CF_OPERATION_PARAMETERS structure (cfapi.h)

Detailed parameters of an operation on a placeholder file or folder. The data provided in this structure is specific to the CF_OPERATION_TYPE of the operation.

Syntax

typedef struct CF_OPERATION_PARAMETERS {
  ULONG ParamSize;
  union {
    struct {
      CF_OPERATION_TRANSFER_DATA_FLAGS Flags;
      NTSTATUS                         CompletionStatus;
      LPCVOID                          Buffer;
      LARGE_INTEGER                    Offset;
      LARGE_INTEGER                    Length;
    } TransferData;
    struct {
      CF_OPERATION_RETRIEVE_DATA_FLAGS Flags;
      LPVOID                           Buffer;
      LARGE_INTEGER                    Offset;
      LARGE_INTEGER                    Length;
      LARGE_INTEGER                    ReturnedLength;
    } RetrieveData;
    struct {
      CF_OPERATION_ACK_DATA_FLAGS Flags;
      NTSTATUS                    CompletionStatus;
      LARGE_INTEGER               Offset;
      LARGE_INTEGER               Length;
    } AckData;
    struct {
      CF_OPERATION_RESTART_HYDRATION_FLAGS Flags;
      const CF_FS_METADATA                 *FsMetadata;
      LPCVOID                              FileIdentity;
      DWORD                                FileIdentityLength;
    } RestartHydration;
    struct {
      CF_OPERATION_TRANSFER_PLACEHOLDERS_FLAGS Flags;
      NTSTATUS                                 CompletionStatus;
      LARGE_INTEGER                            PlaceholderTotalCount;
      CF_PLACEHOLDER_CREATE_INFO               *PlaceholderArray;
      DWORD                                    PlaceholderCount;
      DWORD                                    EntriesProcessed;
    } TransferPlaceholders;
    struct {
      CF_OPERATION_ACK_DEHYDRATE_FLAGS Flags;
      NTSTATUS                         CompletionStatus;
      LPCVOID                          FileIdentity;
      DWORD                            FileIdentityLength;
    } AckDehydrate;
    struct {
      CF_OPERATION_ACK_RENAME_FLAGS Flags;
      NTSTATUS                      CompletionStatus;
    } AckRename;
    struct {
      CF_OPERATION_ACK_DELETE_FLAGS Flags;
      NTSTATUS                      CompletionStatus;
    } AckDelete;
  } DUMMYUNIONNAME;
} CF_OPERATION_PARAMETERS;

Members

ParamSize

DUMMYUNIONNAME

DUMMYUNIONNAME.TransferData

A sync provider performs TRANSFER_DATA to hydrate a placeholder file. This operation can be performed as a response to a FETCH_DATA callback, a VALIDATE_DATA callback, or as part of a preemptive background hydration effort outside of any callback context.

This API is only applicable for sync providers that support a hydration policy that is less than ALWAYS_FULL. Otherwise, the operation will fail with ERROR_CLOUD_FILE_INVALID_REQUEST.

The sync provider is required to have WRITE_DATA or WRITE_DAC access to the file on which the TRANSFER_DATA operation is to be performed. Otherwise the operation will fail with ERROR_CLOUD_FILE_ACCESS_DENIED.

To perform this operation:

• OpInfo.Type must be set to CF_OPERATION_TYPE_TRANSFER_DATA.
• OpParams.ParamSize must be set to the exact size of OpParams.TransferData plus the offset of OpParams.TransferData.
• OpParams.TransferData.Flags must be set to CF_OPERATION_TRANSFER_DATA_FLAG_NONE.
• OpParams.TransferData.CompletionStatus must be set to STATUS_SUCCESS if the sync provider wants to transfer data to the placeholder file. If the sync provider fails to process the hydration request for any reason and does not want to restart the hydration, it must set a STATUS_CLOUD_FILE_* status in CompletionStatus. Any status code returned outside of STATUS_CLOUD_FILE_* range except for STATUS_SUCCESS will be converted to STATUS_CLOUD_FILE_UNSUCCESSFUL by the platform.
• OpParams.TransferData.Buffer must point to a valid user mode buffer if OpParams.TransferData.CompletionStatus is STATUS_SUCCESS and should be of at least OpParams.TransferData.Length bytes; otherwise the buffer field is ignored.
• OpParams.TransferData.Offset and OpParams.TransferData.Length describe a range in the placeholder to which the sync provider is transferring the data. There is no requirement that the sync provider return all data as requested in one shot. It is also OK for a sync provider to return more data than requested. As an example, the sync provider can decide to over-read, for performance or other reasons. The sync provider can also perform multiple TRANSFER_DATA operations repeatedly as a response to the same FETCH_DATA callback. The only requirement is that both offset and length are 4KB aligned unless the range described ends on the logical file size (EoF), in which case, the length is not required to be 4KB aligned as long as the resulting range ends on or beyond the logical file size.

