commit | 587750bce09b9937504f949dfc3d7e63c47408f7 | [log] [tgz] |
---|---|---|
author | Primiano Tucci <primiano@google.com> | Tue Oct 27 15:37:01 2020 +0100 |
committer | Primiano Tucci <primiano@google.com> | Tue Oct 27 15:37:01 2020 +0100 |
tree | 3d63220e76e618243390f81b1f025d3905b7dc35 | |
parent | 037b4922a67ca117f62276995fa03c1924a9b0b2 [diff] |
Kallsyms: fix detection when first lines are zero This CL fixes a couple of issues and re-enables the test on non-cuttlefish: - The first entries of /proc/kallsyms can be zero even when kptr_restrict is lowered. Improve the heuristic for that by looking at the first page. - make the KernelAddressSymbolization run for longer and re-enable it. Because the symbolizer is initialized asynchronously, we need to give time to the traced_probes process to enqueue the PostTask, before sending the stop. Othewise traced_probes might end up running, in sequence: Start, Stop, InitializeSymbolizer. - Re-enable the KernelAddressSymbolization but not on cuttlefish. cf on GCE seems too slow and it would hit the timeout in the production code. Not worth incrementing that to deal with slow emulators. Our CI is fine because on standalone builds we don't use the setprop & wait-with-timeout path but directly lower kptr_restrict. Test: perfetto_integrationtests Change-Id: I9e0933d11f25a9b79198cd70927d88d400454a42
Perfetto is a production-grade open-source stack for performance instrumentation and trace analysis. It offers services and libraries and for recording system-level and app-level traces, native + java heap profiling, a library for analyzing traces using SQL and a web-based UI to visualize and explore multi-GB traces.
See https://perfetto.dev/docs or the /docs/ directory for documentation.