/**
* Copyright (c) 2025 Huawei Technologies Co., Ltd.
* This program is free software, you can redistribute it and/or modify it under the terms and conditions of
* CANN Open Software License Agreement Version 2.0 (the "License").
* Please refer to the License for details. You may not use this file except in compliance with the License.
* THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED,
* INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
* See LICENSE in the root of the software repository for the full text of the License.
*/

/*!
 * \file kernel_check_data_copy_overflow.h
 * \brief
 */
#if !defined(__ASCENDC_INCLUDE_INTERNAL_HEADERS__)
#pragma message("impl/basic_api/utils/kernel_check_data_copy_overflow.h is an internal header file and must not be used directly. Functions or variables defined in this file may be removed in the future. Please use \"#include \"basic_api/kernel_tensor.h\"\" and use public functions or variables defined in interface headers files.")
#define __ASCENDC_INCLUDE_INTERNAL_HEADERS__
#define __UNDEF_ASCENDC_INCLUDE_INTERNAL_HEADERS_KERNEL_CHECK_DATA_COPY_OVERFLOW_H__
#endif
#ifndef ASCENDC_MODULE_CHECK_DATA_COPY_OVERFLOW_H
#define ASCENDC_MODULE_CHECK_DATA_COPY_OVERFLOW_H

#if ASCENDC_CPU_DEBUG
#include "kernel_check_util.h"
#include "kernel_common.h"
#include "kernel_struct_data_copy.h"

namespace AscendC {
/* **************************************************************************************************
 * Check function for CPU debug
 * ************************************************************************************************* */
template <typename T>
bool CheckDataCopyTensorSizeOverflow(const LocalTensor<T>& dst, const GlobalTensor<T>& src,
    const DataCopyParams& repeatParams)
{
    const Hardware srcPos = Hardware::GM;
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    uint64_t srcMaxOffsetBytes = 0;
    uint64_t dstMaxOffsetBytes = 0;
    BlockMode mode = check::GetBlockMode({ srcPos, dstPos });
    uint8_t biasConvFlag = 0;
    std::string apiInfo = "DataCopy from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(TPosition::GM)) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    check::CalculateDataCopyMaxOffset<PrimT<T>, PrimT<T>>(
        repeatParams, srcPos, dstPos, mode, srcMaxOffsetBytes, dstMaxOffsetBytes, DeqScale::DEQ_NONE, biasConvFlag);
    return check::ReportTensorSizeOverflow(srcPos, dstPos, 0, dst.GetSize() * sizeof(PrimT<T>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyTensorSizeOverflow(const LocalTensor<T>& dst, const GlobalTensor<T>& src,
    const Nd2NzParams& intriParams)
{
    if (intriParams.ndNum == 0 || intriParams.nValue == 0 || intriParams.dValue == 0) {
        return true;
    }
    const Hardware srcPos = Hardware::GM;
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    uint64_t srcMaxOffsetBytes = 0;
    uint64_t dstMaxOffsetBytes = (intriParams.ndNum - 1) * intriParams.dstNzMatrixStride * sizeof(PrimT<T>) +
        (intriParams.nValue - 1) * intriParams.dstNzNStride * DEFAULT_C0_SIZE +
        (DivCeil(intriParams.dValue * sizeof(PrimT<T>), DEFAULT_C0_SIZE) - 1) *
        intriParams.dstNzC0Stride * DEFAULT_C0_SIZE +
        DEFAULT_C0_SIZE;
    std::string apiInfo = "DataCopy with Nd2NzParams from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(TPosition::GM)) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    return check::ReportTensorSizeOverflow(srcPos, dstPos, 0, dst.GetSize() * sizeof(PrimT<T>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyTensorSizeOverflow(const LocalTensor<T>& dst, const LocalTensor<T>& src,
    const Nd2NzParams& intriParams)
{
    if (intriParams.ndNum == 0 || intriParams.nValue == 0 || intriParams.dValue == 0) {
        return true;
    }
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    uint64_t srcMaxOffsetBytes = ((intriParams.ndNum - 1) * intriParams.srcNdMatrixStride +
        (intriParams.nValue - 1) * intriParams.srcDValue + intriParams.dValue) * sizeof(PrimT<T>);
    uint64_t dstMaxOffsetBytes = (intriParams.ndNum - 1) * intriParams.dstNzMatrixStride * sizeof(PrimT<T>) +
        (intriParams.nValue - 1) * intriParams.dstNzNStride * DEFAULT_C0_SIZE +
        (DivCeil(intriParams.dValue * sizeof(PrimT<T>), DEFAULT_C0_SIZE) - 1) *
        intriParams.dstNzC0Stride * DEFAULT_C0_SIZE +
        DEFAULT_C0_SIZE;
    std::string apiInfo = "DataCopy with Nd2NzParams from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    return check::ReportTensorSizeOverflow(srcPos, dstPos, src.GetSize() * sizeof(PrimT<T>),
        dst.GetSize() * sizeof(PrimT<T>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyTensorSizeOverflow(const GlobalTensor<T>& dst, const LocalTensor<T>& src,
    const DataCopyParams& repeatParams)
{
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = Hardware::GM;
    uint64_t srcMaxOffsetBytes = 0;
    uint64_t dstMaxOffsetBytes = 0;
    BlockMode mode = check::GetBlockMode({ srcPos, dstPos });
    uint8_t biasConvFlag = 0;
    std::string apiInfo = "DataCopy from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(TPosition::GM));

