summaryrefslogtreecommitdiffstats
path: root/fs/fuse
diff options
context:
space:
mode:
authorStefan Richter <stefanr@s5r6.in-berlin.de>2011-05-10 20:52:07 +0200
committerStefan Richter <stefanr@s5r6.in-berlin.de>2011-05-10 22:50:41 +0200
commit020abf03cd659388f94cb328e1e1df0656e0d7ff (patch)
tree40d05011708ad1b4a05928d167eb120420581aa6 /fs/fuse
parentfirewire: ohci: optimize find_branch_descriptor() (diff)
parentLinux 2.6.39-rc7 (diff)
downloadlinux-020abf03cd659388f94cb328e1e1df0656e0d7ff.tar.xz
linux-020abf03cd659388f94cb328e1e1df0656e0d7ff.zip
Merge tag 'v2.6.39-rc7'
in order to pull in changes in drivers/media/dvb/firewire/ and sound/firewire/.
Diffstat (limited to 'fs/fuse')
-rw-r--r--fs/fuse/cuse.c14
-rw-r--r--fs/fuse/dev.c183
-rw-r--r--fs/fuse/dir.c101
-rw-r--r--fs/fuse/file.c186
-rw-r--r--fs/fuse/fuse_i.h34
-rw-r--r--fs/fuse/inode.c54
6 files changed, 453 insertions, 119 deletions
diff --git a/fs/fuse/cuse.c b/fs/fuse/cuse.c
index 3e87cce5837d..b6cca47f7b07 100644
--- a/fs/fuse/cuse.c
+++ b/fs/fuse/cuse.c
@@ -305,7 +305,7 @@ static void cuse_gendev_release(struct device *dev)
static void cuse_process_init_reply(struct fuse_conn *fc, struct fuse_req *req)
{
struct cuse_conn *cc = fc_to_cc(fc);
- struct cuse_init_out *arg = &req->misc.cuse_init_out;
+ struct cuse_init_out *arg = req->out.args[0].value;
struct page *page = req->pages[0];
struct cuse_devinfo devinfo = { };
struct device *dev;
@@ -384,6 +384,7 @@ static void cuse_process_init_reply(struct fuse_conn *fc, struct fuse_req *req)
dev_set_uevent_suppress(dev, 0);
kobject_uevent(&dev->kobj, KOBJ_ADD);
out:
+ kfree(arg);
__free_page(page);
return;
@@ -405,6 +406,7 @@ static int cuse_send_init(struct cuse_conn *cc)
struct page *page;
struct fuse_conn *fc = &cc->fc;
struct cuse_init_in *arg;
+ void *outarg;
BUILD_BUG_ON(CUSE_INIT_INFO_MAX > PAGE_SIZE);
@@ -419,6 +421,10 @@ static int cuse_send_init(struct cuse_conn *cc)
if (!page)
goto err_put_req;
+ outarg = kzalloc(sizeof(struct cuse_init_out), GFP_KERNEL);
+ if (!outarg)
+ goto err_free_page;
+
arg = &req->misc.cuse_init_in;
arg->major = FUSE_KERNEL_VERSION;
arg->minor = FUSE_KERNEL_MINOR_VERSION;
@@ -429,7 +435,7 @@ static int cuse_send_init(struct cuse_conn *cc)
req->in.args[0].value = arg;
req->out.numargs = 2;
req->out.args[0].size = sizeof(struct cuse_init_out);
- req->out.args[0].value = &req->misc.cuse_init_out;
+ req->out.args[0].value = outarg;
req->out.args[1].size = CUSE_INIT_INFO_MAX;
req->out.argvar = 1;
req->out.argpages = 1;
@@ -440,6 +446,8 @@ static int cuse_send_init(struct cuse_conn *cc)
return 0;
+err_free_page:
+ __free_page(page);
err_put_req:
fuse_put_request(fc, req);
err:
@@ -458,7 +466,7 @@ static void cuse_fc_release(struct fuse_conn *fc)
* @file: file struct being opened
*
* Userland CUSE server can create a CUSE device by opening /dev/cuse
- * and replying to the initilaization request kernel sends. This
+ * and replying to the initialization request kernel sends. This
* function is responsible for handling CUSE device initialization.
