- 08 Sep, 2023 3 commits
-
-
Björn Töpel authored
When kselftest is built/installed with the 'gen_tar' target, rsync is used for the installation step to copy files. Extra care is needed for tests that have symlinks. Commit ae108c48 ("selftests: net: Fix cross-tree inclusion of scripts") added '-L' (transform symlink into referent file/dir) to rsync, to fix dangling links. However, that broke some tests where the symlink (being a symlink) is part of the test (e.g. exec:execveat). Use rsync's '--copy-unsafe-links' that does right thing. Fixes: ae108c48 ("selftests: net: Fix cross-tree inclusion of scripts") Signed-off-by:
Björn Töpel <bjorn@rivosinc.com> Reviewed-by:
Benjamin Poirier <bpoirier@nvidia.com> Signed-off-by:
Shuah Khan <skhan@linuxfoundation.org>
-
Ricardo B. Marliere authored
This patch fixes inconsistencies in the parsing rules of the levels 1 and 2 of the kselftest_deps.sh. It was added the levels 4 and 5 to account for a few edge cases that are present in some tests, also some minor identation styling have been fixed (s/ /\t/g). Signed-off-by:
Ricardo B. Marliere <rbmarliere@gmail.com> Signed-off-by:
-
Björn Töpel authored
Timeouts in kselftest are done using the "timeout" command with the "--foreground" option. Without the "foreground" option, it is not possible for a user to cancel the runner using SIGINT, because the signal is not propagated to timeout which is running in a different process group. The "forground" options places the timeout in the same process group as its parent, but only sends the SIGTERM (on timeout) signal to the forked process. Unfortunately, this does not play nice with all kselftests, e.g. "net:fcnal-test.sh", where the child processes will linger because timeout does not send SIGTERM to the group. Some users have noted these hangs [1]. Fix this by nesting the timeout with an additional timeout without the foreground option. Link: https://lore.kernel.org/all/7650b2eb-0aee-a2b0-2e64-c9bc63210f67@alu.unizg.hr/ # [1] Fixes: 651e0d88 ("kselftest/runner: allow to properly deliver signals to tests") Signed-off-by:
-
- 05 Sep, 2023 1 commit
-
-
Zheng Yejian authored
Function instance_set() expects to enable event 'sched_switch', so we should set 1 to its 'enable' file. Testcase passed after this patch: # ./ftracetest test.d/instances/instance-event.tc === Ftrace unit tests === [1] Test creation and deletion of trace instances while setting an event [PASS] # of passed: 1 # of failed: 0 # of unresolved: 0 # of untested: 0 # of unsupported: 0 # of xfailed: 0 # of undefined(test bug): 0 Signed-off-by:
Zheng Yejian <zhengyejian1@huawei.com> Acked-by:
Masami Hiramatsu (Google) <mhiramat@kernel.org> Acked-by:
Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by:
-
- 16 Aug, 2023 6 commits
-
-
Andre Przywara authored
At the moment the cachestat syscall number is hard coded into the test source code. Remove that and replace it with the proper __NR_cachestat macro. That ensures compatibility should other architectures pick a different number. Signed-off-by:
Andre Przywara <andre.przywara@arm.com> Reviewed-by:
Nhat Pham <nphamcs@gmail.com> Signed-off-by:
-
Andre Przywara authored
Libraries should be listed last on the compiler's command line, so that the linker can look for and find still unresolved symbols. The librt library, required for the shm_* functions, was announced using CFLAGS, which puts the library *before* the source files, and fails compilation on my system: ====================== gcc -isystem /src/linux-selftests/usr/include -Wall -lrt test_cachestat.c -o /src/linux-selftests/kselftest/cachestat/test_cachestat /usr/bin/ld: /tmp/cceQWO3u.o: in function `test_cachestat_shmem': test_cachestat.c:(.text+0x890): undefined reference to `shm_open' /usr/bin/ld: test_cachestat.c:(.text+0x99c): undefined reference to `shm_unlink' collect2: error: ld returned 1 exit status make[4]: *** [../lib.mk:181: /src/linux-selftests/kselftest/cachestat/test_cachestat] Error 1 ====================== Announce the library using the LDLIBS variable, which ensures the proper ordering on the command line. Signed-off-by:
-
Nysal Jan K.A authored
Observed occassional failures in the futex_wait_timeout test: ok 1 futex_wait relative succeeds ok 2 futex_wait_bitset realtime succeeds ok 3 futex_wait_bitset monotonic succeeds ok 4 futex_wait_requeue_pi realtime succeeds ok 5 futex_wait_requeue_pi monotonic succeeds not ok 6 futex_lock_pi realtime returned 0 ...... The test expects the child thread to complete some steps before the parent thread gets to run. There is an implicit expectation of the order of invocation of futex_lock_pi between the child thread and the parent thread. Make this order explicit. If the order is not met, the futex_lock_pi call in the parent thread succeeds and will not timeout. Fixes: f4addd54 ("selftests: futex: Expand timeout test") Signed-off-by:
