Changes in Makefile

d5859510 - kselftest/arm64: Include kernel mode NEON in fp-stress

Tue, 21 May 2024 18:18:26 +0000

kselftest/arm64: Include kernel mode NEON in fp-stressCurrently fp-stress only covers userspace use of floating point, it doesnot cover any kernel mode uses. Since currently kernel mode floatingpoint usage can't be preempted and there are explicit preemption points inthe existing implementations this isn't so important for fp-stress butwhen we readd preemption it will be good to try to exercise it.When the arm64 accelerated crypto operations are implemented we canrelatively straightforwardly trigger kernel mode floating point usage byusing the crypto userspace API to hash data, using the splice() supportin an effort to minimise copying. We use /proc/crypto to check whichaccelerated implementations are available, picking the first symmetrichash we find. We run the kernel mode test unconditionally, replacing thesecond copy of the FPSIMD testcase for systems with FPSIMD only. If wedon't think there are any suitable kernel mode implementations we fall backto running another copy of fpsimd-stress.There are a number issues with this approach, we don't actually verifythat we are using an accelerated (or even CPU) implementation of thealgorithm being tested and even with attempting to use splice() tominimise copying there are sizing limits on how much data gets splicedat once.Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20240521-arm64-fp-stress-kernel-v1-1-e38f107baad4@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

c745b15c - kselftest/arm64: Test that ptrace takes effect in the target process

Mon, 22 Jan 2024 21:05:03 +0000

kselftest/arm64: Test that ptrace takes effect in the target processWhile we have test coverage for the ptrace interface in our selfteststhe current programs have a number of gaps. The testing is done perregset so does not cover interactions and at no point do any of thetests actually run the traced processes meaning that there is novalidation that anything we read or write corresponds to register valuesthe process actually sees. Let's add a new program which attempts to coverthese gaps.Each test we do performs a single ptrace write. For each test we generatesome random initial register data in memory and then fork() and trace achild. The child will load the generated data into the registers thentrigger a breakpoint. The parent waits for the breakpoint then reads theentire child register state via ptrace, verifying that the values expectedwere actually loaded by the child. It then does the write being testedand resumes the child. Once resumed the child saves the register stateit sees to memory and executes another breakpoint. The parent usesprocess_vm_readv() to get these values from the child and verifies thatthe values were as expected before cleaning up the child.We generate configurations with combinations of vector lengths and SVCRvalues and then try every ptrace write which will implement thetransition we generated. In order to control execution time (especiallyin emulation) we only cover the minimum and maximum VL for each of SVEand SME, this will ensure we generate both increasing and decreasingchanges in vector length. In order to provide a baseline test we alsocheck the case where we resume the child without doing a ptrace write.In order to simplify the generation of the test count for kselftest wewill report but skip a substantial number of tests that can't actuallybe expressed via a single ptrace write, several times more than weactually run. This is noisy and will add some overhead but is very muchsimpler so is probably worth the tradeoff.Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20240122-arm64-test-ptrace-regs-v1-1-0897f822d73e@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

266679ff - kselftest/arm64: Convert za-fork to use kselftest.h

Thu, 06 Apr 2023 16:19:12 +0000

kselftest/arm64: Convert za-fork to use kselftest.hNow that kselftest.h can be used with nolibc convert the za-fork test touse it. We do still have to open code ksft_print_msg() but that's not theend of the world. Some of the advantage comes from using printf() which wecould have been using already.This does change the output when tests are skipped, bringing it in linewith the standard kselftest output by removing the test name - we movefrom ok 0 skippedto ok 1 # SKIP fork_testThe old output was not following KTAP format for skips, and thenumbering was not standard or consistent with the reported plan.Signed-off-by: Mark Brown Signed-off-by: Shuah Khan List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

7482c191 - selftests: arm64: Fix incorrect kernel headers search path

Fri, 27 Jan 2023 13:57:22 +0000

selftests: arm64: Fix incorrect kernel headers search pathUse $(KHDR_INCLUDES) as lookup path for kernel headers. This preventsbuilding against kernel headers from the build environment in scenarioswhere kernel headers are installed into a specific output directory(O=...).Signed-off-by: Mathieu Desnoyers Cc: Shuah Khan Cc: linux-kselftest@vger.kernel.orgCc: Ingo Molnar Cc: # 5.18+Acked-by: Shuah Khan Acked-by: Catalin Marinas Signed-off-by: Shuah Khan List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

4e1aa1a1 - kselftest/arm64: Add coverage of the ZT ptrace regset

Mon, 16 Jan 2023 16:04:55 +0000

kselftest/arm64: Add coverage of the ZT ptrace regsetAdd coverage of the ZT ptrace interface.Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20221208-arm64-sme2-v4-20-f2fa0aef982f@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

