Skip to content

V
V8
Switch branch/tag
  1. 20 Mar, 2018 38 commits
  3. 19 Mar, 2018 2 commits
    • Peter Marshall's avatar
      [runtime] Use JSBuiltinsConstructStub for all builtins where possible · 4af7efc1
      Peter Marshall authored 
      In order to remove the construct_stub field of the SFI we need all
construct stubs to be the same, and do any branching at runtime
instead. For builtins we don't need to set the construct stub because
the builtins construct stub will call into it for us.

There should only be two builtins left without the builtins construct
stub: Array and InternalArray, which are special cases that need to
be dealt with in another CL.

Bug: v8:7503
Change-Id: If0d419399a9ee22c09cf2a5a3d3dbea7a04dee77
Reviewed-on: https://chromium-review.googlesource.com/968862Reviewed-by: 's avatarBenedikt Meurer <bmeurer@chromium.org>
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52046}
      4af7efc1
    • Michael Stanton's avatar
      Revert "[turbofan] Masking/poisoning in codegen (optimized code, mips & mips64)" · 11831378
      Michael Stanton authored 
      This reverts commit 46a3c772.

Reason for revert: This is actually not quite ready. What we need is a speculation free poisoning, and if we do another branch, then I think that won't happen.

Original change's description:
> [turbofan] Masking/poisoning in codegen (optimized code, mips & mips64)
> 
> This introduces masking of loads with speculation bit during code generation.
> At the moment, this is done only under the
> --branch-load-poisoning flag, and this CL enlarges the set of supported
> platforms from {x64, arm, arm64} to {x64, arm, arm64, mips, mips64}.
> 
> Overview of changes:
> - new register configuration configuration with one register reserved for
>   the speculation poison/mask (kSpeculationPoisonRegister).
> - in codegen, we introduce an update to the poison register at the starts
>   of all successors of branches (and deopts) that are marked as safety
>   branches (deopts).
> - in memory optimizer, we lower all field and element loads to PoisonedLoads.
> - poisoned loads are then masked in codegen with the poison register.
>   * only integer loads are masked at the moment.
> 
> Bug: chromium:798964
> Change-Id: I211395b8305ed0ad9288d6da48fa159fa970c827
> Reviewed-on: https://chromium-review.googlesource.com/951382
> Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
> Reviewed-by: Ivica Bogosavljevic <ivica.bogosavljevic@mips.com>
> Commit-Queue: Ivica Bogosavljevic <ivica.bogosavljevic@mips.com>
> Cr-Commit-Position: refs/heads/master@{#52042}

TBR=mvstanton@chromium.org,mstarzinger@chromium.org,ivica.bogosavljevic@mips.com

Change-Id: Ief4d9ef56d918172f0b545d321a64b1ab5b46915
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: chromium:798964
Reviewed-on: https://chromium-review.googlesource.com/969041Reviewed-by: 's avatarMichael Stanton <mvstanton@chromium.org>
Commit-Queue: Michael Stanton <mvstanton@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52045}
      11831378