Linux 7.3-rc2 Released with Hybrid CPU Improvements and Numerous Fixes
Linux 7.3-rc2 released with hybrid CPU scheduling and Nouveau Blackwell display fixes, heading toward stable in late October.
Development of Linux 7.3 has entered its second phase. One week after the end of the merge window, Linux 7.3-rc2, the second release candidate, has been released. The stable release is expected in late October. In reporting by Phoronix’s Michael Larabel, it was described as a busy weekly candidate.
A busy weekly candidate just after the merge window, with an influx of AI- and LLM-generated material running in parallel with fixes for issues found in rc1 testing
Fixes have accumulated across the entire source tree, which now approaches 41 million lines. The focus is on addressing regressions found in early testing. Greg Kroah-Hartman, the deputy maintainer of development, commented on the turbulent nature of this cycle. He cited the high volume of AI- and LLM-derived activity as a factor.
Role of the Second Candidate Toward Stable
The Linux 7.3 merge window closed last weekend. Issues detected during rc1 testing are being triaged. rc2 is the stage for consolidating those fixes. The emphasis is on stabilization rather than adding new features.
The release is available from kernel.org. Distributors and testers are asked to perform regression testing on real hardware. As shown in Linux 7.2-rc2 Strengthens BPF JIT Hardening, rc2 serves as a consolidation point for safety fixes. It plays the same role this time. Weekly candidates are expected to continue iterating toward the stable release in late October.
Including the graphics-related support reported in Linux 7.3 Brings Stable Support for Intel Nova Lake S Integrated GPU, the functional foundation is taking shape. From rc2 onward, refining quality will be the challenge. Operators need to verify bootability and performance on their target configurations.
Scheduling Fixes for Hybrid CPUs
One highlight is a fix to Cache Aware Scheduling. It addresses a misfit handling issue on hybrid CPUs. The use of performance cores and efficiency cores will be better balanced. Task placement decisions will become more appropriate.
In hybrid configurations, differences in core characteristics directly affect performance. Improper placement causes latency and wasted power. This fix aims to reduce those losses. Measuring its effect under real workloads will be the focus going forward.
As reported in Linux Cache Aware Scheduling Expansion Speeds Up MySQL by up to 360%, the relaxed scheduling extension showed strong gains under commercial workloads. This fix is positioned as continued stabilization along that line. The scheduling infrastructure proves its worth under large-scale concurrency. Comparative measurements in test environments are essential.
Tree-Wide Cleanup of Memory Allocation
As a tree-wide cleanup, replacement of kmalloc() calls has progressed. This includes migration to the more explicit kmalloc_obj(). The aim is to clarify the correspondence between types and usage. Improved maintainability and verifiability are expected.
Memory allocation is a fundamental operation used throughout the kernel. Unifying the call format improves the accuracy of static analysis. Misuse and leaks become easier to detect. This should lead to lower defect density over the long term.
The replacement may look mechanical. But because the scope is broad, care is needed to avoid regressions. Build and boot verification is required in each subsystem. The rc2 test items span a wide range.
Nouveau and Error Detection Correction Updates
For graphics, a fix to the Nouveau driver was included. It addresses a display issue with NVIDIA Blackwell. Work continues despite limited public specifications. Improved display stability is expected for users.
Together with the Intel support covered in Linux 7.3 Brings Stable Support for Intel Nova Lake S Integrated GPU, the graphics infrastructure is being improved. Maintaining the open drivers remains highly constrained. Sharing test results is supporting progress in development.
Changes to EDAC, error detection and correction, were also merged. Additions that missed the merge window were included. They concern the accuracy of memory error reporting and logging. This is an area that affects operational stability for infrastructure machines.
Rust Support and Security Feature Adjustments
A default change was also made when Rust support is present. It disables the RandStruct security feature by default. This applies to configurations where the Rust toolchain exists. The measure is intended to keep the build system consistent.
RandStruct is a feature that randomizes struct layouts. It aims to make prediction by attackers more difficult. On the other hand, it imposes constraints on toolchains and cross-language interoperability. The default adjustment strikes a balance with practicality.
The introduction of Rust is expanding options for driver development. Ensuring memory safety is the main motivation. Linking build environment detection with defaults is important. Distributors need to check documentation and configuration updates.
Increased AI-Generated Contributions and
Strain on Development
A defining feature of this cycle is the increased influx of AI- and LLM-generated material. rc2 also contains much related material. The growing volume of changes increases the burden of review and testing. This is the background to the deputy maintainer’s description of a turbulent cycle.
Generated contributions speed up writing. However, they risk unclear provenance and duplication. Maintainers’ review capacity becomes a constraint. The design of quality assurance processes is being forced into review.
In the case of Linux 7.2-rc2 Strengthens BPF JIT Hardening, scrutiny of security measures took effort. The rise in auto-generated contributions reinforces that trend. Test automation and unified reporting formats are essential. This is an issue concerning the sustainability of collaborative development.
Editorial Opinion
In the short term, we expect the increased verification burden to spill over to operations. Reviewing fixes from rc2 onward and the effort for regression testing will increase. Deployment decisions just before the stable release will require caution. In the long term, we expect acceptance criteria for AI-generated changes to be established. Provenance management for generated contributions and test automation will become development norms. How quality assurance costs are borne may change. The remaining issues are how to identify generated contributions and allocate responsibility. The effect of hybrid CPU optimization needs to be backed by actual measurements. The impact of the security feature default change on operational policy is in question.
References
- “Linux 7.3-rc2 Released Following Another Busy Week”, by Michael Larabel — Phoronix, 2026-09-06T22:22:25.000Z (ARR)
- Source URL: https://www.phoronix.com/news/Linux-7.3-rc2-Released
Frequently Asked Questions
- What is the status of the Linux 7.3-rc2 release?
- It is the second weekly test release after the end of the merge window. Rather than adding new features, it focuses on fixing issues found in rc1 and stabilization. The stable release is expected in late October.
- What are the main fixes in this release?
- They include a misfit fix for scheduling on hybrid CPUs, a tree-wide replacement with kmalloc_obj(), a Blackwell display fix in Nouveau, EDAC additions, and disabling RandStruct by default when Rust is present. Fixes were consolidated across the entire source tree of nearly 41 million lines.
- Why is the increase in AI- and LLM-generated material a challenge?
- Because the growing volume of changes increases the burden of review and testing. Without provenance management and deduplication mechanisms keeping pace, regressions may be missed. Acceptance criteria and automated testing need to be established.
Comments