#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/vserver/xid.h>
+#include <linux/vs_tag.h>
#include "nfs4_fs.h"
#include "delegation.h"
return res;
}
-typedef u32 * (*decode_dirent_t)(u32 *, struct nfs_entry *, int);
+typedef __be32 * (*decode_dirent_t)(__be32 *, struct nfs_entry *, int);
typedef struct {
struct file *file;
struct page *page;
unsigned long page_index;
- u32 *ptr;
+ __be32 *ptr;
u64 *dir_cookie;
loff_t current_index;
struct nfs_entry *entry;
int nfs_readdir_filler(nfs_readdir_descriptor_t *desc, struct page *page)
{
struct file *file = desc->file;
- struct inode *inode = file->f_dentry->d_inode;
+ struct inode *inode = file->f_path.dentry->d_inode;
struct rpc_cred *cred = nfs_file_cred(file);
unsigned long timestamp;
int error;
again:
timestamp = jiffies;
- error = NFS_PROTO(inode)->readdir(file->f_dentry, cred, desc->entry->cookie, page,
+ 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 */
* 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);
+ 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:
static inline
int dir_decode(nfs_readdir_descriptor_t *desc)
{
- u32 *p = desc->ptr;
+ __be32 *p = desc->ptr;
p = desc->decode(p, desc->entry, desc->plus);
if (IS_ERR(p))
return PTR_ERR(p);
static inline
int find_dirent_page(nfs_readdir_descriptor_t *desc)
{
- struct inode *inode = desc->file->f_dentry->d_inode;
+ struct inode *inode = desc->file->f_path.dentry->d_inode;
struct page *page;
int status;
struct file *file = desc->file;
struct nfs_entry *entry = desc->entry;
struct dentry *dentry = NULL;
- u64 fileid;
+ unsigned long fileid;
int loop_count = 0,
res;
unsigned d_type = DT_UNKNOWN;
/* Note: entry->prev_cookie contains the cookie for
* retrieving the current dirent on the server */
- fileid = entry->ino;
+ fileid = nfs_fileid_to_ino_t(entry->ino);
/* Get a dentry if we have one */
if (dentry != NULL)
/* Use readdirplus info */
if (dentry != NULL && dentry->d_inode != NULL) {
d_type = dt_type(dentry->d_inode);
- fileid = NFS_FILEID(dentry->d_inode);
+ fileid = dentry->d_inode->i_ino;
}
res = filldir(dirent, entry->name, entry->len,
filldir_t filldir)
{
struct file *file = desc->file;
- struct inode *inode = file->f_dentry->d_inode;
+ struct inode *inode = file->f_path.dentry->d_inode;
struct rpc_cred *cred = nfs_file_cred(file);
struct page *page = NULL;
int status;
status = -ENOMEM;
goto out;
}
- desc->error = NFS_PROTO(inode)->readdir(file->f_dentry, cred, *desc->dir_cookie,
+ desc->error = NFS_PROTO(inode)->readdir(file->f_path.dentry, cred, *desc->dir_cookie,
page,
NFS_SERVER(inode)->dtsize,
desc->plus);
*/
static int nfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
{
- struct dentry *dentry = filp->f_dentry;
+ struct dentry *dentry = filp->f_path.dentry;
struct inode *inode = dentry->d_inode;
nfs_readdir_descriptor_t my_desc,
*desc = &my_desc;
lock_kernel();
- res = nfs_revalidate_mapping(inode, filp->f_mapping);
+ res = nfs_revalidate_mapping_nolock(inode, filp->f_mapping);
if (res < 0) {
unlock_kernel();
return res;
loff_t nfs_llseek_dir(struct file *filp, loff_t offset, int origin)
{
- mutex_lock(&filp->f_dentry->d_inode->i_mutex);
+ mutex_lock(&filp->f_path.dentry->d_inode->i_mutex);
switch (origin) {
case 1:
offset += filp->f_pos;
((struct nfs_open_context *)filp->private_data)->dir_cookie = 0;
}
out:
- mutex_unlock(&filp->f_dentry->d_inode->i_mutex);
+ mutex_unlock(&filp->f_path.dentry->d_inode->i_mutex);
return offset;
}
nfs_inode_return_delegation(inode);
