Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi...
[safe/jmp/linux-2.6] / fs / fuse / inode.c
index 437d61c..91f7c85 100644 (file)
@@ -1,6 +1,6 @@
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <miklos@szeredi.hu>
+  Copyright (C) 2001-2008  Miklos Szeredi <miklos@szeredi.hu>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
@@ -17,6 +17,9 @@
 #include <linux/parser.h>
 #include <linux/statfs.h>
 #include <linux/random.h>
+#include <linux/sched.h>
+#include <linux/exportfs.h>
+#include <linux/smp_lock.h>
 
 MODULE_AUTHOR("Miklos Szeredi <miklos@szeredi.hu>");
 MODULE_DESCRIPTION("Filesystem in Userspace");
@@ -28,15 +31,17 @@ DEFINE_MUTEX(fuse_mutex);
 
 #define FUSE_SUPER_MAGIC 0x65735546
 
+#define FUSE_DEFAULT_BLKSIZE 512
+
 struct fuse_mount_data {
        int fd;
        unsigned rootmode;
        unsigned user_id;
        unsigned group_id;
-       unsigned fd_present : 1;
-       unsigned rootmode_present : 1;
-       unsigned user_id_present : 1;
-       unsigned group_id_present : 1;
+       unsigned fd_present:1;
+       unsigned rootmode_present:1;
+       unsigned user_id_present:1;
+       unsigned group_id_present:1;
        unsigned flags;
        unsigned max_read;
        unsigned blksize;
@@ -55,6 +60,12 @@ static struct inode *fuse_alloc_inode(struct super_block *sb)
        fi->i_time = 0;
        fi->nodeid = 0;
        fi->nlookup = 0;
+       fi->attr_version = 0;
+       fi->writectr = 0;
+       INIT_LIST_HEAD(&fi->write_files);
+       INIT_LIST_HEAD(&fi->queued_writes);
+       INIT_LIST_HEAD(&fi->writepages);
+       init_waitqueue_head(&fi->page_waitq);
        fi->forget_req = fuse_request_alloc();
        if (!fi->forget_req) {
                kmem_cache_free(fuse_inode_cachep, inode);
@@ -67,18 +78,15 @@ static struct inode *fuse_alloc_inode(struct super_block *sb)
 static void fuse_destroy_inode(struct inode *inode)
 {
        struct fuse_inode *fi = get_fuse_inode(inode);
+       BUG_ON(!list_empty(&fi->write_files));
+       BUG_ON(!list_empty(&fi->queued_writes));
        if (fi->forget_req)
                fuse_request_free(fi->forget_req);
        kmem_cache_free(fuse_inode_cachep, inode);
 }
 
-static void fuse_read_inode(struct inode *inode)
-{
-       /* No op */
-}
-
 void fuse_send_forget(struct fuse_conn *fc, struct fuse_req *req,
-                     unsigned long nodeid, u64 nlookup)
+                     u64 nodeid, u64 nlookup)
 {
        struct fuse_forget_in *inarg = &req->misc.forget_in;
        inarg->nlookup = nlookup;
@@ -87,7 +95,7 @@ void fuse_send_forget(struct fuse_conn *fc, struct fuse_req *req,
        req->in.numargs = 1;
        req->in.args[0].size = sizeof(struct fuse_forget_in);
        req->in.args[0].value = inarg;
-       request_send_noreply(fc, req);
+       fuse_request_send_noreply(fc, req);
 }
 
 static void fuse_clear_inode(struct inode *inode)
@@ -108,20 +116,28 @@ static int fuse_remount_fs(struct super_block *sb, int *flags, char *data)
        return 0;
 }
 
-void fuse_change_attributes(struct inode *inode, struct fuse_attr *attr)
+void fuse_truncate(struct address_space *mapping, loff_t offset)
+{
+       /* See vmtruncate() */
+       unmap_mapping_range(mapping, offset + PAGE_SIZE - 1, 0, 1);
+       truncate_inode_pages(mapping, offset);
+       unmap_mapping_range(mapping, offset + PAGE_SIZE - 1, 0, 1);
+}
+
+void fuse_change_attributes_common(struct inode *inode, struct fuse_attr *attr,
+                                  u64 attr_valid)
 {
        struct fuse_conn *fc = get_fuse_conn(inode);
-       if (S_ISREG(inode->i_mode) && i_size_read(inode) != attr->size)
-               invalidate_inode_pages(inode->i_mapping);
+       struct fuse_inode *fi = get_fuse_inode(inode);
+
+       fi->attr_version = ++fc->attr_version;
+       fi->i_time = attr_valid;
 
