dir.c

linux 内核源代码

/*
 *  linux/fs/nfs/dir.c
 *
 *  Copyright (C) 1992  Rick Sladkey
 *
 *  nfs directory handling functions
 *
 * 10 Apr 1996	Added silly rename for unlink	--okir
 * 28 Sep 1996	Improved directory cache --okir
 * 23 Aug 1997  Claus Heine [e-mail] 
 *              Re-implemented silly rename for unlink, newly implemented
 *              silly rename for nfs_rename() following the suggestions
 *              of Olaf Kirch (okir) found in this file.
 *              Following Linus comments on my original hack, this version
 *              depends only on the dcache stuff and doesn't touch the inode
 *              layer (iput() and friends).
 *  6 Jun 1999	Cache readdir lookups in the page cache. -DaveM
 */

#include <linux/time.h>
#include <linux/errno.h>
#include <linux/stat.h>
#include <linux/fcntl.h>
#include <linux/string.h>
#include <linux/kernel.h>
#include <linux/slab.h>
#include <linux/mm.h>
#include <linux/sunrpc/clnt.h>
#include <linux/nfs_fs.h>
#include <linux/nfs_mount.h>
#include <linux/pagemap.h>
#include <linux/smp_lock.h>
#include <linux/pagevec.h>
#include <linux/namei.h>
#include <linux/mount.h>
#include <linux/sched.h>

#include "nfs4_fs.h"
#include "delegation.h"
#include "iostat.h"
#include "internal.h"

/* #define NFS_DEBUG_VERBOSE 1 */

static int nfs_opendir(struct inode *, struct file *);
static int nfs_readdir(struct file *, void *, filldir_t);
static struct dentry *nfs_lookup(struct inode *, struct dentry *, struct nameidata *);
static int nfs_create(struct inode *, struct dentry *, int, struct nameidata *);
static int nfs_mkdir(struct inode *, struct dentry *, int);
static int nfs_rmdir(struct inode *, struct dentry *);
static int nfs_unlink(struct inode *, struct dentry *);
static int nfs_symlink(struct inode *, struct dentry *, const char *);
static int nfs_link(struct dentry *, struct inode *, struct dentry *);
static int nfs_mknod(struct inode *, struct dentry *, int, dev_t);
static int nfs_rename(struct inode *, struct dentry *,
		      struct inode *, struct dentry *);
static int nfs_fsync_dir(struct file *, struct dentry *, int);
static loff_t nfs_llseek_dir(struct file *, loff_t, int);

const struct file_operations nfs_dir_operations = {
	.llseek		= nfs_llseek_dir,
	.read		= generic_read_dir,
	.readdir	= nfs_readdir,
	.open		= nfs_opendir,
	.release	= nfs_release,
	.fsync		= nfs_fsync_dir,
};

const struct inode_operations nfs_dir_inode_operations = {
	.create		= nfs_create,
	.lookup		= nfs_lookup,
	.link		= nfs_link,
	.unlink		= nfs_unlink,
	.symlink	= nfs_symlink,
	.mkdir		= nfs_mkdir,
	.rmdir		= nfs_rmdir,
	.mknod		= nfs_mknod,
	.rename		= nfs_rename,
	.permission	= nfs_permission,
	.getattr	= nfs_getattr,
	.setattr	= nfs_setattr,
};

#ifdef CONFIG_NFS_V3
const struct inode_operations nfs3_dir_inode_operations = {
	.create		= nfs_create,
	.lookup		= nfs_lookup,
	.link		= nfs_link,
	.unlink		= nfs_unlink,
	.symlink	= nfs_symlink,
	.mkdir		= nfs_mkdir,
	.rmdir		= nfs_rmdir,
	.mknod		= nfs_mknod,
	.rename		= nfs_rename,
	.permission	= nfs_permission,
	.getattr	= nfs_getattr,
	.setattr	= nfs_setattr,
	.listxattr	= nfs3_listxattr,
	.getxattr	= nfs3_getxattr,
	.setxattr	= nfs3_setxattr,
	.removexattr	= nfs3_removexattr,
};
#endif  /* CONFIG_NFS_V3 */