* Because the fd opened by this function is used during
* initialization, this function only creates cuse_conn and sends
diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c
index 6e07696308dc..640fc229df10 100644
--- a/fs/fuse/dev.c
+++ b/fs/fuse/dev.c
@@ -251,6 +251,20 @@ static void queue_request(struct fuse_conn *fc, struct fuse_req *req)
kill_fasync(&fc->fasync, SIGIO, POLL_IN);
}
+void fuse_queue_forget(struct fuse_conn *fc, struct fuse_forget_link *forget,
+ u64 nodeid, u64 nlookup)
+{
+ forget->forget_one.nodeid = nodeid;
+ forget->forget_one.nlookup = nlookup;
+
+ spin_lock(&fc->lock);
+ fc->forget_list_tail->next = forget;
+ fc->forget_list_tail = forget;
+ wake_up(&fc->waitq);
+ kill_fasync(&fc->fasync, SIGIO, POLL_IN);
+ spin_unlock(&fc->lock);
+}
+
static void flush_bg_queue(struct fuse_conn *fc)
{
while (fc->active_background < fc->max_background &&
@@ -438,12 +452,6 @@ static void fuse_request_send_nowait(struct fuse_conn *fc, struct fuse_req *req)
}
}
-void fuse_request_send_noreply(struct fuse_conn *fc, struct fuse_req *req)
-{
- req->isreply = 0;
- fuse_request_send_nowait(fc, req);
-}
-
void fuse_request_send_background(struct fuse_conn *fc, struct fuse_req *req)
{
req->isreply = 1;
@@ -729,14 +737,12 @@ static int fuse_try_move_page(struct fuse_copy_state *cs, struct page **pagep)
if (WARN_ON(PageMlocked(oldpage)))
goto out_fallback_unlock;
- remove_from_page_cache(oldpage);
- page_cache_release(oldpage);
-
- err = add_to_page_cache_locked(newpage, mapping, index, GFP_KERNEL);
+ err = replace_page_cache_page(oldpage, newpage, GFP_KERNEL);
if (err) {
- printk(KERN_WARNING "fuse_try_move_page: failed to add page");
- goto out_fallback_unlock;
+ unlock_page(newpage);
+ return err;
}
+
page_cache_get(newpage);
if (!(buf->flags & PIPE_BUF_FLAG_LRU))
@@ -896,9 +902,15 @@ static int fuse_copy_args(struct fuse_copy_state *cs, unsigned numargs,
return err;
}
+static int forget_pending(struct fuse_conn *fc)
+{
+ return fc->forget_list_head.next != NULL;
+}
+
static int request_pending(struct fuse_conn *fc)
{
- return !list_empty(&fc->pending) || !list_empty(&fc->interrupts);
+ return !list_empty(&fc->pending) || !list_empty(&fc->interrupts) ||
+ forget_pending(fc);
}
/* Wait until a request is available on the pending list */
@@ -960,6 +972,120 @@ __releases(fc->lock)
return err ? err : reqsize;
}
+static struct fuse_forget_link *dequeue_forget(struct fuse_conn *fc,
+ unsigned max,
+ unsigned *countp)
+{
+ struct fuse_forget_link *head = fc->forget_list_head.next;
+ struct fuse_forget_link **newhead = &head;
+ unsigned count;
+
+ for (count = 0; *newhead != NULL && count < max; count++)
+ newhead = &(*newhead)->next;
+
+ fc->forget_list_head.next = *newhead;
+ *newhead = NULL;
+ if (fc->forget_list_head.next == NULL)
+ fc->forget_list_tail = &fc->forget_list_head;
+
+ if (countp != NULL)
+ *countp = count;
+
+ return head;
+}
+
+static int fuse_read_single_forget(struct fuse_conn *fc,
+ struct fuse_copy_state *cs,
+ size_t nbytes)
+__releases(fc->lock)
+{
+ int err;
+ struct fuse_forget_link *forget = dequeue_forget(fc, 1, NULL);
+ struct fuse_forget_in arg = {
+ .nlookup = forget->forget_one.nlookup,
+ };
+ struct fuse_in_header ih = {
+ .opcode = FUSE_FORGET,
+ .nodeid = forget->forget_one.nodeid,