        inode->i_ino     = attr->ino;
-       inode->i_mode    = (inode->i_mode & S_IFMT) + (attr->mode & 07777);
+       inode->i_mode    = (inode->i_mode & S_IFMT) | (attr->mode & 07777);
        inode->i_nlink   = attr->nlink;
        inode->i_uid     = attr->uid;
        inode->i_gid     = attr->gid;
-       spin_lock(&fc->lock);
-       i_size_write(inode, attr->size);
-       spin_unlock(&fc->lock);
        inode->i_blocks  = attr->blocks;
        inode->i_atime.tv_sec   = attr->atime;
        inode->i_atime.tv_nsec  = attr->atimensec;
@@ -129,6 +145,46 @@ void fuse_change_attributes(struct inode *inode, struct fuse_attr *attr)
        inode->i_mtime.tv_nsec  = attr->mtimensec;
        inode->i_ctime.tv_sec   = attr->ctime;
        inode->i_ctime.tv_nsec  = attr->ctimensec;
+
+       if (attr->blksize != 0)
+               inode->i_blkbits = ilog2(attr->blksize);
+       else
+               inode->i_blkbits = inode->i_sb->s_blocksize_bits;
+
+       /*
+        * Don't set the sticky bit in i_mode, unless we want the VFS
+        * to check permissions.  This prevents failures due to the
+        * check in may_delete().
+        */
+       fi->orig_i_mode = inode->i_mode;
+       if (!(fc->flags & FUSE_DEFAULT_PERMISSIONS))
+               inode->i_mode &= ~S_ISVTX;
+}
+
+void fuse_change_attributes(struct inode *inode, struct fuse_attr *attr,
+                           u64 attr_valid, u64 attr_version)
+{
+       struct fuse_conn *fc = get_fuse_conn(inode);
+       struct fuse_inode *fi = get_fuse_inode(inode);
+       loff_t oldsize;
+
+       spin_lock(&fc->lock);
+       if (attr_version != 0 && fi->attr_version > attr_version) {
+               spin_unlock(&fc->lock);
+               return;
+       }
+
+       fuse_change_attributes_common(inode, attr, attr_valid);
+
+       oldsize = inode->i_size;
+       i_size_write(inode, attr->size);
+       spin_unlock(&fc->lock);
+
+       if (S_ISREG(inode->i_mode) && oldsize != attr->size) {
+               if (attr->size < oldsize)
+                       fuse_truncate(inode->i_mapping, attr->size);
+               invalidate_inode_pages2(inode->i_mapping);
+       }
 }
 
 static void fuse_init_inode(struct inode *inode, struct fuse_attr *attr)
@@ -153,7 +209,7 @@ static void fuse_init_inode(struct inode *inode, struct fuse_attr *attr)
 
 static int fuse_inode_eq(struct inode *inode, void *_nodeidp)
 {
-       unsigned long nodeid = *(unsigned long *) _nodeidp;
+       u64 nodeid = *(u64 *) _nodeidp;
        if (get_node_id(inode) == nodeid)
                return 1;
        else
@@ -162,13 +218,14 @@ static int fuse_inode_eq(struct inode *inode, void *_nodeidp)
 
 static int fuse_inode_set(struct inode *inode, void *_nodeidp)
 {
-       unsigned long nodeid = *(unsigned long *) _nodeidp;
+       u64 nodeid = *(u64 *) _nodeidp;
        get_fuse_inode(inode)->nodeid = nodeid;
        return 0;
 }
 
-struct inode *fuse_iget(struct super_block *sb, unsigned long nodeid,
-                       int generation, struct fuse_attr *attr)
+struct inode *fuse_iget(struct super_block *sb, u64 nodeid,
+                       int generation, struct fuse_attr *attr,
+                       u64 attr_valid, u64 attr_version)
 {
        struct inode *inode;
        struct fuse_inode *fi;
@@ -194,16 +251,18 @@ struct inode *fuse_iget(struct super_block *sb, unsigned long nodeid,
 
        fi = get_fuse_inode(inode);
        spin_lock(&fc->lock);
-       fi->nlookup ++;
+       fi->nlookup++;
        spin_unlock(&fc->lock);
-       fuse_change_attributes(inode, attr);
+       fuse_change_attributes(inode, attr, attr_valid, attr_version);
+
        return inode;
 }
 
