/**************************************************************************** * * BSD LICENSE * * Copyright(c) 2007-2023 Intel Corporation. All rights reserved. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * * * Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * * Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in * the documentation and/or other materials provided with the * distribution. * * Neither the name of Intel Corporation nor the names of its * contributors may be used to endorse or promote products derived * from this software without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. * * ***************************************************************************/ /* ***************************************************************************** * Doxygen group definitions ****************************************************************************/ /** ***************************************************************************** * @file cpa_dc_bp.h * * @defgroup cpaDcBp Data Compression Batch and Pack API * * @ingroup cpaDc * * @description * These functions specify the API for Data Compression operations related * to the 'Batch and Pack' mode of operation. * * @remarks * * *****************************************************************************/ #ifndef CPA_DC_BP_H #define CPA_DC_BP_H #ifdef __cplusplus extern"C" { #endif #include "cpa_dc.h" /** ***************************************************************************** * @ingroup cpaDcBp * Batch request input parameters. * @description * This structure contains the request information for use with batched * compression operations. * * ****************************************************************************/ typedef struct _CpaDcBatchOpData { CpaDcOpData opData; /**< Compression input parameters */ CpaBufferList *pSrcBuff; /**< Input buffer list containing the data to be compressed. */ CpaBoolean resetSessionState; /**< Reset the session state at the beginning of this request within * the batch. Only applies to stateful sessions. When this flag is * set, the history from previous requests in this session will not be * used when compressing the input data for this request in the batch. * */ } CpaDcBatchOpData ; /** ***************************************************************************** * @ingroup cpaDcBp * Submit a batch of requests to compress a batch of input buffers into * a common output buffer. The same output buffer is used for each request * in the batch. This is termed 'batch and pack'. * * @description * This API consumes data from the input buffer and generates compressed * data in the output buffer. * This API compresses a batch of input buffers and concatenates the * compressed data into the output buffer. A results structure is also * generated for each request in the batch. * * @context * When called as an asynchronous funnction it cannot sleep. It can be * executed in a context that does not permit sleeping. * When called as a synchronous function it may sleep. It MUST NOT be * executed in a context that DOES NOT permit sleeping. * @assumptions * None * @sideEffects * None * @blocking * Yes when configured to operate in synchronous mode. * @reentrant * No * @threadSafe * Yes * * @param[in] dcInstance Target service instance. * @param[in,out] pSessionHandle Session handle. * @param[in] numRequests Number of requests in the batch. * @param[in] pBatchOpData Pointer to an array of CpaDcBatchOpData * structures which contain the input buffers * and parameters for each request in the * batch. There should be numRequests entries * in the array. * @param[in] pDestBuff Pointer to buffer space for data after * compression. * @param[in,out] pResults Pointer to an array of results structures. * There should be numRequests entries in the * array. * @param[in] callbackTag User supplied value to help correlate * the callback with its associated * request. * * @retval CPA_STATUS_SUCCESS Function executed successfully. * @retval CPA_STATUS_FAIL Function failed. * @retval CPA_STATUS_RETRY Resubmit the request. * @retval CPA_STATUS_INVALID_PARAM Invalid parameter passed in. * @retval CPA_STATUS_RESOURCE Error related to system resources. * @retval CPA_DC_BAD_DATA The input data was not properly formed. * @retval CPA_STATUS_UNSUPPORTED Function is not supported. * @retval CPA_STATUS_RESTARTING API implementation is restarting. Resubmit * the request. * * @pre * pSessionHandle has been setup using cpaDcInitSession() * Session must be setup as a stateless session. * @note * This function passes control to the compression service for processing * * In synchronous mode the function returns the error status returned from the * service. In asynchronous mode the status is returned by the callback * function. * * This function may be called repetitively with input until all of the input * has been consumed by the compression service and all the output has been * produced. * * When this function returns, it may be that all of the available buffers in * the input list has not been compressed. This situation will occur when * there is insufficient space in the output buffer. The calling application * should note the amount of buffers processed, and then submit the request * again, with a new output buffer and with the input buffer list containing * the buffers that were not previously compressed. * * Relationship between input buffers and results buffers. * -# Implementations of this API must not modify the individual * flat buffers of the input buffer list. * -# The implementation communicates the number of buffers * consumed from the source