During the recent Rust Week 2026 conference held in Utrecht, Greg Kroah-Hartman, one of the main people responsible for maintaining the Linux kernel, took to the stage with a very clear message: Rust is the salvation that the project needs.
With a staggering thirteen CVEs (Common Vulnerabilities) published daily , the development team faces an almost unmanageable volume of bugs. According to Kroah-Hartman, who has dealt with these vulnerabilities since 2005, the blame lies with the classic shortcomings of the C language in handling memory and errors —a problem that Rust promises to drastically alleviate.
The message was direct and straightforward before an audience packed with developers. The veteran programmer confessed that the team has complete faith in the Rust community to safeguard the foundations of modern computing. For them, the testing phase is over, as Rust's use in the kernel is now a reality, and they will move forward at full speed to integrate it into the most critical components.
The nightmare of C code and Rust's automatic solution
To illustrate the real problem they face daily , the developer recalled a flaw in the Bluetooth code that went unnoticed for fifteen years , caused simply by failing to check a memory pointer . He also mentioned a bug in the Xen code where someone forgot to release a lock . These small human oversights accumulate over time and end up causing system crashes or opening doors to cyberattacks. Rust's great advantage is that it catches these errors while the code is being compiled , before it reaches the review stage. If a programmer forgets to check for an error or release a resource, the program simply refuses to compile.
Thanks to this automatic verification , code reviewers no longer have to waste time searching for minute details and can focus on the program's actual logic. Kroah-Hartman estimates that these features alone will eliminate sixty percent of current kernel bugs. Furthermore, this new way of working has already influenced C code, forcing programmers to clean up their interfaces and adopt safer techniques, demonstrating that Rust's mere existence has already improved Linux.
Beyond memory locks, the talk addressed a vital security philosophy based on an old Microsoft maxim: assume that any data entering the system is malicious . Kroah-Hartman explained that they are working on a method within Rust to explicitly label data as "untrusted" within the type system itself.
In this way, the compiler will force the programmer to validate that information before it can be used or accessed . This is especially important today, as developers not only distrust user programs but also the hardware itself, which often comes with factory defects or may even be maliciously designed. By concentrating all validation in a single point, security checks will be much faster and more accurate.
Evolution without rewriting from scratch
Despite his enthusiasm, the expert made it clear that Rust isn't magic . As an example, he mentioned a small piece of code written in Rust designed to display a QR code when the system crashed, which ended up causing memory problems because no one thought to check the data buffer size. Rust can also fail if not used carefully and with common sense. For this reason, no one in the organization is considering rewriting the 36 million lines of C code that Linux already has. The rule is clear: the old code stays as is, and the new tools will only be used to create modern drivers.
Currently, the kernel consists of only about 113.000 lines in Rust , mostly temporary adaptations to bridge the gap between the two languages. However, massive projects like Android's internal communication system (Binder) are already in the midst of the transition , meaning that billions of phones will soon be running this code daily.
The project assumes this change as a natural evolution, where old C drivers will simply fall out of use over the years as new hardware demands programs created from scratch with modern security guarantees.