Speed up ID precompile

[lu-pinto]

PR description

If one looks at Bytes it is supposed to be an immutable structure so there wouldn't be any need for copying it in IDPrecompileContract. However, looking closely at where this data comes from is required to know if the backing byte[] can also be mutated from other sources.
Currently inputData that is passed to the ID precompile can only come from 3 sources:

1. Transaction data
2. From memory during a CALL
3. From memory during a CREATE

To me CALL seems to be the only place that is passing a Bytes instance that can be mutated later, so by making a copy of it before returning we ensure that modifying memory does not change the returned Bytes instance.
I'm also cutting off a bit of fat in MessageFrame.readMemory because we are actually copying the same data twice.

This needs to go through performance benchmarks before gets pulled in and a full block import with mainnet to be safe.

Fixed Issue(s)

Thanks for sending a pull request! Have you done the following?

• Checked out our contribution guidelines?
• Considered documentation and added the doc-change-required label to this PR if updates are required.
• Considered the changelog and included an update if required.
• For database changes (e.g. KeyValueSegmentIdentifier) considered compatibility and performed forwards and backwards compatibility tests

Locally, you can run these tests to catch failures early:

• spotless: ./gradlew spotlessApply
• unit tests: ./gradlew build
• acceptance tests: ./gradlew acceptanceTest
• integration tests: ./gradlew integrationTest
• reference tests: ./gradlew ethereum:referenceTests:referenceTests

[github-actions]

This pr is stale because it has been open for 30 days with no activity.