Nysal Jan K.A <nysal@linux.ibm.com> Signed-off-by:
-
Mark Brown authored
We have some dmabuf-heaps and perf_events tests but they are not hooked up to the kselftest build infrastructure which is a bit of an obstacle to running them in systems with generic infrastructure for selftests. Add them to the top level kselftest Makefile so they get built as standard. Signed-off-by:
Mark Brown <broonie@kernel.org> Signed-off-by:
-
Mark Brown authored
The user_events selftests were removed from the standard set of selftests due to the uapi header it relies on having been temporarily removed. That header is now reinstated so we can reenable the tests. Signed-off-by:
-
Hui Min Mina Chou authored
In busybox, the mktemp requires that the generated filename be suffixed with at least six consecutive 'X' characters. Otherwise, it will return an "Invalid argument" error. Signed-off-by:
Hui Min Mina Chou <minachou@andestech.com> Reviewed-by:
Javier Martinez Canillas <javierm@redhat.com> Signed-off-by:
-
- 08 Aug, 2023 4 commits
-
-
Mathieu Desnoyers authored
Use rseq_unqual_scalar_typeof() rather than typeof() in macros to remove the volatile qualifier (if there is one in the input argument), thus generating better assembly code in those scenarios. Also add extra brackets around the "p" parameter in RSEQ_READ_ONCE(), RSEQ_WRITE_ONCE(), and rseq_unqual_scalar_typeof() across architectures to preserve expectations of operator priority. Here is an example that shows how operator priority may be an issue with missing parentheses: #define m(p) \ do { \ __typeof__(*p) v = 0; \ } while (0) void fct(unsigned long long *p1) { m(p1 + 1); /* works */ m(1 + p1); /* broken */ } Signed-off-by:Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: Peter Zijlstra <peterz@infradead.org> Signed-off-by:
-
Mathieu Desnoyers authored
The arm64 load-acquire/store-release macros from the Linux kernel rseq selftests are buggy. Remplace them by a working implementation. Signed-off-by:
-
Mathieu Desnoyers authored
Allow defining variables and perform cast with a typeof which removes the volatile and const qualifiers. This prevents declaring a stack variable with a volatile qualifier within a macro, which would generate sub-optimal assembler. This is imported from the "librseq" project. Signed-off-by:
-
Mathieu Desnoyers authored
Ensure that the basic percpu ops tests are effectively built against mm_cid. Signed-off-by:
-
- 25 Jul, 2023 21 commits
-
-
Shuah Khan authored
Add newly addded set-process-name test to .gitignore Signed-off-by: -
Shuah Khan authored
Remove clean target in Makefile to fix the following warning and use the one in common lib.mk Makefile:14: warning: overriding recipe for target 'clean' ../lib.mk:160: warning: ignoring old recipe for target 'clean' Signed-off-by: -
Ilpo Järvinen authored
cat_val() is only used during CAT test but it checks for test type. Remove test type checks and the unused else branch from cat_val(). Signed-off-by:
Ilpo Järvinen <ilpo.jarvinen@linux.intel.com> Reviewed-by:
Reinette Chatre <reinette.chatre@intel.com> Tested-by:
Babu Moger <babu.moger@amd.com> Tested-by:
Shaopeng Tan (Fujitsu) <tan.shaopeng@fujitsu.com> Signed-off-by:
-
Ilpo Järvinen authored
Results include warm-up test which is discarded before passing the sum to show_cache_info(). show_cache_info() handles this by subtracting one from the number of tests in divisor. It is a trappy construct to have sum and number of tests parameters to disagree like this. A more logical place for subtracting the skipped tests is where the sum is calculated so move it there. Pass the correct number of tests to show_cache_info() so it can be used directly as the divisor for calculating the average. Signed-off-by:
-
Ilpo Järvinen authored
CAT and CMT tests have count_of_bits, long_mask, cbm_mask, and cache_size global variables that can be moved into the sole using function. Make the global variables local variables of the relevant function to scope them better. While at it, move cache_size initialization into the declaration line. Signed-off-by:
-
Ilpo Järvinen authored
struct resctrl_val_param has ->setup() function that accepts variable argument list. All test cases use only 1 argument as input and it's the struct resctrl_val_param pointer. Instead of variable argument list, directly pass struct resctrl_val_param pointer as the only parameter to ->setup(). Co-developed-by:
Fenghua Yu <fenghua.yu@intel.com> Signed-off-by:
-
Ilpo Järvinen authored
