Import 2.3.7pre3

[davej-history.git] / fs / nfs / dir.c

/*
 *  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 claus@momo.math.rwth-aachen.de 
 *              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
 */

#define NFS_NEED_XDR_TYPES
#include <linux/sched.h>
#include <linux/errno.h>
#include <linux/stat.h>
#include <linux/fcntl.h>
#include <linux/string.h>
#include <linux/kernel.h>
#include <linux/malloc.h>
#include <linux/mm.h>
#include <linux/sunrpc/clnt.h>
#include <linux/nfs_fs.h>
#include <linux/nfs.h>
#include <linux/pagemap.h>

#include <asm/segment.h>        /* for fs functions */

#define NFS_PARANOIA 1
/* #define NFS_DEBUG_VERBOSE 1 */

static int nfs_safe_remove(struct dentry *);

static ssize_t nfs_dir_read(struct file *, char *, size_t, loff_t *);
static int nfs_readdir(struct file *, void *, filldir_t);
static struct dentry *nfs_lookup(struct inode *, struct dentry *);
static int nfs_create(struct inode *, struct dentry *, int);
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, int);
static int nfs_rename(struct inode *, struct dentry *,
                      struct inode *, struct dentry *);

static struct file_operations nfs_dir_operations = {
        NULL,                   /* lseek - default */
        nfs_dir_read,           /* read - bad */
        NULL,                   /* write - bad */
        nfs_readdir,            /* readdir */
        NULL,                   /* select - default */
        NULL,                   /* ioctl - default */
        NULL,                   /* mmap */
        nfs_open,               /* open */
        NULL,                   /* flush */
        nfs_release,            /* release */
        NULL                    /* fsync */
};

struct inode_operations nfs_dir_inode_operations = {
        &nfs_dir_operations,    /* default directory file-ops */
        nfs_create,             /* create */
        nfs_lookup,             /* lookup */
        nfs_link,               /* link */
        nfs_unlink,             /* unlink */
        nfs_symlink,            /* symlink */
        nfs_mkdir,              /* mkdir */
        nfs_rmdir,              /* rmdir */
        nfs_mknod,              /* mknod */
        nfs_rename,             /* rename */
        NULL,                   /* readlink */
        NULL,                   /* follow_link */
        NULL,                   /* readpage */
        NULL,                   /* writepage */
        NULL,                   /* bmap */
        NULL,                   /* truncate */
        NULL,                   /* permission */
        NULL,                   /* smap */
        NULL,                   /* updatepage */
        nfs_revalidate,         /* revalidate */
};

static ssize_t
nfs_dir_read(struct file *filp, char *buf, size_t count, loff_t *ppos)
{
        return -EISDIR;
}

/* Each readdir response is composed of entries which look
 * like the following, as per the NFSv2 RFC:
 *
 *      __u32   not_end                 zero if end of response
 *      __u32   file ID                 opaque ino_t
 *      __u32   namelen                 size of name string
 *      VAR     name string             the string, padded to modulo 4 bytes
 *      __u32   cookie                  opaque ID of next entry
 *
 * When you hit not_end being zero, the next __u32 is non-zero if
 * this is the end of the complete set of readdir entires for this
 * directory.  This can be used, for example, to initiate pre-fetch.
 *
 * In order to know what to ask the server for, we only need to know
 * the final cookie of the previous page, and offset zero has cookie
 * zero, so we cache cookie to page offset translations in chunks.
 */
#define COOKIES_PER_CHUNK (8 - ((sizeof(void *) / sizeof(__u32))))
struct nfs_cookie_table {
        struct nfs_cookie_table *next;
        __u32   cookies[COOKIES_PER_CHUNK];
};
static kmem_cache_t *nfs_cookie_cachep;

/* This whole scheme relies on the fact that dirent cookies
 * are monotonically increasing.
 *
 * Another invariant is that once we have a valid non-zero
 * EOF marker cached, we also have the complete set of cookie
 * table entries.
 *
 * We return the page offset assosciated with the page where
 * cookie must be if it exists at all, however if we can not
 * figure that out conclusively, we return < 0.
 */
static long __nfs_readdir_offset(struct inode *inode, __u32 cookie)
{
        struct nfs_cookie_table *p;
        unsigned long ret = 0;

        for(p = NFS_COOKIES(inode); p != NULL; p = p->next) {
                int i;

                for (i = 0; i < COOKIES_PER_CHUNK; i++) {
                        __u32 this_cookie = p->cookies[i];

                        /* End of known cookies, EOF is our only hope. */
                        if (!this_cookie)
                                goto check_eof;

                        /* Next cookie is larger, must be in previous page. */
                        if (this_cookie > cookie)
                                return ret;

                        ret += 1;

                        /* Exact cookie match, it must be in this page :-) */
                        if (this_cookie == cookie)
                                return ret;
                }
        }
check_eof:
        if (NFS_DIREOF(inode) != 0)
                return ret;

        return -1L;
}

static __inline__ long nfs_readdir_offset(struct inode *inode, __u32 cookie)
{
        /* Cookie zero is always at page offset zero.   Optimize the
         * other common case since most directories fit entirely
         * in one page.
         */
        if (!cookie || (!NFS_COOKIES(inode) && NFS_DIREOF(inode)))
                return 0;
        return __nfs_readdir_offset(inode, cookie);
}