+ .unique = fuse_get_unique(fc),
+ .len = sizeof(ih) + sizeof(arg),
+ };
+
+ spin_unlock(&fc->lock);
+ kfree(forget);
+ if (nbytes < ih.len)
+ return -EINVAL;
+
+ err = fuse_copy_one(cs, &ih, sizeof(ih));
+ if (!err)
+ err = fuse_copy_one(cs, &arg, sizeof(arg));
+ fuse_copy_finish(cs);
+
+ if (err)
+ return err;
+
+ return ih.len;
+}
+
+static int fuse_read_batch_forget(struct fuse_conn *fc,
+ struct fuse_copy_state *cs, size_t nbytes)
+__releases(fc->lock)
+{
+ int err;
+ unsigned max_forgets;
+ unsigned count;
+ struct fuse_forget_link *head;
+ struct fuse_batch_forget_in arg = { .count = 0 };
+ struct fuse_in_header ih = {
+ .opcode = FUSE_BATCH_FORGET,
+ .unique = fuse_get_unique(fc),
+ .len = sizeof(ih) + sizeof(arg),
+ };
+
+ if (nbytes < ih.len) {
+ spin_unlock(&fc->lock);
+ return -EINVAL;
+ }
+
+ max_forgets = (nbytes - ih.len) / sizeof(struct fuse_forget_one);
+ head = dequeue_forget(fc, max_forgets, &count);
+ spin_unlock(&fc->lock);
+
+ arg.count = count;
+ ih.len += count * sizeof(struct fuse_forget_one);
+ err = fuse_copy_one(cs, &ih, sizeof(ih));
+ if (!err)
+ err = fuse_copy_one(cs, &arg, sizeof(arg));
+
+ while (head) {
+ struct fuse_forget_link *forget = head;
+
+ if (!err) {
+ err = fuse_copy_one(cs, &forget->forget_one,
+ sizeof(forget->forget_one));
+ }
+ head = forget->next;
+ kfree(forget);
+ }
+
+ fuse_copy_finish(cs);
+
+ if (err)
+ return err;
+
+ return ih.len;
+}
+
+static int fuse_read_forget(struct fuse_conn *fc, struct fuse_copy_state *cs,
+ size_t nbytes)
+__releases(fc->lock)
+{
+ if (fc->minor < 16 || fc->forget_list_head.next->next == NULL)
+ return fuse_read_single_forget(fc, cs, nbytes);
+ else
+ return fuse_read_batch_forget(fc, cs, nbytes);
+}
+
/*
* Read a single request into the userspace filesystem's buffer. This
* function waits until a request is available, then removes it from
@@ -998,6 +1124,14 @@ static ssize_t fuse_dev_do_read(struct fuse_conn *fc, struct file *file,
return fuse_read_interrupt(fc, cs, nbytes, req);
}
+ if (forget_pending(fc)) {
+ if (list_empty(&fc->pending) || fc->forget_batch-- > 0)
+ return fuse_read_forget(fc, cs, nbytes);
+
+ if (fc->forget_batch <= -8)
+ fc->forget_batch = 16;
+ }
+
req = list_entry(fc->pending.next, struct fuse_req, list);
req->state = FUSE_REQ_READING;
list_move(&req->list, &fc->io);
@@ -1090,7 +1224,7 @@ static ssize_t fuse_dev_splice_read(struct file *in, loff_t *ppos,
if (!fc)
return -EPERM;
- bufs = kmalloc(pipe->buffers * sizeof (struct pipe_buffer), GFP_KERNEL);
+ bufs = kmalloc(pipe->buffers * sizeof(struct pipe_buffer), GFP_KERNEL);
if (!bufs)
return -ENOMEM;
@@ -1626,7 +1760,7 @@ static ssize_t fuse_dev_splice_write(struct pipe_inode_info *pipe,
if (!fc)
return -EPERM;
- bufs = kmalloc(pipe->buffers * sizeof (struct pipe_buffer), GFP_KERNEL);
+ bufs = kmalloc(pipe->buffers * sizeof(struct pipe_buffer), GFP_KERNEL);
if (!bufs)
return -ENOMEM;
@@ -1770,6 +1904,23 @@ __acquires(fc->lock)
flush_bg_queue(fc);
end_requests(fc, &fc->pending);
end_requests(fc, &fc->processing);
+ while (forget_pending(fc))
+ kfree(dequeue_forget(fc, 1, NULL));
+}
+
+static void end_polls(struct fuse_conn *fc)
+{
+ struct rb_node *p;
+
+ p = rb_first(&fc->polled_files);
+
+ while (p) {
+ struct fuse_file *ff;
