| Previous | Next |
| ERROR_SIGNAL_REFUSED | ERROR_NOT_LOCKED |
ERROR_DISCARDED
What ERROR_DISCARDED means
ERROR_DISCARDED is Win32 system error code associated with the documented message: “The segment is already discarded and cannot be locked.” The significant condition is that a movable legacy memory segment has already been discarded, so a later lock request has no backing storage to pin. Preserve the symbolic name together with the numeric value because older diagnostic tools may display only one form.
The term “segment” here refers to movable memory used by historical Windows and OS/2-style programming models. In a the case, this is not the same as a modern virtual-memory region reported by tools such as VirtualQuery.
Likely causes
- a 16-bit component retained a stale selector or segment handle after memory pressure
- cleanup discarded a movable segment while another path still expected to lock it
- an emulator or compatibility layer exposed lifecycle behavior that the old application did not anticipate
Diagnostic procedure
- capture the segment handle at allocation, discard, and lock points
- verify whether the failure occurs only after memory pressure or task switching
- test the component in the intended compatibility subsystem and compare its segment-lifetime trace
During a this result investigation, capture the first operation returning it. In the result timeline, a later cleanup failure can be easier to notice while no longer describing the original defect. The evidence set for it should record executable path, file version, architecture, Windows build, compatibility settings, and non-secret inputs used by the failing operation.
How to interpret the result in modern software
This result belongs to the 0–499 system-error range, but its wording may describe a historical subsystem. Do not classify it automatically as a current Windows kernel defect. First determine whether the value came directly from GetLastError(), was translated from another status domain, arrived over a protocol, or was stored by an old application. Because it can be propagated through wrappers, a missing producer API leaves translation errors and stale last-error values indistinguishable.
Code-specific investigation notes
A discarded movable segment has lost its backing contents but may still be represented by a handle. Code that treats the handle alone as proof of valid storage can therefore fail later.
Look for memory-pressure callbacks and discardable-resource policies in addition to explicit discard calls. The segment may have been released by the compatibility memory manager.
A correct repair re-establishes the data or prevents premature discard; repeatedly locking the stale handle cannot reconstruct the segment contents.
Developer guidance
Code handling this result should check the exact API return first, copy the last-error value immediately, and avoid intervening calls before logging. If it is raised by a loader or compatibility helper, collect parent-process diagnostics because the child may never initialize its own logger. Validate format-specific fields with an appropriate parser; byte-level edits made only to suppress it can turn a clean rejection into corruption or unsafe execution.
Administrator and support guidance
To recover from this result, prefer a matched trusted binary/configuration set over individual DLL downloads or global compatibility changes. Before replacing an artifact associated with it, retain it and calculate a cryptographic hash. When it is isolated to one account or session, compare mappings, namespaces, environment, current directory, and policy before considering system-wide reinstall.
Example incident
A compatibility component retains a stale segment lifecycle state across an error path. The next memory operation encounters that state and returns it.
Difference from related errors
ERROR_LOCKED says a segment is still locked and cannot be reallocated; it is the opposite lifecycle problem—the segment contents are already gone when code tries to lock them.
Evidence to collect
- the raw decimal and hexadecimal value corresponding to it
- the exact API, command, or loader action that first produced the result
- paths after normalization and redirection, plus hashes and versions of involved binaries
- process architecture, session identity, compatibility mode, and relevant virtualization details
- a timestamp that can be correlated with application logs, Process Monitor traces, and Windows Event Log
Recovery and verification
A it repair is verified only when the original operation succeeds with equivalent inputs and produces usable output. After correcting it, repeat the action in a fresh process and again in the same workflow to test both stale-state removal and repeatable cleanup. If reboot alone removes it, collect enough evidence to identify which mapping, lock, process, or compatibility state the reboot cleared.
References
Looking for a different code? Search another status or error code.