Linus Torvalds released Linux 7.1-rc6 on Sunday. The candidate arrives with another sizable batch of fixes. He called it smaller than rc5 yet still larger than he’d wish for at this stage.
Developers and testers now have fresh code to run through its paces. The stable Linux 7.1 release remains targeted for mid-June. But the persistent flow of changes keeps the schedule from feeling certain.
Torvalds pulled in updates across drivers for GPUs, networking, USB, serial, sound and SCSI. Core networking code saw touches. So did selftests, architecture work on x86, MIPS and arm64 KVM, filesystems including SMB and NFS, memory management and live update features. The spread shows how late-stage polishing touches many corners of the kernel at once.
Persistent Fixes and the Shadow of AI Assistance
This marks the latest in a pattern. Earlier candidates in the 7.1 cycle also arrived heavier than historical norms. Phoronix reported that rc5 carried lots of AI and LLM-assisted fixes. Networking pull requests stood out as particularly large. The trend raises questions about how automated code generation and review now shape kernel development.
But the changes themselves don’t trigger alarm. Most count as small, targeted corrections for issues users reported or scanners uncovered. Torvalds has noted the volume without signaling any show-stopping bugs. Still, the consistency across recent release candidates leaves him cautious about declaring the cycle fully settled.
One concrete late addition hides documentation for the clearcpuid feature. Another brings official support for the ASUS ROG RAIKIRI II and Nova 2 Lite game controllers. These arrive alongside broader driver maintenance.
USB updates stand out in the immediate run-up to rc6. Greg Kroah-Hartman sent a pull request filled with tiny fixes and new device IDs. “A set of USB fixes and new device ids for 7.1-rc6. Nothing major in here, just lots of tiny fixes for reported issues found by users and some older patches found by some scanning tools,” he wrote, according to Phoronix coverage of the request.
Specific USB quirks address real hardware headaches. The PNY Elite Portable SSD relies on a Phison UAS bridge. Older firmware failed on READ CAPACITY commands and blocked full UAS throughput. The new quirk corrects enumeration so the drive performs as expected. Lenovo ThinkPad USB-C Dock Gen 2 users saw repeated disconnect and reconnect cycles tied to USB Link Power Management instability. Disabling LPM resolves it. Device Tree adjustments also fix USB reset behavior on the EIC7700 RISC-V SoC.
And the cycle began with bigger shifts. The 7.1-rc1 announcement from Torvalds on April 26 closed the merge window after roughly 13,000 non-merge commits plus 1,000 merges. LWN.net carried the full posting. Changes touched drivers, architecture, filesystems, networking, core kernel and documentation. The kernel dropped support for ancient Intel 486 processors. It introduced a rewritten in-kernel NTFS driver that finally offers reliable write performance without depending on the old FUSE-based solution. Tom’s Hardware highlighted the storage upgrade as one of the more significant under-the-radar improvements for users who move data between Linux and Windows systems.
Networking maintainer Jakub Kicinski even dealt with an “LLM-pocalypse” of sorts. AI-generated bug reports prompted the deletion of 138,000 lines of code in one pull request that Torvalds merged into rc1, according to reporting on the episode.
So what does this all mean for the kernel’s future? The 7.1 series builds on the just-released 7.0. That earlier cycle saw similar late-stage patch spikes. Torvalds observed that an apparent calm at rc5 for 7.0 turned out to be temporary. The rc6 for 7.0 carried far more fixes than typical. He linked the phenomenon in part to improved AI coding tools that surface more edge cases and generate more patches.
Kernel developers now work in an environment where large language models scan codebases, suggest fixes and sometimes generate reports that force real code deletions. The volume of small fixes in late release candidates may reflect better tooling rather than worsening quality. Yet the numbers still make Torvalds pause. He offers no firm promises on exact timing even as rc7 approaches.
Testers should treat rc6 as exactly that. A candidate. Production systems stay on stable releases. But for those tracking mainline, the tree at git.kernel.org now holds the latest. Hardware vendors, driver authors and filesystem teams will continue to watch for regressions in their areas.
The Linux kernel’s development process has always balanced new features against stability. This cycle shows the balance shifting under the weight of automation. More eyes, more tools and more patches arrive earlier and persist longer. Whether that leads to stronger final releases or simply longer stabilization periods remains an open question the next few weeks will help answer.
Recent coverage reinforces the trend. 9to5Linux noted the rc1 launch and the expected mid-June stable target. Discussions on X this weekend echoed the release announcement from accounts tracking kernel news.
By the time 7.1 ships, the accumulated fixes should produce a tighter kernel. The question is how many more Sundays of sizable updates stand between here and that point. Torvalds will tag the next candidate soon. The industry will keep watching the size of the diff.
WebProNews is an iEntry Publication