Upon operation completion:

• If the sync provider does not specify VALIDATION_REQUIRED at the sync root registration time.

In the successful transfer case, any pending user IO requests n the placeholder file that have received all needed bytes as a result of the transfer will be completed; otherwise the incomplete user IO requests will be updated to reflect the latest hydration state. In a failed transfer case, any pending user IO requests on the placeholder file that overlap with the range as described by the offset and length will be failed with OpParams.TransferData.CompletionStatus.

If the sync provider specifies VALIDATION_REQUIRED at the sync root registration time then no more processing will be done by the platform until the data in the range is acknowledged positively by the sync provider via an ACK_DATA operation.

DUMMYUNIONNAME.TransferData.Flags

Flags for transferring data. This must be set to CF_OPERATION_TRANSFER_DATA_FLAG_NONE. See CF_OPERATION_TRANSFER_DATA_FLAGS for more information.

DUMMYUNIONNAME.TransferData.CompletionStatus

Status for transferring data. This must be set to STATUS_SUCCESS if the sync provider wants to transfer data to the placeholder file. If the sync provider fails to process the hydration request for any reason and does not want to restart the hydration, it must set a STATUS_CLOUD_FILE_* status in CompletionStatus. Any status code returned outside of STATUS_CLOUD_FILE_* range except for STATUS_SUCCESS will be converted to STATUS_CLOUD_FILE_UNSUCCESSFUL by the platform.

DUMMYUNIONNAME.TransferData.Buffer

A valid user mode buffer. This must point to a valid user mode buffer if CompletionStatus is STATUS_SUCCESS and should be of at least Length bytes. Otherwise, the buffer field is ignored.

DUMMYUNIONNAME.TransferData.Offset

The offset used with the Length to describe a range in the placeholder to which the data is transferred. It describes a range in the placeholder to which the sync provider is transferring the data. There is no requirement that the sync provider return all data as requested in one shot. It is also OK for a sync provider to return more data than requested. As an example, the sync provider can decide to over-read, for performance or other reasons. The sync provider can also perform multiple TRANSFER_DATA operations repeatedly as a response to the same FETCH_DATA callback. Offset must be 4KB aligned.

DUMMYUNIONNAME.TransferData.Length

The Length in bytes of the Buffer. The length must be 4KB aligned unless the range described ends on the logical file size (EoF), in which case, the Length is not required to be 4KB aligned as long as the resulting range ends on or beyond the logical file size. Even if the CompletionStatus is not STATUS_SUCCESS, this field should be set to a valid value.

DUMMYUNIONNAME.RetrieveData

A sync provider performs a RETRIEVE_DATA operation as part of the placeholder hydration in order to validate the integrity of the data that was previously transferred to the placeholder. This operation can be performed as a response to a FETCH_DATA callback, a VALIDATE_DATA callback, or as part of a preemptive background hydration effort outside of any callback context.

This API is only applicable for sync providers that support a hydration policy that is less than ALWAYS_FULL. Otherwise, the operation will be failed with STATUS_CLOUD_FILE_NOT_SUPPORTED.

The sync provider is required to specify the hydration policy modifier VALIDATE_REQUIRED at the sync root registration time in order to perform this operation. Otherwise the operation will be failed with STATUS_CLOUD_FILE_NOT_SUPPORTED.

The sync provider is required to have READ_DATA or WRITE_DAC access to the file on which the RETRIEVE_DATA operation is to be performed. Otherwise the operation will be failed with STATUS_CLOUD_FILE_ACCESS_DENIED.