/* Since a cookie of zero is declared special by the NFS
 * protocol, we easily can tell if a cookie in an existing
 * table chunk is valid or not.
 *
 * NOTE: The cookies are indexed off-by-one because zero
 *       need not an entry.
 */
static __inline__ __u32 *find_cookie(struct inode *inode, unsigned long off)
{
        static __u32 cookie_zero = 0;
        struct nfs_cookie_table *p;
        __u32 *ret;

        if (!off)
                return &cookie_zero;
        off -= 1;
        p = NFS_COOKIES(inode);
        while(off >= COOKIES_PER_CHUNK && p) {
                off -= COOKIES_PER_CHUNK;
                p = p->next;
        }
        ret = NULL;
        if (p) {
                ret = &p->cookies[off];
                if (!*ret)
                        ret = NULL;
        }
        return ret;
}

#define NFS_NAMELEN_ALIGN(__len) ((((__len)+3)>>2)<<2)
static int create_cookie(__u32 cookie, unsigned long off, struct inode *inode)
{
        struct nfs_cookie_table **cpp;

        cpp = (struct nfs_cookie_table **) &NFS_COOKIES(inode);
        while (off >= COOKIES_PER_CHUNK && *cpp) {
                off -= COOKIES_PER_CHUNK;
                cpp = &(*cpp)->next;
        }
        if (*cpp) {
                (*cpp)->cookies[off] = cookie;
        } else {
                struct nfs_cookie_table *new;
                int i;

                new = kmem_cache_alloc(nfs_cookie_cachep, SLAB_ATOMIC);
                if(!new)
                        return -1;
                *cpp = new;
                new->next = NULL;
                for(i = 0; i < COOKIES_PER_CHUNK; i++) {
                        if (i == off) {
                                new->cookies[i] = cookie;
                        } else {
                                new->cookies[i] = 0;
                        }
                }
        }
        return 0;
}

static struct page *try_to_get_dirent_page(struct file *, __u32, int);

/* Recover from a revalidation flush.  The case here is that
 * the inode for the directory got invalidated somehow, and
 * all of our cached information is lost.  In order to get
 * a correct cookie for the current readdir request from the
 * user, we must (re-)fetch older readdir page cache entries.
 *
 * Returns < 0 if some error occurrs, else it is the page offset
 * to fetch.
 */
static long refetch_to_readdir_cookie(struct file *file, struct inode *inode)
{
        struct page *page;
        u32 goal_cookie = file->f_pos;
        long cur_off, ret = -1L;

again:
        cur_off = 0;
        for (;;) {
                page = find_get_page(inode, cur_off);
                if (page) {
                        if (!Page_Uptodate(page))
                                goto out_error;
                } else {
                        __u32 *cp = find_cookie(inode, cur_off);

                        if (!cp)
                                goto out_error;

                        page = try_to_get_dirent_page(file, *cp, 0);
                        if (!page) {
                                if (!cur_off)
                                        goto out_error;

                                /* Someone touched the dir on us. */
                                goto again;
                        }
                }
                page_cache_release(page);

                if ((ret = nfs_readdir_offset(inode, goal_cookie)) >= 0)
                        goto out;

                cur_off += 1;
        }
out:
        return ret;

out_error:
        if (page)
                page_cache_release(page);
        goto out;
}

/* 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 struct page *try_to_get_dirent_page(struct file *file, __u32 cookie, int refetch_ok)
{
        struct nfs_readdirargs rd_args;
        struct nfs_readdirres rd_res;
        struct dentry *dentry = file->f_dentry;
        struct inode *inode = dentry->d_inode;
        struct page *page, **hash;
        unsigned long page_cache;
        long offset;
        __u32 *cookiep;

        page = NULL;
        page_cache = page_cache_alloc();
        if (!page_cache)
                goto out;

        if ((offset = nfs_readdir_offset(inode, cookie)) < 0) {
                if (!refetch_ok ||
                    (offset = refetch_to_readdir_cookie(file, inode)) < 0) {
                        page_cache_free(page_cache);
                        goto out;
                }
        }

        cookiep = find_cookie(inode, offset);
        if (!cookiep) {
                /* Gross fatal error. */
                page_cache_free(page_cache);
                goto out;
        }

        hash = page_hash(inode, offset);
repeat:
        page = __find_lock_page(inode, offset, *hash);
        if (page) {
                page_cache_free(page_cache);
                goto unlock_out;
        }

        page = page_cache_entry(page_cache);
        if (add_to_page_cache_unique(page, inode, offset, hash)) {
                page_cache_release(page);
                goto repeat;
        }

        rd_args.fh = NFS_FH(dentry);
        rd_res.buffer = (char *)page_cache;
        rd_res.bufsiz = PAGE_CACHE_SIZE;
        rd_res.cookie = *cookiep;
        do {
                rd_args.buffer = rd_res.buffer;
                rd_args.bufsiz = rd_res.bufsiz;
                rd_args.cookie = rd_res.cookie;
                if (rpc_call(NFS_CLIENT(inode),
                             NFSPROC_READDIR, &rd_args, &rd_res, 0) < 0)
                        goto error;
        } while(rd_res.bufsiz > 0);

        if (rd_res.bufsiz < 0)
                NFS_DIREOF(inode) = rd_res.cookie;
        else if (create_cookie(rd_res.cookie, offset, inode))
                goto error;