Test name is passed to fill_buf functions so that they can loop around buffer only once. This is required for CAT test case. To loop around buffer only once, caller doesn't need to let fill_buf know which test case it is. Instead, pass a boolean argument 'once' which makes fill_buf more generic. As run_benchmark() no longer needs to pass the test name to run_fill_buf(), a few test running functions can be simplified to not write the test name into the default benchmark_cmd. The has_ben argument can also be removed now from those test running functions. Co-developed-by:
-
Ilpo Järvinen authored
fill_buf's arguments can be improved in multiple ways: - Multiple functions in fill_buf have start_ptr as one of their argument which is a bit long and the extra "start" is pretty obvious when it comes to pointers. - Some of the functions take end_ptr and others size_t to indicate the end of the buffer. - Some arguments meaning buffer size are called just 's' - mem_flush() takes void * but immediately converts it to char * Cleanup the parameters to make things simpler and more consistent: - Rename start_ptr to simply buf as it's shorter. - Replace end_ptr and s parameters with buf_size and only calculate end_ptr in the functions that truly use it. - Make mem_flush() parameters to follow the same convention as the other functions in fill_buf. - convert mem_flush() char * to unsigned char *. While at it, fix also a typo in a comment. Signed-off-by: -
Ilpo Järvinen authored
fill_buf stores buffer pointer into global variable startptr that is only used in fill_cache(). Remove startptr as global variable, the local variable in fill_cache() is enough to keep the pointer. Signed-off-by:
-
Ilpo Järvinen authored
run_fill_buf()'s malloc_and_init_memory parameter is always 1. There's also duplicated memory init code for malloc_and_init_memory == 0 case in fill_buf() which is unused. Remove the malloc_and_init_memory parameter and the duplicated mem init code. While at it, fix also a typo in run_fill_buf() prototype's argument. Co-developed-by:
-
Ilpo Järvinen authored
When no benchmark_cmd is given, benchmark_cmd[1] is set to span in main(). There's no need to do it again in run_mba_test(). Remove the duplicated preparation for span argument into benchmark_cmd[1] from run_mba_test(). After this, the has_ben and span arguments to run_mba_test() can be removed. Co-developed-by:
-
Ilpo Järvinen authored
MBA and MBM tests to use megabytes to represent span. CMT test uses bytes. The difference requires run_benchmark() to size the buffer differently based on the test name, which in turn requires passing the test name into run_benchmark(). Convert MBA and MBM tests to use span internally in bytes like CMT test to remove the internal inconsistency between the tests. Remove the test dependent buffer sizing from run_benchmark(). This change eliminates one of the reasons why the test name has to be passed around but there are still other users too so the test name passing cannot yet be removed. Co-developed-by:
-
Ilpo Järvinen authored
Span is defined either as unsigned long or int. Consistently use size_t everywhere for span as it refers to size of the memory block. Co-developed-by:
-
Ilpo Järvinen authored
Resctrl FS mount/umount are now cleanly paired leaving .mum_resctrlfs in the struct resctrl_val_param unused. Remove .mum_resctrlfs from struct resctrl_val_param. Signed-off-by:
-
Ilpo Järvinen authored
Mount/umount of the resctrl FS is now paired nicely per test. Rename remount_resctrl(bool mum_resctrlfs) to mount_resctrl(). Make it unconditionally try to mount the resctrl FS and return error if resctrl FS was mounted already. While at it, group the mount/umount prototypes in the header. Signed-off-by:
-
Ilpo Järvinen authored
A few places currently lack umounting resctrl FS on error paths: - cmt_resctrl_val() has multiple error paths with direct return. - cat_perf_miss_val() has multiple error paths with direct return. In addition, validate_resctrl_feature_request() is called by run_mbm_test() and run_mba_test(). Neither MBA nor MBM test tries to umount resctrl FS. Each and every test does require resctrl FS to be present already for feature check. Thus, it makes sense to just mount it on higher level in resctrl_tests.c and properly pair it with umount. Move resctrl FS (re)mount/unmount into each test function in resctrl_tests.c. Make feature validation to simply check that resctrl FS is mounted. As there's the final umount in main() before this change, the selftest should not leave resctrl FS behind after the tests even if one of the forementioned paths is taken (thus, no fixes tag). Signed-off-by:
-
Ilpo Järvinen authored
Resctrl FS mount/remount/umount code is hard to track. Better approach is to use mount/umount pair for each test but that assumes resctrl FS is not mounted beforehand. Change umount_resctrlfs() so that it can unmount resctrl FS from any path, and enable further simplifications into mount/remount/umount logic by unmounting resctrl FS at the start if a pre-existing mountpoint is found. Suggested-by:
