d87ae0fa - selftests: Add benchmark for syscall user dispatch

Fri, 27 Nov 2020 19:32:37 +0000

selftests: Add benchmark for syscall user dispatchThis is the patch I'm using to evaluate the impact syscall user dispatchhas on native syscall (syscalls not redirected to userspace) whenenabled for the process and submiting syscalls though the unblockeddispatch selector. It works by running a step to define a baseline ofthe cost of executing sysinfo, then enabling SUD, and rerunning thatstep.On my test machine, an AMD Ryzen 5 1500X, I have the following resultswith the latest version of syscall user dispatch patches.root@olga:~# syscall_user_dispatch/sud_benchmark Calibrating test set to last ~5 seconds... test iterations = 37500000 Avg syscall time 134ns. Caught sys_ff00 trapped_call_count 1, native_call_count 0. Avg syscall time 147ns. Interception overhead: 9.7% (+13ns).Signed-off-by: Gabriel Krisman Bertazi Signed-off-by: Thomas Gleixner Reviewed-by: Kees Cook Reviewed-by: Andy Lutomirski Acked-by: Peter Zijlstra (Intel) Link: https://lore.kernel.org/r/20201127193238.821364-7-krisman@collabora.com List of files: /linux-6.15/tools/testing/selftests/syscall_user_dispatch/Makefile

179ef035 - selftests: Add kselftest for syscall user dispatch

Fri, 27 Nov 2020 19:32:36 +0000

selftests: Add kselftest for syscall user dispatchImplement functionality tests for syscall user dispatch. In order tomake the test portable, refrain from open coding syscall dispatchers andcalculating glibc memory ranges.Signed-off-by: Gabriel Krisman Bertazi Signed-off-by: Thomas Gleixner Reviewed-by: Andy Lutomirski Acked-by: Kees Cook Acked-by: Peter Zijlstra (Intel) Link: https://lore.kernel.org/r/20201127193238.821364-6-krisman@collabora.com List of files: /linux-6.15/tools/testing/selftests/syscall_user_dispatch/Makefile

Changes in Makefile

d87ae0fa - selftests: Add benchmark for syscall user dispatch

179ef035 - selftests: Add kselftest for syscall user dispatch