Update EIP-4762: tweaks to support 7702 #8896

gballet · 2024-09-20T13:49:55Z

This is a proposed tweak to the stateless gas costs EIP in order to take into account the extra cases introduced in EIP 4762.

Signed-off-by: Guillaume Ballet <3272758+gballet@users.noreply.github.com>

eth-bot · 2024-09-20T13:50:30Z

✅ All reviewers have approved.

gballet · 2024-09-20T13:56:25Z

EIPS/eip-4762.md

@@ -62,10 +70,10 @@ If a `*CALL` or `SELFDESTRUCT` is value-bearing (ie. it transfers nonzero wei),

 Note: when checking for the existence of the `callee`, the existence check is done by validating that there is an extension-and-suffix tree at the corresponding stem, and does not rely on `CODEHASH_LEAF_KEY`.

-When calling `EXTCODEHASH`, process the access event:
+When calling `*CALL`, `SELFDESTRUCT`, `EXTCODESIZE`, `EXTCODECOPY` or `EXTCODEHASH` on an EIP-7702 delegating contract, process the additional access event:


@lightclient in the case of a (same-tx) selfdestruct, and assuming it is executed by a delegating EoA, what is being destroyed? the EoA or the delegated contract? Presumably the former?

Suggested change

When calling `*CALL`, `SELFDESTRUCT`, `EXTCODESIZE`, `EXTCODECOPY` or `EXTCODEHASH` on an EIP-7702 delegating contract, process the additional access event:

When calling `*CALL`, `EXTCODESIZE`, `EXTCODECOPY` or `EXTCODEHASH` on an EIP-7702 delegating contract, process the additional access event:

gballet · 2024-09-20T14:05:32Z

EIPS/eip-4762.md

 ### Witness gas costs

 Remove the following gas costs:

 * Increased gas cost of `CALL` if it is nonzero-value-sending
 * [EIP-2200](./eip-2200.md) `SSTORE` gas costs except for the `SLOAD_GAS`
 * 200 per byte contract code cost
+ * `PER_EMPTY_ACCOUNT_COST` and `PER_AUTH_BASE_COST`


This is potentially a controversial move, but since we have the stateless costs to cover for the writes to the state, that should match. Currently, they are covered by the 21000 intrinsic gas. Are these extra costs related to the transaction itself, or the state being written?

If it's the state being written, then I think we should drop them in verkle

if it's to disincentivize growing the transaction payload, then this line needs to be removed.

also, its done is a different way and i dont understand the reason behind that
we charge the PER_EMPTY_ACCOUNT_COST always and then refund PER_AUTH_BASE_COST if the account exists
looking at how the gas costs are implemented in this EIP. I think we should keep the gas costs as it is and deal with them in the way we deal with intrisic gas costs for transactions right now

gballet · 2024-09-20T14:08:13Z

EIPS/eip-4762.md

+(address, 0, CODEHASH_LEAF_KEY)
+```
+
+Note: this latter access is mean to check whether the contract is a [EIP-7702](./eip-7702.md) contract, or an EOF contract.


So this is the weird bit, I am using the code hash to determine if it's a 7702 contract. Come to think of it, it should be reading the first code chunk directly because then you will also know what is the delegation address.

So yeah, the question is: apart from EXTCODEHASH, is there any reason to read the code hash for all these extra instructions?

Yes, I believe we should ask to read the first code-chunk to check if there's a delegation. Also, it might worth clarifying (as 7702 does) that we shouldn't follow a chain of delegations.

So yeah, the question is: apart from EXTCODEHASH, is there any reason to read the code hash for all these extra instructions?

I don't think so, since if there's a delegation that's irrelevant?

