Skip to content

V
V8
Switch branch/tag
  1. 08 Aug, 2017 2 commits
    • Jakob Gruber's avatar
      [snapshot] Simplify deserializer APIs · 95f86859
      Jakob Gruber authored 
      This moves deserializer setup boilerplate inside the deserializers, and
makes improper usage less likely. For instance:

  ObjectDeserializer deserializer(&scd);
  /* ... deserializer setup ... */
  MaybeHandle<HeapObject> obj = deserializer.Deserialize(isolate);
  /* ... result checks and casts ... */

has now become:

  /* All setup and casts inside deserializer, impossible to illegally
     use the same instance multiple times. */
  MaybeHandle<SharedFunctionInfo> maybe_result =
      ObjectDeserializer::DeserializeSharedFunctionInfo(
          isolate, &scd, source);

Bug: v8:6624
Change-Id: Id5a1848e024e89cf86e5292389ba7c89f31d8e6b
Reviewed-on: https://chromium-review.googlesource.com/604791
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: 's avatarYang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47219}
      95f86859
    • Jakob Gruber's avatar
      [snapshot] Continue splitting up Deserializer · 2e2069cd
      Jakob Gruber authored 
      Another pure refactoring CL. This extracts methods used only by Deserializer
subclasses.

Bug: v8:6624
Change-Id: Ib4dd7cdc591dff217e282e68a490c8c7129b9c96
Reviewed-on: https://chromium-review.googlesource.com/602188
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: 's avatarYang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47211}
      2e2069cd
  3. 03 Aug, 2017 2 commits
  5. 02 Aug, 2017 1 commit
  7. 01 Aug, 2017 1 commit
  9. 27 Jul, 2017 2 commits
  11. 25 Jul, 2017 1 commit
  13. 24 Jul, 2017 3 commits
  15. 21 Jul, 2017 1 commit
    • Ulan Degenbaev's avatar
      [heap, deserializer] Restore marking invariant for deserialized maps · 250ba28a
      Ulan Degenbaev authored 
      when black allocation is on.

The scenario:
1) Incremental marking is off.
2) Partial deserialization starts and calls Heap::ReserveSpace.
2) ReserveSpace creates (white) reservations in old space.
3) ReserveSpace allocates map placeholders. One of these allocations
starts incremental marking, which starts black allocation (currently
when concurrent marking is on). Subsequent maps are black allocated.
4) ReserveSpace succeeds without triggering a GC.
5) Deserialization continues. Some maps are black. Note that
deserialization emits only old->new write barriers and skips
marking write barriers.
6) Deserialization finishes and re-visits the black allocated
reservations and large object. This misses black allocated maps.
7) There is black->white descriptor array pointer in one of these map.

BUG=chromium:723600

Change-Id: Ifffe46f22a7d7dbc5cff2e882190234fcc722ccb
Reviewed-on: https://chromium-review.googlesource.com/581187
Commit-Queue: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: 's avatarMichael Lippautz <mlippautz@chromium.org>
Reviewed-by: 's avatarYang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46816}
      250ba28a
  17. 20 Jul, 2017 1 commit
  19. 18 Jul, 2017 1 commit
  21. 14 Jul, 2017 1 commit
  23. 13 Jul, 2017 1 commit
  25. 10 Jul, 2017 1 commit
  27. 26 Jun, 2017 1 commit
    • hans's avatar
      Make some functions that are hit during renderer startup available for inlining · 777da354
      hans authored 
      This is towards closing the perf gap between the MSVC build (which uses link-
time optimization) and Clang (where LTO isn't ready on Windows yet). We did
a study (see bug) to see which non-inlined functions are hit a lot during render
start-up, and which would be inlined during LTO. This should benefit performance
in all builds which currently don't use LTO (Android, Linux, Mac) as well as
the Win/Clang build.

The binary size of chrome_child.dll increases by 2KB with this.

BUG=chromium:728324
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_chromium_compile_dbg_ng;master.tryserver.chromium.mac:mac_chromium_compile_dbg_ng

Review-Url: https://codereview.chromium.org/2950993002
Cr-Commit-Position: refs/heads/master@{#46229}
      777da354
  29. 25 Jun, 2017 1 commit
  31. 23 Jun, 2017 1 commit
    • hans's avatar
      Make some functions that are hit during renderer startup available for inlining · d00d52be
      hans authored 
      This is towards closing the perf gap between the MSVC build (which uses link-
time optimization) and Clang (where LTO isn't ready on Windows yet). We did
a study (see bug) to see which non-inlined functions are hit a lot during render
start-up, and which would be inlined during LTO. This should benefit performance
in all builds which currently don't use LTO (Android, Linux, Mac) as well as
the Win/Clang build.

The binary size of chrome_child.dll increases by 2KB with this.

BUG=chromium:728324

Review-Url: https://codereview.chromium.org/2950993002
Cr-Commit-Position: refs/heads/master@{#46191}
      d00d52be
  33. 16 Jun, 2017 1 commit
  35. 13 Jun, 2017 4 commits
  37. 12 Jun, 2017 1 commit
  39. 02 Jun, 2017 1 commit
    • thakis's avatar
      v8: Fix unaligned access when deserializing snapshots. · 13af45c9
      thakis authored 
      The code was already careful to use memcpy() here, but then it
added needless casts to wider types that made the compiler think
that it can do 4-byte-aligned accesses when it couldn't.

(It's also a bug that the snapshot got loaded at an unaligned
address, but we can fix both bugs.)

BUG=chromium:729059

Review-Url: https://codereview.chromium.org/2915323002
Cr-Commit-Position: refs/heads/master@{#45698}
      13af45c9
  41. 27 May, 2017 1 commit
  43. 22 May, 2017 1 commit
  45. 10 May, 2017 1 commit
    • jkummerow's avatar
      [deserializer] Make large object deserialization GC safe · 6bfee50e
      jkummerow authored 
      When black allocation is turned on at deserialization time, then
slots in deserialized objects have to be visited by the incremental
marker. For spaces with reservations, this has always been done; for
large object space with its special handling, this patch adds it.

Additionally, we must ensure that no incremental steps that might
cause incremental marking to finish are performed while there is an
AlwaysAllocateScope around.

BUG=chromium:718859

Review-Url: https://codereview.chromium.org/2868103002
Cr-Commit-Position: refs/heads/master@{#45231}
      6bfee50e
  47. 28 Apr, 2017 1 commit
  49. 27 Apr, 2017 1 commit
  51. 26 Apr, 2017 2 commits
  53. 25 Apr, 2017 5 commits