-static void fuse_umount_begin(struct vfsmount *vfsmnt, int flags)
+static void fuse_umount_begin(struct super_block *sb)
 {
-       if (flags & MNT_FORCE)
-               fuse_abort_conn(get_fuse_conn_super(vfsmnt->mnt_sb));
+       lock_kernel();
+       fuse_abort_conn(get_fuse_conn_super(sb));
+       unlock_kernel();
 }
 
 static void fuse_send_destroy(struct fuse_conn *fc)
@@ -213,7 +272,7 @@ static void fuse_send_destroy(struct fuse_conn *fc)
                fc->destroy_req = NULL;
                req->in.h.opcode = FUSE_DESTROY;
                req->force = 1;
-               request_send(fc, req);
+               fuse_request_send(fc, req);
                fuse_put_request(fc, req);
        }
 }
@@ -231,10 +290,12 @@ static void fuse_put_super(struct super_block *sb)
        kill_fasync(&fc->fasync, SIGIO, POLL_IN);
        wake_up_all(&fc->waitq);
        wake_up_all(&fc->blocked_waitq);
+       wake_up_all(&fc->reserved_req_waitq);
        mutex_lock(&fuse_mutex);
        list_del(&fc->entry);
        fuse_ctl_remove_conn(fc);
        mutex_unlock(&fuse_mutex);
+       bdi_destroy(&fc->bdi);
        fuse_conn_put(fc);
 }
 
@@ -260,6 +321,11 @@ static int fuse_statfs(struct dentry *dentry, struct kstatfs *buf)
        struct fuse_statfs_out outarg;
        int err;
 
+       if (!fuse_allow_task(fc, current)) {
+               buf->f_type = FUSE_SUPER_MAGIC;
+               return 0;
+       }
+
        req = fuse_get_req(fc);
        if (IS_ERR(req))
                return PTR_ERR(req);
@@ -272,7 +338,7 @@ static int fuse_statfs(struct dentry *dentry, struct kstatfs *buf)
        req->out.args[0].size =
                fc->minor < 4 ? FUSE_COMPAT_STATFS_SIZE : sizeof(outarg);
        req->out.args[0].value = &outarg;
-       request_send(fc, req);
+       fuse_request_send(fc, req);
        err = req->out.h.error;
        if (!err)
                convert_fuse_statfs(buf, &outarg.st);
@@ -292,7 +358,7 @@ enum {
        OPT_ERR
 };
 
-static match_table_t tokens = {
+static const match_table_t tokens = {
        {OPT_FD,                        "fd=%u"},
        {OPT_ROOTMODE,                  "rootmode=%o"},
        {OPT_USER_ID,                   "user_id=%u"},
@@ -309,7 +375,7 @@ static int parse_fuse_opt(char *opt, struct fuse_mount_data *d, int is_bdev)
        char *p;
        memset(d, 0, sizeof(struct fuse_mount_data));
        d->max_read = ~0;
-       d->blksize = 512;
+       d->blksize = FUSE_DEFAULT_BLKSIZE;
 
        while ((p = strsep(&opt, ",")) != NULL) {
                int token;
@@ -330,6 +396,8 @@ static int parse_fuse_opt(char *opt, struct fuse_mount_data *d, int is_bdev)
                case OPT_ROOTMODE:
                        if (match_octal(&args[0], &value))
                                return 0;
+                       if (!fuse_valid_type(value))
+                               return 0;
                        d->rootmode = value;
                        d->rootmode_present = 1;
                        break;
@@ -392,33 +460,75 @@ static int fuse_show_options(struct seq_file *m, struct vfsmount *mnt)
                seq_puts(m, ",allow_other");
        if (fc->max_read != ~0)
                seq_printf(m, ",max_read=%u", fc->max_read);
+       if (mnt->mnt_sb->s_bdev &&
+           mnt->mnt_sb->s_blocksize != FUSE_DEFAULT_BLKSIZE)
+               seq_printf(m, ",blksize=%lu", mnt->mnt_sb->s_blocksize);
        return 0;
 }
 
-static struct fuse_conn *new_conn(void)
+int fuse_conn_init(struct fuse_conn *fc, struct super_block *sb)
 {
-       struct fuse_conn *fc;
+       int err;
 