Also, I'd recommend removing the EOF mentions in this PR.

jsign · 2024-09-20T14:31:45Z

EIPS/eip-4762.md


 ```
-(address, 0, CODEHASH_LEAF_KEY)
+(delegated_address, 0, CODEHASH_LEAF_KEY)


I'm a bit confused by this change.
For example, if you're executing an EXTCODESIZE as mentioned above:

We should provide the first-code chunk to allow a stateless client to detect if there's a delegation and setup the right context of the target.

If there is a delegation, we'll include the BASIC_DATA of the delegated address which contains the code size.

Why there is a need for this CODEHASH_LEAF_KEY of the delegated address?

More generally, I think we should only explain impact on the witness for the delegation resolving step. Then that might change the context of which address is used for the EXTCODEHASH/EXTCODESIZE/etc and simply apply already existing rules to include the witness for that address (which now can be address or delegated_address)

As in, the new 7702 rule is about right context setup of the target address. After that is done, the corresponding instructions should do what we already explained in the current version of the EIP.

We could try doing some renames to make this clear and always use (a new) target_address as the placeholder for 99% of witness addition rules. And explain what target_address is e.g: address if there's no delegation, or delegated_address. Food for thought.

Why there is a need for this CODEHASH_LEAF_KEY of the delegated address?

that's an open question for @lightclient but the idea was that you wouldn't read the code directly, but the code hash. I don't think it's a good idea after all, but I'm waiting on his input.

jsign · 2024-09-20T14:37:41Z

EIPS/eip-4762.md

+```
+(address, 0, BASIC_DATA_LEAF_KEY)
+(address, 0, CODEHASH_LEAF_KEY)
+(address, 0, CODE_OFFSET)


What is CODE_OFFSET?

CODE_OFFSET is defined in eip-6800 at 128: it's the start of the code.

i.e. the first code chunk is being read

bc1qdfs6vkttw7fj7t4m73humszxmgeq78a8we4ncf that's my address wallet

beage666 · 2024-09-22T17:03:53Z

What dose this mean

Scamreno · 2024-09-24T04:21:26Z

Okay 👍

…

On Mon, Sep 23, 2024, 10:05 PM beage666 ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In EIPS/eip-4762.md <#8896 (comment)>: > @@ -186,13 +196,37 @@ If `value` is non-zero: (tx.target, 0, BASIC_DATA_LEAF_KEY) ``` +#### Setcode transactions + +For EIP-7702 transactions, for each `[[chain_id, address, nonce, y_parity, r, s], ...]` entry in the authorization list, process the additional access events: + +``` +(address, 0, BASIC_DATA_LEAF_KEY) +(address, 0, CODEHASH_LEAF_KEY) +(address, 0, CODE_OFFSET) bc1qdfs6vkttw7fj7t4m73humszxmgeq78a8we4ncf that's my address wallet — Reply to this email directly, view it on GitHub <#8896 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BJB7T4MVJRVNOBI2I4W7GZLZYDCF3AVCNFSM6AAAAABOSC4FAOVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMZDGMRTHA2DONJXGI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

tanishqjasoria · 2024-09-24T11:44:45Z

EIPS/eip-4762.md

+For EIP-7702 transactions, for each `[[chain_id, address, nonce, y_parity, r, s], ...]` entry in the authorization list, process the additional access events:
+
+```
+(address, 0, BASIC_DATA_LEAF_KEY)
+(address, 0, CODEHASH_LEAF_KEY)
+(address, 0, CODE_OFFSET)


this does not look right to me, it should be authority, insted of address

tanishqjasoria · 2024-09-24T11:48:36Z

EIPS/eip-4762.md

+(address, 0, CODE_OFFSET)
+```
+
+If `address` is not present in the state, also process the write event:


we should always process the write event
(authority, 0, CODEHASH_LEAF_KEY)

tanishqjasoria · 2024-09-24T11:52:51Z

EIPS/eip-4762.md

 ### Witness gas costs

 Remove the following gas costs:

 * Increased gas cost of `CALL` if it is nonzero-value-sending
 * [EIP-2200](./eip-2200.md) `SSTORE` gas costs except for the `SLOAD_GAS`
 * 200 per byte contract code cost
