#
358b763e |
|
09-Aug-2023 |
Mark Brown <broonie@kernel.org> |
kselftest/arm64: Size sycall-abi buffers for the actual maximum VL Our ABI opts to provide future proofing by defining a much larger SVE_VQ_MAX than the architecture actually supports. Since we use this define to control the size of our vector data buffers this results in a lot of overhead when we initialise which can be a very noticable problem in emulation, we fill buffers that are orders of magnitude larger than we will ever actually use even with virtual platforms that provide the full range of architecturally supported vector lengths. Define and use the actual architecture maximum to mitigate this. Signed-off-by: Mark Brown <broonie@kernel.org> Link: https://lore.kernel.org/r/20230810-arm64-syscall-abi-perf-v1-1-6a0d7656359c@kernel.org Signed-off-by: Will Deacon <will@kernel.org>
|
#
daac8353 |
|
18-Dec-2022 |
Zenghui Yu <yuzenghui@huawei.com> |
kselftest/arm64: Correct buffer size for SME ZA storage It looks like a copy-paste error to describe the ZA buffer size using (the number of P registers * the maximum size of a Z register). This doesn't have practical impact though as we're always allocating enough space even for the architectural maximum ZA storage, with SVL equals to 2048 bits. Switch to use ZA_SIG_REGS_SIZE(SVE_VQ_MAX). setup_za() will need to initialize two 64MB arraies with this change and can be optimized later (if someone complain). Signed-off-by: Zenghui Yu <yuzenghui@huawei.com> Reviewed-by: Mark Brown <broonie@kernel.org> Link: https://lore.kernel.org/r/20221218092942.1940-2-yuzenghui@huawei.com Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
|
#
00598857 |
|
18-Dec-2022 |
Zenghui Yu <yuzenghui@huawei.com> |
kselftest/arm64: Remove the local NUM_VL definition It was introduced in commit b77e995e3b96 ("kselftest/arm64: Add a test program to exercise the syscall ABI") but never actually used. Remove it. Signed-off-by: Zenghui Yu <yuzenghui@huawei.com> Reviewed-by: Mark Brown <broonie@kernel.org> Link: https://lore.kernel.org/r/20221218092942.1940-1-yuzenghui@huawei.com Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
|
#
49886aa9 |
|
16-Jan-2023 |
Mark Brown <broonie@kernel.org> |
kselftest/arm64: Add SME2 coverage to syscall-abi Verify that ZT0 is preserved over syscalls when it is present and PSTATE.ZA is set. Signed-off-by: Mark Brown <broonie@kernel.org> Link: https://lore.kernel.org/r/20221208-arm64-sme2-v4-19-f2fa0aef982f@kernel.org Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
|
#
10f326fb |
|
27-Dec-2022 |
Mark Brown <broonie@kernel.org> |
kselftest/arm64: Only enumerate power of two VLs in syscall-abi As documented in issue C215 in the known issues list for DDI0487I.a [1] Arm will be making a retroactive change to SVE to remove the possibility of selecting non power of two vector lengths. This has no impact on existing physical implementations but most virtual implementations have implemented the full range of permissible vector lengths. Since virtual implementations are noticeably slow in general and the larger vector lengths amplify the issue there's a useful improvement in runtime from only covering the vector lengths that will exist in practical systems, adjust our enumeration accordingly. We have other tests that aim to cover the enumeration interfaces. For symmetry we apply the same change to the eumeration for SME vector lengths, though the power of two restriction was already present for SME so there is no impact on the set of vector lengths tested. [1] https://developer.arm.com/documentation/102105/ia-00/ Signed-off-by: Mark Brown <broonie@kernel.org> Link: https://lore.kernel.org/r/20221223-arm64-syscall-abi-sme-only-v1-4-4fabfbd62087@kernel.org Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
|
#
024e4a15 |
|
27-Dec-2022 |
Mark Brown <broonie@kernel.org> |
kselftest/arm64: Verify SME only ABI in syscall-abi Currently syscall-abi only covers SME in the case where the system supports SVE however it is architecturally valid to support SME without SVE. Update the program to cover this case, this requires adjustments in the code to check for SVCR.SM being set when deciding if we're handling the FPSIMD or SVE registers and the addition of new test cases for the SME only case. Note that in the SME only case we should not save the SVE registers after a syscall since even if we were in streaming mode and therefore set them the syscall should have exited streaming mode, we check that we have done so by looking at SVCR. Signed-off-by: Mark Brown <broonie@kernel.org> Link: https://lore.kernel.org/r/20221223-arm64-syscall-abi-sme-only-v1-3-4fabfbd62087@kernel.org Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
|
#
fae491e5 |
|
27-Dec-2022 |
Mark Brown <broonie@kernel.org> |
kselftest/arm64: Only enumerate VLs once in syscall-abi Currently syscall-abi not only enumerates the SVE VLs twice while working out how many tests are planned, it also repeats the enumeration process while doing the actual tests. Record the VLs when we enumerate and use that list when we are performing the tests, removing some duplicated logic. Signed-off-by: Mark Brown <broonie@kernel.org> Link: https://lore.kernel.org/r/20221223-arm64-syscall-abi-sme-only-v1-2-4fabfbd62087@kernel.org Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
