Send all gas in call

[chriseth]

Function calls usually provide all available gas to the callee because it is extremely hard if not impossible to come up with a good estimate of how much gas the callee will need.

At he same time, this mechanism relies on actual gas costs inside the EVM and is very fragile:

The amount of gas left can be retrieved via the gas instruction - the problem is that some gas is already spent again between this instruction and the actual call, most notably the gas for the call instruction itself (which can even vary depending on its arguments).

The usual way to cope with this is to subtract a certain constant from the value of the gas instruction (this constant includes the costs for performing this subtraction...).

This is of course all very fragile and means that if we ever increase the gas costs of the instructions call, gas, sub or of pushing constants, we will break existing contracts.

Furthermore, sometimes some gas has to be retained by the caller because it cannot efficiently estimate how costly the call opcode itself will be.

This is less an actual EIP but more a request for ideas about how to handle that situation.

Possibility 1: Use the special case of 2**256-1 gas to mean "all available gas". This would at least make future contracts more flexible.

Possibility 2: Change the semantics of push1 50 gas sub call (similar for other constants) to mean "all available gas" regardless of the result of the actual computation. That should at least keep the intended semantics of existing contracts.

[wanderer]

why not just add another opcode? call_all_gas or something?

[wanderer]

related, call's with relative gas source #61

[gavofyork]

dislike possibility 2; it would introduce an extremely opaque meaning to a particular opcode string.

there's also possibility 3: the gas limit given is capped at the amount of gas remaining. this is a similar to possibility 1, but simpler as it doesn't introduce a constant (2^256-1) or a conditional, but merely a trivial "limit" operation.

[chfast]

I like @gavofyork's idea, where call(G, ...) means call(min(G, gas), ...). With the current gas limit of 2^63-1 it is enough to use call(2^63-1, ...) to always use all available gas.

[karalabe]

Another bit more flexible alternative (if we're going with the constant approach) is to interpret 2-s complement as the amount of gas to retain. So 2^256-1 would mean 0 to retain (essentially the original proposal), but it would also allow me to specify 2^256-1 - 21000 to send all the gas, except the 21K. This could be useful to guarantee that no matter what an external contract does, the caller still has some minimum gas to meaningfully finish its execution.

[chriseth]

I guess this can be closed now.

[chfast]

I think we still should create the EIP document (by a PR) with status "obsolated by".