Farewell ISDN, Ham Radio & Old Network Drivers: Linus Torvalds Merges 138k L.O.C. Removal

Linus Torvalds has done what many kernel developers have likely wanted for years: he merged the pull request to eliminate the old Integrated Services Digital Network (ISDN) subsystem and various legacy network drivers primarily from the PCMCIA era. This substantial code cleanup represents one of the largest single-line-of-code reductions in recent kernel history, trimming a massive 138,161 lines from the Linux codebase.

The timing of this removal isn't coincidental. With the recent influx of AI/LLM-generated bug reports against this dated code that likely has no active upstream users remaining, maintaining these drivers for obsolete hardware has become an unsustainable burden. The kernel community has traditionally kept drivers for old/obsolete hardware as long as there are active users and the code doesn't impose maintenance burdens. However, the combination of large language models and increased code fuzzing has turned these decades-old drivers into a liability.

Performance Impact Analysis

Removing this legacy code provides several measurable benefits:

1. Kernel Size Reduction: The 138k line reduction represents a significant decrease in kernel binary size, potentially improving load times and memory footprint.

2. Attack Surface Reduction: Fewer lines of code mean fewer potential security vulnerabilities, a critical consideration for enterprise deployments.

3. Build Time Improvement: The kernel compilation process will now be faster due to the reduced codebase.

The removed components include:

• ISDN subsystem: Once crucial for early digital telephony, now completely obsolete
• Legacy ATM device drivers: Asynchronous Transfer Mode hardware that never gained mainstream adoption
• Amateur ham radio support: While still relevant to hobbyists, the implementation was outdated
• Bus mouse support: Hardware that hasn't been common since the early 2000s

Specific Network Drivers Removed

The networking drivers removed represent a who's-who of 1990s networking hardware:

Manufacturer	Models	Era	Common Use Case
3com	3c509/3c515/3c574/3c589	Early 1990s	ISA/PnP Ethernet adapters
AMD	Lance/NMCLAN	Early 1990s	Workstation networking
SMSC	SMC9194/SMC91C92	Mid-1990s	PCMCIA Ethernet cards
Fujitsu	FMVJ18X	Mid-1990s	Laptop networking
Various	8390 AX88190/Ultra/WD80X3	1990s	Early Ethernet controllers

Build Recommendations for Modern Systems

For users building new systems or upgrading existing kernels, this cleanup has several implications:

1. Legacy Hardware Compatibility: Systems using these old drivers will need alternative solutions. For vintage hardware enthusiasts, consider:

   • Using older kernel versions that still support these drivers
   • Finding modern replacements that can be adapted with custom drivers
   • Moving to emulation solutions like QEMU for legacy networking

2. Performance Gains: Modern systems will benefit from the reduced kernel size, particularly in:

   • Boot time improvements
   • Memory footprint reduction
   • Faster kernel compilation times

3. Security Posture: The reduced attack surface provides inherent security benefits, especially for critical infrastructure.

Broader Kernel Cleanup Trends

This removal is part of a larger trend in the Linux kernel to streamline and modernize the codebase. Linux 7.1 also includes:

• Phasing out Intel 486 CPU support
• Removing support for Russia's Baikal CPUs
• Eliminating other legacy components that no longer serve active user bases

The kernel community faces an ongoing challenge: balancing support for legacy systems against the need to maintain a lean, secure, and modern codebase. As AI tools become more prevalent in identifying potential issues in older code, we may see more such cleanups in the future.

For kernel developers and maintainers, this merge sets an important precedent: when legacy code becomes a maintenance burden without active users, removal is a reasonable and beneficial option. The substantial line count reduction demonstrates that sometimes, less is indeed more in kernel development.

The kernel's evolution continues to prioritize modern hardware and use cases while strategically pruning code that no longer serves the community. This approach ensures Linux remains a competitive, efficient, and secure operating system for current and future computing needs.