EIP 7951 Hub side

[amkCha]

• make a command to list the addresses of precompiles (which is now a union of ranges)
• add a PRC scenario column for P256_VERIFY
• update CALL/precompile ≡ 1 in table unexcept and unaborted call type
• update precompile related shorthands (common except id, common, weighted sum) + table recap
• add P256_VERIFY to precompile classification liste + table
• add P256_VERIFY to nsr and flag sum II def + table
• explicit P256_VERIFY cannot have empty call data
• add P256_VERIFY success, FKTH, FKTR case

---

Note

Introduces P256_VERIFY precompile across HUB/OOB/MMU with full scenarios, shorthands, and tables; unifies precompile address range and updates BLS/PRECOMPILE docs and diagrams.

• Precompiles (HUB/OOB/MMU):
  • Add P256_VERIFY support: new scenario \scenPVerify, success/Failure (FKTH/FKTR) flows, result size, exo-phase wiring, and data transfer logic.
  • Extend common precompile processing and shorthands to include P256_VERIFY; enforce non-empty call data like BLS/point-evaluation.
  • Update OOB common handling to recognize \oobInstPVerify and related predictions.
• Addressing and constants:
  • Introduce \prcAddressRange and switch references to it (incl. CALL handling and Tx-skip rules).
  • Define/normalize precompile address constants via \numConst and add \prcAddressPVerifyValue.
• Scenarios, shorthands, and tables:
  • Update weighted sums (\scenPrecompileWeightedSum) to use address constants; add P256_VERIFY to classification, NSR/flag-sum defs, and tables.
  • Expand precompile scenario columns and recap tables to include P256_VERIFY.
• Docs/diagrams:
  • Replace/introduce post-Cancun setup diagrams for fixed/variable-size precompiles; integrate P256_VERIFY alongside BLS.
  • Remove obsolete BLS-only representations; minor clarifications and typos fixed (e.g., "known" typos, OOB list adds CALL→precompile).

^{Written by Cursor Bugbot for commit dbf81e4. This will update automatically on new commits. Configure here.}

[cursor]

Bug: Empty Call Data Check Inconsistency

The sanity check enforcing \locOobResultEmptyCallData = 0 is conditioned on \scenPrecompileCommonBls _{i} = 1, but \scenPVerify is not part of \scenPrecompileCommonBls. Since P256_VERIFY actively disallows empty call data (per ecadd_ecmul_ecpairing_bls_pverify/empty.tex), this constraint should also apply to P256_VERIFY but currently won't be enforced. The sanity check needs to include P256_VERIFY explicitly or be restructured to cover all precompiles that disallow empty call data.

hub/instruction_handling/call/precompiles/common/generalities.tex#L139-L145

linea-specification/hub/instruction_handling/call/precompiles/common/generalities.tex

Lines 139 to 145 in a944488

	anchorRow = i ,
	relOffset = \prcCommonMiscRowOffset ,
	sourceId = \cn_{i} ,
	targetId = 1 + \hubStamp_{i} ,
	sourceOffsetLo = \locPrcCdo ,
	size = \locPrcCds ,
	referenceOffset = 0 ,

 

---

[amkCha]

I may have missed it but there is a binarity constraints that should be added to \scenPVerify. Overall a lot of good stuff, a few comments are in place.

Okay, after checking, I think it's included here

[OlivierBBB]

Only one concern about a macro containing numerals in its name. LGTM otherwise.

[OlivierBBB]

It this accepted by Latex ? I believe you can't have numerals in latex macros.

[amkCha]

Agreed that it's underlining it in my IDE, but it's still showing nicely

[amkCha]

I'll change it to be consistent with other naming

[OlivierBBB]

Apparently it is. I had no idea ... how many macros I created would be improved by using numerals 😱

[OlivierBBB]

This is ok. We may (not sure at this point) change this back as what we discovered this afternoon is that P256_VERIFY can only fail for OOG reasons, which are FKTH.

[OlivierBBB]

This we will have to change in light of our new understanding of P256_VERIFY

All requested changed are done :)