VM tracing seems broken at least in some circumstances on newer perfs.
$ magic-trace attach -p 647317
Intel PT support not found. magic-trace will continue and use sampling instead.
Warning: [-callgraph-mode] is defaulting to [Dwarf] which may have high overhead and decoding time. For more info: https://magic-trace.org/w/b
[ Attached. Press Ctrl-C to stop recording. ]
^C[ Got signal, detaching... ]
[ Snapshot taken. ]
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.570 MB /tmp/magic_trace.tmp.5d6b24/perf.data (82 samples) ]
[ Finished recording. ]
[ Decoding, this takes a while... ]
(monitor.ml.Error
("BUG: exception raised while parsing perf output. Please report this to https://github.com/janestreet/magic-trace/issues/"
(exn
("Unexpected event type when parsing perf output" (event_name cpu-clock)))
(perf_output
("647317/647317 529215.810279538: 100000 cpu-clock:uH: "
"\t 687bd57 check_action_pending+0x7 (redacted)"
"\t 684ddfd caml_thread_restore_runtime_state+0x151 (redacted)"
"\t 684de81 caml_thread_leave_blocking_section+0xd (redacted)")))
("Raised at Base__Error.raise in file \"src/error.ml\" (inlined), line 9, characters 14-30"
"Called from Base__Error.raise_s in file \"src/error.ml\", line 10, characters 19-40"
"Called from Async_kernel__Pipe.map.(fun) in file \"src/pipe.ml\", line 1066, characters 65-70"
"Called from Async_kernel__Job_queue.run_jobs in file \"src/job_queue.ml\", line 180, characters 6-47"
"Caught by monitor Monitor.protect"))
VM tracing seems broken at least in some circumstances on newer perfs.