if (dentry->d_flags & DCACHE_NFSFS_RENAMED) {
lock_kernel();
- inode->i_nlink--;
+ drop_nlink(inode);
nfs_complete_unlink(dentry);
unlock_kernel();
}
res = (struct dentry *)inode;
if (IS_ERR(res))
goto out_unlock;
- vx_propagate_xid(nd, inode);
+ dx_propagate_tag(nd, inode);
no_entry:
res = d_materialise_unique(dentry, inode);
- if (res != NULL)
+ if (res != NULL) {
+ struct dentry *parent;
+ if (IS_ERR(res))
+ goto out_unlock;
+ /* Was a directory renamed! */
+ parent = dget_parent(res);
+ if (!IS_ROOT(parent))
+ nfs_mark_for_revalidate(parent->d_inode);
+ dput(parent);
dentry = res;
+ }
nfs_renew_times(dentry);
nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
out_unlock:
static struct dentry *nfs_readdir_lookup(nfs_readdir_descriptor_t *desc)
{
- struct dentry *parent = desc->file->f_dentry;
+ struct dentry *parent = desc->file->f_path.dentry;
struct inode *dir = parent->d_inode;
struct nfs_entry *entry = desc->entry;
struct dentry *dentry, *alias;
alias = d_materialise_unique(dentry, inode);
if (alias != NULL) {
dput(dentry);
+ if (IS_ERR(alias))
+ return NULL;
dentry = alias;
}
struct inode *dir = dentry->d_parent->d_inode;
error = NFS_PROTO(dir)->lookup(dir, &dentry->d_name, fhandle, fattr);
if (error)
- goto out_err;
+ return error;
}
if (!(fattr->valid & NFS_ATTR_FATTR)) {
struct nfs_server *server = NFS_SB(dentry->d_sb);
error = server->nfs_client->rpc_ops->getattr(server, fhandle, fattr);
if (error < 0)
- goto out_err;
+ return error;
}
inode = nfs_fhget(dentry->d_sb, fhandle, fattr);
error = PTR_ERR(inode);
if (IS_ERR(inode))
- goto out_err;
+ return error;
d_instantiate(dentry, inode);
if (d_unhashed(dentry))
d_rehash(dentry);
return 0;
-out_err:
- d_drop(dentry);
- return error;
}
/*
error = NFS_PROTO(dir)->rmdir(dir, &dentry->d_name);
/* Ensure the VFS deletes this inode */
if (error == 0 && dentry->d_inode != NULL)
- dentry->d_inode->i_nlink = 0;
+ clear_nlink(dentry->d_inode);
nfs_end_data_update(dir);
unlock_kernel();
error = NFS_PROTO(dir)->remove(dir, &dentry->d_name);
/* The VFS may want to delete this inode */
if (error == 0)
- inode->i_nlink--;
+ drop_nlink(inode);
nfs_mark_for_revalidate(inode);
nfs_end_data_update(inode);
} else
return error;
}
-static int
-nfs_symlink(struct inode *dir, struct dentry *dentry, const char *symname)
+/*
+ * To create a symbolic link, most file systems instantiate a new inode,
+ * add a page to it containing the path, then write it out to the disk
+ * using prepare_write/commit_write.
+ *
+ * Unfortunately the NFS client can't create the in-core inode first
+ * because it needs a file handle to create an in-core inode (see
+ * fs/nfs/inode.c:nfs_fhget). We only have a file handle *after* the
+ * symlink request has completed on the server.
+ *
+ * So instead we allocate a raw page, copy the symname into it, then do
+ * the SYMLINK request with the page as the buffer. If it succeeds, we
+ * now have a new file handle and can instantiate an in-core NFS inode
+ * and move the raw page into its mapping.
+ */
+static int nfs_symlink(struct inode *dir, struct dentry *dentry, const char *symname)
{
+ struct pagevec lru_pvec;
+ struct page *page;
+ char *kaddr;
struct iattr attr;
- struct nfs_fattr sym_attr;
- struct nfs_fh sym_fh;
- struct qstr qsymname;
+ unsigned int pathlen = strlen(symname);
int error;
dfprintk(VFS, "NFS: symlink(%s/%ld, %s, %s)\n", dir->i_sb->s_id,
dir->i_ino, dentry->d_name.name, symname);
-#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!