+ ff = rb_entry(p, struct fuse_file, polled_node);
+ wake_up_interruptible_all(&ff->poll_wait);
+
+ p = rb_next(p);
+ }
}
/*
@@ -1799,6 +1950,7 @@ void fuse_abort_conn(struct fuse_conn *fc)
fc->blocked = 0;
end_io_requests(fc);
end_queued_requests(fc);
+ end_polls(fc);
wake_up_all(&fc->waitq);
wake_up_all(&fc->blocked_waitq);
kill_fasync(&fc->fasync, SIGIO, POLL_IN);
@@ -1815,6 +1967,7 @@ int fuse_dev_release(struct inode *inode, struct file *file)
fc->connected = 0;
fc->blocked = 0;
end_queued_requests(fc);
+ end_polls(fc);
wake_up_all(&fc->blocked_waitq);
spin_unlock(&fc->lock);
fuse_conn_put(fc);
diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c
index c9627c95482d..c6ba49bd95b3 100644
--- a/fs/fuse/dir.c
+++ b/fs/fuse/dir.c
@@ -10,9 +10,9 @@
#include <linux/pagemap.h>
#include <linux/file.h>
-#include <linux/gfp.h>
#include <linux/sched.h>
#include <linux/namei.h>
+#include <linux/slab.h>
#if BITS_PER_LONG >= 64
static inline void fuse_dentry_settime(struct dentry *entry, u64 time)
@@ -156,8 +156,9 @@ u64 fuse_get_attr_version(struct fuse_conn *fc)
*/
static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd)
{
- struct inode *inode = entry->d_inode;
+ struct inode *inode;
+ inode = ACCESS_ONCE(entry->d_inode);
if (inode && is_bad_inode(inode))
return 0;
else if (fuse_dentry_time(entry) < get_jiffies_64()) {
@@ -165,7 +166,7 @@ static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd)
struct fuse_entry_out outarg;
struct fuse_conn *fc;
struct fuse_req *req;
- struct fuse_req *forget_req;
+ struct fuse_forget_link *forget;
struct dentry *parent;
u64 attr_version;
@@ -173,13 +174,16 @@ static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd)
if (!inode)
return 0;
+ if (nd->flags & LOOKUP_RCU)
+ return -ECHILD;
+
fc = get_fuse_conn(inode);
req = fuse_get_req(fc);
if (IS_ERR(req))
return 0;
- forget_req = fuse_get_req(fc);
- if (IS_ERR(forget_req)) {
+ forget = fuse_alloc_forget();
+ if (!forget) {
fuse_put_request(fc, req);
return 0;
}
@@ -199,15 +203,14 @@ static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd)
if (!err) {
struct fuse_inode *fi = get_fuse_inode(inode);
if (outarg.nodeid != get_node_id(inode)) {
- fuse_send_forget(fc, forget_req,
- outarg.nodeid, 1);
+ fuse_queue_forget(fc, forget, outarg.nodeid, 1);
return 0;
}
spin_lock(&fc->lock);
fi->nlookup++;
spin_unlock(&fc->lock);
}
- fuse_put_request(fc, forget_req);
+ kfree(forget);
if (err || (outarg.attr.mode ^ inode->i_mode) & S_IFMT)
return 0;
@@ -259,7 +262,7 @@ int fuse_lookup_name(struct super_block *sb, u64 nodeid, struct qstr *name,
{
struct fuse_conn *fc = get_fuse_conn_super(sb);
struct fuse_req *req;
- struct fuse_req *forget_req;
+ struct fuse_forget_link *forget;
u64 attr_version;
int err;
@@ -273,9 +276,9 @@ int fuse_lookup_name(struct super_block *sb, u64 nodeid, struct qstr *name,
if (IS_ERR(req))
goto out;
- forget_req = fuse_get_req(fc);
- err = PTR_ERR(forget_req);
- if (IS_ERR(forget_req)) {
+ forget = fuse_alloc_forget();
+ err = -ENOMEM;
+ if (!forget) {
fuse_put_request(fc, req);
goto out;
}
@@ -301,13 +304,13 @@ int fuse_lookup_name(struct super_block *sb, u64 nodeid, struct qstr *name,
attr_version);
err = -ENOMEM;
if (!*inode) {
- fuse_send_forget(fc, forget_req, outarg->nodeid, 1);
