Skip to content

V
V8
Switch branch/tag
  1. 18 Feb, 2022 1 commit
    • Dominik Inführ's avatar
      [compiler] Always use full version of RecordWrite builtin · 1b437aa8
      Dominik Inführ authored 
      When we know that the value in a write barrier is a map, we know that
we are not going to have an old-to-new reference (maps are always in
old generation). Therefore we also don't really need the generational
barrier in RecordWrite. While this is technically correct, we don't
gain much from this optimization. The inline and out-of-line generated
code for the barrier is still the same as in all other cases. Which
means that outside marking we don't even reach the RecordWrite builtin.
Most write barrier executions happen outside incremental marking, hence
performance of the incremental marking barrier isn't critical. This CL
always uses the full RecordWrite builtin using a flag in order to
allow for an easy revert.

This CL is motivated by the shared heap work, which needs an additional
always-on barrier in the future (similar to OLD_TO_NEW) to keep a
OLD_TO_SHARED remembered set up-to-date. While maps are always in the
old generation, they maybe by located in the shared heap.

Bug: v8:11708
Change-Id: I71a6ded2547a0b2bbb9bbbd796dbcae0987b2232
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3471854Reviewed-by: 's avatarMichael Lippautz <mlippautz@chromium.org>
Reviewed-by: 's avatarTobias Tebbi <tebbi@chromium.org>
Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79160}
      1b437aa8
  3. 17 Feb, 2022 1 commit
  5. 15 Feb, 2022 5 commits
  7. 11 Feb, 2022 1 commit
  9. 10 Feb, 2022 1 commit
  11. 07 Feb, 2022 1 commit
  13. 04 Feb, 2022 2 commits
    • Junliang Yan's avatar
      s390x: [baseline] enable sparkplug on s390x · 78bc7852
      Junliang Yan authored 
      Change-Id: I4646bb0f3f6291c97bb4b397d6248b9bdaa2059a
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3439641Reviewed-by: 's avatarVictor Gomes <victorgomes@chromium.org>
Commit-Queue: Junliang Yan <junyan@redhat.com>
Cr-Commit-Position: refs/heads/main@{#78959}
      78bc7852
    • Seth Brenith's avatar
      Add verifier for retaining paths in heap snapshots · d937a0bb
      Seth Brenith authored 
      The web app owner who notified me about bugs v8:12112 and v8:12126 asked
me a reasonable question: "how am I ever supposed to trust the retaining
paths in the devtools, if the heap snapshot is generated by a different
component than the actual marking code?". This change is my attempt to
answer that question. If verification is enabled, the heap snapshot
generator will visit each heap object with a realistic marking visitor
to find all references from that object. It will then check that those
references match the HeapGraphEdges in the snapshot.

I also considered the idea that we could collect retaining information
during the last GC cycle before taking the heap snapshot, or during an
extra GC cycle immediately after. However, running the full GC provides
the embedder with the opportunity to run arbitrary code (including JS)
both before and after PerformGarbageCollection, so there is no clear
guarantee that the heap state during the snapshot actually matches the
heap state during marking.

Bug: v8:12112, v8:12126
Change-Id: Id29e75ecf9eee19e35daedbdb4a3e1df64785380
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3299590Reviewed-by: 's avatarMichael Lippautz <mlippautz@chromium.org>
Reviewed-by: 's avatarCamillo Bruni <cbruni@chromium.org>
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Cr-Commit-Position: refs/heads/main@{#78952}
      d937a0bb
  15. 01 Feb, 2022 1 commit
  17. 31 Jan, 2022 1 commit
  19. 28 Jan, 2022 1 commit
    • Dominik Inführ's avatar
      [heap] Compact map space with --compact-map-space · fff5ed12
      Dominik Inführ authored 
      Enable compaction of objects in the map space during a full GC. So far
pages in the map space were never chosen as evacuation candidates. We
might be able to improve memory usage a bit by also compacting map
space. Luckily for us the marking barrier was already emitted when
updating an object's map word.

This CL adds a new flag FLAG_compact_map_space to easily turn off this
feature again. For now we keep this flag (and with that map space
compaction) disabled by default. So GC behavior does not change with
this CL.

Bug: v8:12578
Change-Id: I99c0cd826bd824af5383fb3ce64796693a59d1ff
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3404775Reviewed-by: 's avatarMichael Lippautz <mlippautz@chromium.org>
Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78844}
      fff5ed12
  21. 27 Jan, 2022 1 commit
  23. 25 Jan, 2022 1 commit
  25. 24 Jan, 2022 1 commit
  27. 20 Jan, 2022 1 commit
  29. 17 Jan, 2022 1 commit
  31. 15 Jan, 2022 1 commit
  33. 13 Jan, 2022 1 commit
  35. 11 Jan, 2022 1 commit
  37. 10 Jan, 2022 2 commits
  39. 07 Jan, 2022 1 commit
  41. 05 Jan, 2022 1 commit
  43. 16 Dec, 2021 1 commit
    • Manos Koukoutos's avatar
      [wasm][turbofan] Inline before loop unrolling · b9855173
      Manos Koukoutos authored 
      We switch the order of inlining and loop unrolling optimizations. This
gives small improvements to wasm-gc benchmarks.
Changes:
- Change the loop analysis algorithm to accept loops directly connected
  to the graph's end. This is required because some nodes in an inlined
  function, such as tail calls, might be directly connected to the outer
  function's end without an intervening LoopExit node.
- Based on the above, skip emitting loop exits for some Throw nodes in
  WasmGraphBuildingInterface.
- Introduce WasmInliningPhase, add it before loop unrolling. Remove
  inlining from WasmOptimizationPhase.
- Handle graph terminators in loop unrolling.
- Add loops in the inlined function to the callers loop_infos.
Drive-by:
- Allow more wasm builtins in unrolled loops.
- Reduce inlining parameters to reflect that functions are now slightly
  smaller during inlining, as no unrolling has taken place yet.

Bug: v8:12166
Change-Id: Iadd6b2f75170aa153ca1efb47fbb0d185c2b8371
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3329783Reviewed-by: 's avatarJakob Kummerow <jkummerow@chromium.org>
Reviewed-by: 's avatarNico Hartmann <nicohartmann@chromium.org>
Commit-Queue: Manos Koukoutos <manoskouk@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78394}
      b9855173
  45. 15 Dec, 2021 2 commits
  47. 14 Dec, 2021 2 commits
  49. 13 Dec, 2021 1 commit
  51. 08 Dec, 2021 4 commits
  53. 07 Dec, 2021 1 commit
    • Dominik Inführ's avatar
      [heap] Fix marking state regression · e7fbe09d
      Dominik Inführ authored 
      Concurrent marking got accidentally disabled because
V8_ATOMIC_MARKING_STATE got removed.

Bug: v8:12470, chromium:1277264, chromium:1277330, chromium:1277392
Bug: chromium:1277287, chromium:1277260, chromium:1277259
Bug: chromium:1277256, chromium:1277252, chromium:1277250
Bug: chromium:1277254, chromium:1277251
Change-Id: I21211e8f6d3a65714ba3204e87863b71af42d82d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3318663
Auto-Submit: Dominik Inführ <dinfuehr@chromium.org>
Reviewed-by: 's avatarMichael Lippautz <mlippautz@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#78269}
      e7fbe09d
  55. 06 Dec, 2021 2 commits