To perform this operation:

• OpInfo.Type must be set to CF_OPERATION_TYPE_RETRIEVE_DATA.
• OpParams.ParamSize must be set to the exact size of OpParams.RetrieveData plus the offset of OpParams.RetrieveData.
• OpParams.RetrieveData.Flags must be set to CF_OPERATION_RETRIEVE_DATA_FLAG_NONE.
• OpParams.RetrieveData.Buffer must point to a valid user mode buffer and be of at least OpParams.RetrieveData.Length bytes. Upon a successful completion of the operation, it receives the data previously transferred to the placeholder via TRANSFER_DATA.
• OpParams.RetrieveData.Offset and OpParams.RetrieveData.Length describe a range in the placeholder from which the sync provider is retrieving data. The range being requested must have been fully hydrated by a TRANSFER_DATA operation prior to the RETRIEVE_DATA operation otherwise the operation will be failed with STATUS_CLOUD_FILE_INVALID_REQUEST. Both offset and length are 4KB aligned unless the range described ends on the logical file size (EoF), in which case, the length is not required to be 4KB aligned as long as it ends on or beyond the logical file size.
• OpParams.RetrieveData.ReturnedLength receives the actual number of bytes retrieved upon a successful completion of the operation.

DUMMYUNIONNAME.RetrieveData.Flags

Flags for retrieving data. This must be set to CF_OPERATION_RETRIEVE_DATA_FLAG_NONE. See CF_OPERATION_RETRIEVE_DATA_FLAGS for more information.

DUMMYUNIONNAME.RetrieveData.Buffer

This must point to a valid user mode buffer and be of at least Length bytes. Upon a successful completion of the operation, it receives the data previously transferred to the placeholder via TRANSFER_DATA.

DUMMYUNIONNAME.RetrieveData.Offset

The offset used with the Length to describe the range of data retrieved from a placeholder. It describes a range in the placeholder from which the sync provider is retrieving data. The range being requested must have been fully hydrated by a TRANSFER_DATA operation prior to the RETRIEVE_DATA operation otherwise the operation will be failed with STATUS_CLOUD_FILE_INVALID_REQUEST. Offset must be 4KB aligned.

DUMMYUNIONNAME.RetrieveData.Length

The length in bytes of the Buffer. This is 4KB aligned unless the range described ends on the logical file size (EoF), in which case, Length is not required to be 4KB aligned as long as it ends on or beyond the logical file size.

DUMMYUNIONNAME.RetrieveData.ReturnedLength

The number of bytes retrieved on successful completion of the operation.

DUMMYUNIONNAME.AckData

A sync provider performs an ACK_DATA operation as part of the placeholder hydration after validating the integrity of the data that was previously transferred to the platform. This operation can be performed as a response to a FETCH_DATA callback, a VALIDATE_DATA callback, or as part of a preemptive background hydration effort outside of any callback context.

This API is only applicable for sync providers that support a hydration policy that is less than ALWAYS_FULL. Otherwise, the operation will be failed with STATUS_CLOUD_FILE_NOT_SUPPORTED.

The sync provider is required to specify the hydration policy modifier VALIDATE_REQUIRED at the sync root registration time to perform this operation. Otherwise the operation will be failed with STATUS_CLOUD_FILE_NOT_SUPPORTED.

The sync provider is required to have READ_DATA or WRITE_DAC access to the file on which the ACK_DATA operation is to be performed. Otherwise the operation will be failed with STATUS_CLOUD_FILE_ACCESS_DENIED.

To perform this operation:

• OpInfo.Type must be set to CF_OPERATION_TYPE_ACK_DATA.
• OpParams.ParamSize must be set to the exact size of OpParams.AckData plus the Offset of OpParams.AckData.
• OpParams.AckData.Flags must be set to CF_OPERATION_ACK_DATA_FLAG_NONE.
• OpParams.AckData.CompletionStatus must be set to STATUS_SUCCESS if the sync provider validates the data within the range to be acknowledged is good. If the sync provider fails to validate the data for any reason and does not want to restart the hydration, it must set a STATUS_CLOUD_FILE_* status in CompletionStatus. Any status code returned outside of STATUS_CLOUD_FILE_* range except for STATUS_SUCCESS will be converted to STATUS_CLOUD_FILE_UNSUCCESSFUL by the platform.
• OpParams.AckData.Offset and OpParams.AckData.Length describe a range in the placeholder whose data is going to be acknowledged by the sync provider. The range being requested does not need to be fully hydrated by a TRANSFER_DATA operation prior to the operation. Both Offset and Length are 4KB aligned unless the range described ends on the logical file size (EoF), in which case, Length is not required to be 4KB aligned as long as it ends on or beyond the logical file size.
• A Length of -1, denoted as CF_EOF, means infinity (i.e. to end of file).

Upon a successful ACK_DATA operation, any pending user IO requests on the placeholder file that have received all needed bytes as a result of the ACK_DATA operation will be completed; otherwise the incomplete user IO requests will be updated to reflect the latest hydration state. In a failed ACK_DATA case, any pending user IO requests on the placeholder file that overlap with the range as described by the Offset and Length will be failed with CompletionStatus.