|
#
97ec597b |
|
27-Dec-2022 |
Mark Brown <broonie@kernel.org> |
kselftest/arm64: Fix syscall-abi for systems without 128 bit SME SME does not mandate any specific VL so we may not have 128 bit SME but the algorithm used for enumerating VLs assumes that we will. Add the required check to ensure that the algorithm terminates. Fixes: 43e3f85523e4 ("kselftest/arm64: Add SME support to syscall ABI test") Signed-off-by: Mark Brown <broonie@kernel.org> Link: https://lore.kernel.org/r/20221223-arm64-syscall-abi-sme-only-v1-1-4fabfbd62087@kernel.org Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
|
#
af3ce550 |
|
29-Aug-2022 |
Mark Brown <broonie@kernel.org> |
kselftest/arm64: Enforce actual ABI for SVE syscalls Currently syscall-abi permits the bits in Z registers not shared with the V registers as well as all of the predicate registers to be preserved on syscall but the actual implementation has always cleared them and our documentation has now been updated to make that the documented ABI so update the syscall-abi test to match. Signed-off-by: Mark Brown <broonie@kernel.org> Acked-by: Catalin Marinas <catalin.marinas@arm.com> Link: https://lore.kernel.org/r/20220829162502.886816-4-broonie@kernel.org Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
|
#
bb7852c1 |
|
29-Aug-2022 |
Mark Brown <broonie@kernel.org> |
kselftest/arm64: Correct buffer allocation for SVE Z registers The buffer used for verifying SVE Z registers allocated enough space for 16 maximally sized registers rather than 32 due to using the macro for the number of P registers. In practice this didn't matter since for historical reasons the maximum VQ defined in the ABI is greater the architectural maximum so we will always allocate more space than is needed even with emulated platforms implementing the architectural maximum. Still, we should use the right define. Signed-off-by: Mark Brown <broonie@kernel.org> Acked-by: Catalin Marinas <catalin.marinas@arm.com> Link: https://lore.kernel.org/r/20220829162502.886816-2-broonie@kernel.org Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
|
#
43e3f855 |
|
18-Apr-2022 |
Mark Brown <broonie@kernel.org> |
kselftest/arm64: Add SME support to syscall ABI test For every possible combination of SVE and SME vector length verify that for each possible value of SVCR after a syscall we leave streaming mode and ZA is preserved. We don't need to take account of any streaming/non streaming SVE vector length changes in the assembler code since the store instructions will handle the vector length for us. We log if the system supports FA64 and only try to set FFR in streaming mode if it does. Signed-off-by: Mark Brown <broonie@kernel.org> Reviewed-by: Shuah Khan <skhan@linuxfoundation.org> Acked-by: Catalin Marinas <catalin.marinas@arm.com> Link: https://lore.kernel.org/r/20220419112247.711548-39-broonie@kernel.org Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
|
#
3673d4b9 |
|
24-Jan-2022 |
Mark Brown <broonie@kernel.org> |
kselftest/arm64: Remove local ARRAY_SIZE() definitions An ARRAY_SIZE() has been added to kselftest.h so remove the local versions in some of the arm64 selftests. Signed-off-by: Mark Brown <broonie@kernel.org> Reviewed-by: Shuah Khan <skhan@linuxfoundation.org> Acked-by: Catalin Marinas <catalin.marinas@arm.com> Link: https://lore.kernel.org/r/20220124171748.2195875-1-broonie@kernel.org Signed-off-by: Will Deacon <will@kernel.org>
|
#
b77e995e |
|
10-Dec-2021 |
Mark Brown <broonie@kernel.org> |
kselftest/arm64: Add a test program to exercise the syscall ABI Currently we don't have any coverage of the syscall ABI so let's add a very dumb test program which sets up register patterns, does a sysscall and then checks that the register state after the syscall matches what we expect. The program is written in an extremely simplistic fashion with the goal of making it easy to verify that it's doing what it thinks it's doing, it is not a model of how one should write actual code. Currently we validate the general purpose, FPSIMD and SVE registers. There are other thing things that could be covered like FPCR and flags registers, these can be covered incrementally - my main focus at the minute is covering the ABI for the SVE registers. The program repeats the tests for all possible SVE vector lengths in case some vector length specific optimisation causes issues, as well as testing FPSIMD only. It tries two syscalls, getpid() and sched_yield(), in an effort to cover both immediate return to userspace and scheduling another task though there are no guarantees which cases will be hit. A new test directory "abi" is added to hold the test, it doesn't seem to fit well into any of the existing directories. Signed-off-by: Mark Brown <broonie@kernel.org> Link: https://lore.kernel.org/r/20211210184133.320748-7-broonie@kernel.org Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
|