    check::CalculateDataCopyMaxOffset<PrimT<T>, PrimT<T>>(
        repeatParams, srcPos, dstPos, mode, srcMaxOffsetBytes, dstMaxOffsetBytes, DeqScale::DEQ_NONE, biasConvFlag);
    return check::ReportTensorSizeOverflow(srcPos, dstPos, src.GetSize() * sizeof(PrimT<T>), 0,
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyTensorSizeOverflow(const LocalTensor<T> &dst, const LocalTensor<T> &src,
    const DataCopyParams &repeatParams)
{
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    uint64_t srcMaxOffsetBytes = 0;
    uint64_t dstMaxOffsetBytes = 0;
    BlockMode mode = check::GetBlockMode({ srcPos, dstPos });
    uint8_t biasConvFlag = 0;
    std::string apiInfo = "DataCopy from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    check::CalculateDataCopyMaxOffset<PrimT<T>, PrimT<T>>(
        repeatParams, srcPos, dstPos, mode, srcMaxOffsetBytes, dstMaxOffsetBytes, DeqScale::DEQ_NONE, biasConvFlag);
    return check::ReportTensorSizeOverflow(srcPos, dstPos,
        src.GetSize() * sizeof(PrimT<T>), dst.GetSize() * sizeof(PrimT<T>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T, typename U>
bool CheckDataCopyTensorSizeOverflow(const LocalTensor<T> &dst, const LocalTensor<U> &src,
    const DataCopyParams &repeatParams)
{
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    uint64_t srcMaxOffsetBytes = 0;
    uint64_t dstMaxOffsetBytes = 0;
    BlockMode mode = check::GetBlockMode({ srcPos, dstPos });
    uint8_t biasConvFlag = check::IsBiasConv({ srcPos, dstPos }) && (sizeof(PrimT<U>) != sizeof(PrimT<T>));
    std::string apiInfo = "DataCopy from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    check::CalculateDataCopyMaxOffset<PrimT<U>, PrimT<T>>(
        repeatParams, srcPos, dstPos, mode, srcMaxOffsetBytes, dstMaxOffsetBytes, DeqScale::DEQ_NONE, biasConvFlag);
    return check::ReportTensorSizeOverflow(srcPos, dstPos,
        src.GetSize() * sizeof(PrimT<U>), dst.GetSize() * sizeof(PrimT<T>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyTensorSizeOverflow(const LocalTensor<T> &dst, const GlobalTensor<T> &src,
    const SliceInfo dstSliceInfo[], const SliceInfo srcSliceInfo[], const uint32_t dimValue)
{
    const Hardware srcPos = Hardware::GM;
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    uint64_t srcMaxOffsetBytes = 0;
    uint64_t dstMaxOffsetBytes = 1;
    std::string apiInfo = "DataCopy with SliceInfo from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(TPosition::GM)) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    for (uint32_t i = 0; i < dimValue; i++) {
        dstMaxOffsetBytes *= dstSliceInfo[i].shapeValue;
    }
    dstMaxOffsetBytes *= sizeof(PrimT<T>);
    return check::ReportTensorSizeOverflow(srcPos, dstPos, 0, dst.GetSize() * sizeof(PrimT<T>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyTensorSizeOverflow(const GlobalTensor<T> &dst, const LocalTensor<T> &src,
    const SliceInfo dstSliceInfo[], const SliceInfo srcSliceInfo[], const uint32_t dimValue)
{
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = Hardware::GM;
    uint64_t srcMaxOffsetBytes = 1;
    uint64_t dstMaxOffsetBytes = 0;
    std::string apiInfo = "DataCopy with SliceInfo from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(TPosition::GM));

