zlramfs模块实现

参考ramfs实现一个Linux ramfs内核模块，作为Linux文件系统分析的一个脚手架。

Makefile：

ifneq ($(KERNELRELEASE),)
obj-m += zlramfs.o
else
KERNELDIR:=/lib/modules/$(shell uname -r)/build
PWD:=$(shell pwd)
all:
	$(MAKE) -C $(KERNELDIR) M=$(PWD) modules
 
clean:
	rm -rf *.o *.mod.c *.mod.o *.ko *.symvers *.mod .*.cmd *.order *.a
endif

源码文件zlramfs.c：

#include <linux/fs.h>
#include <linux/pagemap.h>
#include <linux/highmem.h>
#include <linux/time.h>
#include <linux/init.h>
#include <linux/string.h>
#include <linux/backing-dev.h>
#include <linux/sched.h>
#include <linux/parser.h>
#include <linux/magic.h>
#include <linux/slab.h>
#include <linux/uaccess.h>
#include <linux/fs_context.h>
#include <linux/fs_parser.h>

static unsigned long zlramfs_mmu_get_unmapped_area(struct file *file,
		unsigned long addr, unsigned long len, unsigned long pgoff,
		unsigned long flags)
{
	return current->mm->get_unmapped_area(file, addr, len, pgoff, flags);
}

const struct file_operations zlramfs_file_operations = {
	.read_iter	= generic_file_read_iter,
	.write_iter	= generic_file_write_iter,
	.mmap		= generic_file_mmap,
	.fsync		= noop_fsync,
	.splice_read	= generic_file_splice_read,
	.splice_write	= iter_file_splice_write,
	.llseek		= generic_file_llseek,
	.get_unmapped_area	= zlramfs_mmu_get_unmapped_area,
};

const struct inode_operations zlramfs_file_inode_operations = {
	.setattr	= simple_setattr,
	.getattr	= simple_getattr,
};
struct zlramfs_mount_opts {
	umode_t mode;
};

struct zlramfs_fs_info {
	struct zlramfs_mount_opts mount_opts;
};

#define RAMFS_DEFAULT_MODE	0755

static const struct super_operations zlramfs_ops;
static const struct inode_operations zlramfs_dir_inode_operations;

struct inode *zlramfs_get_inode(struct super_block *sb,
				const struct inode *dir, umode_t mode, dev_t dev)
{
	struct inode * inode = new_inode(sb);

	if (inode) {
		inode->i_ino = get_next_ino();
		inode_init_owner(&init_user_ns, inode, dir, mode);
		inode->i_mapping->a_ops = &ram_aops;
		mapping_set_gfp_mask(inode->i_mapping, GFP_HIGHUSER);
		mapping_set_unevictable(inode->i_mapping);
		inode->i_atime = inode->i_mtime = inode->i_ctime = current_time(inode);
		switch (mode & S_IFMT) {
		default:
			init_special_inode(inode, mode, dev);
			break;
		case S_IFREG:
			inode->i_op = &zlramfs_file_inode_operations;
			inode->i_fop = &zlramfs_file_operations;
			break;
		case S_IFDIR:
			inode->i_op = &zlramfs_dir_inode_operations;
			inode->i_fop = &simple_dir_operations;

			/* directory inodes start off with i_nlink == 2 (for "." entry) */
			inc_nlink(inode);
			break;
		case S_IFLNK:
			inode->i_op = &page_symlink_inode_operations;
			inode_nohighmem(inode);
			break;
		}
	}
	return inode;
}