- */
- attr.ia_valid = ATTR_MODE;
- attr.ia_mode = S_IFLNK | S_IRWXUGO;
+ if (pathlen > PAGE_SIZE)
+ return -ENAMETOOLONG;
- qsymname.name = symname;
- qsymname.len = strlen(symname);
+ attr.ia_mode = S_IFLNK | S_IRWXUGO;
+ attr.ia_valid = ATTR_MODE;
lock_kernel();
+
+ page = alloc_page(GFP_KERNEL);
+ if (!page) {
+ unlock_kernel();
+ return -ENOMEM;
+ }
+
+ kaddr = kmap_atomic(page, KM_USER0);
+ memcpy(kaddr, symname, pathlen);
+ if (pathlen < PAGE_SIZE)
+ memset(kaddr + pathlen, 0, PAGE_SIZE - pathlen);
+ kunmap_atomic(kaddr, KM_USER0);
+
nfs_begin_data_update(dir);
- error = NFS_PROTO(dir)->symlink(dir, &dentry->d_name, &qsymname,
- &attr, &sym_fh, &sym_attr);
+ error = NFS_PROTO(dir)->symlink(dir, dentry, page, pathlen, &attr);
nfs_end_data_update(dir);
- if (!error) {
- error = nfs_instantiate(dentry, &sym_fh, &sym_attr);
- } else {
- if (error == -EEXIST)
- printk("nfs_proc_symlink: %s/%s already exists??\n",
- dentry->d_parent->d_name.name, dentry->d_name.name);
+ if (error != 0) {
+ dfprintk(VFS, "NFS: symlink(%s/%ld, %s, %s) error %d\n",
+ dir->i_sb->s_id, dir->i_ino,
+ dentry->d_name.name, symname, error);
d_drop(dentry);
+ __free_page(page);
+ unlock_kernel();
+ return error;
}
+
+ /*
+ * No big deal if we can't add this page to the page cache here.
+ * READLINK will get the missing page from the server if needed.
+ */
+ pagevec_init(&lru_pvec, 0);
+ if (!add_to_page_cache(page, dentry->d_inode->i_mapping, 0,
+ GFP_KERNEL)) {
+ pagevec_add(&lru_pvec, page);
+ pagevec_lru_add(&lru_pvec);
+ SetPageUptodate(page);
+ unlock_page(page);
+ } else
+ __free_page(page);
+
unlock_kernel();
- return error;
+ return 0;
}
static int
goto out;
}
} else
- new_inode->i_nlink--;
+ drop_nlink(new_inode);
go_ahead:
/*
* ... prune child dentries and writebacks if needed.
*/
if (atomic_read(&old_dentry->d_count) > 1) {
- nfs_wb_all(old_inode);
+ if (S_ISREG(old_inode->i_mode))
+ nfs_wb_all(old_inode);
shrink_dcache_parent(old_dentry);
}
nfs_inode_return_delegation(old_inode);
if (rehash)
d_rehash(rehash);
if (!error) {
- if (!S_ISDIR(old_inode->i_mode))
- d_move(old_dentry, new_dentry);
+ d_move(old_dentry, new_dentry);
nfs_renew_times(new_dentry);
nfs_set_verifier(new_dentry, nfs_save_change_attribute(new_dir));
}
return error;
}
+static DEFINE_SPINLOCK(nfs_access_lru_lock);
+static LIST_HEAD(nfs_access_lru_list);
+static atomic_long_t nfs_access_nr_entries;
+
+static void nfs_access_free_entry(struct nfs_access_entry *entry)
+{
+ put_rpccred(entry->cred);
+ kfree(entry);
+ smp_mb__before_atomic_dec();
+ atomic_long_dec(&nfs_access_nr_entries);
+ smp_mb__after_atomic_dec();
+}
+
+int nfs_access_cache_shrinker(int nr_to_scan, gfp_t gfp_mask)
+{
+ LIST_HEAD(head);
+ struct nfs_inode *nfsi;
+ struct nfs_access_entry *cache;
+
+ spin_lock(&nfs_access_lru_lock);
+restart:
+ list_for_each_entry(nfsi, &nfs_access_lru_list, access_cache_inode_lru) {
+ struct inode *inode;
+
+ if (nr_to_scan-- == 0)
+ break;