    for (uint32_t i = 0; i < dimValue; i++) {
        srcMaxOffsetBytes *= dstSliceInfo[i].shapeValue;
    }
    srcMaxOffsetBytes *= sizeof(PrimT<T>);
    return check::ReportTensorSizeOverflow(srcPos, dstPos, src.GetSize() * sizeof(PrimT<T>), 0,
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyTensorSizeOverflow(const LocalTensor<T>& dst, const GlobalTensor<T>& src,
    const uint32_t count)
{
    const Hardware srcPos = Hardware::GM;
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    uint64_t srcMaxOffsetBytes = 0;
    uint64_t dstMaxOffsetBytes = count * sizeof(PrimT<T>);
    std::string apiInfo = "DataCopy from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(TPosition::GM)) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    return check::ReportTensorSizeOverflow(srcPos, dstPos, 0, dst.GetSize() * sizeof(PrimT<T>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyTensorSizeOverflow(const LocalTensor<T>& dst, const LocalTensor<T>& src,
    const uint32_t count)
{
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    uint64_t srcMaxOffsetBytes = count * sizeof(PrimT<T>);
    uint64_t dstMaxOffsetBytes = count * sizeof(PrimT<T>);
    std::string apiInfo = "DataCopy from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    return check::ReportTensorSizeOverflow(srcPos, dstPos,
        src.GetSize() * sizeof(PrimT<T>), dst.GetSize() * sizeof(PrimT<T>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyTensorSizeOverflow(const GlobalTensor<T>& dst, const LocalTensor<T>& src,
    const uint32_t count)
{
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = Hardware::GM;
    uint64_t srcMaxOffsetBytes = count * sizeof(PrimT<T>);
    uint64_t dstMaxOffsetBytes = 0;
    std::string apiInfo = "DataCopy from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(TPosition::GM));

    return check::ReportTensorSizeOverflow(srcPos, dstPos,
        src.GetSize() * sizeof(PrimT<T>), 0,
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyTensorSizeOverflow(const GlobalTensor<T>& dst, const LocalTensor<T>& src,
    const Nz2NdParamsFull& intriParams)
{
    if (intriParams.ndNum == 0 || intriParams.nValue == 0 || intriParams.dValue == 0) {
        return true;
    }
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = Hardware::GM;
    const int nzWidth = 16;
    const int srcMatrixStrideUnit = 256;
    uint64_t srcMaxOffsetBytes = (intriParams.ndNum - 1) * intriParams.srcNdMatrixStride * srcMatrixStrideUnit *
        sizeof(PrimT<T>) + (intriParams.dValue / nzWidth - 1) * intriParams.srcNStride * nzWidth * sizeof(PrimT<T>) +
        nzWidth * intriParams.nValue * sizeof(PrimT<T>);
    uint64_t dstMaxOffsetBytes = 0;
    std::string apiInfo = "DataCopy with Nz2NdParamsFull from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(TPosition::GM));

    return check::ReportTensorSizeOverflow(srcPos, dstPos,
        src.GetSize() * sizeof(PrimT<T>), 0,
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyTensorSizeOverflow(const LocalTensor<T>& dst, const GlobalTensor<T>& src,
    const DataCopyParams& intriParams, const DataCopyEnhancedParams& enhancedParams)
{
    if (intriParams.blockCount == 0 || intriParams.blockLen == 0) {
        return true;
    }
    const Hardware srcPos = Hardware::GM;
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    uint64_t srcMaxOffsetBytes = 0;
    uint64_t dstMaxOffsetBytes = 0;
    BlockMode mode = check::GetBlockMode({ srcPos, dstPos }, enhancedParams.blockMode);
    uint8_t biasConvFlag = 0;
    std::string apiInfo = "DataCopy with DataCopyEnhancedParams from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(TPosition::GM)) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    check::CalculateDataCopyMaxOffset<PrimT<T>, PrimT<T>>(
        intriParams, srcPos, dstPos, mode, srcMaxOffsetBytes, dstMaxOffsetBytes,
        enhancedParams.deqScale, biasConvFlag);
    return check::ReportTensorSizeOverflow(srcPos, dstPos, 0, dst.GetSize() * sizeof(PrimT<T>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyTensorSizeOverflow(const GlobalTensor<T>& dst, const LocalTensor<T>& src,
    const DataCopyParams& intriParams, const DataCopyEnhancedParams& enhancedParams)
{
    if (intriParams.blockCount == 0 || intriParams.blockLen == 0) {
        return true;
    }
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = Hardware::GM;
    uint64_t srcMaxOffsetBytes = 0;
    uint64_t dstMaxOffsetBytes = 0;
    BlockMode mode = check::GetBlockMode({ srcPos, dstPos }, enhancedParams.blockMode);
    uint8_t biasConvFlag = 0;
    std::string apiInfo = "DataCopy with DataCopyEnhancedParams from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(TPosition::GM));