-       fc = kzalloc(sizeof(*fc), GFP_KERNEL);
-       if (fc) {
-               spin_lock_init(&fc->lock);
-               mutex_init(&fc->inst_mutex);
-               atomic_set(&fc->count, 1);
-               init_waitqueue_head(&fc->waitq);
-               init_waitqueue_head(&fc->blocked_waitq);
-               INIT_LIST_HEAD(&fc->pending);
-               INIT_LIST_HEAD(&fc->processing);
-               INIT_LIST_HEAD(&fc->io);
-               INIT_LIST_HEAD(&fc->interrupts);
-               atomic_set(&fc->num_waiting, 0);
-               fc->bdi.ra_pages = (VM_MAX_READAHEAD * 1024) / PAGE_CACHE_SIZE;
-               fc->bdi.unplug_io_fn = default_unplug_io_fn;
-               fc->reqctr = 0;
-               fc->blocked = 1;
-               get_random_bytes(&fc->scramble_key, sizeof(fc->scramble_key));
+       memset(fc, 0, sizeof(*fc));
+       spin_lock_init(&fc->lock);
+       mutex_init(&fc->inst_mutex);
+       atomic_set(&fc->count, 1);
+       init_waitqueue_head(&fc->waitq);
+       init_waitqueue_head(&fc->blocked_waitq);
+       init_waitqueue_head(&fc->reserved_req_waitq);
+       INIT_LIST_HEAD(&fc->pending);
+       INIT_LIST_HEAD(&fc->processing);
+       INIT_LIST_HEAD(&fc->io);
+       INIT_LIST_HEAD(&fc->interrupts);
+       INIT_LIST_HEAD(&fc->bg_queue);
+       INIT_LIST_HEAD(&fc->entry);
+       atomic_set(&fc->num_waiting, 0);
+       fc->bdi.ra_pages = (VM_MAX_READAHEAD * 1024) / PAGE_CACHE_SIZE;
+       fc->bdi.unplug_io_fn = default_unplug_io_fn;
+       /* fuse does it's own writeback accounting */
+       fc->bdi.capabilities = BDI_CAP_NO_ACCT_WB;
+       fc->khctr = 0;
+       fc->polled_files = RB_ROOT;
+       fc->dev = sb->s_dev;
+       err = bdi_init(&fc->bdi);
+       if (err)
+               goto error_mutex_destroy;
+       if (sb->s_bdev) {
+               err = bdi_register(&fc->bdi, NULL, "%u:%u-fuseblk",
+                                  MAJOR(fc->dev), MINOR(fc->dev));
+       } else {
+               err = bdi_register_dev(&fc->bdi, fc->dev);
        }
-       return fc;
+       if (err)
+               goto error_bdi_destroy;
+       /*
+        * For a single fuse filesystem use max 1% of dirty +
+        * writeback threshold.
+        *
+        * This gives about 1M of write buffer for memory maps on a
+        * machine with 1G and 10% dirty_ratio, which should be more
+        * than enough.
+        *
+        * Privileged users can raise it by writing to
+        *
+        *    /sys/class/bdi/<bdi>/max_ratio
+        */
+       bdi_set_max_ratio(&fc->bdi, 1);
+       fc->reqctr = 0;
+       fc->blocked = 1;
+       fc->attr_version = 1;
+       get_random_bytes(&fc->scramble_key, sizeof(fc->scramble_key));
+
+       return 0;
+
+ error_bdi_destroy:
+       bdi_destroy(&fc->bdi);
+ error_mutex_destroy:
+       mutex_destroy(&fc->inst_mutex);
+       return err;
 }
+EXPORT_SYMBOL_GPL(fuse_conn_init);
 
 void fuse_conn_put(struct fuse_conn *fc)
 {
@@ -426,7 +536,7 @@ void fuse_conn_put(struct fuse_conn *fc)
                if (fc->destroy_req)
                        fuse_request_free(fc->destroy_req);
                mutex_destroy(&fc->inst_mutex);
-               kfree(fc);
+               fc->release(fc);
        }
 }
 
@@ -436,21 +546,182 @@ struct fuse_conn *fuse_conn_get(struct fuse_conn *fc)
        return fc;
 }
 
-static struct inode *get_root_inode(struct super_block *sb, unsigned mode)
+static struct inode *fuse_get_root_inode(struct super_block *sb, unsigned mode)
 {
        struct fuse_attr attr;
        memset(&attr, 0, sizeof(attr));
 
        attr.mode = mode;
        attr.ino = FUSE_ROOT_ID;
-       return fuse_iget(sb, 1, 0, &attr);
+       attr.nlink = 1;
+       return fuse_iget(sb, 1, 0, &attr, 0, 0);
 }
 