        SetPageUptodate(page);
unlock_out:
        UnlockPage(page);
out:
        return page;

error:
        SetPageError(page);
        goto unlock_out;
}

/* Seek up to dirent assosciated with the passed in cookie,
 * then fill in dirents found.  Return the last cookie
 * actually given to the user, to update the file position.
 */
static __inline__ u32 nfs_do_filldir(__u32 *p, u32 cookie,
                                     void *dirent, filldir_t filldir)
{
        u32 end;

        while((end = *p++) != 0) {
                __u32 fileid, len, skip, this_cookie;
                char *name;

                fileid = *p++;
                len = *p++;
                name = (char *) p;
                skip = NFS_NAMELEN_ALIGN(len);
                p += (skip >> 2);
                this_cookie = *p++;

                if (this_cookie < cookie)
                        continue;

                cookie = this_cookie;
                if (filldir(dirent, name, len, cookie, fileid) < 0)
                        break;
        }

        return cookie;
}

/* The file offset position is represented in pure bytes, to
 * make the page cache interface straight forward.
 *
 * However, some way is needed to make the connection between the
 * opaque NFS directory entry cookies and our offsets, so a per-inode
 * cookie cache table is used.
 */
static int nfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
{
        struct dentry *dentry = filp->f_dentry;
        struct inode *inode = dentry->d_inode;
        struct page *page, **hash;
        long offset;
        int res;

        res = nfs_revalidate_inode(NFS_DSERVER(dentry), dentry);
        if (res < 0)
                return res;

        if (NFS_DIREOF(inode) && filp->f_pos >= NFS_DIREOF(inode))
                return 0;

        if ((offset = nfs_readdir_offset(inode, filp->f_pos)) < 0)
                goto no_dirent_page;

        hash = page_hash(inode, offset);
        page = __find_get_page(inode, offset, *hash);
        if (!page)
                goto no_dirent_page;
        if (!Page_Uptodate(page))
                goto dirent_read_error;
success:
        filp->f_pos = nfs_do_filldir((__u32 *) page_address(page),
                                     filp->f_pos, dirent, filldir);
        page_cache_release(page);
        return 0;

no_dirent_page:
        page = try_to_get_dirent_page(filp, filp->f_pos, 1);
        if (!page)
                goto no_page;

        if (Page_Uptodate(page))
                goto success;
dirent_read_error:
        page_cache_release(page);
no_page:
        return -EIO;
}

/* Flush directory cookie and EOF caches for an inode.
 * So we don't thrash allocating/freeing cookie tables,
 * we keep the cookies around until the inode is
 * deleted/reused.
 */
__inline__ void nfs_flush_dircache(struct inode *inode)
{
        struct nfs_cookie_table *p = NFS_COOKIES(inode);

        while (p != NULL) {
                int i;

                for(i = 0; i < COOKIES_PER_CHUNK; i++)
                        p->cookies[i] = 0;

                p = p->next;
        }
        NFS_DIREOF(inode) = 0;
}

/* Free up directory cache state, this happens when
 * nfs_delete_inode is called on an NFS directory.
 */
void nfs_free_dircache(struct inode *inode)
{
        struct nfs_cookie_table *p = NFS_COOKIES(inode);

        while (p != NULL) {
                struct nfs_cookie_table *next = p->next;
                kmem_cache_free(nfs_cookie_cachep, p);
                p = next;
        }
        NFS_COOKIES(inode) = NULL;
        NFS_DIREOF(inode) = 0;
}

/*
 * Whenever an NFS operation succeeds, we know that the dentry
 * is valid, so we update the revalidation timestamp.
 */
static inline void nfs_renew_times(struct dentry * dentry)
{
        dentry->d_time = jiffies;
}

static inline int nfs_dentry_force_reval(struct dentry *dentry, int flags)
{
        struct inode *inode = dentry->d_inode;
        unsigned long timeout = NFS_ATTRTIMEO(inode);

        /*
         * If it's the last lookup in a series, we use a stricter
         * cache consistency check by looking at the parent mtime.
         *
         * If it's been modified in the last hour, be really strict.
         * (This still means that we can avoid doing unnecessary
         * work on directories like /usr/share/bin etc which basically
         * never change).
         */
        if (!(flags & LOOKUP_CONTINUE)) {
                long diff = CURRENT_TIME - dentry->d_parent->d_inode->i_mtime;

                if (diff < 15*60)
                        timeout = 0;
        }
        
        return time_after(jiffies,dentry->d_time + timeout);
}

/*
 * We judge how long we want to trust negative
 * dentries by looking at the parent inode mtime.
 *
 * If mtime is close to present time, we revalidate
 * more often.
 */
static inline int nfs_neg_need_reval(struct dentry *dentry)
{
        unsigned long timeout = 30 * HZ;
        long diff = CURRENT_TIME - dentry->d_parent->d_inode->i_mtime;

        if (diff < 5*60)
                timeout = 1 * HZ;

        return time_after(jiffies, dentry->d_time + timeout);
}

/*
 * This is called every time the dcache has a lookup hit,
 * and we should check whether we can really trust that
 * lookup.
 *
 * NOTE! The hit can be a negative hit too, don't assume
 * we have an inode!
 *
 * If the dentry is older than the revalidation interval, 
 * we do a new lookup and verify that the dentry is still
 * correct.
 */
static int nfs_lookup_revalidate(struct dentry * dentry, int flags)
{
        struct dentry * parent = dentry->d_parent;
        struct inode * inode = dentry->d_inode;
        int error;
        struct nfs_fh fhandle;
        struct nfs_fattr fattr;

        /*
         * If we don't have an inode, let's look at the parent
         * directory mtime to get a hint about how often we
         * should validate things..
         */
        if (!inode) {
                if (nfs_neg_need_reval(dentry))
                        goto out_bad;
                goto out_valid;
        }

        if (is_bad_inode(inode)) {
                dfprintk(VFS, "nfs_lookup_validate: %s/%s has dud inode\n",
                        parent->d_name.name, dentry->d_name.name);
                goto out_bad;
        }

        if (IS_ROOT(dentry))
                goto out_valid;

        if (!nfs_dentry_force_reval(dentry, flags))
                goto out_valid;