    check::CalculateDataCopyMaxOffset<PrimT<T>, PrimT<T>>(
        intriParams, srcPos, dstPos, mode, srcMaxOffsetBytes, dstMaxOffsetBytes,
        enhancedParams.deqScale, biasConvFlag);
    return check::ReportTensorSizeOverflow(srcPos, dstPos, src.GetSize() * sizeof(PrimT<T>), 0,
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyTensorSizeOverflow(const LocalTensor<T> &dst, const LocalTensor<T> &src,
    const DataCopyParams &intriParams, const DataCopyEnhancedParams &enhancedParams)
{
    if (intriParams.blockCount == 0 || intriParams.blockLen == 0) {
        return true;
    }
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    uint64_t srcMaxOffsetBytes = 0;
    uint64_t dstMaxOffsetBytes = 0;
    BlockMode mode = check::GetBlockMode({ srcPos, dstPos }, enhancedParams.blockMode);
    uint8_t biasConvFlag = 0;
    std::string apiInfo = "DataCopy with DataCopyEnhancedParams from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    check::CalculateDataCopyMaxOffset<PrimT<T>, PrimT<T>>(
        intriParams, srcPos, dstPos, mode, srcMaxOffsetBytes, dstMaxOffsetBytes,
        enhancedParams.deqScale, biasConvFlag);
    return check::ReportTensorSizeOverflow(srcPos, dstPos,
        src.GetSize() * sizeof(PrimT<T>), dst.GetSize() * sizeof(PrimT<T>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T, typename U>
bool CheckDataCopyTensorSizeOverflow(const LocalTensor<T>& dst, const LocalTensor<U>& src,
    const DataCopyCO12DstParams& intriParams)
{
    if (intriParams.nSize == 0) {
        return true;
    }
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    uint64_t srcMaxOffsetBytes = 0;
    uint64_t dstMaxOffsetBytes = 0;
    std::string apiInfo = "DataCopy with DataCopyCO12DstParams from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    check::CalculateDataCopyMaxOffset<PrimT<U>, PrimT<T>>(
        srcPos, dstPos, intriParams, srcMaxOffsetBytes, dstMaxOffsetBytes);
    return check::ReportTensorSizeOverflow(srcPos, dstPos,
        src.GetSize() * sizeof(PrimT<U>), dst.GetSize() * sizeof(PrimT<T>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T, typename U>
bool CheckDataCopyTensorSizeOverflow(const GlobalTensor<T>& dst, const LocalTensor<U>& src,
    const DataCopyCO12DstParams& intriParams)
{
    if (intriParams.nSize == 0) {
        return true;
    }
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = Hardware::GM;
    uint64_t srcMaxOffsetBytes = 0;
    uint64_t dstMaxOffsetBytes = 0;
    std::string apiInfo = "DataCopy with DataCopyCO12DstParams from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(TPosition::GM));