buffer list via pResults->consumed arg. * -# The implementation communicates the amount of data in the * destination buffer list via pResults->produced arg. * * Source Buffer Setup Rules * -# The buffer list must have the correct number of flat buffers. This * is specified by the numBuffers element of the CpaBufferList. * -# Each flat buffer must have a pointer to contiguous memory that has * been allocated by the calling application. The number of octets to be * compressed or decompressed must be stored in the dataLenInBytes element * of the flat buffer. * -# It is permissible to have one or more flat buffers with a zero length * data store. This function will process all flat buffers until the * destination buffer is full or all source data has been processed. * If a buffer has zero length, then no data will be processed from * that buffer. * * Source Buffer Processing Rules. * -# The buffer list is processed in index order - SrcBuff->pBuffers[0] * will be completely processed before SrcBuff->pBuffers[1] begins to * be processed. * -# The application must drain the destination buffers. * If the source data was not completely consumed, the application * must resubmit the request. * -# On return, the pResults->consumed will indicate the number of buffers * consumed from the input buffer list. * * Destination Buffer Setup Rules * -# The destination buffer list must have storage for processed data and * for the packed header information. * This means that least two flat buffer must exist in the buffer list. * The first buffer entry will be used for the header information. * Subsequent entries will be used for the compressed data. * -# For each flat buffer in the buffer list, the dataLenInBytes element * must be set to the size of the buffer space. * -# It is permissible to have one or more flat buffers with a zero length * data store. * If a buffer has zero length, then no data will be added to * that buffer. * * Destination Buffer Processing Rules. * -# The buffer list is processed in index order. * -# On return, the pResults->produced will indicate the number of bytes * of compressed data written to the output buffers. Note that this * will not include the header information buffer. * -# If processing has not been completed, the application must drain the * destination buffers and resubmit the request. The application must reset * the dataLenInBytes for each flat buffer in the destination buffer list. * * Synchronous or Asynchronous operation of the API is determined by * the value of the callbackFn parameter passed to cpaDcInitSession() * when the sessionHandle was setup. If a non-NULL value was specified * then the supplied callback function will be invoked asynchronously * with the response of this request. * * Response ordering: * For each session, the implementation must maintain the order of * responses. That is, if in asynchronous mode, the order of the callback * functions must match the order of jobs submitted by this function. * In a simple synchronous mode implementation, the practice of submitting * a request and blocking on its completion ensure ordering is preserved. * * This limitation does not apply if the application employs multiple * threads to service a single session. * * If this API is invoked asynchronous, the return code represents * the success or not of asynchronously scheduling the request. * The results of the operation, along with the amount of data consumed * and produced become available when the callback function is invoked. * As such, pResults->consumed and pResults->produced are available * only when the operation is complete. * * The application must not use either the source or destination buffers * until the callback has completed. * * @see * None * *****************************************************************************/ CpaStatus cpaDcBPCompressData( CpaInstanceHandle dcInstance, CpaDcSessionHandle pSessionHandle, const Cpa32U numRequests, CpaDcBatchOpData *pBatchOpData, CpaBufferList *pDestBuff, CpaDcRqResults *pResults, void *callbackTag ); /** ***************************************************************************** * @ingroup cpaDcBp * Function to return the size of the memory which must be allocated for * the pPrivateMetaData member of CpaBufferList contained within * CpaDcBatchOpData. * * @description * This function is used to obtain the size (in bytes) required to allocate * a buffer descriptor for the pPrivateMetaData member in the * CpaBufferList structure when Batch and Pack API are used. * Should the function return zero then no meta data is required for the * buffer list. * * @context * This function may be called from any context. * @assumptions * None * @sideEffects * None * @blocking * No * @reentrant * No * @threadSafe * Yes * * @param[in] instanceHandle Handle to an instance of this API. * @param[in] numJobs The number of jobs defined in the CpaDcBatchOpData * table. * @param[out] pSizeInBytes Pointer to the size in bytes of memory to be * allocated when the client wishes to allocate * a cpaFlatBuffer and the Batch and Pack OP data. * * @retval CPA_STATUS_SUCCESS Function executed successfully. * @retval CPA_STATUS_FAIL Function failed. * @retval CPA_STATUS_INVALID_PARAM Invalid parameter passed in. * * @pre * None * @post * None * @note * None * @see * cpaDcBPCompressData() * *****************************************************************************/ CpaStatus cpaDcBnpBufferListGetMetaSize(const CpaInstanceHandle instanceHandle, Cpa32U numJobs, Cpa32U *pSizeInBytes); #ifdef __cplusplus } /* close the extern "C" { */ #endif #endif /* CPA_DC_BP_H */