我会错误地集中这些字符数组吗？

Question

我一直在尝试使用c ++从头开始创建一个midi文件。我正在将不同的块（和部分块）分成不同的字符数组。

我的代码：

#include <iostream>
#include <fstream>
#include <cstring>

int main(int, char* []) {
    std::ofstream MIDIfile ("example.mid")
    char header[] = { /* hex data */ };
    char track_header[] = { /* hex data */ };
    char track_data[] = { /* hex data */ };
    MIDIfile << strcat(header, strcat(track_header, track_data));
    MIDIfile.close();
    return 0;
}

我唯一的问题是，在写入文件时，仅写入了81个字节中的8个字节。是否有一个原因？我在做错什么吗？

致谢，shadowstalker75

Answer 1

您将必须了解testMemory的用途。这条线永远行不通。实际上，更好的是，永远不要使用#include <linux/init.h> #include <linux/module.h> #include <linux/kernel.h> #include <linux/platform_device.h> #include <linux/device.h> #include <linux/dma-mapping.h> MODULE_LICENSE("GPL"); MODULE_AUTHOR("Louis Jenkins"); MODULE_DESCRIPTION("Persistent Memory Analysis Tool - Kernel Module"); MODULE_VERSION("0.1"); static uint64_t *testMemory; static uint64_t *readMemory; static struct resource res[] = ; static struct platform_device *pdev; static int __init dummymodule_init(void){ void *addr; int i; int ret; size_t size; dma_addr_t dma_handle; // Allocate memory... testMemory = kvzalloc(sizeof(*testMemory), GFP_KERNEL); readMemory = kvzalloc(sizeof(*readMemory), GFP_KERNEL); // Give testMemory a deterministic value *testMemory = 1; // Initialize resource to point to the readMemory (so that it will do a DMA memory-to-memory // transfer from testMemory into readMemory... res[0] = (struct resource) { .start = virt_to_phys(readMemory), .end = virt_to_phys(readMemory) + sizeof(*readMemory), .flags = IORESOURCE_MEM, .name = "readMemory" }; // Some debug information... printk(KERN_INFO ".start = %u, .end = %u\n", res[0].start, res[0].end); // Allocate platform device pdev = platform_device_alloc("Non-Coherent DMA Memory Prodding Tool", 0); // Ensure that it did not return NULL printk(KERN_INFO "Allocated pdev: %x", pdev); BUG_ON(pdev == NULL); // Configure resources pdev->num_resources = 1; pdev->resource = res; // Add device ret = platform_device_add(pdev); if (ret != 0) { printk(KERN_ERR "Could not add device! error=%d", ret); BUG(); } // Perform DMA memory-to-memory of testMemory into (hopefully) readMemory dma_handle = dma_map_single(&pdev->dev, testMemory, sizeof(*testMemory), DMA_TO_DEVICE); BUG_ON(dma_mapping_error(&pdev->dev, dma_handle)); for(i = 0; i < 1000000; i++) { // I don't know if DMA will be performed immediately, so I poll on it a bit... if (readMemory == 0) { __asm__ __volatile__ ("NOP"); } else { break; } } // Debug information... Prints out whether we read the value from CPU Cache or not printk(KERN_INFO "testMemory = %d, readMemory = %d", *testMemory, *readMemory); dma_unmap_single(&pdev->dev, dma_handle, size, DMA_TO_DEVICE); return 0; } static void __exit dummymodule_exit(void){ kvfree(testMemory); kvfree(readMemory); platform_device_unregister(pdev); } module_init(dummymodule_init); module_exit(dummymodule_exit); 。废话。

strcat()

您有十六进制数据的二进制缓冲区，只需使用strcat()函数：

MIDIfile << strcat(header, strcat(track_header, track_data));

并一次写入一个缓冲区。

Answer 2

formatted operator<< function for char arrays or pointers用于打印 空终止 字符串。如果您的“十六进制数据”包含任何零（二进制零，0，而不是字符'0'），那么它将用作终止符。

更不用说缓冲区溢出了，因为您将附加到固定大小的数组中，该数组专门针对初始化数据而设置。

解决方案是首先不要将任意二进制数据用作字符串（strcat函数 希望数据为以空值结尾的字符串），而不能用作原始数据。其次，您需要使用the write function写入任意数据。