+ fuse_queue_forget(fc, forget, outarg->nodeid, 1);
goto out;
}
err = 0;
out_put_forget:
- fuse_put_request(fc, forget_req);
+ kfree(forget);
out:
return err;
}
@@ -347,7 +350,6 @@ static struct dentry *fuse_lookup(struct inode *dir, struct dentry *entry,
}
entry = newent ? newent : entry;
- entry->d_op = &fuse_dentry_operations;
if (outarg_valid)
fuse_change_entry_timeout(entry, &outarg);
else
@@ -374,7 +376,7 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, int mode,
struct inode *inode;
struct fuse_conn *fc = get_fuse_conn(dir);
struct fuse_req *req;
- struct fuse_req *forget_req;
+ struct fuse_forget_link *forget;
struct fuse_create_in inarg;
struct fuse_open_out outopen;
struct fuse_entry_out outentry;
@@ -388,9 +390,9 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, int mode,
if (flags & O_DIRECT)
return -EINVAL;
- forget_req = fuse_get_req(fc);
- if (IS_ERR(forget_req))
- return PTR_ERR(forget_req);
+ forget = fuse_alloc_forget();
+ if (!forget)
+ return -ENOMEM;
req = fuse_get_req(fc);
err = PTR_ERR(req);
@@ -448,10 +450,10 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, int mode,
if (!inode) {
flags &= ~(O_CREAT | O_EXCL | O_TRUNC);
fuse_sync_release(ff, flags);
- fuse_send_forget(fc, forget_req, outentry.nodeid, 1);
+ fuse_queue_forget(fc, forget, outentry.nodeid, 1);
return -ENOMEM;
}
- fuse_put_request(fc, forget_req);
+ kfree(forget);
d_instantiate(entry, inode);
fuse_change_entry_timeout(entry, &outentry);
fuse_invalidate_attr(dir);
@@ -469,7 +471,7 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, int mode,
out_put_request:
fuse_put_request(fc, req);
out_put_forget_req:
- fuse_put_request(fc, forget_req);
+ kfree(forget);
return err;
}
@@ -483,12 +485,12 @@ static int create_new_entry(struct fuse_conn *fc, struct fuse_req *req,
struct fuse_entry_out outarg;
struct inode *inode;
int err;
- struct fuse_req *forget_req;
+ struct fuse_forget_link *forget;
- forget_req = fuse_get_req(fc);
- if (IS_ERR(forget_req)) {
+ forget = fuse_alloc_forget();
+ if (!forget) {
fuse_put_request(fc, req);
- return PTR_ERR(forget_req);
+ return -ENOMEM;
}
memset(&outarg, 0, sizeof(outarg));
@@ -515,10 +517,10 @@ static int create_new_entry(struct fuse_conn *fc, struct fuse_req *req,
inode = fuse_iget(dir->i_sb, outarg.nodeid, outarg.generation,
&outarg.attr, entry_attr_timeout(&outarg), 0);
if (!inode) {
- fuse_send_forget(fc, forget_req, outarg.nodeid, 1);
+ fuse_queue_forget(fc, forget, outarg.nodeid, 1);
return -ENOMEM;
}
- fuse_put_request(fc, forget_req);
+ kfree(forget);
if (S_ISDIR(inode->i_mode)) {
struct dentry *alias;
@@ -541,7 +543,7 @@ static int create_new_entry(struct fuse_conn *fc, struct fuse_req *req,
return 0;
out_put_forget_req:
- fuse_put_request(fc, forget_req);
+ kfree(forget);
return err;
}
@@ -968,6 +970,14 @@ static int fuse_access(struct inode *inode, int mask)
return err;
}
+static int fuse_perm_getattr(struct inode *inode, int flags)
+{
+ if (flags & IPERM_FLAG_RCU)
+ return -ECHILD;
+
+ return fuse_do_getattr(inode, NULL, NULL);
+}
+
/*
* Check permission. The two basic access models of FUSE are:
*
@@ -981,7 +991,7 @@ static int fuse_access(struct inode *inode, int mask)
* access request is sent. Execute permission is still checked
* locally based on file mode.
