How to understand the difference between transfer and function call using assembly call(g,a,v,in,insize,out,outsize)

Question

I understand that if there is an external function call, it will use all 7 parameters of the call(g,a,v,in,insize,out,outsize). However, I got a question I know low-level transfers like .transfer() will also use the instruction call(g,a,v,in,insize,out,outsize), if wei > 0, I can treat it as transfers, does that make sense? Meanwhile, should there still be 7 parameters for the instruction call?

Answer 1

CALL can be used for multiple purposes: mainly function calls and variants of sending ethers: transfer payable and call. To distinguish between function calls and ether transfer, the simplest way is to check if the input data is blank (the 4th item from the stack's top means input length so it's zero). If the CALL is sending ether it has zero input length. Otherwise, the length of the input follows abi encoding (e.g. 4 bytes if calling function without argument)

I used the ethervm.io annotation for the above :

Opcode CALL

Stack : gas | addr | value | argsOffset | argsLength | retOffset | retLength

• Value > 0 is not an indicator of transfer because it can be a call to payable function in the external contract (when argsLength > 0).
• With argsLength == 0, to distinguished between different variants of sending ethers, you can look at the gas forwarded (the first item from the stack). If it's 0x8fc, it's probably transfer or send. Otherwise it's direct call. Look at this article from solidity docs for the differences between variants for sending ether