        /*
         * Do a new lookup and check the dentry attributes.
         */
        error = nfs_proc_lookup(NFS_DSERVER(parent), NFS_FH(parent),
                                dentry->d_name.name, &fhandle, &fattr);
        if (error)
                goto out_bad;

        /* Inode number matches? */
        if (fattr.fileid != inode->i_ino)
                goto out_bad;

        /* Filehandle matches? */
        if (memcmp(dentry->d_fsdata, &fhandle, sizeof(struct nfs_fh))) {
                if (dentry->d_count < 2)
                        goto out_bad;
        }

        /* Ok, remeber that we successfully checked it.. */
        nfs_renew_times(dentry);
        nfs_refresh_inode(inode, &fattr);

out_valid:
        return 1;
out_bad:
        /* Purge readdir caches. */
        if (dentry->d_parent->d_inode) {
                invalidate_inode_pages(dentry->d_parent->d_inode);
                nfs_flush_dircache(dentry->d_parent->d_inode);
        }
        if (inode && S_ISDIR(inode->i_mode)) {
                invalidate_inode_pages(inode);
                nfs_flush_dircache(inode);
        }
        return 0;
}

/*
 * This is called from dput() when d_count is going to 0.
 * We use it to clean up silly-renamed files.
 */
static void nfs_dentry_delete(struct dentry *dentry)
{
        dfprintk(VFS, "NFS: dentry_delete(%s/%s, %x)\n",
                dentry->d_parent->d_name.name, dentry->d_name.name,
                dentry->d_flags);

        if (dentry->d_flags & DCACHE_NFSFS_RENAMED) {
                int error;
                
                dentry->d_flags &= ~DCACHE_NFSFS_RENAMED;
                /* Unhash it first */
                d_drop(dentry);
                error = nfs_safe_remove(dentry);
                if (error)
                        printk("NFS: can't silly-delete %s/%s, error=%d\n",
                                dentry->d_parent->d_name.name,
                                dentry->d_name.name, error);
        }

#ifdef NFS_PARANOIA
        /*
         * Sanity check: if the dentry has been unhashed and the
         * inode still has users, we could have problems ...
         */
        if (list_empty(&dentry->d_hash) && dentry->d_inode) {
                struct inode *inode = dentry->d_inode;
                int max_count = (S_ISDIR(inode->i_mode) ? 1 : inode->i_nlink);
                if (inode->i_count > max_count) {
printk("nfs_dentry_delete: %s/%s: ino=%ld, count=%d, nlink=%d\n",
dentry->d_parent->d_name.name, dentry->d_name.name,
inode->i_ino, inode->i_count, inode->i_nlink);
                }
        }
#endif
}

static kmem_cache_t *nfs_fh_cachep;

__inline__ struct nfs_fh *nfs_fh_alloc(void)
{
        return kmem_cache_alloc(nfs_fh_cachep, SLAB_KERNEL);
}

__inline__ void nfs_fh_free(struct nfs_fh *p)
{
        kmem_cache_free(nfs_fh_cachep, p);
}

/*
 * Called when the dentry is being freed to release private memory.
 */
static void nfs_dentry_release(struct dentry *dentry)
{
        if (dentry->d_fsdata)
                nfs_fh_free(dentry->d_fsdata);
}

struct dentry_operations nfs_dentry_operations = {
        nfs_lookup_revalidate,  /* d_revalidate(struct dentry *, int) */
        NULL,                   /* d_hash */
        NULL,                   /* d_compare */
        nfs_dentry_delete,      /* d_delete(struct dentry *) */
        nfs_dentry_release,     /* d_release(struct dentry *) */
        NULL                    /* d_iput */
};

#ifdef NFS_PARANOIA
/*
 * Display all dentries holding the specified inode.
 */
static void show_dentry(struct list_head * dlist)
{
        struct list_head *tmp = dlist;

        while ((tmp = tmp->next) != dlist) {
                struct dentry * dentry = list_entry(tmp, struct dentry, d_alias);
                const char * unhashed = "";

                if (list_empty(&dentry->d_hash))
                        unhashed = "(unhashed)";

                printk("show_dentry: %s/%s, d_count=%d%s\n",
                        dentry->d_parent->d_name.name,
                        dentry->d_name.name, dentry->d_count,
                        unhashed);
        }
}
#endif

static struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry)
{
        struct inode *inode;
        int error;
        struct nfs_fh fhandle;
        struct nfs_fattr fattr;

        dfprintk(VFS, "NFS: lookup(%s/%s)\n",
                dentry->d_parent->d_name.name, dentry->d_name.name);

        error = -ENAMETOOLONG;
        if (dentry->d_name.len > NFS_MAXNAMLEN)
                goto out;

        error = -ENOMEM;
        if (!dentry->d_fsdata) {
                dentry->d_fsdata = nfs_fh_alloc();
                if (!dentry->d_fsdata)
                        goto out;
        }
        dentry->d_op = &nfs_dentry_operations;

        error = nfs_proc_lookup(NFS_SERVER(dir), NFS_FH(dentry->d_parent), 
                                dentry->d_name.name, &fhandle, &fattr);
        inode = NULL;
        if (error == -ENOENT)
                goto no_entry;
        if (!error) {
                error = -EACCES;
                inode = nfs_fhget(dentry, &fhandle, &fattr);
                if (inode) {
#ifdef NFS_PARANOIA
if (inode->i_count > (S_ISDIR(inode->i_mode) ? 1 : inode->i_nlink)) {
printk("nfs_lookup: %s/%s ino=%ld in use, count=%d, nlink=%d\n",
dentry->d_parent->d_name.name, dentry->d_name.name,
inode->i_ino, inode->i_count, inode->i_nlink);
show_dentry(&inode->i_dentry);
}
#endif
            no_entry:
                        d_add(dentry, inode);
                        nfs_renew_times(dentry);
                        error = 0;
                }
        }
out:
        return ERR_PTR(error);
}

