format.h

Linux 的最重要的特徵之一是它的为许多不同的文件系统的支持。这使其非常灵活从而与许多另外的操作系统可以很好的共存。linux采用的是树型结构。最上层是根目录

format()
{
 	struct inode *inode;
	struct direct dir_buf[BLOCKSIZ/(DIRSIZ+2)];
	struct pwd passwd[BLOCKSIZ/(PWDSIZ+4)];
	struct filsys filsys;
	unsigned int block_buf[BLOCKSIZ/sizeof(int)];
	char *buf;
	char *empty;
	int i,j,k;

	fd=fopen("filesystem","w+b");
	buf=(char*)malloc((DINODEBLK+FILEBLK+2)*BLOCKSIZ*sizeof(char));
	if(fd==NULL)
	{
		printf("\nfile system file creat failed!!\n");
		exit(0);
	}
	fseek(fd,0,SEEK_SET);
	fwrite(buf,1,(DINODEBLK+FILEBLK+2)*BLOCKSIZ*sizeof(char),fd);
	
	free(buf);
	
	passwd[0].p_uid=2116;
	passwd[0].p_gid=03;
	strcpy(passwd[0].password,"don1");
	passwd[1].p_uid=2117;
	passwd[1].p_gid=03;
	strcpy(passwd[1].password,"don2");
	passwd[2].p_uid=2118;
	passwd[2].p_gid=04;
	strcpy(passwd[2].password,"abcd");
	passwd[3].p_uid=2119;
	passwd[3].p_gid=04;
	strcpy(passwd[3].password,"don4");
	passwd[4].p_uid=2220;
	passwd[4].p_gid=05;
	strcpy(passwd[4].password,"don5");
	
	
	inode=iget(0);
	inode->di_mode=DIEMPTY;
	iput(inode);

	inode=iget(1);
	inode->di_number=1;
	inode->di_mode=DEFAULTMODE|DIDIR;
	inode->di_size=3*(DIRSIZ+2);
	inode->di_addr[0]=0;
	
	strcpy(dir_buf[0].d_name,"..");
	dir_buf[0].d_ino=1;
	strcpy(dir_buf[1].d_name,".");
	dir_buf[1].d_ino=1;
	strcpy(dir_buf[2].d_name,"etc");
	dir_buf[2].d_ino=2;

	fseek(fd,DATASTART,SEEK_SET);
	fwrite(dir_buf,1,3*(DIRSIZ+2),fd);
	
	iput(inode);
			fseek(fd,1056,SEEK_SET);
			fread(inode,DINODESIZ,1,fd);

	inode=iget(1);
	iput(inode);
	
	inode=iget(2);
	inode->di_number=1;
	inode->di_mode=DEFAULTMODE|DIDIR;
	inode->di_size=3*(DIRSIZ+2);
	inode->di_addr[0]=1;
	strcpy(dir_buf[0].d_name,"..");
	dir_buf[0].d_ino=1;
	strcpy(dir_buf[1].d_name,".");
	dir_buf[1].d_ino=2;
	strcpy(dir_buf[2].d_name,"password");
	dir_buf[2].d_ino=3;

	fseek(fd,DATASTART+BLOCKSIZ*1,SEEK_SET);
	fwrite(dir_buf,1,3*(DIRSIZ+2),fd);
	iput(inode);
	inode=iget(3);
	inode->di_number=1;
	inode->di_mode=DEFAULTMODE|DIFILE;
	inode->di_size=BLOCKSIZ;
	inode->di_addr[0]=2;
	for(i=5;i<PWDNUM;i++)
	{
		passwd[i].p_uid=0;
		passwd[i].p_gid=0;
		strcpy(passwd[i].password,"       ");
	}

	fseek(fd,DATASTART+2*BLOCKSIZ,SEEK_SET);
	fwrite(passwd,1,BLOCKSIZ,fd);
	iput(inode);
	filsys.s_isize=DINODEBLK;
	filsys.s_fsize=FILEBLK;
	filsys.s_ninode=DINODEBLK*BLOCKSIZ/DINODESIZ-4;
	filsys.s_nfree=FILEBLK-3;
	for(i=0;i<NICINOD;i++)
	{
		filsys.s_inode[i]=4+i;
	}
	filsys.s_pinode=0;
	filsys.s_rinode=NICINOD+4;
	
	for(i=NICFREE+2;i<FILEBLK;i+=50)//为何要加2，参看149行的注释
	{
		for(j=0;j<NICFREE;j++)
		{
			block_buf[NICFREE-1-j]=i-j;
		}
		fseek(fd,DATASTART+BLOCKSIZ*(i-49),SEEK_SET);
		fwrite(block_buf,1,BLOCKSIZ,fd);
	}//当i=502之后，完成文件块502～453的写入；
	 //之后文件块512～503不能进行，需要特殊处理
	for(i=503;i<512;i++)
		block_buf[i-503]=i;
	fseek(fd,DATASTART+BLOCKSIZ*503,SEEK_SET);
	fwrite(block_buf,1,BLOCKSIZ,fd);//至此，才完成512块文件块的定位
	
	for(i=0;i<NICFREE;i++)
	{
		filsys.s_free[i]=i+3;//从DATASTART的第一个BLOCK作为MAIN DIRECTORY
		                                  //第二个BLOCK作为etc目录
		                                  //第三个BLOCK作为password文件
		                     //故此i要加3
	}
	//每50个BLOCK成组，在每个BLOCK组当中的第一个BLOCK（称为地址块）放有整个BLOCK
	//组的地址。这样，filsys.s_free[0]指向该组的地址块，而filsys.s_free[49]＋1
	//则指向下一个组的地址块。
	j=1;
	filsys.s_pfree=NICFREE-j;
	filsys.s_pinode=0;
	fseek(fd,BLOCKSIZ,SEEK_SET);
	fwrite(&filsys,1,sizeof(struct filsys),fd);
	fclose(fd);
}