Robert O'Callahan

Scalability to high numbers of threads suffers due to `try_wait()` calls

5

For an application with thousands of threads of equal priority, rr context switching can get very slow due to `find_next_runnable_task`, calling `is_task_runnable` on each thread, which in many cases (e.g....

Add zen_workaround to our repo

3

It previously lived at https://gist.github.com/glandium/01d54cefdb70561b5f6675e08f2990f2 Resolves #3690 @glandium can you verify that this is OK? I'm not sure what to do about licensing, authorship etc.

Make breakpoint.run test run against LLDB as well as GDB

6

Test PR to see if CI works

Chaos mode can fail with glibc <= 2.27 and executables with large unmapped regions

1

This can happen in glibc 2.27: * An executable is mapped into memory with a large unmapped hole in it * The dynamic linker loads `libdl.so` into that hole (due...

Make rr work with `perf_event_paranoid`=2

9

I think this might be possible. It would make using rr a bit simpler for most users, and there are environments where changing `perf_event_paranoid` is really hard. The basic idea...

Intermittent counter interrupts missed on AWS Aarch64 Graviton instances

28

``` ESC[?2004h(rr) break abort_ip break abort_ip ESC[?2004l^MBreakpoint 1 at 0xaaaacde00ef4 c ESC[?2004h(rr) c ESC[?2004l^MContinuing. [FATAL src/PerfCounters.cc:824:read_ticks()] (task 2569834 (rec:2567804) at time 189) -> Assertion `!counting_period || interrupt_val

Support `io_uring`

8

Here's a possible approach. * On `io_uring_setup`, create a file monitor identifying the fd as an io_uring fd. * When that fd is mapped, remove any `MAP_FIXED` flag and set...

Missed/delayed ticks interrupt in CI

1

See https://buildkite.com/julialang/rr/builds/1417#018ad0a0-afaf-4aa7-be67-23f12ace08aa ``` Test 'unexpected_exit_pid_ns_32' FAILED: : error during replay: -------------------------------------------------- [FATAL src/PerfCounters.cc:824:read_ticks()] (task 245379 (rec:244028) at time 202) -> Assertion `!counting_period || interrupt_val

Commit 3ef59d96be8a4b620c34663a6f2d043543ec4f00 does not really work because there are cases where `t` has actually been reaped but we haven't noticed and with that commit we are unable to make forward...

Rework trace storage

5

The current trace storage approach has some issues that we need to fix: * Traces are written directly by the rr process and buffered data is not always flushed, so...