Recent patches to the Linux kernel introduce support for larger block sizes in the EXT4 file system, exceeding the typical 4KB page size. Submitted by contributor Matthew Wilcox, these changes aim to boost performance in storage-intensive applications by up to 30%. The development builds on prior Virtual File System modifications and follows similar updates in XFS and Btrfs.
The patches enable EXT4 to handle block sizes up to 64KB or beyond, aligning with modern hardware like NVMe drives. This reduces overhead in read and write operations, particularly beneficial for data centers managing massive datasets in cloud providers and AI training clusters. Tests referenced in reports indicate "stupendous performance" gains, with up to 30% improvements in throughput for large-file workloads.
EXT4, which evolved from EXT3 in 2008, was initially constrained by page-size norms for compatibility. The enhancements stem from multi-year efforts on the Virtual File System (VFS) layer, initiated last year, involving maintainers like Theodore Ts’o. Engineers from companies such as Google and IBM have advocated for these changes to overcome bottlenecks on x86 architectures.
While promising for enterprise SSDs by minimizing fragmentation and improving sequential access, challenges include potential space waste for small files. Developers have addressed this with optional configurations for tailored use cases, incorporating safeguards against data corruption via EXT4’s journaling.
Integration is expected in upcoming kernels, potentially Linux 6.13 or later, with production availability in distributions like Red Hat and Ubuntu by mid-2026. A kernel expert noted in discussions that this represents a step toward "fully utilizing hardware potential without software bottlenecks." These updates support hybrid storage setups and could accelerate adoption in financial services and scientific computing, enhancing efficiency for virtualization and containerized workloads.