Execution Layer Meeting 195

[timbeiko]

Meeting Info

• Aug 29, 2024, 14:00-15:30 UTC
• Stream: https://youtube.com/live/HJ9WxAOwTTA
• Zoom: shared on #allcoredevs Discord channel shortly before the call
• Ethereum Protocol Calls Calendar subscription

Agenda

• Pectra
  • devnet-2 & devnet-3 updates
  • EIP-7702
    • Update EIP-7702: selfdestruct opcode behaviour change in eip-7702 EIPs#8832
    • Update EIP-7702: Change MAGIC from 0x05 to 0x1a EIPs#8835
    • Update EIP-7702: Restrict chain_id to chain's id or zero EIPs#8833
  • EOF
    • EXTCODESIZE and DELEGATECALL issues
    • Devnet updates
  • EIP-7251: Update correlation penalty computation consensus-specs#3882
  • engine: unify request objects execution-apis#565
  • Pectra EIP proposals
    • EIP-7623
    • EIP-7742
• Raising the blob reserve price (context)

[MaxResnick]

Would like a few minutes to discuss raising the blob reserve price to lower the ramp up time when blobs enter price discovery.

[yperbasis]

Would like to discuss ethereum/EIPs#8832 (EIP-7702)

[timbeiko]

@MaxResnick @yperbasis added! Max, if there's anything you can point people to for them to get to speed on the issue prior to the call, please link it here. Thanks!

[nerolation]

I would like to discuss the inclusion of EIP-7623: Increase Calldata Cost for Pectra.
The eip is currently CFI'd and it be great to have a decission on it to then plan including it in devnet 4.

[lightclient]

Any EL teams opposed to this engine API change for requests: ethereum/execution-apis#565

[ralexstokes]

I'd like to propose EIP-7742: https://eips.ethereum.org/EIPS/eip-7742 for inclusion in pectra (in support of peerdas)

[frangio]

I'd like to discuss problems caused by EOF's lack of EXTCODESIZE and restricted DELEGATECALL, and whether we should lift these EOF restrictions: https://hackmd.io/@frangio/S1VvatXiR

[MaxResnick]

@MaxResnick @yperbasis added! Max, if there's anything you can point people to for them to get to speed on the issue prior to the call, please link it here. Thanks!

Working on a draft EIP but just wanted to get a gutcheck on whether the order of magnitude I was thinking about for the parameter was ok with everyone. The goal for choosing the number is not to increase the price of blob gas too much but more to make it so that the time it takes to update to a reasonable price when blobs do enter price discovery is reduced.

i'll start by suggesting we do

+ MIN_BASE_FEE_PER_BLOB_GAS  =  160217286
- MIN_BASE_FEE_PER_BLOB_GAS = 1

the reasoning behind this is if we look at the cost of a simple transfer when the base fee is 1 GWEI ~5c usd at today's prices and try to peg the min price of a blob to that, this is the number that pops out.

to reach this number today, excess blob gas has to be at 63070646. When you calculate how long this would take to reach from 0 excess blob gas you get:

63070646/(3 * 2**17) = 160.396947225

So it takes 160 blocks or 32 mins to get up to this price today.

I did this math in floating point so its a bit off (by a few WEI) but i'll do it in fixed point before I submit the official EIP draft PR.

[abcoathup]

@timbeiko stream link shows Video unavailable: This video is private.

There currently isn't an ACDE stream on YouTube:
https://www.youtube.com/@EthereumProtocol/streams

[etan-status]

not on the call, but have two minor things for EIP-7702. putting here for awareness as EIP-7702 is already an agenda item:

• https://ethereum-magicians.org/t/eip-7702-set-eoa-account-code-for-one-transaction/19923/195?u=etan-status

1. Change signing domain of authorizations from 0x05 to 0x1a, so that future txtype can use 0x05
2. Marking chain_id=0 as reserved (EIP-7702 currently uses those semantics for chain agnostic authorizations)

[timbeiko]

@nerolation @lightclient @ralexstokes @frangio @etan-status, I've added each of your items to the agenda. @abcoathup thanks for the flag: the stream link should be public now!

[mkalinin]

I can’t attend the call, but would like to make the last call on merging the following PR:

• EIP-7251: Update correlation penalty computation consensus-specs#3882

If anyone is opposed to get this one merged, please, leave your comment, thank you!