/*
 * Code common to create, mkdir, and mknod.
 */
static int nfs_instantiate(struct dentry *dentry, struct nfs_fh *fhandle,
                                struct nfs_fattr *fattr)
{
        struct inode *inode;
        int error = -EACCES;

        inode = nfs_fhget(dentry, fhandle, fattr);
        if (inode) {
#ifdef NFS_PARANOIA
if (inode->i_count > (S_ISDIR(inode->i_mode) ? 1 : inode->i_nlink)) {
printk("nfs_instantiate: %s/%s ino=%ld in use, count=%d, nlink=%d\n",
dentry->d_parent->d_name.name, dentry->d_name.name,
inode->i_ino, inode->i_count, inode->i_nlink);
show_dentry(&inode->i_dentry);
}
#endif
                d_instantiate(dentry, inode);
                nfs_renew_times(dentry);
                error = 0;
        }
        return error;
}

/*
 * Following a failed create operation, we drop the dentry rather
 * than retain a negative dentry. This avoids a problem in the event
 * that the operation succeeded on the server, but an error in the
 * reply path made it appear to have failed.
 */
static int nfs_create(struct inode *dir, struct dentry *dentry, int mode)
{
        int error;
        struct nfs_sattr sattr;
        struct nfs_fattr fattr;
        struct nfs_fh fhandle;

        dfprintk(VFS, "NFS: create(%x/%ld, %s\n",
                dir->i_dev, dir->i_ino, dentry->d_name.name);

        sattr.mode = mode;
        sattr.uid = sattr.gid = sattr.size = (unsigned) -1;
        sattr.atime.seconds = sattr.mtime.seconds = (unsigned) -1;

        /*
         * Invalidate the dir cache before the operation to avoid a race.
         */
        invalidate_inode_pages(dir);
        nfs_flush_dircache(dir);
        error = nfs_proc_create(NFS_SERVER(dir), NFS_FH(dentry->d_parent),
                        dentry->d_name.name, &sattr, &fhandle, &fattr);
        if (!error)
                error = nfs_instantiate(dentry, &fhandle, &fattr);
        if (error)
                d_drop(dentry);
        return error;
}

/*
 * See comments for nfs_proc_create regarding failed operations.
 */
static int nfs_mknod(struct inode *dir, struct dentry *dentry, int mode, int rdev)
{
        int error;
        struct nfs_sattr sattr;
        struct nfs_fattr fattr;
        struct nfs_fh fhandle;

        dfprintk(VFS, "NFS: mknod(%x/%ld, %s\n",
                dir->i_dev, dir->i_ino, dentry->d_name.name);

        sattr.mode = mode;
        sattr.uid = sattr.gid = sattr.size = (unsigned) -1;
        if (S_ISCHR(mode) || S_ISBLK(mode))
                sattr.size = rdev; /* get out your barf bag */
        sattr.atime.seconds = sattr.mtime.seconds = (unsigned) -1;

        invalidate_inode_pages(dir);
        nfs_flush_dircache(dir);
        error = nfs_proc_create(NFS_SERVER(dir), NFS_FH(dentry->d_parent),
                                dentry->d_name.name, &sattr, &fhandle, &fattr);
        if (!error)
                error = nfs_instantiate(dentry, &fhandle, &fattr);
        if (error)
                d_drop(dentry);
        return error;
}

/*
 * See comments for nfs_proc_create regarding failed operations.
 */
static int nfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
{
        int error;
        struct nfs_sattr sattr;
        struct nfs_fattr fattr;
        struct nfs_fh fhandle;

        dfprintk(VFS, "NFS: mkdir(%x/%ld, %s\n",
                dir->i_dev, dir->i_ino, dentry->d_name.name);

        sattr.mode = mode | S_IFDIR;
        sattr.uid = sattr.gid = sattr.size = (unsigned) -1;
        sattr.atime.seconds = sattr.mtime.seconds = (unsigned) -1;

        /*
         * Always drop the dentry, we can't always depend on
         * the fattr returned by the server (AIX seems to be
         * broken). We're better off doing another lookup than
         * depending on potentially bogus information.
         */
        d_drop(dentry);
        invalidate_inode_pages(dir);
        nfs_flush_dircache(dir);
        error = nfs_proc_mkdir(NFS_DSERVER(dentry), NFS_FH(dentry->d_parent),
                                dentry->d_name.name, &sattr, &fhandle, &fattr);
        return error;
}

static int nfs_rmdir(struct inode *dir, struct dentry *dentry)
{
        int error;

        dfprintk(VFS, "NFS: rmdir(%x/%ld, %s\n",
                dir->i_dev, dir->i_ino, dentry->d_name.name);

#ifdef NFS_PARANOIA
if (dentry->d_inode->i_count > 1)
printk("nfs_rmdir: %s/%s inode busy?? i_count=%d, i_nlink=%d\n",
dentry->d_parent->d_name.name, dentry->d_name.name,
dentry->d_inode->i_count, dentry->d_inode->i_nlink);
#endif

        invalidate_inode_pages(dir);
        nfs_flush_dircache(dir);
        error = nfs_proc_rmdir(NFS_SERVER(dir), NFS_FH(dentry->d_parent),
                                dentry->d_name.name);