    check::CalculateDataCopyMaxOffset<PrimT<U>, PrimT<T>>(
        srcPos, dstPos, intriParams, srcMaxOffsetBytes, dstMaxOffsetBytes);
    return check::ReportTensorSizeOverflow(srcPos, dstPos,
        src.GetSize() * sizeof(PrimT<U>), 0,
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T, typename U>
bool CheckDataCopyTensorSizeOverflow(const LocalTensor<U> &dst, const LocalTensor<T> &src,
    const DataCopyParams &intriParams, const DataCopyEnhancedParams &enhancedParams)
{
    if (intriParams.blockCount == 0 || intriParams.blockLen == 0) {
        return true;
    }
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    uint64_t srcMaxOffsetBytes = 0;
    uint64_t dstMaxOffsetBytes = 0;
    BlockMode mode = check::GetBlockMode({ srcPos, dstPos }, enhancedParams.blockMode);
    uint8_t biasConvFlag = check::IsBiasConv({ srcPos, dstPos }) && (sizeof(PrimT<T>) != sizeof(PrimT<U>));
    std::string apiInfo = "DataCopy with DataCopyEnhancedParams from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    check::CalculateDataCopyMaxOffset<PrimT<T>, PrimT<U>>(
        intriParams, srcPos, dstPos, mode, srcMaxOffsetBytes, dstMaxOffsetBytes,
        enhancedParams.deqScale, biasConvFlag, enhancedParams.sidStoreMode);
    return check::ReportTensorSizeOverflow(srcPos, dstPos,
        src.GetSize() * sizeof(PrimT<T>), dst.GetSize() * sizeof(PrimT<U>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyPadTensorSizeOverflow(const LocalTensor<T> &dst,
    const GlobalTensor<T> &src, const DataCopyParams &dataCopyParams, const DataCopyPadParams &padParams)
{
    const Hardware srcPos = Hardware::GM;
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    if (dstPos != Hardware::UB) {
        return true;
    }
    uint64_t srcMaxOffsetBytes = 0;
    uint64_t paddingSize = (padParams.leftPadding + padParams.rightPadding) * sizeof(PrimT<T>);
    uint64_t dstMaxOffsetBytes = dataCopyParams.blockCount *
        AlignUp(dataCopyParams.blockLen + paddingSize, DEFAULT_C0_SIZE) +
        (dataCopyParams.blockCount - 1) * dataCopyParams.dstStride * DEFAULT_C0_SIZE;
    std::string apiInfo = "DataCopyPad with DataCopyParams and DataCopyPadParams from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(TPosition::GM)) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    return check::ReportTensorSizeOverflow(srcPos, dstPos,
        0, dst.GetSize() * sizeof(PrimT<T>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyPadTensorSizeOverflow(const GlobalTensor<T> &dst,
    const LocalTensor<T> &src, const DataCopyParams &dataCopyParams)
{
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = Hardware::GM;
    if (srcPos != Hardware::UB) {
        return true;
    }
    uint64_t srcMaxOffsetBytes = dataCopyParams.blockCount * AlignUp(dataCopyParams.blockLen, DEFAULT_C0_SIZE) +
        (dataCopyParams.blockCount - 1) * dataCopyParams.srcStride * DEFAULT_C0_SIZE;
    uint64_t dstMaxOffsetBytes = 0;
    std::string apiInfo = "DataCopyPad with DataCopyParams from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(TPosition::GM));

    return check::ReportTensorSizeOverflow(srcPos, dstPos,
        src.GetSize() * sizeof(PrimT<T>), 0,
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T, typename U>
bool CheckDataCopyPadTensorSizeOverflow(const LocalTensor<T> &dst,
    const GlobalTensor<T> &src,
    const DataCopyExtParams &dataCopyParams, const DataCopyPadExtParams<U> &padParams)
{
    const Hardware srcPos = Hardware::GM;
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    if (dstPos != Hardware::UB) {
        return true;
    }
    uint64_t srcMaxOffsetBytes = 0;
    uint64_t paddingSize = (padParams.leftPadding + padParams.rightPadding) * sizeof(PrimT<T>);
    uint64_t dstMaxOffsetBytes = dataCopyParams.blockCount *
        AlignUp(dataCopyParams.blockLen + paddingSize, DEFAULT_C0_SIZE) +
        (dataCopyParams.blockCount - 1) * dataCopyParams.dstStride * DEFAULT_C0_SIZE;
    std::string apiInfo = "DataCopyPad with DataCopyExtParams and DataCopyPadExtParams from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(TPosition::GM)) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    return check::ReportTensorSizeOverflow(srcPos, dstPos,
        0, dst.GetSize() * sizeof(PrimT<T>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyPadTensorSizeOverflow(const GlobalTensor<T> &dst,
    const LocalTensor<T> &src, const DataCopyExtParams &dataCopyParams)
{
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = Hardware::GM;
    if (srcPos != Hardware::UB) {
        return true;
    }
    uint64_t srcMaxOffsetBytes = dataCopyParams.blockCount * AlignUp(dataCopyParams.blockLen, DEFAULT_C0_SIZE) +
        (dataCopyParams.blockCount - 1) * dataCopyParams.srcStride * DEFAULT_C0_SIZE;
    uint64_t dstMaxOffsetBytes = 0;
    std::string apiInfo = "DataCopyPad with DataCopyExtParams from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(TPosition::GM));