-static struct super_operations fuse_super_operations = {
+struct fuse_inode_handle {
+       u64 nodeid;
+       u32 generation;
+};
+
+static struct dentry *fuse_get_dentry(struct super_block *sb,
+                                     struct fuse_inode_handle *handle)
+{
+       struct fuse_conn *fc = get_fuse_conn_super(sb);
+       struct inode *inode;
+       struct dentry *entry;
+       int err = -ESTALE;
+
+       if (handle->nodeid == 0)
+               goto out_err;
+
+       inode = ilookup5(sb, handle->nodeid, fuse_inode_eq, &handle->nodeid);
+       if (!inode) {
+               struct fuse_entry_out outarg;
+               struct qstr name;
+
+               if (!fc->export_support)
+                       goto out_err;
+
+               name.len = 1;
+               name.name = ".";
+               err = fuse_lookup_name(sb, handle->nodeid, &name, &outarg,
+                                      &inode);
+               if (err && err != -ENOENT)
+                       goto out_err;
+               if (err || !inode) {
+                       err = -ESTALE;
+                       goto out_err;
+               }
+               err = -EIO;
+               if (get_node_id(inode) != handle->nodeid)
+                       goto out_iput;
+       }
+       err = -ESTALE;
+       if (inode->i_generation != handle->generation)
+               goto out_iput;
+
+       entry = d_obtain_alias(inode);
+       if (!IS_ERR(entry) && get_node_id(inode) != FUSE_ROOT_ID) {
+               entry->d_op = &fuse_dentry_operations;
+               fuse_invalidate_entry_cache(entry);
+       }
+
+       return entry;
+
+ out_iput:
+       iput(inode);
+ out_err:
+       return ERR_PTR(err);
+}
+
+static int fuse_encode_fh(struct dentry *dentry, u32 *fh, int *max_len,
+                          int connectable)
+{
+       struct inode *inode = dentry->d_inode;
+       bool encode_parent = connectable && !S_ISDIR(inode->i_mode);
+       int len = encode_parent ? 6 : 3;
+       u64 nodeid;
+       u32 generation;
+
+       if (*max_len < len)
+               return  255;
+
+       nodeid = get_fuse_inode(inode)->nodeid;
+       generation = inode->i_generation;
+
+       fh[0] = (u32)(nodeid >> 32);
+       fh[1] = (u32)(nodeid & 0xffffffff);
+       fh[2] = generation;
+
+       if (encode_parent) {
+               struct inode *parent;
+
+               spin_lock(&dentry->d_lock);
+               parent = dentry->d_parent->d_inode;
+               nodeid = get_fuse_inode(parent)->nodeid;
+               generation = parent->i_generation;
+               spin_unlock(&dentry->d_lock);
+
+               fh[3] = (u32)(nodeid >> 32);
+               fh[4] = (u32)(nodeid & 0xffffffff);
+               fh[5] = generation;
+       }
+
+       *max_len = len;
+       return encode_parent ? 0x82 : 0x81;
+}
+
+static struct dentry *fuse_fh_to_dentry(struct super_block *sb,
+               struct fid *fid, int fh_len, int fh_type)
+{
+       struct fuse_inode_handle handle;
+
+       if ((fh_type != 0x81 && fh_type != 0x82) || fh_len < 3)
+               return NULL;
+
+       handle.nodeid = (u64) fid->raw[0] << 32;
+       handle.nodeid |= (u64) fid->raw[1];
+       handle.generation = fid->raw[2];
+       return fuse_get_dentry(sb, &handle);
+}
+
+static struct dentry *fuse_fh_to_parent(struct super_block *sb,
+               struct fid *fid, int fh_len, int fh_type)
+{
+       struct fuse_inode_handle parent;
+
+       if (fh_type != 0x82 || fh_len < 6)
+               return NULL;
+
+       parent.nodeid = (u64) fid->raw[3] << 32;
+       parent.nodeid |= (u64) fid->raw[4];
+       parent.generation = fid->raw[5];
+       return fuse_get_dentry(sb, &parent);
+}
+
+static struct dentry *fuse_get_parent(struct dentry *child)
+{
+       struct inode *child_inode = child->d_inode;