1c07425e - kselftest/arm64: Add a stress test program for ZT0

Mon, 16 Jan 2023 16:04:49 +0000

kselftest/arm64: Add a stress test program for ZT0Following the pattern for the other register sets add a stress test programfor ZT0 which continually loads and verifies patterns in the register inan effort to discover context switching problems.Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20221208-arm64-sme2-v4-14-f2fa0aef982f@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

fd5c2c6f - kselftest/arm64: kselftest harness for FP stress tests

Mon, 29 Aug 2022 15:44:52 +0000

kselftest/arm64: kselftest harness for FP stress testsCurrently the stress test programs for floating point context switching arerun by hand, there are extremely simplistic harnesses which run some copiesof each test individually but they are not integrated into kselftest andwith SVE and SME they only run with whatever vector length the process hasby default. This is hassle when running the tests and means that they'renot being run at all by CI systems picking up kselftest.In order to improve our coverage and provide a more convenient interfaceprovide a harness program which starts enough stress test programs up tocause context switching and runs them for a set period. If only FPSIMD isavailable in the system we start two copies of the FPSIMD stress test perCPU, otherwise we start one copy of the FPSIMD and then start the SVE,streaming SVE and ZA tests once per CPU for each available VL they haveto run on. We then run for a set period monitoring for any errorsreported by the test programs before cleanly terminating them.In order to provide additional coverage of signal handling and some extranoise in the scheduling we send a SIGUSR2 to the stress tests once asecond, the tests will count the number of signals they get.Since kselftest is generally expected to run quickly we by default only runfor ten seconds. This is enough to show if there is anything cripplinglywrong but not exactly a thorough soak test, for interactive and morefocused use a command line option -t N is provided which overrides thelength of time to run for (specified in seconds) and if 0 is specified thenthere is no timeout and the test must be manually terminated. The timeoutis counted in seconds with no output, this is done to account for thepotentially slow startup time for the test programs on virtual platformswhich tend to struggle during startup as they are both slow and tend tosupport a wide range of vector lengths.Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20220829154452.824870-5-broonie@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

aca43ad5 - selftests/arm64: Fix O= builds for the floating point tests

Wed, 27 Apr 2022 18:19:54 +0000

selftests/arm64: Fix O= builds for the floating point testsCurrently the arm64 floating point tests don't support out of tree buildsdue to two quirks of the kselftest build system. One is that when buildinga program from multiple files we shouldn't separately compile the mainprogram to an object file as that will result in the pattern rule notmatching when adjusted for the output directory. The other is that we alsoneed to include $(OUTPUT) in the names of the binaries when specifying thedependencies in order to ensure that they get picked up with O=.Rewrite the dependencies for the executables to fix these issues. Thekselftest build system will ensure OUTPUT is always defined.Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20220427181954.357975-5-broonie@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

399cf0a3 - selftests/arm64: Clean the fp helper libraries

Wed, 27 Apr 2022 18:19:53 +0000

selftests/arm64: Clean the fp helper librariesWe provide a couple of object files with helpers linked into several ofthe test programs, ensure they are cleaned.Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20220427181954.357975-4-broonie@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

3a23a42d - selftests/arm64: Define top_srcdir for the fp tests

Wed, 27 Apr 2022 18:19:52 +0000

selftests/arm64: Define top_srcdir for the fp testsSome of the rules in lib.mk use a top_srcdir variable to figure out wherethe top of the kselftest tree is, provide it.Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20220427181954.357975-3-broonie@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

a59f7a7f - selftests/arm64: Use TEST_GEN_PROGS_EXTENDED in the FP Makefile

Wed, 27 Apr 2022 18:19:51 +0000

selftests/arm64: Use TEST_GEN_PROGS_EXTENDED in the FP MakefileThe kselftest lib.mk provides a default all target which builds additionalprograms from TEST_GEN_PROGS_EXTENDED, use that rather than usingTEST_PROGS_EXTENDED which is for programs that don't need to be built likeshell scripts. Leave fpsimd-stress and sve-stress there since they arescripts.Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20220427181954.357975-2-broonie@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

212b0426 - selftests/arm64: Add a testcase for handling of ZA on clone()

Tue, 19 Apr 2022 11:22:47 +0000

selftests/arm64: Add a testcase for handling of ZA on clone()Add a small testcase that attempts to do a clone() with ZA enabled andverifies that it remains enabled with the same contents. We only checkone word in one horizontal vector of ZA since there's already other teststhat check for data corruption more broadly, we're just looking to makesure that ZA is still enabled and it looks like the data got copied.Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20220419112247.711548-40-broonie@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

86c8888f - kselftest/arm64: Add coverage for the ZA ptrace interface

Tue, 19 Apr 2022 11:22:45 +0000

kselftest/arm64: Add coverage for the ZA ptrace interfaceAdd some basic coverage for the ZA ptrace interface, including walkingthrough all the vector lengths supported in the system. Unlike SVEdoing syscalls does not discard the ZA state so when we set data in ZAwe run the child process briefly, having it add one to each byte in ZAin order to validate that both the vector size and data are being readand written as expected when the process runs.Signed-off-by: Mark Brown Reviewed-by: Shuah Khan Acked-by: Catalin Marinas Link: https://lore.kernel.org/r/20220419112247.711548-38-broonie@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

