deps: backport v8 commits useful for snapshot integration

[joyeecheung]

The first and last commits are fairly new and contain conflicts with our current version of V8, they may need to be reapplied the next time we update V8

deps: V8: backport ea0719b8ed08

Original commit message:

[snapshot] Do not defer ArrayBuffers during snapshotting

ArrayBuffer instances are serialized by first re-assigning a index
to the backing store field, then serializing the object, and then
storing the actual backing store address again (and the same for the
ArrayBufferExtension). If serialization of the object itself is deferred,
the real backing store address is written into the snapshot, which cannot be
processed when deserializing, leading to a crash.

This fixes this by not deferring ArrayBuffer serialization and adding a DCHECK
for the crash that previously occurred.

Change-Id: Id9bea8268061bd0770cde7bfeb6695248978f994
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2144123
Commit-Queue: Jakob Gruber <[email protected]>
Reviewed-by: Dan Elphick <[email protected]>
Cr-Commit-Position: refs/heads/master@{#67114}

Refs: v8/v8@ea0719b

deps: V8: backport bb9f0c2b2fe9

Original commit message:

[snapshot] Improve snapshot docs and error printing

- Minor improvements to the documentation for snapshotting.
- Add newlines to printed errors where necessary.

Change-Id: I822e7e850adb67eae73b51c23cf34e40ba3106f0
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2144954
Reviewed-by: Toon Verwaest <[email protected]>
Commit-Queue: Toon Verwaest <[email protected]>
Cr-Commit-Position: refs/heads/master@{#67111}

Refs: v8/v8@bb9f0c2

deps: V8: backport 22014de00115

Original commit message:

Reland "[snapshot] rehash JSMap and JSSet during deserialization"

This is a reland of 8374feed55a5b3010f2e9593560a2d84f9f6725f.

Fixed rehashing of global proxy keys by creating its identity hash
early, before the deserialization of the context snapshot.

Original change's description:
> [snapshot] rehash JSMap and JSSet during deserialization
>
> To rehash JSMap and JSSet, we simply replace the backing store
> with a new one created with the new hash.
>
> Bug: v8:9187
> Change-Id: I90c25b18b33b7bc2b6ffe1b89fe17aa5f978b517
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2143983
> Commit-Queue: Joyee Cheung <[email protected]>
> Reviewed-by: Jakob Gruber <[email protected]>
> Reviewed-by: Camillo Bruni <[email protected]>
> Cr-Commit-Position: refs/heads/master@{#67663}

Bug: v8:9187, v8:10523
Change-Id: I7a0319b1d10ff07644de902fec43e7c2b1dd8da9
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2212085
Reviewed-by: Leszek Swirski <[email protected]>
Reviewed-by: Camillo Bruni <[email protected]>
Reviewed-by: Jakob Gruber <[email protected]>
Commit-Queue: Joyee Cheung <[email protected]>
Cr-Commit-Position: refs/heads/master@{#67999}

Refs: v8/v8@22014de

Refs: #17058

Checklist

• make -j4 test (UNIX), or vcbuild test (Windows) passes
• tests and/or benchmarks are included
• documentation is changed or added
• commit message follows commit guidelines

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/31235/

[joyeecheung]

https://ci.nodejs.org/job/node-test-commit-v8-linux/3099/

[targos]

@joyeecheung the first commit is empty. it only bumps the embedder string

[joyeecheung]

@targos Thanks for catching that, the last commit was reverted in the upstream because it broke a bot, I will get back to this after I fix it in the upstream

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/31625/

[joyeecheung]

Replaced the last commit with the reland merged in the upstream. This can wait until next week in case the fuzzer in the upstream finds anything.

V8 CI: https://ci.nodejs.org/job/node-test-commit-v8-linux/3141/

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/31633/

[targos]

Did you manually backport the commits (i.e. without git node v8 backport ?). They look like clean cherry-picks to me.

[joyeecheung]

@targos I think the first two should be clean. The last one definitely isn't because the upstream changed the file names and there are conflicts with off-thread deserialization support changes not in any releases.

[joyeecheung]

Landed in 178efdf...0cb194e, thanks!