+       struct fuse_conn *fc = get_fuse_conn(child_inode);
+       struct inode *inode;
+       struct dentry *parent;
+       struct fuse_entry_out outarg;
+       struct qstr name;
+       int err;
+
+       if (!fc->export_support)
+               return ERR_PTR(-ESTALE);
+
+       name.len = 2;
+       name.name = "..";
+       err = fuse_lookup_name(child_inode->i_sb, get_node_id(child_inode),
+                              &name, &outarg, &inode);
+       if (err) {
+               if (err == -ENOENT)
+                       return ERR_PTR(-ESTALE);
+               return ERR_PTR(err);
+       }
+
+       parent = d_obtain_alias(inode);
+       if (!IS_ERR(parent) && get_node_id(inode) != FUSE_ROOT_ID) {
+               parent->d_op = &fuse_dentry_operations;
+               fuse_invalidate_entry_cache(parent);
+       }
+
+       return parent;
+}
+
+static const struct export_operations fuse_export_operations = {
+       .fh_to_dentry   = fuse_fh_to_dentry,
+       .fh_to_parent   = fuse_fh_to_parent,
+       .encode_fh      = fuse_encode_fh,
+       .get_parent     = fuse_get_parent,
+};
+
+static const struct super_operations fuse_super_operations = {
        .alloc_inode    = fuse_alloc_inode,
        .destroy_inode  = fuse_destroy_inode,
-       .read_inode     = fuse_read_inode,
        .clear_inode    = fuse_clear_inode,
+       .drop_inode     = generic_delete_inode,
        .remount_fs     = fuse_remount_fs,
        .put_super      = fuse_put_super,
        .umount_begin   = fuse_umount_begin,
@@ -473,6 +744,15 @@ static void process_init_reply(struct fuse_conn *fc, struct fuse_req *req)
                                fc->async_read = 1;
                        if (!(arg->flags & FUSE_POSIX_LOCKS))
                                fc->no_lock = 1;
+                       if (arg->flags & FUSE_ATOMIC_O_TRUNC)
+                               fc->atomic_o_trunc = 1;
+                       if (arg->minor >= 9) {
+                               /* LOOKUP has dependency on proto version */
+                               if (arg->flags & FUSE_EXPORT_SUPPORT)
+                                       fc->export_support = 1;
+                       }
+                       if (arg->flags & FUSE_BIG_WRITES)
+                               fc->big_writes = 1;
                } else {
                        ra_pages = fc->max_read / PAGE_CACHE_SIZE;
                        fc->no_lock = 1;
@@ -481,9 +761,9 @@ static void process_init_reply(struct fuse_conn *fc, struct fuse_req *req)
                fc->bdi.ra_pages = min(fc->bdi.ra_pages, ra_pages);
                fc->minor = arg->minor;
                fc->max_write = arg->minor < 5 ? 4096 : arg->max_write;
+               fc->max_write = max_t(unsigned, 4096, fc->max_write);
                fc->conn_init = 1;
        }
-       fuse_put_request(fc, req);
        fc->blocked = 0;
        wake_up_all(&fc->blocked_waitq);
 }
@@ -495,7 +775,8 @@ static void fuse_send_init(struct fuse_conn *fc, struct fuse_req *req)
        arg->major = FUSE_KERNEL_VERSION;
        arg->minor = FUSE_KERNEL_MINOR_VERSION;
        arg->max_readahead = fc->bdi.ra_pages * PAGE_CACHE_SIZE;
-       arg->flags |= FUSE_ASYNC_READ | FUSE_POSIX_LOCKS;
+       arg->flags |= FUSE_ASYNC_READ | FUSE_POSIX_LOCKS | FUSE_ATOMIC_O_TRUNC |
+               FUSE_EXPORT_SUPPORT | FUSE_BIG_WRITES;
        req->in.h.opcode = FUSE_INIT;
        req->in.numargs = 1;
        req->in.args[0].size = sizeof(*arg);
@@ -508,13 +789,12 @@ static void fuse_send_init(struct fuse_conn *fc, struct fuse_req *req)
        req->out.args[0].size = sizeof(struct fuse_init_out);
        req->out.args[0].value = &req->misc.init_out;
        req->end = process_init_reply;
-       request_send_background(fc, req);
+       fuse_request_send_background(fc, req);
 }
 