*/
-static int fuse_permission(struct inode *inode, int mask)
+static int fuse_permission(struct inode *inode, int mask, unsigned int flags)
{
struct fuse_conn *fc = get_fuse_conn(inode);
bool refreshed = false;
@@ -995,21 +1005,28 @@ static int fuse_permission(struct inode *inode, int mask)
*/
if ((fc->flags & FUSE_DEFAULT_PERMISSIONS) ||
((mask & MAY_EXEC) && S_ISREG(inode->i_mode))) {
- err = fuse_update_attributes(inode, NULL, NULL, &refreshed);
- if (err)
- return err;
+ struct fuse_inode *fi = get_fuse_inode(inode);
+
+ if (fi->i_time < get_jiffies_64()) {
+ refreshed = true;
+
+ err = fuse_perm_getattr(inode, flags);
+ if (err)
+ return err;
+ }
}
if (fc->flags & FUSE_DEFAULT_PERMISSIONS) {
- err = generic_permission(inode, mask, NULL);
+ err = generic_permission(inode, mask, flags, NULL);
/* If permission is denied, try to refresh file
attributes. This is also needed, because the root
node will at first have no permissions */
if (err == -EACCES && !refreshed) {
- err = fuse_do_getattr(inode, NULL, NULL);
+ err = fuse_perm_getattr(inode, flags);
if (!err)
- err = generic_permission(inode, mask, NULL);
+ err = generic_permission(inode, mask,
+ flags, NULL);
}
/* Note: the opposite of the above test does not
@@ -1017,13 +1034,16 @@ static int fuse_permission(struct inode *inode, int mask)
noticed immediately, only after the attribute
timeout has expired */
} else if (mask & (MAY_ACCESS | MAY_CHDIR)) {
+ if (flags & IPERM_FLAG_RCU)
+ return -ECHILD;
+
err = fuse_access(inode, mask);
} else if ((mask & MAY_EXEC) && S_ISREG(inode->i_mode)) {
if (!(inode->i_mode & S_IXUGO)) {
if (refreshed)
return -EACCES;
- err = fuse_do_getattr(inode, NULL, NULL);
+ err = fuse_perm_getattr(inode, flags);
if (!err && !(inode->i_mode & S_IXUGO))
return -EACCES;
}
@@ -1277,8 +1297,11 @@ static int fuse_do_setattr(struct dentry *entry, struct iattr *attr,
if (err)
return err;
- if ((attr->ia_valid & ATTR_OPEN) && fc->atomic_o_trunc)
- return 0;
+ if (attr->ia_valid & ATTR_OPEN) {
+ if (fc->atomic_o_trunc)
+ return 0;
+ file = NULL;
+ }
if (attr->ia_valid & ATTR_SIZE)
is_truncate = true;
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index 9242d294fe90..82a66466a24c 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -13,6 +13,7 @@
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/module.h>
+#include <linux/compat.h>
static const struct file_operations fuse_direct_io_file_operations;
@@ -85,18 +86,52 @@ struct fuse_file *fuse_file_get(struct fuse_file *ff)
return ff;
}
+static void fuse_release_async(struct work_struct *work)
+{
+ struct fuse_req *req;
+ struct fuse_conn *fc;
+ struct path path;
+
+ req = container_of(work, struct fuse_req, misc.release.work);
+ path = req->misc.release.path;
+ fc = get_fuse_conn(path.dentry->d_inode);
+
+ fuse_put_request(fc, req);
+ path_put(&path);
+}
+
static void fuse_release_end(struct fuse_conn *fc, struct fuse_req *req)
{
- path_put(&req->misc.release.path);
+ if (fc->destroy_req) {
+ /*
+ * If this is a fuseblk mount, then it's possible that
+ * releasing the path will result in releasing the
+ * super block and sending the DESTROY request. If
+ * the server is single threaded, this would hang.