+ inode = igrab(&nfsi->vfs_inode);
+ if (inode == NULL)
+ continue;
+ spin_lock(&inode->i_lock);
+ if (list_empty(&nfsi->access_cache_entry_lru))
+ goto remove_lru_entry;
+ cache = list_entry(nfsi->access_cache_entry_lru.next,
+ struct nfs_access_entry, lru);
+ list_move(&cache->lru, &head);
+ rb_erase(&cache->rb_node, &nfsi->access_cache);
+ if (!list_empty(&nfsi->access_cache_entry_lru))
+ list_move_tail(&nfsi->access_cache_inode_lru,
+ &nfs_access_lru_list);
+ else {
+remove_lru_entry:
+ list_del_init(&nfsi->access_cache_inode_lru);
+ clear_bit(NFS_INO_ACL_LRU_SET, &nfsi->flags);
+ }
+ spin_unlock(&inode->i_lock);
+ iput(inode);
+ goto restart;
+ }
+ spin_unlock(&nfs_access_lru_lock);
+ while (!list_empty(&head)) {
+ cache = list_entry(head.next, struct nfs_access_entry, lru);
+ list_del(&cache->lru);
+ nfs_access_free_entry(cache);
+ }
+ return (atomic_long_read(&nfs_access_nr_entries) / 100) * sysctl_vfs_cache_pressure;
+}
+
+static void __nfs_access_zap_cache(struct inode *inode)
+{
+ struct nfs_inode *nfsi = NFS_I(inode);
+ struct rb_root *root_node = &nfsi->access_cache;
+ struct rb_node *n, *dispose = NULL;
+ struct nfs_access_entry *entry;
+
+ /* Unhook entries from the cache */
+ while ((n = rb_first(root_node)) != NULL) {
+ entry = rb_entry(n, struct nfs_access_entry, rb_node);
+ rb_erase(n, root_node);
+ list_del(&entry->lru);
+ n->rb_left = dispose;
+ dispose = n;
+ }
+ nfsi->cache_validity &= ~NFS_INO_INVALID_ACCESS;
+ spin_unlock(&inode->i_lock);
+
+ /* Now kill them all! */
+ while (dispose != NULL) {
+ n = dispose;
+ dispose = n->rb_left;
+ nfs_access_free_entry(rb_entry(n, struct nfs_access_entry, rb_node));
+ }
+}
+
+void nfs_access_zap_cache(struct inode *inode)
+{
+ /* Remove from global LRU init */
+ if (test_and_clear_bit(NFS_INO_ACL_LRU_SET, &NFS_FLAGS(inode))) {
+ spin_lock(&nfs_access_lru_lock);
+ list_del_init(&NFS_I(inode)->access_cache_inode_lru);
+ spin_unlock(&nfs_access_lru_lock);
+ }
+
+ spin_lock(&inode->i_lock);
+ /* This will release the spinlock */
+ __nfs_access_zap_cache(inode);
+}
+
+static struct nfs_access_entry *nfs_access_search_rbtree(struct inode *inode, struct rpc_cred *cred)
+{
+ struct rb_node *n = NFS_I(inode)->access_cache.rb_node;
+ struct nfs_access_entry *entry;
+
+ while (n != NULL) {
+ entry = rb_entry(n, struct nfs_access_entry, rb_node);
+
+ if (cred < entry->cred)
+ n = n->rb_left;
+ else if (cred > entry->cred)
+ n = n->rb_right;
+ else
+ return entry;
+ }
+ return NULL;
+}
+
int nfs_access_get_cached(struct inode *inode, struct rpc_cred *cred, struct nfs_access_entry *res)
{
struct nfs_inode *nfsi = NFS_I(inode);
- struct nfs_access_entry *cache = &nfsi->cache_access;
+ struct nfs_access_entry *cache;
+ int err = -ENOENT;
- if (cache->cred != cred
- || time_after(jiffies, cache->jiffies + NFS_ATTRTIMEO(inode))
- || (nfsi->cache_validity & NFS_INO_INVALID_ACCESS))
- return -ENOENT;
- memcpy(res, cache, sizeof(*res));
- return 0;
