Skip to content

[BugVirtualBox 7.2.10: Rapid consecutive AHCI HBA resets during guest boot cause Windows 11 guest BSOD (CACHE_MANAGER, 0x34) via storport.sys]: #725

Description

@DEHughes

Version

7.2.10

Host OS Type

Linux

Host OS name + version

SuSE Leap 16.0

Host Architecture

x86

Guest OS Type

Windows

Guest Architecture

x86

Guest OS name + version

Windows 11

Component

Host Support

What happened?

On a Windows 11 Pro (build 26100) guest running under VirtualBox 7.2.10 on Linux (openSUSE Leap 16.0, kernel 6.12.0, AMD Zen 4 host), the AHCI controller emulation issues two full HBA resets within approximately 1.1 seconds of each other, very early in the guest boot sequence (under 2 minutes after VM start, before the desktop session was fully established). Each reset cancels all in-flight requests across all 5 attached virtual disks. Shortly afterward the guest BSODs with bugcheck 0x34 (CACHE_MANAGER), with storport.sys identified as the only driver referenced in the resulting crash dump.

This is one instance of a recurring pattern of single-port and full-HBA AHCI resets observed across many sessions over several weeks, with guest-side consequences ranging from silent freezes to BSODs with several different bugcheck codes (0x7E, 0xDE, 0xA, 0x1A, 0x1E, 0x34). This report documents the most precise and reproducible instance captured to date.

Ruled-out variables

The following were tested and do not appear to be necessary conditions for the AHCI reset pattern or the resulting guest crashes:

AHCI BlockCache: Disabled via VBoxInternal/Devices/ahci/0/LUN#n/Config/BlockCache=0 for all 5 LUNs and run for a 31-hour session. Same reset pattern and eventual host-level lockup occurred regardless.
Host filesystem fragmentation: The instance documented in this report occurred immediately after the host-side virtual disk image files were defragmented (system disk image extent count reduced from 719,602 to 7,227 via btrfs filesystem defragment) and with the host's btrfs discard queue essentially empty (<1 GB pending) at the time of the crash. Fragmentation may be a contributing/accelerating factor in other sessions but is not necessary to reproduce this failure.
Third-party software: Audio passthrough, antivirus (Vipre/SBAMSvc), and backup software (JungleDisk) were all idle or absent during multiple reproductions; none of these are required to trigger the pattern.
btrfs filesystem corruption: btrfs scrub on the underlying filesystem reported no errors at the time of this and other incidents.
VirtualBox package source/version: Confirmed running current 7.2.10; the issue predates and persists across at least one point-release upgrade.

How can we reproduce this?

062426-5062-01.dmp

Wren11B_freeze_20260624_122519.zip

Did you upload all of your necessary log files, screenshots, etc.?

  • Yes, I've uploaded all pertinent files to this issue.

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions