Linus Torvalds creator and leader of the development of the Linux kernel, has once again demonstrated that quality and punctuality are non-negotiable in the development of the Linux Kernel.
And is that during the Linux 6.17 development cycle, Google engineer Palmer Dabbelt submitted a set of patches for the RISC-V architecture that did not pass the filter from the top developer. The reason: they were submitted too late and contained changes considered "garbage" that didn't belong in the RISC-V tree.
This decision means that the proposed improvements will not arrive in Linux until version 6.18, which could slow the adoption of RISC-V in certain environments and hinder developers who rely on these updates.
For those who are still unfamiliar with RISC-V, you should know that this It is a free and open instruction set architecture Born at the University of California, Berkeley, in 2010, RISC-V has since gained ground against proprietary architectures such as ARM and x86, thanks to its flexibility and a growing developer community. RISC-V International currently has more than 4500 members.
Despite this growth, the architecture faces challenges of maturity and stability. Cases like Chimera Linux, which temporarily suspended support due to performance issues on available hardware, reflect the difficulties of establishing itself in production environments.
A rejection marked by time and code quality
Torvalds had requested that pull requests be submitted in advance because he would be traveling. However, the update arrived on August 8, 2025, just two days before the Linux 6.17 merge window closed, scheduled for the release of version -rc1.
The project leader not only criticized the delay, but also the contentHe pointed out that the changes added unnecessary modifications outside the scope of RISC-V, affecting generic files and adding no value. In his own words, this "makes the world worse."
Reactions and the unmistakable style of Linus Torvalds
The episode with the RISC-V patches is not an isolated one. In June 2025, Torvalds had already been involved in a controversy by removing the Bcachefs file system from the kernel due to concerns about its quality and maintainability.
This time, his message was straightforward: no more late submissions or irrelevant changes. Palmer Dabbelt apologized and promised to correct the process in future releases.
The community has reacted with mixed opinions. Some value his frankness as an effective way to maintain the kernel's quality; others, however, believe his tone can discourage new contributors. However, Torvalds maintains his position: the stability and security of the kernel come before all other considerations.
In more than three decades at the helm of the Linux kernel, Linus Torvalds has gained fame for his direct, sometimes aggressive, communication style.Although he has acknowledged the need to moderate his tone in the past, he continues to prioritize technical rigor over diplomacy.
For him, decisions such as introducing case-insensitive file systems or out-of-context changes to the codebase represent unnecessary risks that could compromise the system's long-term stability.
This new rejection of RISC-V patches is, in essence, a reaffirmation of the philosophy that has guided the Linux kernel for decades: quality is non-negotiable, and deadlines are met.
Btrfs performance improvements and other new features for Linux 6.17
Other new features and improvements that have been prepared for Linux 6.17 are in performance and functionality of the btrfs file system. Among the most notable new features is experimental support for large folios, an optimization that promises to reduce overhead and accelerate operations in multiple scenariosAlthough this feature is technically ready for wider use, the developers caution that it hasn't yet been tested enough to be considered completely stable.
The new version of Btrfs tIt also introduces a denser placement of keys in the XArray structure., which increases the compactness of the storage spanning tree nodes. This change can reduce the number of end nodes by 50% to 70%, thus optimizing the internal structure and reducing resource consumption. Furthermore, Defragmentation adjustments have been applied via ioctl, along with the activation of a previously designed mechanism to prevent critical file system corruption, which limits writing to block devices with a mounted file system.
In addition to this, it also highlights the implementation of a caching system for requests to free space-allocated bitmaps. In tests using empty file creation, this optimization enabled performance increases of up to 20%, as well as notable improvements in environments with heavy metadata loads. Similarly, read-ahead has been improved on systems using data compression, improving data access speed.
Finally, if you are interested in knowing more about it, you can consult the details in the following link