Ciphertext-policy attribute-based encryption (CP-ABE) can offer fine-grained access control over encrypted data, which is suitable for complex commercial applications. However, since the same decryption privileges could be shared by multiple users in the one-to-many encryption mechanism, it is dangerous that a malicious user misuses his secret key but cannot be traced. In addition to further security, when the malicious user has been caught, it is required to revoke him from the system. To address these problems, we propose a novel updatable CP-ABE scheme supporting white-box traceability and traitor revocation. In the proposed scheme, a “fixed point” is embedded into the user's secret key to achieving the traceability and each user is assigned with a unique identifier for revocation. Moreover, the secret exponent used to encrypt a message is divided into two parts: one is assigned to access policy and the other to the revocation list. Therefore, only a part of the ciphertext components needs to be updated when the revocation list is changed, which greatly simplifies the process of ciphertext update. Compared to the previous works, our scheme is more efficient, and can achieve valid revocation and ciphertext update. In addition, the traceability of the proposed scheme is depended on the l-Strong Difffie-Hellman assumption, and the indistinguishability security under selective access policy and chosen-plaintext attacks in the standard model is reduced to the Decisional q-Bilinear Diffie-Hellman assumption. Furthermore, the experimental results show that the proposed scheme is efficient.