Skip to content

V
V8
Switch branch/tag
  1. 27 May, 2019 2 commits
    • Peter Marshall's avatar
      [tracing] Add a way to test perfetto traces. · d365f62e
      Peter Marshall authored 
      Add a new abstract class TraceEventListener which is just an interface
for consuming trace events. This separates the V8-specific stuff that
an actual perfetto consumer needs to do e.g. handling the has_more flag
and signalling back to the controller with a semaphore.

This is a change from the previous plan of making the PerfettoConsumer
class sub-classable to implement custom consumption of trace events.
This will be difficult when the consumer is created outside of the
PerfettoTracingController as we can't hook up the
consumer_finished_semaphore_ that belongs to the controller.

Now the PerfettoTracingController is responsible for the Consumer life-
cycle and hides it entirely from callers. We add the
AddTraceEventListener() method to allow callers to register a listener
either for testing or a JSON listener for real tracing.

This lets us write tests that can store all the trace events in memory
without first converting them to JSON, letting us write test more
easily. There's an example test add to test-tracing - more tests using
this style will follow.

Cq-Include-Trybots: luci.v8.try:v8_linux64_perfetto_dbg_ng
Bug: v8:8339
Change-Id: I2d2b0f408b1c7bed954144163e1968f40d772c1b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1628789
Commit-Queue: Ulan Degenbaev <ulan@chromium.org>
Auto-Submit: Peter Marshall <petermarshall@chromium.org>
Reviewed-by: 's avatarUlan Degenbaev <ulan@chromium.org>
Reviewed-by: 's avatarJakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#61854}
      d365f62e
    • Clemens Hammacher's avatar
      [cleanup] Replace simple typedefs by using · a335f2ae
      Clemens Hammacher authored 
      This replaces all typedefs that define types and not functions by the
equivalent "using" declaration.

This was done mostly automatically using this command:
ag -l '\btypedef\b' src test | xargs -L1 \
     perl -i -p0e 's/typedef ([^*;{}]+) (\w+);/using \2 = \1;/sg'

Patchset 2 then adds some manual changes for typedefs for pointer types,
where the regular expression did not match.

R=mstarzinger@chromium.org
TBR=yangguo@chromium.org, jarin@chromium.org

Bug: v8:9183
Change-Id: I6f6ee28d1793b7ac34a58f980b94babc21874b78
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1631409
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: 's avatarMichael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#61849}
      a335f2ae
  3. 23 May, 2019 2 commits
    • Peter Marshall's avatar
      [tracing] Add a perfetto consumer base class · 9a2e4693
      Peter Marshall authored 
      Split the JSONConsumer into a base case and subclass so that we can
implement different consumers without needing to repeat the code that
deals with the finished semaphore and endpoint setters/getters.

This makes it possible to add a test consumer that just holds the trace
events in memory in a later CL.

This CL re-uses the existing perfetto-consumer.h file to some extent,
but it was accidentally leftover from a previous CL where we stopped
using it and switched to the JSON consumer instead.

Cq-Include-Trybots: luci.v8.try:v8_linux64_perfetto_dbg_ng
Bug: v8:8339
Change-Id: I56c305833c8937da76a8e6defbe7a52c59b25106
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1621938
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Reviewed-by: 's avatarJakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#61799}
      9a2e4693
    • Peter Marshall's avatar
      Reland "[tracing] Add a configurable output stream for perfetto tracing" · d5688b1f
      Peter Marshall authored 
      This is a reland of a03ed626

Removed the added test which was failing on win32. The test was unrelated
to the CL; we can add it later.

Original change's description:
> [tracing] Add a configurable output stream for perfetto tracing
>
> Add the ability to provide perfetto with an output stream for the JSON
> consumer rather than hardcode it. D8 will use this interface exclusively
> once the old trace controller is removed.
>
> Also add a test for scope-managed trace events and their duration - this
> was leftover from a previous CL.
>
> Cq-Include-Trybots: luci.v8.try:v8_linux64_perfetto_dbg_ng
> Bug: v8:8339
> Change-Id: I1c45e17e528b549a4cfdaecabd33c7ac4ab4af77
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1611801
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Commit-Queue: Peter Marshall <petermarshall@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#61753}

TBR=jgruber@chromium.org, ulan@chromium.org

Bug: v8:8339
Change-Id: I3442a4d111e12947c107e7d0c226ae934acd06e4
Cq-Include-Trybots: luci.v8.try:v8_linux64_perfetto_dbg_ng
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1627334Reviewed-by: 's avatarPeter Marshall <petermarshall@chromium.org>
Reviewed-by: 's avatarUlan Degenbaev <ulan@chromium.org>
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#61780}
      d5688b1f
  5. 22 May, 2019 2 commits
  7. 20 May, 2019 2 commits
    • Peter Marshall's avatar
      [tracing] Use ChromeTracePacket to save binary size with perfetto · 76f85199
      Peter Marshall authored 
      Previously, compiling with perfetto would add 2.3 MiB to the d8 release
binary. With this change it's 472 KiB.