DUMMYUNIONNAME.AckData.Flags

Flags for acknowledging data. This must be set to CF_OPERATION_ACK_DATA_FLAG_NONE. See CF_OPERATION_ACK_DATA_FLAGS for more information.

DUMMYUNIONNAME.AckData.CompletionStatus

Completion status of data acknowledgment. This must be set to STATUS_SUCCESS if the sync provider validates the data within the range to be acknowledged is good. If the sync provider fails to validate the data for any reason and does not want to restart the hydration, it must set a STATUS_CLOUD_FILE_* status in CompletionStatus. Any status code returned outside of STATUS_CLOUD_FILE_* range except for STATUS_SUCCESS will be converted to STATUS_CLOUD_FILE_UNSUCCESSFUL by the platform.

DUMMYUNIONNAME.AckData.Offset

The offset in bytes of the placeholder data to be acknowledged. Offset describes a range in the placeholder whose data is going to be acknowledged by the sync provider. The range being requested does not need to be fully hydrated by a TRANSFER_DATA operation prior to the operation. Offset must be 4KB aligned.

DUMMYUNIONNAME.AckData.Length

The length in bytes of data in the placeholder to be acknowledged. Must be 4KB aligned unless the range described ends on the logical file size (EoF), in which case, Length is not required to be 4KB aligned as long as it ends on or beyond the logical file size. A Length of -1, denoted as CF_EOF, means infinity (i.e. to end of file).

DUMMYUNIONNAME.RestartHydration

A sync provider performs a RESTART_HYDRATION operation to restart an ongoing hydration. This operation can be performed as a response to a FETCH_DATA callback, a VALIDATE_DATA callback, or as part of a preemptive background hydration effort outside of any callback context.

This API is only applicable for sync providers that support FULL hydration policy. Otherwise the operation will be failed with STATUS_CLOUD_FILE_NOT_SUPPORTED.

The sync provider is required to have WRITE_DATA or WRITE_DAC access to the file on which the RESTART_HYDRATION operation is to be performed. Otherwise, the operation will be failed with STATUS_CLOUD_FILE_ACCESS_DENIED.

To perform this operation:

• OpInfo.Type must be set to CF_OPERATION_TYPE_RESTART_HYDRATION.
• OpParams.ParamSize must be set to the exact size of OpParams.RestartHydration plus the Offset of OpParams.RestartHydration.
• OpParams.RestartHydration.Flags must be set to either CF_OPERATION_RESTART_HYDRATION_FLAG_NONE or CF_OPERATION_RESTART_HYDRATION_FLAG_MARK_IN_SYNC.
  • If CF_OPERATION_RESTART_HYDRATION_FLAG_MARK_IN_SYNC is specified, the placeholder will be marked in-sync upon a successful RESTART_HYDRATION operation.
• OpParams.RestartHydration.FsMetadata is optional. When provided:
  • A 0 value in the timestamp field (CreationTime, LastAccessTime, LastWriteTime, and ChangeTime) means no change to the current timestamp on the file.
  • A 0 value in FileAttributes means there is no change to the current file attributes on the file.
  • There is no special value in FileSize. A 0 value in FileSize truncates the file size to 0.
• OpParams.RestartHydration.FileIdentity and OpParams.RestartHydration.FileIdentityLength are optional. When provided, the new identity will be persisted on the file. Otherwise, the current file identity remains.

The sync provider calls this API in the event that it determines the on-disk data in a placeholder is in fact invalid and not suitable for satisfying I/O requests. The typical scenario is that the retrieved data for some reason failed a checksum verification. Another usage scenario occurs when over the course of hydration, the file contents have actually been updated in the cloud (and the sync provider is not able to retrieve historical contents from the cloud corresponding to the local placeholder version).

Upon a successful restart, the placeholder will be completely dehydrated and updated with the new metadata, if any. Pending user IO requests, if there are any, will be reprocessed as if they just arrived and as a result, the sync provider will receive new callbacks that look exactly the same as the ones that the sync provider received earlier.

If, however, the restart failed for whatever reason, the placeholder might be left in a non-deterministic state depending on where the error was encountered and all pending user IO requests on the placeholder will be failed with the same error. The only thing the platform guarantees in this case is that future user access to the placeholder file will trigger the same callbacks.

