Comment 2 for bug 1893024

Revision history for this message
In , colin.king (colin.king-linux-kernel-bugs) wrote :

Ftrace regression tests in tools/testing/selftests/ftrace are picking up regressions in 5.7+ - not sure if these are kernel regressions or bugs in ftrace selftests.

I'm seeing the following ftrace test regresions:

5.9-rc2 test failures with ARM64 in a 4GM VM + 8 CPUs

Test 12, "Generic dynamic event - selective clear (compatibility)"
Test 39, "Kprobe event with comm arguments"
Test 40, "Kprobe event string type argument"
Test 41, "Kprobe event symbol argument"
Test 43, "Kprobe dynamic event with arguments"
Test 44, "Kprobes event arguments with types"
Test 45, "Kprobe event user-memory access"
Test 47, "Kprobe dynamic event with function tracer"
Test 48, "Kprobe dynamic event - probing module"
Test 51, "Kretprobe dynamic event with arguments"

5.8.3, 5.7.17 test failures:
Test 45, "Kprobe event user-memory access"
Test 48, "Kprobe dynamic event - probing module"
Test 51, "Kretprobe dynamic event with arguments"

These tests were working fine when I tested with 5.4.60, 5.5.19, 5.6.19, so these are definite regressions