This is because trace:lite pulls in many more compiled proto classes
than we need, e.g. a bunch of stuff under ftrace/ which is only used
on Android.

Chrome uses a 'mirror proto' ChromeTrackPacket to only provide
compiled protos for the types of packets it will actually see, which on
non-android devices does not include anything under ftrace/.

We use the same trick here in the JSON consumer.

Cq-Include-Trybots: luci.v8.try:v8_linux64_perfetto_dbg_ng
Bug: v8:8339
Change-Id: Iae1f74eec3bd93b18e9f069701fc016440d3ce5e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1619759
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Auto-Submit: Peter Marshall <petermarshall@chromium.org>
Reviewed-by: 's avatarJakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#61662}
      76f85199
    • Peter Marshall's avatar
      [tracing] Convert COMPLETE events to BEGIN/END pairs for perfetto · d5c4d577
      Peter Marshall authored 
      COMPLETE events are being deprecated and don't work well with perfetto
because we can't give out a handle into the buffer like we used to so
that the caller can update the existing event with the duration.

BEGIN/END pairs should be used instead to add two separate trace events
to the buffer which can be associated with one another by the trace
processor e.g. the trace viewer UI.

Cq-Include-Trybots: luci.v8.try:v8_linux64_perfetto_dbg_ng
Bug: v8:8339
Change-Id: Ib73c19d77ad58456ce23d15f0b658c26f3dc3d53
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1615257Reviewed-by: 's avatarJakob Gruber <jgruber@chromium.org>
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#61641}
      d5c4d577
  9. 16 May, 2019 2 commits
    • Peter Marshall's avatar
      [tracing] Roll perfetto · 59cbcb7f
      Peter Marshall authored 
      Roll perfetto to get the fix for a TSAN race.

We need to override a new method on the producer, ClearIncrementalState.
It is only called if the producer opts into it, which we do not so we
add UNREACHABLE() in the implementation.

Cq-Include-Trybots: luci.v8.try:v8_linux64_perfetto_dbg_ng
Bug: v8:8339
Change-Id: I782989f32b511e98479f6637148e3bdd78a6578a
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1615461Reviewed-by: 's avatarJakob Gruber <jgruber@chromium.org>
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#61584}
      59cbcb7f
    • Peter Marshall's avatar
      [tracing] Don't delete TraceWriters in Perfetto · d2061b34
      Peter Marshall authored 
      Perfetto has a mechanism for getting trace data from buffers held alive
by TraceWriters. We enable this with the SetSMBScrapingEnabled() call.

This means we don't need to delete TraceWriter objects and can just leak
them for now, and eventually use thread-local destructors to ensure they
are eventually deleted on thread exit.

Cq-Include-Trybots: luci.v8.try:v8_linux64_perfetto_dbg_ng
Bug: v8:8339
Change-Id: I690bc435c48db6d91eab2fbc7a49ec9feae3b45a
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1607653
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Reviewed-by: 's avatarJakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#61575}
      d2061b34
  11. 10 May, 2019 1 commit
  13. 09 May, 2019 1 commit
  15. 08 May, 2019 1 commit
  17. 03 May, 2019 1 commit
  19. 13 Mar, 2019 1 commit
  21. 21 Feb, 2019 1 commit
    • Peter Marshall's avatar
      [tracing] Fix races in TracingController implementation · 51e80efd
      Peter Marshall authored 
      The default TracingController (used by d8 and Node) has some concurrency
issues. The new test flushes these out, when a second thread logs trace
events while the main thread calls StopTracing().