-
Ilpo Järvinen authored
Perf event fd (fd_lm) is not closed when run_fill_buf() returns error. Close fd_lm only in cat_val() to make it easier to track it is always closed. Fixes: 790bf585 ("selftests/resctrl: Add Cache Allocation Technology (CAT) selftest") Signed-off-by:
-
Ilpo Järvinen authored
A child calls PARENT_EXIT() when it fails to run a benchmark to kill the parent process. PARENT_EXIT() lacks unmount for the resctrl FS and the parent won't be there to unmount it either after it gets killed. Add the resctrl FS unmount also to PARENT_EXIT(). Fixes: 591a6e85 ("selftests/resctrl: Add basic resctrl file system operations and data") Signed-off-by:
-
Ilpo Järvinen authored
The error path in fill_cache() does return before the allocated buffer is freed leaking the buffer. The leak was introduced when fill_cache_read() started to return errors in commit c7b607fa ("selftests/resctrl: Fix null pointer dereference on open failed"), before that both fill functions always returned 0. Move free() earlier to prevent the mem leak. Fixes: c7b607fa ("selftests/resctrl: Fix null pointer dereference on open failed") Signed-off-by:
-
Ilpo Järvinen authored
Makefile only lists *.c as build dependencies for the resctrl_tests executable which excludes resctrl.h. Add *.h to wildcard() to include resctrl.h. Fixes: 591a6e85 ("selftests/resctrl: Add basic resctrl file system operations and data") Signed-off-by:
-
- 24 Jul, 2023 1 commit
-
-
Osama Muhammad authored
This patch will add the new test, which covers the prctl call PR_SET_NAME command. The test tries to give a name using the PR_SET_NAME call and then confirm it that it changed correctly by using PR_GET_NAME. It also tries to rename it with empty name.In the test PR_GET_NAME is tested by passing null pointer to it and check its behaviour. Signed-off-by:
Osama Muhammad <osmtendev@gmail.com> Reviewed-by:
Muhammad Usama Anjum <usama.anjum@collabora.com> Tested-by:
-
- 23 Jul, 2023 4 commits
-
-
Linus Torvalds authored
-
git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-traceLinus Torvalds authored
Pull tracing fixes from Steven Rostedt: - Swapping the ring buffer for snapshotting (for things like irqsoff) can crash if the ring buffer is being resized. Disable swapping when this happens. The missed swap will be reported to the tracer - Report error if the histogram fails to be created due to an error in adding a histogram variable, in event_hist_trigger_parse() - Remove unused declaration of tracing_map_set_field_descr() * tag 'trace-v6.5-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace: tracing/histograms: Return an error if we fail to add histogram to hist_vars list ring-buffer: Do not swap cpu_buffer during resize process tracing: Remove unused extern declaration tracing_map_set_field_descr()
-
Linus Torvalds authored
Merge tag 'kbuild-fixes-v6.5' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild Pull Kbuild fixes from Masahiro Yamada: - Fix stale help text in gconfig - Support *.S files in compile_commands.json - Flatten KBUILD_CFLAGS - Fix external module builds with Rust so that temporary files are created in the modules directories instead of the kernel tree * tag 'kbuild-fixes-v6.5' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild: kbuild: rust: avoid creating temporary files kbuild: flatten KBUILD_CFLAGS gen_compile_commands: add assembly files to compilation database kconfig: gconfig: correct program name in help text kconfig: gconfig: drop the Show Debug Info help text
-
Miguel Ojeda authored
`rustc` outputs by default the temporary files (i.e. the ones saved by `-Csave-temps`, such as `*.rcgu*` files) in the current working directory when `-o` and `--out-dir` are not given (even if `--emit=x=path` is given, i.e. it does not use those for temporaries). Since out-of-tree modules are compiled from the `linux` tree, `rustc` then tries to create them there, which may not be accessible. Thus pass `--out-dir` explicitly, even if it is just for the temporary files. Similarly, do so for Rust host programs too. Reported-by:
Raphael Nestler <raphael.nestler@gmail.com> Closes: https://github.com/Rust-for-Linux/linux/issues/1015Reported-by:
Andrea Righi <andrea.righi@canonical.com> Tested-by: Raphael Nestler <raphael.nestler@gmail.com> # non-hostprogs Tested-by: Andrea Righi <andrea.righi@canonical.com> # non-hostprogs Fixes: 295d8398 ("kbuild: specify output names separately for each emission type from rustc") Cc: stable@vger.kernel.org Signed-off-by:
Miguel Ojeda <ojeda@kernel.org> Tested-by:
Martin Rodriguez Reboredo <yakoyoku@gmail.com> Signed-off-by:
Masahiro Yamada <masahiroy@kernel.org>
-