-static u64 conn_id(void)
+static void fuse_free_conn(struct fuse_conn *fc)
 {
-       static u64 ctr = 1;
-       return ctr++;
+       kfree(fc);
 }
 
 static int fuse_fill_super(struct super_block *sb, void *data, int silent)
@@ -528,15 +808,19 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent)
        int err;
        int is_bdev = sb->s_bdev != NULL;
 
+       err = -EINVAL;
        if (sb->s_flags & MS_MANDLOCK)
-               return -EINVAL;
+               goto err;
 
        if (!parse_fuse_opt((char *) data, &d, is_bdev))
-               return -EINVAL;
+               goto err;
 
        if (is_bdev) {
+#ifdef CONFIG_BLOCK
+               err = -EINVAL;
                if (!sb_set_blocksize(sb, d.blksize))
-                       return -EINVAL;
+                       goto err;
+#endif
        } else {
                sb->s_blocksize = PAGE_CACHE_SIZE;
                sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
@@ -544,35 +828,45 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent)
        sb->s_magic = FUSE_SUPER_MAGIC;
        sb->s_op = &fuse_super_operations;
        sb->s_maxbytes = MAX_LFS_FILESIZE;
+       sb->s_export_op = &fuse_export_operations;
 
        file = fget(d.fd);
+       err = -EINVAL;
        if (!file)
-               return -EINVAL;
+               goto err;
 
        if (file->f_op != &fuse_dev_operations)
-               return -EINVAL;
+               goto err_fput;
 
-       fc = new_conn();
+       fc = kmalloc(sizeof(*fc), GFP_KERNEL);
+       err = -ENOMEM;
        if (!fc)
-               return -ENOMEM;
+               goto err_fput;
 
+       err = fuse_conn_init(fc, sb);
+       if (err) {
+               kfree(fc);
+               goto err_fput;
+       }
+
+       fc->release = fuse_free_conn;
        fc->flags = d.flags;
        fc->user_id = d.user_id;
        fc->group_id = d.group_id;
-       fc->max_read = d.max_read;
+       fc->max_read = max_t(unsigned, 4096, d.max_read);
 
        /* Used by get_root_inode() */
        sb->s_fs_info = fc;
 
        err = -ENOMEM;
-       root = get_root_inode(sb, d.rootmode);
+       root = fuse_get_root_inode(sb, d.rootmode);
        if (!root)
-               goto err;
+               goto err_put_conn;
 
        root_dentry = d_alloc_root(root);
        if (!root_dentry) {
                iput(root);
-               goto err;
+               goto err_put_conn;
        }
 
        init_req = fuse_request_alloc();
@@ -582,7 +876,7 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent)
        if (is_bdev) {
                fc->destroy_req = fuse_request_alloc();
                if (!fc->destroy_req)
-                       goto err_put_root;
+                       goto err_free_init_req;
        }
 
        mutex_lock(&fuse_mutex);
@@ -590,7 +884,6 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent)
        if (file->private_data)
                goto err_unlock;
 
-       fc->id = conn_id();
        err = fuse_ctl_add_conn(fc);
        if (err)
                goto err_unlock;
@@ -613,12 +906,16 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent)
 
  err_unlock:
        mutex_unlock(&fuse_mutex);
+ err_free_init_req:
        fuse_request_free(init_req);
  err_put_root:
        dput(root_dentry);
- err:
-       fput(file);
+ err_put_conn:
+       bdi_destroy(&fc->bdi);
        fuse_conn_put(fc);
+ err_fput:
+       fput(file);
+ err:
        return err;
 }
 
@@ -629,6 +926,15 @@ static int fuse_get_sb(struct file_system_type *fs_type,
        return get_sb_nodev(fs_type, flags, raw_data, fuse_fill_super, mnt);
 }
 
+static struct file_system_type fuse_fs_type = {
+       .owner          = THIS_MODULE,
+       .name           = "fuse",
+       .fs_flags       = FS_HAS_SUBTYPE,
+       .get_sb         = fuse_get_sb,
+       .kill_sb        = kill_anon_super,
+};
+
+#ifdef CONFIG_BLOCK
 static int fuse_get_sb_blk(struct file_system_type *fs_type,
                           int flags, const char *dev_name,
                           void *raw_data, struct vfsmount *mnt)
@@ -637,32 +943,39 @@ static int fuse_get_sb_blk(struct file_system_type *fs_type,
                           mnt);
 }
 
