StepRws in EVM circuits

[nullity00]

Description

Similar to #1582 , We use the StepRws struct to automatically track read-write indices.

Issue Link

#1586

Type of change

• New feature (non-breaking change which adds functionality)

Contents

The change has been made to the following files

• - create.rs
• - end_tx.rs
• - error_oog_call.rs
• - error_oog_memory_copy.rs
• - error_oog_sload_sstore.rs
• - error_return_data_oo_bound.rs
• - error_write_protection.rs
• - extcodecopy.rs
• - extcodehash.rs
• - extcodesize.rs
• - logs.rs
• - return_revert.rs
• - returndatacopy.rs
• - sha3.rs
• - sload.rs
• - sstore.rs

[nullity00]

@ChihChengLiang Hey, let me know what you think of this !

[ChihChengLiang]

Hi @nullity00,
Thanks for taking this issue! I did the first found of review and left some comments. Please take a look.

[ChihChengLiang]

It's great we write the intention of the read-write, e.g. TxId, Depth, RwCounterEndOfReversion, and IsPersistent.
But I think it defeats the purpose of using StepRws when adding counters in comments, e.g. 0, 1, 2, 3, or rw_off + x. We should remove the counters in all the changes in this PR.

One purpose of using StepRws is to add new features without changing all the magic numbers that aren't relevant to the feature. If we add the counters on comments, they have to be updated when the new feature is added. It is also to forget to update the comments.

Suggested change

	rws.next(); // 0 - TxId
	rws.next(); // 1 - Depth
	let rw_end_of_reversion = rws.next().call_context_value().as_usize(); // 2 - RwCounterEndOfReversion
	let is_persistent = rws.next().call_context_value().as_usize() != 0; // 3 - IsPersistent
	rws.next(); // TxId
	rws.next(); // Depth
	let rw_end_of_reversion = rws.next().call_context_value().as_usize(); // RwCounterEndOfReversion
	let is_persistent = rws.next().call_context_value().as_usize() != 0; // IsPersistent

[ChihChengLiang]

Hi @nullity00, do you see any blocker in this PR? Is there anything we can help?

Note that you can use block.debug_print_txs_steps_rw_ops(); to debug the read-writes, as #1586 describes.

[ed255]

Closing due to inactivity. @nullity00 if you would like to retake this let us know and we can open the PR again :)