests: remove O_NONBLOCK before waiting for WSTOPPED

2020-10-09T09:56:51Z

Naresh reported that selftests: pidfd: pidfd_wait hangs on linux next kernel on x86_64, i386 and arm64 Juno-r2 These devices are using NFS mounted rootfs. I have tested pidfd testcases independently and all test PASS. The Hang or exit from test run noticed when run by run_kselftest.sh pidfd_wait.c:208:wait_nonblock:Expected sys_waitid(P_PIDFD, pidfd, &info, WSTOPPED, NULL) (-1) == 0 (0) wait_nonblock: Test terminated by assertion metadata: git branch: master git repo: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git git commit: e64997027d5f171148687e58b78c8b3c869a6158 git describe: next-20200922 make_kernelversion: 5.9.0-rc6 kernel-config: http://snapshots.linaro.org/openembedded/lkft/lkft/sumo/intel-core2-32/lkft/linux-next/865/config The reason for this is a simple race in the selftests, that I overlooked and which is more likely to hit when there's a lot of processes running on the system. Basically the child process hasn't SIGSTOPed itself yet but the parent is already calling waitid() on a O_NONBLOCK pidfd. Since it doesn't find a WSTOPPED process it returns -EAGAIN correctly. The fix for this is to move the line where we're removing the O_NONBLOCK property from the fd before the waitid() WSTOPPED call so we hang until the child becomes stopped. Fixes: cd89597bbe5a ("tests: add waitid() tests for non-blocking pidfds") Reported-by: Naresh Kamboju Tested-by: Naresh Kamboju Link: https://lkft.validation.linaro.org/scheduler/job/1813223 Signed-off-by: Christian Brauner