I am experiencing a persistent kernel-level storage lock on my external hard drives immediately upon a cold boot. The standard system tray "Safely Remove Hardware" option fails with a generic "Device in use" error under all conditions, despite being explicitly configured for "Quick Removal" with write-caching fully disabled.
A deep-dive multi-layer analysis reveals a severe architectural regression in the storage stack driver.
- The USB Controller Subsystem (USBSTOR.SYS) successfully communicates with the external bridge chip, identifies the topology as a PNPBus, and marks the hardware status as DN_REMOVABLE.
- However, the physical disk class driver (disk.sys, Version: 10.0.26100.9711, dated October 14, 2026) incorrectly overrides this data. It strips the physical layer of its Removable capability flags and writes "Legacy BusType: Internal" to the configuration state.
- Because the physical disk is falsified as an internal bus asset, the Windows Kernel treats the downstream logical volumes as permanent motherboard infrastructure. It permanently binds system storage threads (PID 4) to the NTFS change journals ($LogFile/$Mft), creating an unresolvable handle lock that causes the PnP Manager to veto all software ejection requests.
This has been confirmed globally across multiple distinct external enclosures and user accounts. The only functional method to drop the handle is manually taking the disk offline.
I have submitted a comprehensive multi-layer trace with logs directly to the Feedback Hub. Please upvote the public ticket so the Core Storage engineering team can address this regression:
Public Feedback Hub Link: https://aka.ms/AA13sxui OR: https://aka.ms/AA13sxui