#ifdef CONFIG_NFS_V4

static struct dentry *nfs_atomic_lookup(struct inode *, struct dentry *, struct nameidata *);
const struct inode_operations nfs4_dir_inode_operations = {
	.create		= nfs_create,
	.lookup		= nfs_atomic_lookup,
	.link		= nfs_link,
	.unlink		= nfs_unlink,
	.symlink	= nfs_symlink,
	.mkdir		= nfs_mkdir,
	.rmdir		= nfs_rmdir,
	.mknod		= nfs_mknod,
	.rename		= nfs_rename,
	.permission	= nfs_permission,
	.getattr	= nfs_getattr,
	.setattr	= nfs_setattr,
	.getxattr       = nfs4_getxattr,
	.setxattr       = nfs4_setxattr,
	.listxattr      = nfs4_listxattr,
};

#endif /* CONFIG_NFS_V4 */

/*
 * Open file
 */
static int
nfs_opendir(struct inode *inode, struct file *filp)
{
	int res;

	dfprintk(VFS, "NFS: opendir(%s/%ld)\n",
			inode->i_sb->s_id, inode->i_ino);

	lock_kernel();
	/* Call generic open code in order to cache credentials */
	res = nfs_open(inode, filp);
	unlock_kernel();
	return res;
}

typedef __be32 * (*decode_dirent_t)(__be32 *, struct nfs_entry *, int);
typedef struct {
	struct file	*file;
	struct page	*page;
	unsigned long	page_index;
	__be32		*ptr;
	u64		*dir_cookie;
	loff_t		current_index;
	struct nfs_entry *entry;
	decode_dirent_t	decode;
	int		plus;
	int		error;
	unsigned long	timestamp;
	int		timestamp_valid;
} nfs_readdir_descriptor_t;

/* Now we cache directories properly, by stuffing the dirent
 * data directly in the page cache.
 *
 * Inode invalidation due to refresh etc. takes care of
 * _everything_, no sloppy entry flushing logic, no extraneous
 * copying, network direct to page cache, the way it was meant
 * to be.
 *
 * NOTE: Dirent information verification is done always by the
 *	 page-in of the RPC reply, nowhere else, this simplies
 *	 things substantially.
 */
static
int nfs_readdir_filler(nfs_readdir_descriptor_t *desc, struct page *page)
{
	struct file	*file = desc->file;
	struct inode	*inode = file->f_path.dentry->d_inode;
	struct rpc_cred	*cred = nfs_file_cred(file);
	unsigned long	timestamp;
	int		error;

	dfprintk(DIRCACHE, "NFS: %s: reading cookie %Lu into page %lu\n",
			__FUNCTION__, (long long)desc->entry->cookie,
			page->index);

 again:
	timestamp = jiffies;
	error = NFS_PROTO(inode)->readdir(file->f_path.dentry, cred, desc->entry->cookie, page,
					  NFS_SERVER(inode)->dtsize, desc->plus);
	if (error < 0) {
		/* We requested READDIRPLUS, but the server doesn't grok it */
		if (error == -ENOTSUPP && desc->plus) {
			NFS_SERVER(inode)->caps &= ~NFS_CAP_READDIRPLUS;
			clear_bit(NFS_INO_ADVISE_RDPLUS, &NFS_FLAGS(inode));
			desc->plus = 0;
			goto again;
		}
		goto error;
	}
	desc->timestamp = timestamp;
	desc->timestamp_valid = 1;
	SetPageUptodate(page);
	/* Ensure consistent page alignment of the data.
	 * Note: assumes we have exclusive access to this mapping either
	 *	 through inode->i_mutex or some other mechanism.
	 */
	if (page->index == 0 && invalidate_inode_pages2_range(inode->i_mapping, PAGE_CACHE_SIZE, -1) < 0) {
		/* Should never happen */
		nfs_zap_mapping(inode, inode->i_mapping);
	}
	unlock_page(page);
	return 0;
 error:
	unlock_page(page);
	desc->error = error;
	return -EIO;
}

static inline
int dir_decode(nfs_readdir_descriptor_t *desc)
{
	__be32	*p = desc->ptr;
	p = desc->decode(p, desc->entry, desc->plus);
	if (IS_ERR(p))
		return PTR_ERR(p);
	desc->ptr = p;
	if (desc->timestamp_valid)
		desc->entry->fattr->time_start = desc->timestamp;
	else
		desc->entry->fattr->valid &= ~NFS_ATTR_FATTR;
	return 0;
}

static inline
void dir_page_release(nfs_readdir_descriptor_t *desc)
{
	kunmap(desc->page);
	page_cache_release(desc->page);
	desc->page = NULL;
	desc->ptr = NULL;
}

