This CL deduplicates entries in constant pools based on handle
location. This works well because we already use CanonicalHandleScope
in the right places.

The CL gives a ~2% speed-up on octane for arm64. The code size is
reduced significantly as well: Before the change all generated code
spend on constant pools during an Octane run on arm64 was 723kb
(90398 64bit words) before this change, and after this change only
spend 189kb (23615 64bit words) were spend on constant pools. This
is a 73.8% reduction.


Change-Id: If0bb83453a45c0df0d1b0fee591c04c621341af1
Bug: v8:8054, v8:8977, v8:7703
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1672924
Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#62324}