        /* Update i_nlink and invalidate dentry. */
        if (!error) {
                d_drop(dentry);
                if (dentry->d_inode->i_nlink)
                        dentry->d_inode->i_nlink --;
        }

        return error;
}


/*  Note: we copy the code from lookup_dentry() here, only: we have to
 *  omit the directory lock. We are already the owner of the lock when
 *  we reach here. And "down(&dir->i_sem)" would make us sleep forever
 *  ('cause WE have the lock)
 * 
 *  VERY IMPORTANT: calculate the hash for this dentry!!!!!!!!
 *  Otherwise the cached lookup DEFINITELY WILL fail. And a new dentry
 *  is created. Without the DCACHE_NFSFS_RENAMED flag. And with d_count
 *  == 1. And trouble.
 *
 *  Concerning my choice of the temp name: it is just nice to have
 *  i_ino part of the temp name, as this offers another check whether
 *  somebody attempts to remove the "silly renamed" dentry itself.
 *  Which is something that I consider evil. Your opinion may vary.
 *  BUT:
 *  Now that I compute the hash value right, it should be possible to simply
 *  check for the DCACHE_NFSFS_RENAMED flag in dentry->d_flag instead of
 *  doing the string compare.
 *  WHICH MEANS:
 *  This offers the opportunity to shorten the temp name. Currently, I use
 *  the hex representation of i_ino + an event counter. This sums up to
 *  as much as 36 characters for a 64 bit machine, and needs 20 chars on 
 *  a 32 bit machine.
 *  QUINTESSENCE
 *  The use of i_ino is simply cosmetic. All we need is a unique temp
 *  file name for the .nfs files. The event counter seemed to be adequate.
 *  And as we retry in case such a file already exists, we are guaranteed
 *  to succeed.
 */

static
struct dentry *nfs_silly_lookup(struct dentry *parent, char *silly, int slen)
{
        struct qstr    sqstr;
        struct dentry *sdentry;
        struct dentry *res;

        sqstr.name = silly;
        sqstr.len  = slen;
        sqstr.hash = full_name_hash(silly, slen);
        sdentry = d_lookup(parent, &sqstr);
        if (!sdentry) {
                sdentry = d_alloc(parent, &sqstr);
                if (sdentry == NULL)
                        return ERR_PTR(-ENOMEM);
                res = nfs_lookup(parent->d_inode, sdentry);
                if (res) {
                        dput(sdentry);
                        return res;
                }
        }
        return sdentry;
}

static int nfs_sillyrename(struct inode *dir, struct dentry *dentry)
{
        static unsigned int sillycounter = 0;
        const int      i_inosize  = sizeof(dir->i_ino)*2;
        const int      countersize = sizeof(sillycounter)*2;
        const int      slen       = strlen(".nfs") + i_inosize + countersize;
        char           silly[slen+1];
        struct dentry *sdentry;
        int            error = -EIO;

        dfprintk(VFS, "NFS: silly-rename(%s/%s, ct=%d)\n",
                dentry->d_parent->d_name.name, dentry->d_name.name, 
                dentry->d_count);

        /*
         * Note that a silly-renamed file can be deleted once it's
         * no longer in use -- it's just an ordinary file now.
         */
        if (dentry->d_count == 1) {
                dentry->d_flags &= ~DCACHE_NFSFS_RENAMED;
                goto out;  /* No need to silly rename. */
        }

#ifdef NFS_PARANOIA
if (!dentry->d_inode)
printk("NFS: silly-renaming %s/%s, negative dentry??\n",
dentry->d_parent->d_name.name, dentry->d_name.name);
#endif
        /*
         * We don't allow a dentry to be silly-renamed twice.
         */
        error = -EBUSY;
        if (dentry->d_flags & DCACHE_NFSFS_RENAMED)
                goto out;

        sprintf(silly, ".nfs%*.*lx",
                i_inosize, i_inosize, dentry->d_inode->i_ino);

        sdentry = NULL;
        do {
                char *suffix = silly + slen - countersize;

                dput(sdentry);
                sillycounter++;
                sprintf(suffix, "%*.*x", countersize, countersize, sillycounter);

                dfprintk(VFS, "trying to rename %s to %s\n",
                         dentry->d_name.name, silly);
                
                sdentry = nfs_silly_lookup(dentry->d_parent, silly, slen);
                /*
                 * N.B. Better to return EBUSY here ... it could be
                 * dangerous to delete the file while it's in use.
                 */
                if (IS_ERR(sdentry))
                        goto out;
        } while(sdentry->d_inode != NULL); /* need negative lookup */

        invalidate_inode_pages(dir);
        nfs_flush_dircache(dir);
        error = nfs_proc_rename(NFS_SERVER(dir),
                                NFS_FH(dentry->d_parent), dentry->d_name.name,
                                NFS_FH(dentry->d_parent), silly);
        if (!error) {
                nfs_renew_times(dentry);
                d_move(dentry, sdentry);
                dentry->d_flags |= DCACHE_NFSFS_RENAMED;
                /* If we return 0 we don't unlink */
        }
        dput(sdentry);
out:
        return error;
}

/*
 * Remove a file after making sure there are no pending writes,
 * and after checking that the file has only one user. 
 *
 * We update inode->i_nlink and free the inode prior to the operation
 * to avoid possible races if the server reuses the inode.
 */
static int nfs_safe_remove(struct dentry *dentry)
{
        struct inode *dir = dentry->d_parent->d_inode;
        struct inode *inode = dentry->d_inode;
        int error, rehash = 0;
                
        dfprintk(VFS, "NFS: safe_remove(%s/%s, %ld)\n",
                dentry->d_parent->d_name.name, dentry->d_name.name,
                inode->i_ino);