/*
 * File creation. Allocate an inode, and we're done..
 */
/* SMP-safe */
static int
zlramfs_mknod(struct user_namespace *mnt_userns, struct inode *dir,
	    struct dentry *dentry, umode_t mode, dev_t dev)
{
	struct inode * inode = zlramfs_get_inode(dir->i_sb, dir, mode, dev);
	int error = -ENOSPC;

	if (inode) {
		d_instantiate(dentry, inode);
		dget(dentry);	/* Extra count - pin the dentry in core */
		error = 0;
		dir->i_mtime = dir->i_ctime = current_time(dir);
	}
	return error;
}

static int zlramfs_mkdir(struct user_namespace *mnt_userns, struct inode *dir,
		       struct dentry *dentry, umode_t mode)
{
	int retval = zlramfs_mknod(&init_user_ns, dir, dentry, mode | S_IFDIR, 0);
	if (!retval)
		inc_nlink(dir);
	return retval;
}

static int zlramfs_create(struct user_namespace *mnt_userns, struct inode *dir,
			struct dentry *dentry, umode_t mode, bool excl)
{
	return zlramfs_mknod(&init_user_ns, dir, dentry, mode | S_IFREG, 0);
}

static int zlramfs_symlink(struct user_namespace *mnt_userns, struct inode *dir,
			 struct dentry *dentry, const char *symname)
{
	struct inode *inode;
	int error = -ENOSPC;

	inode = zlramfs_get_inode(dir->i_sb, dir, S_IFLNK|S_IRWXUGO, 0);
	if (inode) {
		int l = strlen(symname)+1;
		error = page_symlink(inode, symname, l);
		if (!error) {
			d_instantiate(dentry, inode);
			dget(dentry);
			dir->i_mtime = dir->i_ctime = current_time(dir);
		} else
			iput(inode);
	}
	return error;
}

static int zlramfs_tmpfile(struct user_namespace *mnt_userns,
			 struct inode *dir, struct dentry *dentry, umode_t mode)
{
	struct inode *inode;

	inode = zlramfs_get_inode(dir->i_sb, dir, mode, 0);
	if (!inode)
		return -ENOSPC;
	d_tmpfile(dentry, inode);
	return 0;
}

static const struct inode_operations zlramfs_dir_inode_operations = {
	.create		= zlramfs_create,
	.lookup		= simple_lookup,
	.link		= simple_link,
	.unlink		= simple_unlink,
	.symlink	= zlramfs_symlink,
	.mkdir		= zlramfs_mkdir,
	.rmdir		= simple_rmdir,
	.mknod		= zlramfs_mknod,
	.rename		= simple_rename,
	.tmpfile	= zlramfs_tmpfile,
};

/*
 * Display the mount options in /proc/mounts.
 */
static int zlramfs_show_options(struct seq_file *m, struct dentry *root)
{
	struct zlramfs_fs_info *fsi = root->d_sb->s_fs_info;

	if (fsi->mount_opts.mode != RAMFS_DEFAULT_MODE)
		seq_printf(m, ",mode=%o", fsi->mount_opts.mode);
	return 0;
}

static const struct super_operations zlramfs_ops = {
	.statfs		= simple_statfs,
	.drop_inode	= generic_delete_inode,
	.show_options	= zlramfs_show_options,
};

enum zlramfs_param {
	Opt_mode,
};

const struct fs_parameter_spec zlramfs_fs_parameters[] = {
	fsparam_u32oct("mode",	Opt_mode),
	{}
};

static int zlramfs_parse_param(struct fs_context *fc, struct fs_parameter *param)
{
	struct fs_parse_result result;
	struct zlramfs_fs_info *fsi = fc->s_fs_info;
	int opt;

	opt = fs_parse(fc, zlramfs_fs_parameters, param, &result);
	if (opt < 0) {
		/*
		 * We might like to report bad mount options here;
		 * but traditionally zlramfs has ignored all mount options,
		 * and as it is used as a !CONFIG_SHMEM simple substitute
		 * for tmpfs, better continue to ignore other mount options.
		 */
		if (opt == -ENOPARAM)
			opt = 0;
		return opt;
	}

	switch (opt) {
	case Opt_mode:
		fsi->mount_opts.mode = result.uint_32 & S_IALLUGO;
		break;
	}

	return 0;
}

static int zlramfs_fill_super(struct super_block *sb, struct fs_context *fc)
{
	struct zlramfs_fs_info *fsi = sb->s_fs_info;
	struct inode *inode;

	sb->s_maxbytes		= MAX_LFS_FILESIZE;
	sb->s_blocksize		= PAGE_SIZE;
	sb->s_blocksize_bits	= PAGE_SHIFT;
	sb->s_magic		= RAMFS_MAGIC;
	sb->s_op		= &zlramfs_ops;
	sb->s_time_gran		= 1;

