| Previous | Next |
| ERROR_VID_INVALID_PPM_HANDLE | ERROR_VID_MESSAGE_QUEUE_CLOSED |
ERROR_VID_MBPS_ARE_LOCKED
ERROR_VID_MBPS_ARE_LOCKED is HRESULT 0xC0370019 in the VID locked memory-block pages area of the Windows virtualization stack. The built-in message names the immediate result; the useful custom context is the exact boundary: an operation that requires memory-block pages to be unlocked while one or more MBPs remain locked. Record the first returning operation and host-side event before a management layer retries or translates it.
Locate the failing boundary
The requested transition conflicts with active page locks. Identify the lock owner and purpose; forcing the transition can violate DMA, mapping, or notification assumptions.
Hyper-V architecture documentation identifies VID as the component that provides partition, virtual-processor, and memory-management services., many ERROR_VID_* values describe internal host objects rather than a public API that administrators should call directly. Accordingly, diagnose this result through the named object/state, the VMMS or Worker event chain, and the operation that produced it; do not invent a user-mode VID call from the constant name.
Neighboring result: LOCK_COUNT_EXCEEDED is a capacity failure while acquiring locks. MBPS_ARE_LOCKED is a state conflict encountered by another operation.
Build an incident record
| Evidence | Why it changes the diagnosis |
|---|---|
| Locked page set | Record GPA/pages, lock flags, owner, and acquisition time. |
| Conflicting operation | Keep property, notification, destroy, or remap request that needs unlocked pages. |
| Completion dependency | Identify I/O, VP exit, or asynchronous work holding the lock. |
| Unlock path | Capture whether cancellation and error handling release it. |
Preserve identifiers and counts without dumping guest secrets or unrelated memory. Useful this result timestamps include the last successful operation, first failure, any automatic retry, and the object-generation change that followed.
Run narrow checks
- Wait for the owning operation to complete or cancel through its supported path.
- Do not reuse the block while DMA or mapping work is active.
- Add lock-owner diagnostics and maximum age alerts.
- Inject cancellation at each asynchronous stage.
Use a controlled partition or disposable VM when a mapping, lock, or teardown test can be destructive. During the result run, keep page units, GPA boundaries, parent object, handle generation, and release order visible.
Interpret three controls
Across these controls for it, preserve an operation that requires memory-block pages to be unlocked while one or more MBPs remain locked as the boundary under test.
| Control | Interpretation | Hold constant |
|---|---|---|
| Balanced fresh lifecycle — this result | Create, use, release, and destroy one object generation. A different this result exposes leaked references, locks, handles, or mappings. | Preserve sizes and operation order while changing only generation. |
| Same request under lower host pressure — it | If it changes after controlled pressure is removed, quantify host/pool capacity; if not, focus on caller bookkeeping. | Do not modify guest disks or unrelated VM devices during the result comparison. |
| Smaller exact range — it | Reduce only page count or mapped extent. If it moves, range arithmetic, object limits, or reserve capacity is involved. | Keep the same block, GPA base, flags, and partition generation for it. |
Keep neighboring states separate
Restarting can release the exact lock, reference, mapping, or opaque handle behind it without fixing the lifecycle defect. Before recovering from it, preserve the object graph and acquire/release counts.
A handle observed is generation-bound. Keep its parent partition and object type beside the token; never serialize it, copy it to another partition, or reuse it after the terminal transition.
Prove the correction
The conflicting operation runs only after all intended locks release, and no lock survives cancellation or teardown. Repeat the original operation under the original supported conditions and retain one deliberate negative control. A management command succeeding on a different object is not sufficient to close this incident.
Technical references
These sources define it and the public Hyper-V architecture surrounding the internal state. They do not create a public user-mode VID API for the named object.
- Microsoft Open Specifications: HRESULT values — used to interpret the boundary.
- Microsoft: WHvMapGpaRange — used to interpret the boundary.
- Microsoft: WHP memory data types — used to interpret the boundary.
- Microsoft: Hyper-V architecture — used to interpret the boundary.
Looking for a different code? Search another status or error code.