/*
 * Given a pointer to a buffer that has already been filled by a call
 * to readdir, find the next entry with cookie '*desc->dir_cookie'.
 *
 * If the end of the buffer has been reached, return -EAGAIN, if not,
 * return the offset within the buffer of the next entry to be
 * read.
 */
static inline
int find_dirent(nfs_readdir_descriptor_t *desc)
{
	struct nfs_entry *entry = desc->entry;
	int		loop_count = 0,
			status;

	while((status = dir_decode(desc)) == 0) {
		dfprintk(DIRCACHE, "NFS: %s: examining cookie %Lu\n",
				__FUNCTION__, (unsigned long long)entry->cookie);
		if (entry->prev_cookie == *desc->dir_cookie)
			break;
		if (loop_count++ > 200) {
			loop_count = 0;
			schedule();
		}
	}
	return status;
}

/*
 * Given a pointer to a buffer that has already been filled by a call
 * to readdir, find the entry at offset 'desc->file->f_pos'.
 *
 * If the end of the buffer has been reached, return -EAGAIN, if not,
 * return the offset within the buffer of the next entry to be
 * read.
 */
static inline
int find_dirent_index(nfs_readdir_descriptor_t *desc)
{
	struct nfs_entry *entry = desc->entry;
	int		loop_count = 0,
			status;

	for(;;) {
		status = dir_decode(desc);
		if (status)
			break;

		dfprintk(DIRCACHE, "NFS: found cookie %Lu at index %Ld\n",
				(unsigned long long)entry->cookie, desc->current_index);

		if (desc->file->f_pos == desc->current_index) {
			*desc->dir_cookie = entry->cookie;
			break;
		}
		desc->current_index++;
		if (loop_count++ > 200) {
			loop_count = 0;
			schedule();
		}
	}
	return status;
}

/*
 * Find the given page, and call find_dirent() or find_dirent_index in
 * order to try to return the next entry.
 */
static inline
int find_dirent_page(nfs_readdir_descriptor_t *desc)
{
	struct inode	*inode = desc->file->f_path.dentry->d_inode;
	struct page	*page;
	int		status;

	dfprintk(DIRCACHE, "NFS: %s: searching page %ld for target %Lu\n",
			__FUNCTION__, desc->page_index,
			(long long) *desc->dir_cookie);

	/* If we find the page in the page_cache, we cannot be sure
	 * how fresh the data is, so we will ignore readdir_plus attributes.
	 */
	desc->timestamp_valid = 0;
	page = read_cache_page(inode->i_mapping, desc->page_index,
			       (filler_t *)nfs_readdir_filler, desc);
	if (IS_ERR(page)) {
		status = PTR_ERR(page);
		goto out;
	}

	/* NOTE: Someone else may have changed the READDIRPLUS flag */
	desc->page = page;
	desc->ptr = kmap(page);		/* matching kunmap in nfs_do_filldir */
	if (*desc->dir_cookie != 0)
		status = find_dirent(desc);
	else
		status = find_dirent_index(desc);
	if (status < 0)
		dir_page_release(desc);
 out:
	dfprintk(DIRCACHE, "NFS: %s: returns %d\n", __FUNCTION__, status);
	return status;
}

/*
 * Recurse through the page cache pages, and return a
 * filled nfs_entry structure of the next directory entry if possible.
 *
 * The target for the search is '*desc->dir_cookie' if non-0,
 * 'desc->file->f_pos' otherwise
 */
static inline
int readdir_search_pagecache(nfs_readdir_descriptor_t *desc)
{
	int		loop_count = 0;
	int		res;