DUMMYUNIONNAME.RestartHydration.Flags

Flags to restart placeholder hydration. This must be set to either CF_OPERATION_RESTART_HYDRATION_FLAG_NONE or CF_OPERATION_RESTART_HYDRATION_FLAG_MARK_IN_SYNC. If CF_OPERATION_RESTART_HYDRATION_FLAG_MARK_IN_SYNC is specified, the placeholder will be marked in-sync upon a successful RESTART_HYDRATION operation. See CF_OPERATION_RESTART_HYDRATION_FLAGS for more information.

DUMMYUNIONNAME.RestartHydration.FsMetadata

Optional. Contains updates to the files metadata. When specified:

• A 0 value in the timestamp field (CreationTime, LastAccessTime, LastWriteTime, and ChangeTime) means no change to the current timestamp on the file.
• A 0 value in FileAttributes means no change to the current file attributes on the file.
• There is no special value in FileSize. A 0 value in FileSize truncates the file size to 0.

DUMMYUNIONNAME.RestartHydration.FileIdentity

Optional. When provided, the file identity is updated to this value. Otherwise, it remains the same.

DUMMYUNIONNAME.RestartHydration.FileIdentityLength

Optional. This specifies the length of the FileIdentity.

DUMMYUNIONNAME.TransferPlaceholders

DUMMYUNIONNAME.TransferPlaceholders.Flags

Flags for transferring placeholders. See CF_OPERATION_TRANSFER_PLACEHOLDERS_FLAGS for more information.

DUMMYUNIONNAME.TransferPlaceholders.CompletionStatus

The completion status of the transfer placeholders operation.

DUMMYUNIONNAME.TransferPlaceholders.PlaceholderTotalCount

The total number of placeholders.

DUMMYUNIONNAME.TransferPlaceholders.PlaceholderArray

An array of placeholders to be transferred.

DUMMYUNIONNAME.TransferPlaceholders.PlaceholderCount

The number of placeholders being transferred.

DUMMYUNIONNAME.TransferPlaceholders.EntriesProcessed

The placeholder entries that have been processed.

DUMMYUNIONNAME.AckDehydrate

DUMMYUNIONNAME.AckDehydrate.Flags

Dehydrated data acknowledgment flags. See CF_OPERATION_ACK_DEHYDRATE_FLAGS for more information.

DUMMYUNIONNAME.AckDehydrate.CompletionStatus

The completion status of the acknowledge dehydration operation. This must be set to STATUS_SUCCESS if the sync provider is able to allow the dehydration to proceed. If the sync provider intends to disallow the dehydration for any reason it must set a STATUS_CLOUD_FILE_* status in CompletionStatus. Any status code returned outside of STATUS_CLOUD_FILE_* range except for STATUS_SUCCESS will be converted to STATUS_CLOUD_FILE_UNSUCCESSFUL by the platform.

DUMMYUNIONNAME.AckDehydrate.FileIdentity

The file identity of the placeholder file to acknowledge dehydrated data for.

DUMMYUNIONNAME.AckDehydrate.FileIdentityLength

Length, in bytes, of the FileIdentity.

DUMMYUNIONNAME.AckRename

DUMMYUNIONNAME.AckRename.Flags

Acknowledge rename placeholder flags. See CF_OPERATION_ACK_RENAME_FLAGS for more information.

DUMMYUNIONNAME.AckRename.CompletionStatus

The completion status of the acknowledge rename operation. This must be set to STATUS_SUCCESS if the sync provider is able to allow the rename operation to proceed. If the sync provider intends to disallow the rename for any reason it must set a STATUS_CLOUD_FILE_* status in CompletionStatus. Any status code returned outside of STATUS_CLOUD_FILE_* range except for STATUS_SUCCESS will be converted to STATUS_CLOUD_FILE_UNSUCCESSFUL by the platform.

DUMMYUNIONNAME.AckDelete

DUMMYUNIONNAME.AckDelete.Flags

Acknowledge delete flags. See CF_OPERATION_ACK_DELETE_FLAGS for more information.

DUMMYUNIONNAME.AckDelete.CompletionStatus

The completion status of the acknowledge delete operation. This must be set to STATUS_SUCCESS if the sync provider is able to allow the delete operation to proceed. If the sync provider intends to disallow the delete for any reason it must set a STATUS_CLOUD_FILE_* status in CompletionStatus. Any status code returned outside of STATUS_CLOUD_FILE_* range except for STATUS_SUCCESS will be converted to STATUS_CLOUD_FILE_UNSUCCESSFUL by the platform.

Requirements

See also

CfExecute

CF_OPERATION_TRANSFER_DATA_FLAGS

CF_OPERATION_RETRIEVE_DATA_FLAGS