-static struct file_system_type fuse_fs_type = {
-       .owner          = THIS_MODULE,
-       .name           = "fuse",
-       .get_sb         = fuse_get_sb,
-       .kill_sb        = kill_anon_super,
-};
-
 static struct file_system_type fuseblk_fs_type = {
        .owner          = THIS_MODULE,
        .name           = "fuseblk",
        .get_sb         = fuse_get_sb_blk,
        .kill_sb        = kill_block_super,
-       .fs_flags       = FS_REQUIRES_DEV,
+       .fs_flags       = FS_REQUIRES_DEV | FS_HAS_SUBTYPE,
 };
 
-static decl_subsys(fuse, NULL, NULL);
-static decl_subsys(connections, NULL, NULL);
+static inline int register_fuseblk(void)
+{
+       return register_filesystem(&fuseblk_fs_type);
+}
+
+static inline void unregister_fuseblk(void)
+{
+       unregister_filesystem(&fuseblk_fs_type);
+}
+#else
+static inline int register_fuseblk(void)
+{
+       return 0;
+}
+
+static inline void unregister_fuseblk(void)
+{
+}
+#endif
 
-static void fuse_inode_init_once(void *foo, struct kmem_cache *cachep,
-                                unsigned long flags)
+static void fuse_inode_init_once(void *foo)
 {
-       struct inode * inode = foo;
+       struct inode *inode = foo;
 
-       if ((flags & (SLAB_CTOR_VERIFY|SLAB_CTOR_CONSTRUCTOR)) ==
-           SLAB_CTOR_CONSTRUCTOR)
-               inode_init_once(inode);
+       inode_init_once(inode);
 }
 
 static int __init fuse_fs_init(void)
@@ -673,14 +986,14 @@ static int __init fuse_fs_init(void)
        if (err)
                goto out;
 
-       err = register_filesystem(&fuseblk_fs_type);
+       err = register_fuseblk();
        if (err)
                goto out_unreg;
 
        fuse_inode_cachep = kmem_cache_create("fuse_inode",
                                              sizeof(struct fuse_inode),
                                              0, SLAB_HWCACHE_ALIGN,
-                                             fuse_inode_init_once, NULL);
+                                             fuse_inode_init_once);
        err = -ENOMEM;
        if (!fuse_inode_cachep)
                goto out_unreg2;
@@ -688,7 +1001,7 @@ static int __init fuse_fs_init(void)
        return 0;
 
  out_unreg2:
-       unregister_filesystem(&fuseblk_fs_type);
+       unregister_fuseblk();
  out_unreg:
        unregister_filesystem(&fuse_fs_type);
  out:
@@ -698,43 +1011,48 @@ static int __init fuse_fs_init(void)
 static void fuse_fs_cleanup(void)
 {
        unregister_filesystem(&fuse_fs_type);
-       unregister_filesystem(&fuseblk_fs_type);
+       unregister_fuseblk();
        kmem_cache_destroy(fuse_inode_cachep);
 }
 
+static struct kobject *fuse_kobj;
+static struct kobject *connections_kobj;
+
 static int fuse_sysfs_init(void)
 {
        int err;
 
-       kset_set_kset_s(&fuse_subsys, fs_subsys);
-       err = subsystem_register(&fuse_subsys);
-       if (err)
+       fuse_kobj = kobject_create_and_add("fuse", fs_kobj);
+       if (!fuse_kobj) {
+               err = -ENOMEM;
                goto out_err;
+       }
 
-       kset_set_kset_s(&connections_subsys, fuse_subsys);
-       err = subsystem_register(&connections_subsys);
-       if (err)
+       connections_kobj = kobject_create_and_add("connections", fuse_kobj);
+       if (!connections_kobj) {
+               err = -ENOMEM;
                goto out_fuse_unregister;
+       }
 
        return 0;
 
  out_fuse_unregister:
-       subsystem_unregister(&fuse_subsys);
+       kobject_put(fuse_kobj);
  out_err:
        return err;
 }
 
 static void fuse_sysfs_cleanup(void)
 {
-       subsystem_unregister(&connections_subsys);
-       subsystem_unregister(&fuse_subsys);
+       kobject_put(connections_kobj);
+       kobject_put(fuse_kobj);
 }
 
 static int __init fuse_init(void)
 {
        int res;
 
-       printk("fuse init (API version %i.%i)\n",
+       printk(KERN_INFO "fuse init (API version %i.%i)\n",
               FUSE_KERNEL_VERSION, FUSE_KERNEL_MINOR_VERSION);
 
        INIT_LIST_HEAD(&fuse_conn_list);