linux/tools/testing/selftests/pidfd, branch v5.9

Merge tag 'linux-kselftest-5.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest

2020-08-05T17:28:25Z

Pull kselftest updates form Shuah Khan: - TAP output reporting related fixes from Paolo Bonzini and Kees Cook. These fixes make it skip reporting consistent with TAP format. - Cleanup fixes to framework run_tests from Yauheni Kaliuta * tag 'linux-kselftest-5.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest: (23 commits) selftests/harness: Limit step counter reporting selftests/seccomp: Check ENOSYS under tracing selftests/seccomp: Refactor to use fixture variants selftests/harness: Clean up kern-doc for fixtures selftests: kmod: Add module address visibility test Replace HTTP links with HTTPS ones: KMOD KERNEL MODULE LOADER - USERMODE HELPER selftests: fix condition in run_tests selftests: do not use .ONESHELL selftests: pidfd: skip test if unshare fails with EPERM selftests: pidfd: do not use ksft_exit_skip after ksft_set_plan selftests/harness: Report skip reason selftests/harness: Display signed values correctly selftests/harness: Refactor XFAIL into SKIP selftests/harness: Switch to TAP output selftests: Add header documentation and helpers selftests/binderfs: Fix harness API usage selftests: Remove unneeded selftest API headers selftests/clone3: Reorder reporting output selftests: sync_test: do not use ksft_exit_skip after ksft_set_plan selftests: sigaltstack: do not use ksft_exit_skip after ksft_set_plan ...

tests: add CLONE_NEWTIME setns tests

2020-07-08T09:14:22Z

Now that pidfds support CLONE_NEWTIME as well enable testing them in the setns() testuite. Signed-off-by: Christian Brauner Cc: Serge Hallyn Cc: Michael Kerrisk Cc: Dmitry Safonov Cc: Andrei Vagin Link: https://lore.kernel.org/r/20200706154912.3248030-5-christian.brauner@ubuntu.com

selftests: pidfd: skip test if unshare fails with EPERM

2020-07-07T19:28:58Z

Similar to how ENOSYS causes a skip if pidfd_send_signal is not present, we can do the same for unshare if it fails with EPERM. This way, running the test without privileges causes four tests to skip but no early bail out. Acked-by: Christian Brauner Signed-off-by: Paolo Bonzini Signed-off-by: Shuah Khan

selftests: pidfd: do not use ksft_exit_skip after ksft_set_plan

2020-07-07T19:28:44Z

Calling ksft_exit_skip after ksft_set_plan results in executing fewer tests than planned. Use ksft_test_result_skip instead. The plan passed to ksft_set_plan was wrong, too, so fix it while at it. Acked-by: Christian Brauner Signed-off-by: Paolo Bonzini Signed-off-by: Shuah Khan

selftests: Remove unneeded selftest API headers

2020-07-06T21:59:07Z

Remove unused includes of the kselftest.h header. Acked-by: Christian Brauner Signed-off-by: Kees Cook Signed-off-by: Shuah Khan

tests: test for setns() EINVAL regression

2020-06-16T22:48:54Z

Verify that setns() reports EINVAL when an fd is passed that refers to an open file but the file is not a file descriptor useable to interact with namespaces. Cc: Jan Stancek Cc: Cyril Hrubis Link: https://lore.kernel.org/lkml/20200615085836.GR12456@shao2-debian Signed-off-by: Christian Brauner

selftests/pidfd: add pidfd setns tests

2020-05-13T09:41:22Z

This is basically a test-suite for setns() and as of now contains: - test that we can't pass garbage flags - test that we can't attach to the namespaces of task that has already exited - test that we can incrementally setns into all namespaces of a target task using a pidfd - test that we can setns atomically into all namespaces of a target task - test that we can't cross setns into a user namespace outside of our user namespace hierarchy - test that we can't setns into namespaces owned by user namespaces over which we are not privileged Signed-off-by: Christian Brauner Link: https://lore.kernel.org/r/20200505140432.181565-4-christian.brauner@ubuntu.com

Merge tag 'threads-v5.7' of git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux

2020-04-04T17:08:18Z

Pull thread updates from Christian Brauner: "The main change for this cycle was the extension for clone3() to support spawning processes directly into cgroups via CLONE_INTO_CGROUP (commit ef2c41cf38a7: "clone3: allow spawning processes into cgroups"). But since I had to touch kernel/cgroup/ quite a bit I had Tejun route that through his tree this time around to make it easier for him to handle other changes. So here is just the unexciting leftovers: a regression test for the ENOMEM regression we fixed in commit b26ebfe12f34 ("pid: Fix error return value in some cases") verifying that we report ENOMEM when trying to create a new process in a pid namespace whose init process/subreaper has already exited" * tag 'threads-v5.7' of git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux: selftests: add pid namespace ENOMEM regression test

selftests: add pid namespace ENOMEM regression test

2020-03-25T12:50:34Z

We recently regressed (cf. [1] and its corresponding fix in [2]) returning ENOMEM when trying to create a process in a pid namespace whose init process/child subreaper has already died. This has caused confusion at least once before that (cf. [3]). Let's add a simple regression test to catch this in the future. [1]: 49cb2fc42ce4 ("fork: extend clone3() to support setting a PID") [2]: b26ebfe12f34 ("pid: Fix error return value in some cases") [3]: 35f71bc0a09a ("fork: report pid reservation failure properly") Cc: Corey Minyard Cc: Oleg Nesterov Cc: Adrian Reber Cc: Dmitry Safonov <0x7f454c46@gmail.com> Cc: Andrei Vagin Signed-off-by: Christian Brauner

.gitignore: add SPDX License Identifier

2020-03-25T10:50:48Z

Add SPDX License Identifier to all .gitignore files. Signed-off-by: Masahiro Yamada Signed-off-by: Greg Kroah-Hartman