EVM-C: introduce new status codes #173
Conversation
|
cc @ehildenb |
| enum evm_status_code { | ||
| EVM_SUCCESS = 0, ///< Execution finished with success. | ||
| EVM_FAILURE = 1, ///< Generic execution failure. | ||
| EVM_OUT_OF_GAS = 2, | ||
| EVM_BAD_INSTRUCTION = 3, | ||
| EVM_UNDEFINED_INSTRUCTION = 3, ///< Unknown instruction encountered by the VM. |
There was a problem hiding this comment.
We should also add EVM_INVALID_INSTRUCTION, right?
There was a problem hiding this comment.
I knew I've forgotten something 😉
| /// Tried to read outside memory bounds. | ||
| /// | ||
| /// An example is RETURNDATACOPY reading past the available buffer. | ||
| EVM_INVALID_MEMORY_ACCESS = 10, |
There was a problem hiding this comment.
Can you also add a note that this could be thrown by VMs instead of OUT_OF_GAS if something like CODECOPY access outside the range of addressable local memory? (because it uses index + offset addressing). I think having more specific error codes is a plus.
There was a problem hiding this comment.
Can you also add a note that this could be thrown by VMs instead of OUT_OF_GAS if something like CODECOPY access outside the range of addressable local memory?
I can add the comment, but not sure a compliant VM actually can do that, since the specification says memory expansion must be done first and if that fails it is an OOG case.
There was a problem hiding this comment.
Hmmm, I suppose that we would have to make an EIP then? Then never mind, we can worry about that with the design of ewasm which will have more cases where INVALID_MEMORY_ACCESS can happen anyway.
There was a problem hiding this comment.
It definitely makes sense for ewasm and we should preferably try to clean up the mess EVM is as a second task 😉
|
I will update KEVM to use the |
Fixes #168.