FastSort/src/sort/phase_1_write.cpp

/*
     Description: 第一阶段写入中间文件

     Copyright : All right reserved by ICT

     Author : Zhang Zhonghai
     Date : 2026/06/02
*/

#include "phase_1_write.h"

#include <klib/kthread.h>
#include <spdlog/spdlog.h>
#include <zlib.h>

#include <algorithm>
#include <string>

#include "common_data.h"
#include "const_val.h"
#include "phase_1.h"
#include "phase_1_compress.h"
#include "phase_1_write.h"
#include "sort.h"
#include "util/profiling.h"

static void checkCompress(uint8_t *addr, uint64_t len) {
    size_t curReadPos = 0;
    int blockLen = 0;
    int maxBlockLen = 0;
    DataBuffer buf;
    buf.AllocMem(SINGLE_BLOCK_SIZE);

    while (curReadPos + BLOCK_HEADER_LENGTH <= len) { /* 确保能解析block长度 */
        blockLen = unpackInt16(&addr[curReadPos + 16]) + 1;
        if (blockLen > maxBlockLen) {
            maxBlockLen = blockLen;
        }
        if (curReadPos + blockLen <= len) { /* 完整的block数据在buf里 */
            size_t dlen = SINGLE_BLOCK_SIZE;  // 65535
            uint32_t crc = le_to_u32(addr + curReadPos + blockLen - 8);
            int ret = bgzfUncompress(buf.data, &dlen, (Bytef*)(addr + curReadPos) + BLOCK_HEADER_LENGTH, blockLen - BLOCK_HEADER_LENGTH, crc);
            if (ret != 0) {
                spdlog::error("block len: {}, uncompressed len: {}", blockLen, dlen);
                exit(0);
            }

            curReadPos += blockLen;
        } else {
            spdlog::error("not valid compressed block: {}, {}", curReadPos + blockLen, len);
            break; /* 当前block数据不完整，一部分在还没读入的file数据里 */
        }
    }
    if (curReadPos != len) {
        spdlog::error("addr: {}, len: {}", curReadPos, len);
        exit(0);
    }
}

static void doWrite(Phase1PipelineArg& p) {
    PROF_G_BEG(write_mid);
    DataBuffer& compressBuf = p.compressBuf[p.writeOrder % p.COMPRESS_BUF_NUM];
    // checkCompress(compressBuf.data, compressBuf.curLen);
    fwrite(compressBuf.data, 1, compressBuf.curLen, p.midFilePtr);
    PROF_G_END(write_mid);
}

void* phase1Write(void* data) {
    Phase1PipelineArg& p = *(Phase1PipelineArg*)data;
    // for test，写header
    // fwrite(nsgv::gInHdr.compressed.data, 1, nsgv::gInHdr.compressed.curLen, p.midFilePtr);

    /* do the work */
    while (true) {
        // previous dependency
        yarn::DEPENDENCY_NOT_TO_BE(p.compressSig, 0);

        if (p.compressFinish) {
            while (p.writeOrder < p.compressOrder) {
                doWrite(p);
                p.writeOrder += 1;
            }
            break;
        }
        doWrite(p);
        // update status
        yarn::CONSUME_SIGNAL(p.compressSig);
        p.writeOrder += 1;
    }

    // 写结尾，空block，bam文件需要
    //fwrite("\037\213\010\4\0\0\0\0\0\377\6\0\102\103\2\0\033\0\3\0\0\0\0\0\0\0\0\0", 1, 28, p.midFilePtr);

    spdlog::info("End write order: {}", p.writeOrder);
    return nullptr;
}