Abstract Recently, the concept of committing message authentication codes (MACs) was proposed at Crypto 2024, following the definition of committing authenticated encryption (AE). A secure committing MAC requires that it is infeasible for any adversary to find two different key and message pairs that generate the same tag. In this paper, we present committing attacks against PMAC, LightMAC, and UMAC, demonstrating that these MACs are not CMT secure. We also provide a CMT-K security proof for UMAC. Furthermore, we introduce two general methods for constructing one-key and two-key MACs with CMT-K security.

The Learning with Rounding (LWR) problem, introduced as a deterministic variant of Learning with Errors (LWE), has become a promising foundation for post-quantum cryptography. This Systematization of Knowledge (SoK) article presents a comprehensive survey of the theoretical foundations, algorithmic developments, and practical implementations of LWR-based cryptographic schemes. We introduce LWR within the broader landscape of lattice-based cryptography and post-quantum security, highlighting its advantages such as reduced randomness, improved efficiency, and enhanced side-channel resistance. We explore the evolution of security reductions from LWR to LWE, including recent advances that support practical parameter regimes and address challenges in both bounded and unbounded sample settings. This article systematically reviews existing LWR-based schemes — including Saber, Lizard, Florete, Espada, Sable, and SMAUG — analyzing their design choices, parameter sets, and performance tradeoffs. Furthermore, we examine the impact of LWR on side-channel resistance, failure probabilities, and masking efficiency, demonstrating its suitability for secure and efficient implementations. By consolidating the research spanning theory and practice, this SoK aims at guiding future cryptographic design and standardization efforts leveraging LWR.

The Linux capabilities mechanism divides the root privileges to provide more fine-grained access control, but its effectiveness depends on proper implementation and configuration. The scattered enforcement of capabilities in the kernel and its sporadic usage in programs pose challenges in gathering assessment information. To address this, we propose three tools for diagnosing potential problems in its design, implementation, and utilization. First, we employ LLVM/Clang to examine the capabilities enforcement in the kernel to map capabilities checks to files. This is the first attempt to explore the interaction between capabilities and other mechanisms, such as UGO. Second, We propose a pattern-based method to identify the sensitive kernel functions protected by capabilities, quanti-fying the overlap problem of capabilities. Third, we employ a customized fuzzing approach to determine the minimal set of capabilities required by programs, offering insight for secure usage. Additionally, Our study is further guided by international access management standards, providing structured criteria for the assessment. Leveraging data collected by our tools, we identify imperfections of capabilities and reported to stakeholders. To the best of our knowledge, this is the first systematic assessment of Linux capabilities.

Abstract We present practical attacks on IAR-CTR and IAR-CFB, AE modes proposed by Hwang and Gope to offer both privacy and authenticity in a single pass. Then we use the idea of iCOFB, which is originally defined with a tweakable random function, to fix the IAR-CTR mode using the tweakable block cipher, and propose the TIAR-CTR mode, which is a one-pass AEAD mode with provable security. However, the same fixed method does not work for the IAR-CFB mode.