- Use an acquire load in UpdateCategoryGroupEnabledFlags() because this
  was racing with GetCategoryGroupEnabled() where a new category is
  added in the slow path. g_category_groups is append-only, but
  reads/writes to g_category_index need to be correctly ordered so that
  new categories are added and only then is the change to the index
  visible. The relaxed load ignored this and caused unsynchronized
  read/write.
- Use a relaxed load in ~ScopedTracer() to access category_group_enabled
  as this previously used a non-atomic operation which caused a race
  with UpdateCategoryGroupEnabledFlag() which does a relaxed store.
- Replace TracingController::mode_ with an atomic bool as read/writes to
  mode_ were not synchronized and caused TSAN errors. It only has two
  states and it doesn't seem like we will extend this so just convert it
  to bool.
- Take the lock around calling trace_object->Initialize in
  AddTraceEvent(), and around trace_buffer_->Flush() in StopTracing().
  These two raced previously as the underlying TraceBufferRingBuffer
  passes out pointers to TraceObjects in a synchronized way, but the
  caller (AddTraceEvent) then writes into the object without
  synchronization. This leads to races when Flush() is called, at which
  time TraceBufferRingBuffer assumes that all the pointers it handed out
  are to valid, initialized TraceObjects - which is not true because
  AddTraceEvent may still be calling Initialize on them. This could be
  the cause of issues in Node.js where the last line of tracing/logging
  sometimes gets cut off. This is kind of a band-aid solution - access
  to the TraceObjects handed out by the ring buffer really needs proper
  synchronization which at this point would require redesign. It's quite
  likely we will replace this with Perfetto in the near future so not
  much point investing in this code right now.
- Enable TracingCpuProfiler test which was flaky due to these bugs.

Bug: v8:8821
Change-Id: I141296800c6906ac0e7f3f21dd16d861b07dae62
Reviewed-on: https://chromium-review.googlesource.com/c/1477283
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Reviewed-by: 's avatarYang Guo <yangguo@chromium.org>
Reviewed-by: 's avatarAli Ijaz Sheikh <ofrobots@google.com>
Cr-Commit-Position: refs/heads/master@{#59752}
      51e80efd
  23. 19 Feb, 2019 1 commit
  25. 12 Oct, 2018 1 commit
  27. 20 Sep, 2018 1 commit
  29. 18 Sep, 2018 2 commits
  31. 14 Sep, 2018 1 commit
  33. 11 Sep, 2018 1 commit
  35. 24 Aug, 2018 1 commit
  37. 22 Aug, 2018 1 commit
  39. 06 Aug, 2018 1 commit
  41. 03 Aug, 2018 1 commit
    • Ali Ijaz Sheikh's avatar
      [tracing] allow dynamic control of tracing · bf5ea813
      Ali Ijaz Sheikh authored 
      If the trace_buffer_ was null, we were returning a pointer to a static
flag back that permanently disabled that particular trace point.

This implied an assumption that tracing will be statically enabled at
process startup, and once it is disabled, it will never be enabled
again. On Node.js side we want to dynamically enable/disable tracing as per
programmer intent.

Change-Id: Ic7a7839b8450ab5c356d85e8e0826f42824907f4
Reviewed-on: https://chromium-review.googlesource.com/1161518Reviewed-by: 's avatarYang Guo <yangguo@chromium.org>
Commit-Queue: Ali Ijaz Sheikh <ofrobots@google.com>
Cr-Commit-Position: refs/heads/master@{#54903}
      bf5ea813
  43. 07 May, 2018 1 commit
  45. 02 Feb, 2018 1 commit
  47. 12 Jan, 2018 1 commit
  49. 10 Jan, 2018 1 commit
  51. 13 Oct, 2017 1 commit
  53. 03 Jul, 2017 1 commit
    • Jochen Eisinger's avatar
      Reland "Switch tracing to use v8::TracingController" · cc59f8b1
      Jochen Eisinger authored 
      Original change's description:
> Switch tracing to use v8::TracingController
>
> BUG=v8:6511
> R=fmeawad@chromium.org
>
> Cq-Include-Trybots: master.tryserver.chromium.linux:linux_chromium_rel_ng
> Change-Id: I4961e4b61a9ddc98385ed97c3ffcbcaef2d9cba7
> Reviewed-on: https://chromium-review.googlesource.com/543144
> Commit-Queue: Jochen Eisinger <jochen@chromium.org>
> Reviewed-by: Fadi Meawad <fmeawad@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#46307}

BUG=v8:6511
TBR=fmeawad@chromium.org

Cq-Include-Trybots: master.tryserver.chromium.linux:linux_chromium_rel_ng;master.tryserver.v8:v8_linux64_tsan_rel
Change-Id: Ide32b409248dfd466e7c0bae1d8ae61d6a955d98
Reviewed-on: https://chromium-review.googlesource.com/558865
Commit-Queue: Jochen Eisinger <jochen@chromium.org>
Reviewed-by: 's avatarJochen Eisinger <jochen@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46381}
      cc59f8b1
  55. 29 Jun, 2017 2 commits
  57. 22 Jun, 2017 1 commit
  59. 30 May, 2017 1 commit
  61. 05 Jan, 2017 1 commit
  63. 08 Dec, 2016 1 commit
  65. 31 Oct, 2016 1 commit