+ * For this reason do the path_put() in a separate
+ * thread.
+ */
+ atomic_inc(&req->count);
+ INIT_WORK(&req->misc.release.work, fuse_release_async);
+ schedule_work(&req->misc.release.work);
+ } else {
+ path_put(&req->misc.release.path);
+ }
}
-static void fuse_file_put(struct fuse_file *ff)
+static void fuse_file_put(struct fuse_file *ff, bool sync)
{
if (atomic_dec_and_test(&ff->count)) {
struct fuse_req *req = ff->reserved_req;
- req->end = fuse_release_end;
- fuse_request_send_background(ff->fc, req);
+ if (sync) {
+ fuse_request_send(ff->fc, req);
+ path_put(&req->misc.release.path);
+ fuse_put_request(ff->fc, req);
+ } else {
+ req->end = fuse_release_end;
+ fuse_request_send_background(ff->fc, req);
+ }
kfree(ff);
}
}
@@ -187,7 +222,7 @@ static void fuse_prepare_release(struct fuse_file *ff, int flags, int opcode)
rb_erase(&ff->polled_node, &fc->polled_files);
spin_unlock(&fc->lock);
- wake_up_interruptible_sync(&ff->poll_wait);
+ wake_up_interruptible_all(&ff->poll_wait);
inarg->fh = ff->fh;
inarg->flags = flags;
@@ -218,8 +253,12 @@ void fuse_release_common(struct file *file, int opcode)
* Normally this will send the RELEASE request, however if
* some asynchronous READ or WRITE requests are outstanding,
* the sending will be delayed.
+ *
+ * Make the release synchronous if this is a fuseblk mount,
+ * synchronous RELEASE is allowed (and desirable) in this case
+ * because the server can be trusted not to screw up.
*/
- fuse_file_put(ff);
+ fuse_file_put(ff, ff->fc->destroy_req != NULL);
}
static int fuse_open(struct inode *inode, struct file *file)
@@ -484,7 +523,7 @@ static int fuse_readpage(struct file *file, struct page *page)
goto out;
/*
- * Page writeback can extend beyond the liftime of the
+ * Page writeback can extend beyond the lifetime of the
* page-cache page, so make sure we read a properly synced
* page.
*/
@@ -557,7 +596,7 @@ static void fuse_readpages_end(struct fuse_conn *fc, struct fuse_req *req)
page_cache_release(page);
}
if (req->ff)
- fuse_file_put(req->ff);
+ fuse_file_put(req->ff, false);
}
static void fuse_send_readpages(struct fuse_req *req, struct file *file)
@@ -1136,7 +1175,7 @@ static ssize_t fuse_direct_write(struct file *file, const char __user *buf,
static void fuse_writepage_free(struct fuse_conn *fc, struct fuse_req *req)
{
__free_page(req->pages[0]);
- fuse_file_put(req->ff);
+ fuse_file_put(req->ff, false);
}
static void fuse_writepage_finish(struct fuse_conn *fc, struct fuse_req *req)
@@ -1628,6 +1667,94 @@ static int fuse_ioctl_copy_user(struct page **pages, struct iovec *iov,
}
/*
+ * CUSE servers compiled on 32bit broke on 64bit kernels because the
+ * ABI was defined to be 'struct iovec' which is different on 32bit
+ * and 64bit. Fortunately we can determine which structure the server
+ * used from the size of the reply.
