Although it is about to celebrate its 30th anniversary, open-source Linux is still the largest collaborative development project in computer history. The huge community surrounding Linux has allowed it to successfully complete many amazing tasks.
At the same time, developers seem to lack attention to security flaws. Kees Cook, a software engineer on Google’s open-source security team, pointed out in a blog post: There is a strong connection between code robustness and security. Given that bugs are already difficult to find, security flaws are even harder to reveal.
Google calls for the correct face of the security issues of the Linux kernel. For developers, obviously shouldn’t stop here. When the defect does appear, it is also necessary to deal with it effectively. By taking the first step, you can not only fix one bug at a time but also prevent the adverse effects caused by it. Linux written in C language will obviously face this problem for a long time.
Kees Cook pointed out that Linux must take more proactive security measures to protect itself from related risks. For example, the reason why a car is mandatory to be equipped with seat belts is not that we intend to crash the car, but because danger may come at any time.
Even if everyone wants to run a secure kernel on their computer, mobile device, or vehicle, not everyone has the ability to do something. Upstream kernel developers can fix bugs, but they cannot control what downstream vendors choose to integrate into their products. End users can choose the products they need, but usually, they cannot control which bugs are fixed or which kernels are used (sometimes the kernel itself is also a problem). In summary, the best way is to make suppliers responsible for the kernel security of their products.
Although many vendors have chosen to be an ostrich when they are attacked by malware, botnets, and defective software, Kees Cook still hopes that they can help. Under normal circumstances, these manufacturers treat their own devices as a physical products, rather than hybrid service products that require regular updates.
In addition to encouraging vendors to fix all bugs as early as possible, the Google security team also hopes to allow more engineers to participate in code reviews, security testing, toolchain development, and infrastructure improvements. According to their most conservative estimates, the current hundreds of engineers are still not enough to fully support the development of the Linux kernel and its toolchain.
But this upstream first approach to product kernel development and testing has actually proven to be quite effective. For example, on Chrome OS and Android development projects, Google has been implementing it for some time.