5aa45cc5 - kselftest/arm64: Add stress test for SME ZA context switching

Tue, 19 Apr 2022 11:22:42 +0000

kselftest/arm64: Add stress test for SME ZA context switchingAdd a stress test for context switching of the ZA register state based onthe similar tests Dave Martin wrote for FPSIMD and SVE registers. The testloops indefinitely writing a data pattern to ZA then reading it back andverifying that it's what was expected.Unlike the other tests we manually assemble the SME instructions since atpresent no released toolchain has SME support integrated.Signed-off-by: Mark Brown Reviewed-by: Shuah Khan Acked-by: Catalin Marinas Link: https://lore.kernel.org/r/20220419112247.711548-35-broonie@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

4126bde0 - kselftest/arm64: sme: Provide streaming mode SVE stress test

Tue, 19 Apr 2022 11:22:40 +0000

kselftest/arm64: sme: Provide streaming mode SVE stress testOne of the features of SME is the addition of streaming mode, in which wehave access to a set of streaming mode SVE registers at the SME vectorlength. Since these are accessed using the SVE instructions let's reusethe existing SVE stress test for testing with a compile time option forcontrolling the few small differences needed: - Enter streaming mode immediately on starting the program. - In streaming mode FFR is removed so skip reading and writing FFR.Signed-off-by: Mark Brown Reviewed-by: Shuah Khan Acked-by: Catalin Marinas Link: https://lore.kernel.org/r/20220419112247.711548-33-broonie@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

a0f2eb64 - kselftest/arm64: Extend vector configuration API tests to cover SME

Tue, 19 Apr 2022 11:22:39 +0000

kselftest/arm64: Extend vector configuration API tests to cover SMEProvide RDVL helpers for SME and extend the main vector configuration teststo cover SME.Signed-off-by: Mark Brown Reviewed-by: Shuah Khan Acked-by: Catalin Marinas Link: https://lore.kernel.org/r/20220419112247.711548-32-broonie@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

2c94ebed - kselftest/arm64: Add pidbench for floating point syscall cases

Thu, 02 Dec 2021 16:51:07 +0000

kselftest/arm64: Add pidbench for floating point syscall casesSince it's likely to be useful for performance work with SVE let's have apidbench that gives us some numbers for consideration. In order to ensurethat we test exactly the scenario we want this is written in assembly - ifsystem libraries use SVE this would stop us exercising the case where theprocess has never used SVE.We exercise three cases: - Never having used SVE. - Having used SVE once. - Using SVE after each syscall.by spinning running getpid() for a fixed number of iterations with thetime measured using CNTVCT_EL0 reported on the console. This is obviouslya totally unrealistic benchmark which will show the extremes of anyperformance variation but equally given the potential gotchas with use ofFP instructions by system libraries it's good to have some concrete codeshared to make it easier to compare notes on results.Testing over multiple SVE vector lengths will need to be done with vlsetcurrently, the test could be extended to iterate over all of them ifdesired.Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20211202165107.1075259-1-broonie@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

260ea4ba - selftests: arm64: Factor out utility functions for assembly FP tests

Tue, 19 Oct 2021 18:18:51 +0000

selftests: arm64: Factor out utility functions for assembly FP testsThe various floating point test programs written in assembly have a bunchof helper functions and macros which are cut'n'pasted between them. Factorthem out into a separate source file which is linked into all of them.We don't include memcmp() since it isn't as generic as it should be anddirectly branches to report an error in the programs.Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20211019181851.3341232-1-broonie@kernel.orgSigned-off-by: Will Deacon List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

eab281e3 - selftests: arm64: Remove extraneous register setting code

Mon, 13 Sep 2021 12:55:00 +0000

selftests: arm64: Remove extraneous register setting codeFor some reason the SVE ptrace test code starts off by setting values insome of the SVE vector registers in the parent process which it then neverinteracts with when verifying the ptrace interfaces. This is not especiallyrelevant to what's being tested and somewhat confusing when reading thecode so let's remove it.Signed-off-by: Mark Brown Link: https://lore.kernel.org/r/20210913125505.52619-4-broonie@kernel.orgSigned-off-by: Will Deacon List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile

95cf3f23 - kselftest/arm64: Add tests for SVE vector configuration

Tue, 03 Aug 2021 14:04:49 +0000

kselftest/arm64: Add tests for SVE vector configurationWe provide interfaces for configuring the SVE vector length seen byprocesses using prctl and also via /proc for configuring the defaultvalues. Provide tests that exercise all these interfaces and verify thatthey take effect as expected, though at present no test fully enumeratesall the possible vector lengths.A subset of this is already tested via sve-probe-vls but the /procinterfaces are not currently covered at all.In preparation for the forthcoming support for SME, the Scalable MatrixExtension, which has separately but similarly configured vector lengthswhich we expect to offer similar userspace interfaces for, all the actualfiles and prctls used are parameterised and we don't validate that thearchitectural minimum vector length is the minimum we see.Signed-off-by: Mark Brown Reviewed-by: Dave Martin Link: https://lore.kernel.org/r/20210803140450.46624-4-broonie@kernel.orgSigned-off-by: Catalin Marinas List of files: /linux-6.15/tools/testing/selftests/arm64/fp/Makefile