+ * `PER_EMPTY_ACCOUNT_COST` and `PER_AUTH_BASE_COST`


also, its done is a different way and i dont understand the reason behind that
we charge the PER_EMPTY_ACCOUNT_COST always and then refund PER_AUTH_BASE_COST if the account exists
looking at how the gas costs are implemented in this EIP. I think we should keep the gas costs as it is and deal with them in the way we deal with intrisic gas costs for transactions right now

tanishqjasoria · 2024-09-24T11:54:50Z

EIPS/eip-4762.md

+(address, 0, BASIC_DATA_LEAF_KEY)
+(address, 0, CODE_OFFSET)


Suggested change

(address, 0, BASIC_DATA_LEAF_KEY)

(address, 0, CODE_OFFSET)

(authority, 0, BASIC_DATA_LEAF_KEY)

(authority, 0, CODE_OFFSET)

(authority, 0, CODEHASH_LEAF_KEY)

tanishqjasoria · 2024-09-24T11:55:20Z

EIPS/eip-4762.md

+(address, 0, BASIC_DATA_LEAF_KEY)
+(address, 0, CODEHASH_LEAF_KEY)
+(address, 0, CODE_OFFSET)


Suggested change

(address, 0, BASIC_DATA_LEAF_KEY)

(address, 0, CODEHASH_LEAF_KEY)

(address, 0, CODE_OFFSET)

(authority, 0, BASIC_DATA_LEAF_KEY)

(authority, 0, CODEHASH_LEAF_KEY)

(authority, 0, CODE_OFFSET)

EIP-4762: tweaks to support 7702

1a74546

Signed-off-by: Guillaume Ballet <3272758+gballet@users.noreply.github.com>

github-actions bot added c-update Modifies an existing proposal s-draft This EIP is a Draft t-core labels Sep 20, 2024

eth-bot changed the title ~~EIP-4762: tweaks to support 7702~~ Update EIP-4762: tweaks to support 7702 Sep 20, 2024

gballet commented Sep 20, 2024

View reviewed changes

jsign reviewed Sep 20, 2024

View reviewed changes

gballet mentioned this pull request Sep 23, 2024

Stateless Implementers Call #25 ethereum/pm#1159

Closed

Aces42020 approved these changes Sep 24, 2024

View reviewed changes

beage666 approved these changes Sep 24, 2024

View reviewed changes

tanishqjasoria reviewed Sep 24, 2024

View reviewed changes

Skippyy9 approved these changes Dec 7, 2024

View reviewed changes

daniela2000111 approved these changes Dec 7, 2024

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Update EIP-4762: tweaks to support 7702 #8896

Update EIP-4762: tweaks to support 7702 #8896

gballet commented Sep 20, 2024

eth-bot commented Sep 20, 2024 •

edited

Loading

gballet Sep 20, 2024

gballet Sep 20, 2024

beage666 Sep 24, 2024

tanishqjasoria Sep 24, 2024

gballet Sep 20, 2024

jsign Sep 20, 2024 •

edited

Loading

jsign Sep 20, 2024

jsign Sep 20, 2024 •

edited

Loading

gballet Sep 20, 2024

jsign Sep 20, 2024

gballet Sep 20, 2024

gballet Sep 20, 2024

beage666 Sep 24, 2024

Scamreno Sep 24, 2024

beage666 Oct 9, 2024

beage666 commented Sep 22, 2024

Scamreno commented Sep 24, 2024 via email

tanishqjasoria Sep 24, 2024

tanishqjasoria Sep 24, 2024

tanishqjasoria Sep 24, 2024

tanishqjasoria Sep 24, 2024

tanishqjasoria Sep 24, 2024

	When calling `*CALL`, `SELFDESTRUCT`, `EXTCODESIZE`, `EXTCODECOPY` or `EXTCODEHASH` on an EIP-7702 delegating contract, process the additional access event:
	When calling `*CALL`, `EXTCODESIZE`, `EXTCODECOPY` or `EXTCODEHASH` on an EIP-7702 delegating contract, process the additional access event:

-(address, 0, BASIC_DATA_LEAF_KEY)
-(address, 0, CODE_OFFSET)
+(authority, 0, BASIC_DATA_LEAF_KEY)
+(authority, 0, CODE_OFFSET)
+(authority, 0, CODEHASH_LEAF_KEY)

Update EIP-4762: tweaks to support 7702 #8896

Are you sure you want to change the base?

Update EIP-4762: tweaks to support 7702 #8896

Conversation

gballet commented Sep 20, 2024

eth-bot commented Sep 20, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jsign Sep 20, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jsign Sep 20, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

beage666 commented Sep 22, 2024

Scamreno commented Sep 24, 2024 via email

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

eth-bot commented Sep 20, 2024 •

edited

Loading

jsign Sep 20, 2024 •

edited

Loading

jsign Sep 20, 2024 •

edited

Loading