AES-CTR IV broken(!) - IV is not increasing

[DopeforHope]

Hey,

Didn't expected this but the AES-CTR implementation has a flaw. The counter/IV is not increasing at all.

Critical code snippet:

function AESCTR(blocks::Array{UInt8, 1}, key::Array{UInt8, 1},
	iv::Array{UInt8, 1})
	noBlocks = keyStreamCheck(blocks, key, iv)
	curr = copy(iv)
	o = Array{UInt8}(undef, length(blocks))

	for i=1:noBlocks
		indices = blockIndices(blocks, i)
		eb = AESEncrypt(curr, key)
    	o[indices] = xor.(eb[1:length(indices)] , blocks[indices])
		for bi=(length(curr) - 7):length(curr)
			tmp = curr[bi]
			curr[bi] = UInt8(Int(tmp) + 1)
			if curr[bi] > tmp
				# no byte overflow
				break
			end
		end
	end

	return o
end

If one puts a print before the AESEncrypt call:

function AESCTR(blocks::Array{UInt8, 1}, key::Array{UInt8, 1},
	iv::Array{UInt8, 1})
	noBlocks = keyStreamCheck(blocks, key, iv)
	curr = copy(iv)
	o = Array{UInt8}(undef, length(blocks))

	for i=1:noBlocks
		indices = blockIndices(blocks, i)
		println("######LIB Iv: ", curr)
		eb = AESEncrypt(curr, key)
    	o[indices] = xor.(eb[1:length(indices)] , blocks[indices])
		for bi=(length(curr) - 7):length(curr)
			tmp = curr[bi]
			curr[bi] = UInt8(Int(tmp) + 1)
			if curr[bi] > tmp
				# no byte overflow
				break
			end
		end
	end

	return o
end

You will see that the IV is not increasing. The error lies perhaps in the scopes of variables. I'm not sure but this may have been changed from 0.6 to 1.0?

JuliaLang/julia#28750 seems to be something similiar.

Version: Julia 1.0.2 (LTS)

[DopeforHope]

Bullshit