        /* N.B. not needed now that d_delete is done in advance? */
        error = -EBUSY;
        if (!inode) {
#ifdef NFS_PARANOIA
printk("nfs_safe_remove: %s/%s already negative??\n",
dentry->d_parent->d_name.name, dentry->d_name.name);
#endif
        }

        if (dentry->d_count > 1) {
#ifdef NFS_PARANOIA
printk("nfs_safe_remove: %s/%s busy, d_count=%d\n",
dentry->d_parent->d_name.name, dentry->d_name.name, dentry->d_count);
#endif
                goto out;
        }
#ifdef NFS_PARANOIA
if (inode && inode->i_count > inode->i_nlink)
printk("nfs_safe_remove: %s/%s inode busy?? i_count=%d, i_nlink=%d\n",
dentry->d_parent->d_name.name, dentry->d_name.name,
inode->i_count, inode->i_nlink);
#endif
        /*
         * Unhash the dentry while we remove the file ...
         */
        if (!list_empty(&dentry->d_hash)) {
                d_drop(dentry);
                rehash = 1;
        }
        /*
         * Update i_nlink and free the inode before unlinking.
         */
        if (inode) {
                if (inode->i_nlink)
                        inode->i_nlink --;
                d_delete(dentry);
        }
        invalidate_inode_pages(dir);
        nfs_flush_dircache(dir);
        error = nfs_proc_remove(NFS_SERVER(dir), NFS_FH(dentry->d_parent),
                                dentry->d_name.name);
        /*
         * Rehash the negative dentry if the operation succeeded.
         */
        if (!error && rehash)
                d_add(dentry, NULL);
out:
        return error;
}

/*  We do silly rename. In case sillyrename() returns -EBUSY, the inode
 *  belongs to an active ".nfs..." file and we return -EBUSY.
 *
 *  If sillyrename() returns 0, we do nothing, otherwise we unlink.
 */
static int nfs_unlink(struct inode *dir, struct dentry *dentry)
{
        int error;

        dfprintk(VFS, "NFS: unlink(%x/%ld, %s)\n",
                dir->i_dev, dir->i_ino, dentry->d_name.name);

        error = nfs_sillyrename(dir, dentry);
        if (error && error != -EBUSY) {
                error = nfs_safe_remove(dentry);
                if (!error) {
                        nfs_renew_times(dentry);
                }
        }
        return error;
}

static int
nfs_symlink(struct inode *dir, struct dentry *dentry, const char *symname)
{
        struct nfs_sattr sattr;
        int error;

        dfprintk(VFS, "NFS: symlink(%x/%ld, %s, %s)\n",
                dir->i_dev, dir->i_ino, dentry->d_name.name, symname);

        error = -ENAMETOOLONG;
        if (strlen(symname) > NFS_MAXPATHLEN)
                goto out;

#ifdef NFS_PARANOIA
if (dentry->d_inode)
printk("nfs_proc_symlink: %s/%s not negative!\n",
dentry->d_parent->d_name.name, dentry->d_name.name);
#endif
        /*
         * Fill in the sattr for the call.
         * Note: SunOS 4.1.2 crashes if the mode isn't initialized!
         */
        sattr.mode = S_IFLNK | S_IRWXUGO;
        sattr.uid = sattr.gid = sattr.size = (unsigned) -1;
        sattr.atime.seconds = sattr.mtime.seconds = (unsigned) -1;

        /*
         * Drop the dentry in advance to force a new lookup.
         * Since nfs_proc_symlink doesn't return a fattr, we
         * can't instantiate the new inode.
         */
        d_drop(dentry);
        invalidate_inode_pages(dir);
        nfs_flush_dircache(dir);
        error = nfs_proc_symlink(NFS_SERVER(dir), NFS_FH(dentry->d_parent),
                                dentry->d_name.name, symname, &sattr);
        if (!error) {
                nfs_renew_times(dentry->d_parent);
        } else if (error == -EEXIST) {
                printk("nfs_proc_symlink: %s/%s already exists??\n",
                        dentry->d_parent->d_name.name, dentry->d_name.name);
        }

out:
        return error;
}

static int 
nfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *dentry)
{
        struct inode *inode = old_dentry->d_inode;
        int error;

        dfprintk(VFS, "NFS: link(%s/%s -> %s/%s)\n",
                old_dentry->d_parent->d_name.name, old_dentry->d_name.name,
                dentry->d_parent->d_name.name, dentry->d_name.name);

        /*
         * Drop the dentry in advance to force a new lookup.
         * Since nfs_proc_link doesn't return a file handle,
         * we can't use the existing dentry.
         */
        d_drop(dentry);
        invalidate_inode_pages(dir);
        nfs_flush_dircache(dir);
        error = nfs_proc_link(NFS_DSERVER(old_dentry), NFS_FH(old_dentry),
                                NFS_FH(dentry->d_parent), dentry->d_name.name);
        if (!error) {
                /*
                 * Update the link count immediately, as some apps
                 * (e.g. pine) test this after making a link.
                 */
                inode->i_nlink++;
        }
        return error;
}

