io_uring笔记

io_uring 异步

要求内核linux 5.10
异步四元组：1、init（create）2、commit 3、callback 4、destory
fio : 测iops一秒钟读写磁盘的次数

方式	磁盘iops
io_uring	11.6k
libaio	10.7k
psync	5k
spdk(内存操作 , 尽可能都放到用户态)	157k

io_uring 使用

io_uring内核api：
io uring. setup();
io uring register();
io_ uring_ enter();

liburing接口：

1. io_uring. _queue. init. params
2. io_uring_prep_accept
3. io_ uring. prep_recv
4. io_uring_prep_send(uring, )
5. io_ uring_submit0;
6. io_uring wait_cqe
7. io_uring_ peek_ batch_ cqe
8. io_uring_cq_advance//取消事件

猜测运行机制：

io _uring. _setup();
做两件事
1）注册两个队列
2）使用mmap映射内核的队列和用户态的队列：

echo_websever:

// io_uring, tcp server

// multhread, select/poll, epoll, coroutine, iouring
// reactor

// io 

#include <liburing.h>

#include <stdio.h>
#include <string.h>

#include <sys/socket.h>
#include <netinet/in.h>

#include <unistd.h>

#define ENTRIES_LENGTH		4096



enum {

	READ,
	WRITE,
	ACCEPT,

};

struct conninfo {
	int connfd;
	int type;
};


void set_read_event(struct io_uring *ring, int fd, void *buf, size_t len, int flags) {

	struct io_uring_sqe *sqe = io_uring_get_sqe(ring);

	io_uring_prep_recv(sqe, fd, buf, len, flags);

	struct conninfo ci = {
		.connfd = fd,
		.type = READ
	};

	memcpy(&sqe->user_data, &ci, sizeof(struct conninfo));

}



void set_write_event(struct io_uring *ring, int fd, const void *buf, size_t len, int flags) {

	struct io_uring_sqe *sqe = io_uring_get_sqe(ring);

	io_uring_prep_send(sqe, fd, buf, len, flags);

	struct conninfo ci = {
		.connfd = fd,
		.type = WRITE
	};

	memcpy(&sqe->user_data, &ci, sizeof(struct conninfo));

}




void set_accept_event(struct io_uring *ring, int fd,
	struct sockaddr *cliaddr, socklen_t *clilen, unsigned flags) {

	struct io_uring_sqe *sqe = io_uring_get_sqe(ring);

	io_uring_prep_accept(sqe, fd, cliaddr, clilen, flags);

	struct conninfo ci = {
		.connfd = fd,
		.type = ACCEPT
	};

	memcpy(&sqe->user_data, &ci, sizeof(struct conninfo));

}


int main() {

	int listenfd = socket(AF_INET, SOCK_STREAM, 0);  // 
    if (listenfd == -1) return -1;
// listenfd
    struct sockaddr_in servaddr, clientaddr;
    servaddr.sin_family = AF_INET;
    servaddr.sin_addr.s_addr = htonl(INADDR_ANY);
    servaddr.sin_port = htons(9999);

    if (-1 == bind(listenfd, (struct sockaddr*)&servaddr, sizeof(servaddr))) {
            return -2;
    }
	
	listen(listenfd, 10);

	struct io_uring_params params;
	memset(&params, 0, sizeof(params));

// epoll --> 
	struct io_uring ring;
	io_uring_queue_init_params(ENTRIES_LENGTH, &ring, &params);

	socklen_t clilen = sizeof(clientaddr);
	set_accept_event(&ring, listenfd, (struct sockaddr*)&clientaddr, &clilen, 0);
	
	char buffer[1024] = {0};
//
	while (1) {

		struct io_uring_cqe *cqe;//从sqe地址复制地址
		
		io_uring_submit(&ring) ; //提交，在这个阶段进行accpet和recv、write等

		int ret = io_uring_wait_cqe(&ring, &cqe); //回调

		struct io_uring_cqe *cqes[10];
		int cqecount = io_uring_peek_batch_cqe(&ring, cqes, 10);// 10是每次最多返回10个队列元素
		int i = 0;
		unsigned count = 0;
		
		for (i = 0;i < cqecount;i ++) {

			cqe = cqes[i];
			count ++;

			struct conninfo ci;
			memcpy(&ci, &cqe->user_data, sizeof(ci));

			if (ci.type == ACCEPT) {

				int connfd = cqe->res;
				
				set_read_event(&ring, connfd, buffer, 1024, 0);

			} else if (ci.type == READ) {

				int bytes_read = cqe->res;
				if (bytes_read == 0) {
					close(ci.connfd);
				} else if (bytes_read < 0) {

				} else {
					
					printf("buffer : %s\n", buffer);
					set_write_event(&ring, ci.connfd, buffer, bytes_read, 0);
				}
			} else if (ci.type == WRITE) {

				set_read_event(&ring, ci.connfd, buffer, 1024, 0);

			}
			

		}
		
		io_uring_cq_advance(&ring, count);
	}
	

}

对比于epoll的效果

都采用echo_websever:
epoll: 60w
io_uring: 65w

io_uring效果好在哪？

epoll是把io放入了红黑树
io_uring共享内存：存储空间、修改变量