+ spin_lock(&inode->i_lock);
+ if (nfsi->cache_validity & NFS_INO_INVALID_ACCESS)
+ goto out_zap;
+ cache = nfs_access_search_rbtree(inode, cred);
+ if (cache == NULL)
+ goto out;
+ if (time_after(jiffies, cache->jiffies + NFS_ATTRTIMEO(inode)))
+ goto out_stale;
+ res->jiffies = cache->jiffies;
+ res->cred = cache->cred;
+ res->mask = cache->mask;
+ list_move_tail(&cache->lru, &nfsi->access_cache_entry_lru);
+ err = 0;
+out:
+ spin_unlock(&inode->i_lock);
+ return err;
+out_stale:
+ rb_erase(&cache->rb_node, &nfsi->access_cache);
+ list_del(&cache->lru);
+ spin_unlock(&inode->i_lock);
+ nfs_access_free_entry(cache);
+ return -ENOENT;
+out_zap:
+ /* This will release the spinlock */
+ __nfs_access_zap_cache(inode);
+ return -ENOENT;
}
-void nfs_access_add_cache(struct inode *inode, struct nfs_access_entry *set)
+static void nfs_access_add_rbtree(struct inode *inode, struct nfs_access_entry *set)
{
struct nfs_inode *nfsi = NFS_I(inode);
- struct nfs_access_entry *cache = &nfsi->cache_access;
+ struct rb_root *root_node = &nfsi->access_cache;
+ struct rb_node **p = &root_node->rb_node;
+ struct rb_node *parent = NULL;
+ struct nfs_access_entry *entry;
- if (cache->cred != set->cred) {
- if (cache->cred)
- put_rpccred(cache->cred);
- cache->cred = get_rpccred(set->cred);
- }
- /* FIXME: replace current access_cache BKL reliance with inode->i_lock */
spin_lock(&inode->i_lock);
- nfsi->cache_validity &= ~NFS_INO_INVALID_ACCESS;
+ while (*p != NULL) {
+ parent = *p;
+ entry = rb_entry(parent, struct nfs_access_entry, rb_node);
+
+ if (set->cred < entry->cred)
+ p = &parent->rb_left;
+ else if (set->cred > entry->cred)
+ p = &parent->rb_right;
+ else
+ goto found;
+ }
+ rb_link_node(&set->rb_node, parent, p);
+ rb_insert_color(&set->rb_node, root_node);
+ list_add_tail(&set->lru, &nfsi->access_cache_entry_lru);
+ spin_unlock(&inode->i_lock);
+ return;
+found:
+ rb_replace_node(parent, &set->rb_node, root_node);
+ list_add_tail(&set->lru, &nfsi->access_cache_entry_lru);
+ list_del(&entry->lru);
spin_unlock(&inode->i_lock);
+ nfs_access_free_entry(entry);
+}
+
+void nfs_access_add_cache(struct inode *inode, struct nfs_access_entry *set)
+{
+ struct nfs_access_entry *cache = kmalloc(sizeof(*cache), GFP_KERNEL);
+ if (cache == NULL)
+ return;
+ RB_CLEAR_NODE(&cache->rb_node);
cache->jiffies = set->jiffies;
+ cache->cred = get_rpccred(set->cred);
cache->mask = set->mask;
+
+ nfs_access_add_rbtree(inode, cache);
+
+ /* Update accounting */
+ smp_mb__before_atomic_inc();
+ atomic_long_inc(&nfs_access_nr_entries);
+ smp_mb__after_atomic_inc();
+
+ /* Add inode to global LRU list */
+ if (!test_and_set_bit(NFS_INO_ACL_LRU_SET, &NFS_FLAGS(inode))) {
+ spin_lock(&nfs_access_lru_lock);
+ list_add_tail(&NFS_I(inode)->access_cache_inode_lru, &nfs_access_lru_list);
+ spin_unlock(&nfs_access_lru_lock);
+ }
}
static int nfs_do_access(struct inode *inode, struct rpc_cred *cred, int mask)
git://git.onelab.eu / linux-2.6.git / blobdiff