Halfword by halfword

A locker's 64-bit code is 0xbeefcafef00dffff. Build it in code_r; the starter then adds a number n read from input and prints the code, n, and the sum.

One movCopy register or wide immediate into Rd. can only place a 16-bit piece of a number, a halfword, into a register. movzMove wide zero-extended; writes imm << hw*16, zeros the rest. writes one halfword and clears every other bit. movkMove wide, keep. Writes imm << hw*16 but preserves the other halfwords. writes one halfword and keeps the other bits as they were (the k is for keep). Add lsl 16, lsl 32 or lsl 48 to pick which halfword. The code splits into four:

Start with one movzMove wide zero-extended; writes imm << hw*16, zeros the rest., then add three movkMove wide, keep. Writes imm << hw*16 but preserves the other halfwords.. Try an input of 1: the low halfword rolls over from 0xffff to 0x0000 and carries into the next one, because the four pieces are one number.

What is checked

• with input 0 the line reads 0xbeefcafef00dffff + 0 = 0xbeefcafef00dffff
• the program exits cleanly
• the code is built with movzMove wide zero-extended; writes imm << hw*16, zeros the rest. and movkMove wide, keep. Writes imm << hw*16 but preserves the other halfwords., not written as one number
• the program also runs on inputs you do not see, so it has to work for any input, not only this one