vmConfig.EWASMInterpreter is not passed to NewEVM

[cdetrio]

To run geth + EVMC/Hera, geth accepts the path to the Hera .so file in a --vm.ewasm flag. This path to Hera is then passed in vmConfig when calling newBlockchain here:

go-ethereum/eth/backend.go

Lines 155 to 160 in 52dce7d

	EWASMInterpreter: config.EWASMInterpreter,
	EVMInterpreter: config.EVMInterpreter,
	}
	cacheConfig = &core.CacheConfig{Disabled: config.NoPruning, TrieNodeLimit: config.TrieCache, TrieTimeLimit: config.TrieTimeout}
	)
	eth.blockchain, err = core.NewBlockChain(chainDb, cacheConfig, eth.chainConfig, eth.engine, vmConfig, eth.shouldPreserve)

The problem is that another place where vmConfig is passed is from the rpc method eth_estimateGas, originating here: https://github.com/ewasm/go-ethereum/blob/tantalus-evmc6/internal/ethapi/api.go#L714. Here the vmConfig is empty, and it makes its way to NewEVM() here: https://github.com/ewasm/go-ethereum/blob/tantalus-evmc6/core/vm/evm.go#L135

As a temporary fix, in NewEVM() we're reading the Hera.so path from the env var EVMC_PATH, with this patch: bb778d6. But we'd like to read it from the --vm.ewasm flag instead.

[gballet]

have you tried accessing s.v.chainConfig().EWASMInterpreter?

[chfast]

@gballet The ChainConfig object does not contains information about VM config.

The commit ethereum@0e9dad2 of ethereum#17955 should fix the problem. But it is still a workaround rather for a complete solution.

[gballet]

@chfast ChainConfig does contain the info about the VM config, since the PR that you refer to does exactly that: it extracts the vmConfig from the chain config.

Once again, we agreed with @axic @karalabe and @holiman that fork rules were the way to go and your fix would activate WASM on pre-fork blocks, which must be prevented. Just setting EWASMBlock to 0 does the trick and it works fine with wagon.

[gballet]

Merged ethereum#17955 @cdetrio please check if you still see this issue.

[cdetrio]

it was fixed, thanks!