/*
 * RENAME
 * FIXME: Some nfsds, like the Linux user space nfsd, may generate a
 * different file handle for the same inode after a rename (e.g. when
 * moving to a different directory). A fail-safe method to do so would
 * be to look up old_dir/old_name, create a link to new_dir/new_name and
 * rename the old file using the sillyrename stuff. This way, the original
 * file in old_dir will go away when the last process iput()s the inode.
 *
 * FIXED.
 * 
 * It actually works quite well. One needs to have the possibility for
 * at least one ".nfs..." file in each directory the file ever gets
 * moved or linked to which happens automagically with the new
 * implementation that only depends on the dcache stuff instead of
 * using the inode layer
 *
 * Unfortunately, things are a little more complicated than indicated
 * above. For a cross-directory move, we want to make sure we can get
 * rid of the old inode after the operation.  This means there must be
 * no pending writes (if it's a file), and the use count must be 1.
 * If these conditions are met, we can drop the dentries before doing
 * the rename.
 */
static int nfs_rename(struct inode *old_dir, struct dentry *old_dentry,
                      struct inode *new_dir, struct dentry *new_dentry)
{
        struct inode *old_inode = old_dentry->d_inode;
        struct inode *new_inode = new_dentry->d_inode;
        struct dentry *dentry = NULL;
        int error, rehash = 0, update = 1;

        dfprintk(VFS, "NFS: rename(%s/%s -> %s/%s, ct=%d)\n",
                old_dentry->d_parent->d_name.name, old_dentry->d_name.name,
                new_dentry->d_parent->d_name.name, new_dentry->d_name.name,
                new_dentry->d_count);

        /*
         * First check whether the target is busy ... we can't
         * safely do _any_ rename if the target is in use.
         *
         * For files, make a copy of the dentry and then do a 
         * silly-rename. If the silly-rename succeeds, the
         * copied dentry is hashed and becomes the new target.
         *
         * With directories check is done in VFS.
         */
        error = -EBUSY;
        if (new_dentry->d_count > 1 && new_inode) {
                int err;
                /* copy the target dentry's name */
                dentry = d_alloc(new_dentry->d_parent,
                                 &new_dentry->d_name);
                if (!dentry)
                        goto out;

                /* silly-rename the existing target ... */
                err = nfs_sillyrename(new_dir, new_dentry);
                if (!err) {
                        new_dentry = dentry;
                        new_inode = NULL;
                        /* hash the replacement target */
                        d_add(new_dentry, NULL);
                }

                /* dentry still busy? */
                if (new_dentry->d_count > 1) {
#ifdef NFS_PARANOIA
printk("nfs_rename: target %s/%s busy, d_count=%d\n",
new_dentry->d_parent->d_name.name,new_dentry->d_name.name,new_dentry->d_count);
#endif
                        goto out;
                }
        }

        /*
         * Check for within-directory rename ... no complications.
         */
        if (new_dir == old_dir)
                goto do_rename;
        /*
         * Cross-directory move ...
         *
         * ... prune child dentries and writebacks if needed.
         */
        if (old_dentry->d_count > 1) {
                nfs_wb_all(old_inode);
                shrink_dcache_parent(old_dentry);
        }

        /*
         * Now check the use counts ... we can't safely do the
         * rename unless we can drop the dentries first.
         */
        if (old_dentry->d_count > 1) {
#ifdef NFS_PARANOIA
printk("nfs_rename: old dentry %s/%s busy, d_count=%d\n",
old_dentry->d_parent->d_name.name,old_dentry->d_name.name,old_dentry->d_count);
#endif
                goto out;
        }
        if (new_dentry->d_count > 1 && new_inode) {
#ifdef NFS_PARANOIA
printk("nfs_rename: new dentry %s/%s busy, d_count=%d\n",
new_dentry->d_parent->d_name.name,new_dentry->d_name.name,new_dentry->d_count);
#endif
                goto out;
        }

        d_drop(old_dentry);
        update = 0;

do_rename:
        /*
         * To prevent any new references to the target during the rename,
         * we unhash the dentry and free the inode in advance.
         */
#ifdef NFS_PARANOIA
if (new_inode && 
    new_inode->i_count > (S_ISDIR(new_inode->i_mode) ? 1 : new_inode->i_nlink))
printk("nfs_rename: %s/%s inode busy?? i_count=%d, i_nlink=%d\n",
new_dentry->d_parent->d_name.name, new_dentry->d_name.name,
new_inode->i_count, new_inode->i_nlink);
#endif
        if (!list_empty(&new_dentry->d_hash)) {
                d_drop(new_dentry);
                rehash = update;
        }
        if (new_inode) {
                d_delete(new_dentry);
        }

        invalidate_inode_pages(new_dir);
        nfs_flush_dircache(new_dir);
        invalidate_inode_pages(old_dir);
        nfs_flush_dircache(old_dir);
        error = nfs_proc_rename(NFS_DSERVER(old_dentry),
                        NFS_FH(old_dentry->d_parent), old_dentry->d_name.name,
                        NFS_FH(new_dentry->d_parent), new_dentry->d_name.name);
        if (!error && !S_ISDIR(old_inode->i_mode)) {
                /* Update the dcache if needed */
                if (rehash)
                        d_add(new_dentry, NULL);
                if (update)
                        d_move(old_dentry, new_dentry);
        }

out:
        /* new dentry created? */
        if (dentry)
                dput(dentry);
        return error;
}

int nfs_init_fhcache(void)
{
        nfs_fh_cachep = kmem_cache_create("nfs_fh",
                                          sizeof(struct nfs_fh),
                                          0, SLAB_HWCACHE_ALIGN,
                                          NULL, NULL);
        if (nfs_fh_cachep == NULL)
                return -ENOMEM;

        nfs_cookie_cachep = kmem_cache_create("nfs_dcookie",
                                              sizeof(struct nfs_cookie_table),
                                              0, SLAB_HWCACHE_ALIGN,
                                              NULL, NULL);
        if (nfs_cookie_cachep == NULL)
                return -ENOMEM;

        return 0;
}

/*
 * Local variables:
 *  version-control: t
 *  kept-new-versions: 5
 * End:
 */