    return check::ReportTensorSizeOverflow(srcPos, dstPos,
        src.GetSize() * sizeof(PrimT<T>), 0,
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyPadTensorSizeOverflow(const LocalTensor<T> &dst,
    const LocalTensor<T> &src, const DataCopyParams &dataCopyParams, const Nd2NzParams &nd2nzParams)
{
    if (dataCopyParams.blockCount == 0 || nd2nzParams.nValue == 0 ||
        nd2nzParams.dValue == 0 || nd2nzParams.ndNum == 0) {
        return true;
    }
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    uint64_t srcMaxOffsetBytes = dataCopyParams.blockCount * AlignUp(dataCopyParams.blockLen, DEFAULT_C0_SIZE) +
        (dataCopyParams.blockCount - 1) * dataCopyParams.srcStride * DEFAULT_C0_SIZE;
    uint64_t dstMaxOffsetBytes = (nd2nzParams.ndNum - 1) * nd2nzParams.dstNzMatrixStride * sizeof(PrimT<T>) +
        (nd2nzParams.nValue - 1) * nd2nzParams.dstNzNStride * DEFAULT_C0_SIZE +
        (DivCeil(nd2nzParams.dValue * sizeof(PrimT<T>), DEFAULT_C0_SIZE) - 1) *
        nd2nzParams.dstNzC0Stride * DEFAULT_C0_SIZE +
        DEFAULT_C0_SIZE;
    std::string apiInfo = "DataCopyPad with DataCopyParams and Nd2NzParams from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    return check::ReportTensorSizeOverflow(srcPos, dstPos,
        src.GetSize() * sizeof(PrimT<T>), dst.GetSize() * sizeof(PrimT<T>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

template <typename T>
bool CheckDataCopyPadTensorSizeOverflow(const LocalTensor<T> &dst, const LocalTensor<T> &src,
    const DataCopyExtParams &dataCopyParams, const Nd2NzParams &nd2nzParams)
{
    if (dataCopyParams.blockCount == 0 || nd2nzParams.nValue == 0 ||
        nd2nzParams.dValue == 0 || nd2nzParams.ndNum == 0) {
        return true;
    }
    const Hardware srcPos = GetPhyType((TPosition)src.GetPosition());
    const Hardware dstPos = GetPhyType((TPosition)dst.GetPosition());
    uint64_t srcMaxOffsetBytes = dataCopyParams.blockCount * AlignUp(dataCopyParams.blockLen, DEFAULT_C0_SIZE) +
        (dataCopyParams.blockCount - 1) * dataCopyParams.srcStride * DEFAULT_C0_SIZE;
    uint64_t dstMaxOffsetBytes = (nd2nzParams.ndNum - 1) * nd2nzParams.dstNzMatrixStride * sizeof(PrimT<T>) +
        (nd2nzParams.nValue - 1) * nd2nzParams.dstNzNStride * DEFAULT_C0_SIZE +
        (DivCeil(nd2nzParams.dValue * sizeof(PrimT<T>), DEFAULT_C0_SIZE) - 1) *
        nd2nzParams.dstNzC0Stride * DEFAULT_C0_SIZE +
        DEFAULT_C0_SIZE;
    std::string apiInfo = "DataCopyPad with DataCopyExtParams and Nd2NzParams from " +
        ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(src.GetPosition())) +
        " to " + ConstDefiner::Instance().logicNameMap.at(static_cast<uint8_t>(dst.GetPosition()));

    return check::ReportTensorSizeOverflow(srcPos, dstPos,
        src.GetSize() * sizeof(PrimT<T>), dst.GetSize() * sizeof(PrimT<T>),
        srcMaxOffsetBytes, dstMaxOffsetBytes, apiInfo);
}

} // namespace AscendC
#endif

#endif
#if defined(__UNDEF_ASCENDC_INCLUDE_INTERNAL_HEADERS_KERNEL_CHECK_DATA_COPY_OVERFLOW_H__)
#undef __ASCENDC_INCLUDE_INTERNAL_HEADERS__
#undef __UNDEF_ASCENDC_INCLUDE_INTERNAL_HEADERS_KERNEL_CHECK_DATA_COPY_OVERFLOW_H__
#endif