	/* Always search-by-index from the beginning of the cache */
	if (*desc->dir_cookie == 0) {
		dfprintk(DIRCACHE, "NFS: readdir_search_pagecache() searching for offset %Ld\n",
				(long long)desc->file->f_pos);
		desc->page_index = 0;
		desc->entry->cookie = desc->entry->prev_cookie = 0;
		desc->entry->eof = 0;
		desc->current_index = 0;
	} else
		dfprintk(DIRCACHE, "NFS: readdir_search_pagecache() searching for cookie %Lu\n",
				(unsigned long long)*desc->dir_cookie);

	for (;;) {
		res = find_dirent_page(desc);
		if (res != -EAGAIN)
			break;
		/* Align to beginning of next page */
		desc->page_index ++;
		if (loop_count++ > 200) {
			loop_count = 0;
			schedule();
		}
	}

	dfprintk(DIRCACHE, "NFS: %s: returns %d\n", __FUNCTION__, res);
	return res;
}

static inline unsigned int dt_type(struct inode *inode)
{
	return (inode->i_mode >> 12) & 15;
}

static struct dentry *nfs_readdir_lookup(nfs_readdir_descriptor_t *desc);

/*
 * Once we've found the start of the dirent within a page: fill 'er up...
 */
static 
int nfs_do_filldir(nfs_readdir_descriptor_t *desc, void *dirent,
		   filldir_t filldir)
{
	struct file	*file = desc->file;
	struct nfs_entry *entry = desc->entry;
	struct dentry	*dentry = NULL;
	u64		fileid;
	int		loop_count = 0,
			res;

	dfprintk(DIRCACHE, "NFS: nfs_do_filldir() filling starting @ cookie %Lu\n",
			(unsigned long long)entry->cookie);

	for(;;) {
		unsigned d_type = DT_UNKNOWN;
		/* Note: entry->prev_cookie contains the cookie for
		 *	 retrieving the current dirent on the server */
		fileid = entry->ino;

		/* Get a dentry if we have one */
		if (dentry != NULL)
			dput(dentry);
		dentry = nfs_readdir_lookup(desc);

		/* Use readdirplus info */
		if (dentry != NULL && dentry->d_inode != NULL) {
			d_type = dt_type(dentry->d_inode);
			fileid = NFS_FILEID(dentry->d_inode);
		}

		res = filldir(dirent, entry->name, entry->len, 
			      file->f_pos, nfs_compat_user_ino64(fileid),
			      d_type);
		if (res < 0)
			break;
		file->f_pos++;
		*desc->dir_cookie = entry->cookie;
		if (dir_decode(desc) != 0) {
			desc->page_index ++;
			break;
		}
		if (loop_count++ > 200) {
			loop_count = 0;
			schedule();
		}
	}
	dir_page_release(desc);
	if (dentry != NULL)
		dput(dentry);
	dfprintk(DIRCACHE, "NFS: nfs_do_filldir() filling ended @ cookie %Lu; returning = %d\n",
			(unsigned long long)*desc->dir_cookie, res);
	return res;
}

/*
 * If we cannot find a cookie in our cache, we suspect that this is
 * because it points to a deleted file, so we ask the server to return
 * whatever it thinks is the next entry. We then feed this to filldir.
 * If all goes well, we should then be able to find our way round the
 * cache on the next call to readdir_search_pagecache();
 *
 * NOTE: we cannot add the anonymous page to the pagecache because
 *	 the data it contains might not be page aligned. Besides,
 *	 we should already have a complete representation of the
 *	 directory in the page cache by the time we get here.
 */
static inline
int uncached_readdir(nfs_readdir_descriptor_t *desc, void *dirent,
		     filldir_t filldir)
{
	struct file	*file = desc->file;
	struct inode	*inode = file->f_path.dentry->d_inode;
	struct rpc_cred	*cred = nfs_file_cred(file);
	struct page	*page = NULL;
	int		status;
	unsigned long	timestamp;

	dfprintk(DIRCACHE, "NFS: uncached_readdir() searching for cookie %Lu\n",
			(unsigned long long)*desc->dir_cookie);

	page = alloc_page(GFP_HIGHUSER);
	if (!page) {
		status = -ENOMEM;
		goto out;
	}
	timestamp = jiffies;
	desc->error = NFS_PROTO(inode)->readdir(file->f_path.dentry, cred, *desc->dir_cookie,
						page,
						NFS_SERVER(inode)->dtsize,
						desc->plus);
	desc->page = page;