Symmetric key cryptography is an essential part of communication systems, where a secret key is used to protect data confidentiality. Surprisingly, the only way of trusting these ciphers is to perform continuous analysis that update the security margin. With the advent of quantum computers in an arguably near future, the security of nowadays ciphers has been put into question. While most currently used asymmetric primitives would be completely broken, doubling the key size of symmetric constructions provides the same level of security with respect to exhaustive key search. However, we still have a long way to go in the field of quantum cryptography and further cryptanalysis must be carried out to reassure the validity of these emerging ciphers. We have studied the Salsa20 family of ciphers, which has received very little cryptanalysis ever since the most relevant result one decade ago despite the inclusion of this cipher suit in TLS 1.3. Before providing the first quantum attack we tried to improve the best classical ones. The most remarkable contribution to the stated problem is a new faster attack on 8 rounds of 256 bit key Salsa combining conditional cryptanalysis and the novel idea of forward PNBs, which outperforms the actual state-of-the-art with 8 times faster time complexity.