The Corporate Truck Factor: anonymized cregit token-authorship databases for three Linux Kernel subsystems (amdgpu, net, iio)

cregit token-level authorship databases for three Linux Kernel subsystems: amdgpu (drivers/gpu/drm/amd), net (net), and iio (drivers/iio), chosen to span a spectrum of corporate concentration. Companion dataset for the paper 'The Corporate Truck Factor: Firm-Level Knowledge Concentration in Three Linux Kernel Subsystems' (VEM 2026). Generated with cregit (German et al.) token-level authorship attribution over full git history since the kernel 2005 migration to git. Coverage: 12.6M surviving content tokens across about 4,600 source files. Each subsystem ships the standard cregit 4-file SQLite set: blobmap.db, cregit.db, original.db, persons.db. All pass PRAGMA quick_check = ok. See MANIFEST.txt.