BMC/Intel-BMC/linux.git/tools/testing/selftests/arm64/mte, branch dev-5.15-intel

kselftest/arm64: mte: Fix misleading output when skipping tests

2021-08-20T10:11:05+00:00

When skipping the tests due to a lack of system support for MTE we currently print a message saying FAIL which makes it look like the test failed even though the test did actually report KSFT_SKIP, creating some confusion. Change the error message to say SKIP instead so things are clearer. Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20210819172902.56211-1-broonie@kernel.org Signed-off-by: Catalin Marinas

kselftest/arm64: Ignore check_gcr_el1_cswitch binary

2021-08-02T12:21:30+00:00

We added check_gcr_el1_cswitch but did not ignore the generated binary, add it. Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20210728173539.6231-1-broonie@kernel.org Signed-off-by: Catalin Marinas

kselftest/arm64: mte: Report filename on failing temp file creation

2021-03-24T15:43:20+00:00

The MTE selftests create temporary files in /dev/shm, for later mmap-ing them. When there is no tmpfs mounted on /dev/shm, or /dev/shm does not exist in the first place (on minimal filesystems), the error message is not giving good hints: # FAIL: Unable to open temporary file # FAIL: memory allocation not ok 17 Check initial tags with private mapping, ... Add a perror() call, that gives both the filename and the actual error reason, so that users get a chance of correcting that. Signed-off-by: Andre Przywara Reviewed-by: Mark Brown Link: https://lore.kernel.org/r/20210319165334.29213-12-andre.przywara@arm.com Signed-off-by: Catalin Marinas

kselftest/arm64: mte: Fix clang warning

2021-03-24T15:43:20+00:00

if (!prctl(...) == 0) is not only cumbersome to read, it also upsets clang and triggers a warning: ------------ mte_common_util.c:287:6: warning: logical not is only applied to the left hand side of this comparison [-Wlogical-not-parentheses] .... Fix that by just comparing against "not 0" instead. Signed-off-by: Andre Przywara Reviewed-by: Mark Brown Link: https://lore.kernel.org/r/20210319165334.29213-11-andre.przywara@arm.com Signed-off-by: Catalin Marinas

kselftest/arm64: mte: Makefile: Fix clang compilation

2021-03-24T15:43:20+00:00

When clang finds a header file on the command line, it wants to precompile that, which would end up in a separate output file. Specifying -o on that same command line collides with that effort, so the compiler complains: clang: error: cannot specify -o when generating multiple output files Since we are not really after a precompiled header, just drop the header file from the command line, by removing it from the list of source files in the Makefile. Signed-off-by: Andre Przywara Reviewed-by: Mark Brown Link: https://lore.kernel.org/r/20210319165334.29213-10-andre.przywara@arm.com Signed-off-by: Catalin Marinas

kselftest/arm64: mte: Output warning about failing compiler

2021-03-24T15:43:20+00:00

At the moment we check the compiler's ability to compile MTE enabled code, but guard all the Makefile rules by it. As a consequence a broken or not capable compiler just doesn't do anything, and make happily returns without any error message, but with no programs created. Since the MTE feature is only supported by recent aarch64 compilers (not all stable distro compilers support it), having an explicit message seems like a good idea. To not break building multiple targets, we let make proceed without errors. Signed-off-by: Andre Przywara Reviewed-by: Mark Brown Link: https://lore.kernel.org/r/20210319165334.29213-9-andre.przywara@arm.com Signed-off-by: Catalin Marinas

kselftest/arm64: mte: Use cross-compiler if specified

2021-03-24T15:43:20+00:00

At the moment we either need to provide CC explicitly, or use a native machine to get the ARM64 MTE selftest compiled. It seems useful to use the same (cross-)compiler as we use for the kernel, so copy the recipe we use in the pauth selftest. Signed-off-by: Andre Przywara Reviewed-by: Mark Brown Link: https://lore.kernel.org/r/20210319165334.29213-8-andre.przywara@arm.com Signed-off-by: Catalin Marinas

kselftest/arm64: mte: Fix MTE feature detection

2021-03-24T15:43:14+00:00

To check whether the CPU and kernel support the MTE features we want to test, we use an (emulated) CPU ID register read. However we only check against a very particular feature version (0b0010), even though the ARM ARM promises ID register features to be backwards compatible. While this could be fixed by using ">=" instead of "==", we should actually use the explicit HWCAP2_MTE hardware capability, exposed by the kernel via the ELF auxiliary vectors. That moves this responsibility to the kernel, and fixes running the tests on machines with FEAT_MTE3 capability. Signed-off-by: Andre Przywara Reviewed-by: Mark Brown Link: https://lore.kernel.org/r/20210319165334.29213-7-andre.przywara@arm.com Signed-off-by: Catalin Marinas

kselftest/arm64: mte: common: Fix write() warnings

2021-03-24T15:43:03+00:00

Out of the box Ubuntu's 20.04 compiler warns about missing return value checks for write() (sys)calls. Make GCC happy by checking whether we actually managed to write out our buffer. Signed-off-by: Andre Przywara Reviewed-by: Mark Brown Link: https://lore.kernel.org/r/20210319165334.29213-6-andre.przywara@arm.com Signed-off-by: Catalin Marinas

kselftest/arm64: mte: user_mem: Fix write() warning

2021-03-24T15:43:01+00:00

Out of the box Ubuntu's 20.04 compiler warns about missing return value checks for write() (sys)calls. Make GCC happy by checking whether we actually managed to write "val". Signed-off-by: Andre Przywara Reviewed-by: Mark Brown Link: https://lore.kernel.org/r/20210319165334.29213-5-andre.przywara@arm.com Signed-off-by: Catalin Marinas