	inode = zlramfs_get_inode(sb, NULL, S_IFDIR | fsi->mount_opts.mode, 0);
	sb->s_root = d_make_root(inode);
	if (!sb->s_root)
		return -ENOMEM;

	return 0;
}

static int zlramfs_get_tree(struct fs_context *fc)
{
	return get_tree_nodev(fc, zlramfs_fill_super);
}

static void zlramfs_free_fc(struct fs_context *fc)
{
	kfree(fc->s_fs_info);
}

static const struct fs_context_operations zlramfs_context_ops = {
	.free		= zlramfs_free_fc,
	.parse_param	= zlramfs_parse_param,
	.get_tree	= zlramfs_get_tree,
};

int zlramfs_init_fs_context(struct fs_context *fc)
{
	struct zlramfs_fs_info *fsi;

	fsi = kzalloc(sizeof(*fsi), GFP_KERNEL);
	if (!fsi)
		return -ENOMEM;

	fsi->mount_opts.mode = RAMFS_DEFAULT_MODE;
	fc->s_fs_info = fsi;
	fc->ops = &zlramfs_context_ops;
	return 0;
}

static void zlramfs_kill_sb(struct super_block *sb)
{
	kfree(sb->s_fs_info);
	kill_litter_super(sb);
}

static struct file_system_type zlramfs_fs_type = {
	.name		= "zlramfs",
	.init_fs_context = zlramfs_init_fs_context,
	.parameters	= zlramfs_fs_parameters,
	.kill_sb	= zlramfs_kill_sb,
	.fs_flags	= FS_USERNS_MOUNT,
};

static int init_zlramfs_fs(void)
{
	return register_filesystem(&zlramfs_fs_type);
}

static void deinit_zlramfs_fs(void)
{
	unregister_filesystem(&zlramfs_fs_type);

	return;
}

module_init(init_zlramfs_fs);
module_exit(deinit_zlramfs_fs);
MODULE_AUTHOR("zlcao");
MODULE_LICENSE("GPL");

编译：

加载：

/proc/filesystems查看文件系统注册信息：

挂载文件系统：

$ mkdir testmount
$ sudo mount -t zlramfs zlramfs ./testmount/

查看挂载结果：

cat /proc/mounts |tail -6

zlramfs的后端设备为NONE，说明这个文件系统不需要后端磁盘作为介质。

调试：

在创建node的地方添加打印，创建目录和创建文件会经过这里：

dmesg信息：

lsmount输出

lsmount命令也是从/proc/mounts节点获取信息的，以下是strace的输出：

zlramfs挂载图

Ramfs

以上代码参考内核RAMFS的实现，代码路径在fs/ramfs,当MMU打开时，编译file-mmu.c，忽略掉file-nommu.c（注意是赋值语句，而非累加，所以MMU打开情况下忽略前面的赋值）。

ramfs编译文件包括file-mmu.c/inode.c。

struct dentry的引用计数

struct dentry的生命期由引用计数控制，引用计数在分配到时候初始化为1：

通过dget/dput增加/减少引用。

经过调试，发现内核对struct dentry的管理符合如下规律：

1.struct dentry 有被引用（引用计数大于0，就不会被销毁）.

2.struct dentry引用计数被初始化为1，代表当前目录"."对其的引用。

3.每一个子目录都会递增父目录的引用计数。

4.如果应用将目录作为当前工作目录，会递增目录的引用计数，离开目录后（无论去上级还是下一级目录),引用计数将会减1。

5.引用计数只会被子目录递增（包括表示当前目录的.)，不会被孙目录隔代递增。

如下的目录等级：

在没有控制台将此目录树中的目录作为工作目录时,dir1引用计数为2，dir2引用计数为5。

当删除dir2下所有子目录,dir1不变，dir2的引用计数变为1：

---

参考资料

Linux系统的各个mount点以及文件系统挂载分析_linux挂载点_papaofdoudou的博客-CSDN博客

结束