+ */
+static int fuse_copy_ioctl_iovec_old(struct iovec *dst, void *src,
+ size_t transferred, unsigned count,
+ bool is_compat)
+{
+#ifdef CONFIG_COMPAT
+ if (count * sizeof(struct compat_iovec) == transferred) {
+ struct compat_iovec *ciov = src;
+ unsigned i;
+
+ /*
+ * With this interface a 32bit server cannot support
+ * non-compat (i.e. ones coming from 64bit apps) ioctl
+ * requests
+ */
+ if (!is_compat)
+ return -EINVAL;
+
+ for (i = 0; i < count; i++) {
+ dst[i].iov_base = compat_ptr(ciov[i].iov_base);
+ dst[i].iov_len = ciov[i].iov_len;
+ }
+ return 0;
+ }
+#endif
+
+ if (count * sizeof(struct iovec) != transferred)
+ return -EIO;
+
+ memcpy(dst, src, transferred);
+ return 0;
+}
+
+/* Make sure iov_length() won't overflow */
+static int fuse_verify_ioctl_iov(struct iovec *iov, size_t count)
+{
+ size_t n;
+ u32 max = FUSE_MAX_PAGES_PER_REQ << PAGE_SHIFT;
+
+ for (n = 0; n < count; n++) {
+ if (iov->iov_len > (size_t) max)
+ return -ENOMEM;
+ max -= iov->iov_len;
+ }
+ return 0;
+}
+
+static int fuse_copy_ioctl_iovec(struct fuse_conn *fc, struct iovec *dst,
+ void *src, size_t transferred, unsigned count,
+ bool is_compat)
+{
+ unsigned i;
+ struct fuse_ioctl_iovec *fiov = src;
+
+ if (fc->minor < 16) {
+ return fuse_copy_ioctl_iovec_old(dst, src, transferred,
+ count, is_compat);
+ }
+
+ if (count * sizeof(struct fuse_ioctl_iovec) != transferred)
+ return -EIO;
+
+ for (i = 0; i < count; i++) {
+ /* Did the server supply an inappropriate value? */
+ if (fiov[i].base != (unsigned long) fiov[i].base ||
+ fiov[i].len != (unsigned long) fiov[i].len)
+ return -EIO;
+
+ dst[i].iov_base = (void __user *) (unsigned long) fiov[i].base;
+ dst[i].iov_len = (size_t) fiov[i].len;
+
+#ifdef CONFIG_COMPAT
+ if (is_compat &&
+ (ptr_to_compat(dst[i].iov_base) != fiov[i].base ||
+ (compat_size_t) dst[i].iov_len != fiov[i].len))
+ return -EIO;
+#endif
+ }
+
+ return 0;
+}
+
+
+/*
* For ioctls, there is no generic way to determine how much memory
* needs to be read and/or written. Furthermore, ioctls are allowed
* to dereference the passed pointer, so the parameter requires deep
@@ -1687,18 +1814,25 @@ long fuse_do_ioctl(struct file *file, unsigned int cmd, unsigned long arg,
struct fuse_ioctl_out outarg;
struct fuse_req *req = NULL;
struct page **pages = NULL;
- struct page *iov_page = NULL;
+ struct iovec *iov_page = NULL;
struct iovec *in_iov = NULL, *out_iov = NULL;
unsigned int in_iovs = 0, out_iovs = 0, num_pages = 0, max_pages;
size_t in_size, out_size, transferred;
int err;
+#if BITS_PER_LONG == 32
+ inarg.flags |= FUSE_IOCTL_32BIT;
+#else
+ if (flags & FUSE_IOCTL_COMPAT)
+ inarg.flags |= FUSE_IOCTL_32BIT;
+#endif
+
/* assume all the iovs returned by client always fits in a page */
- BUILD_BUG_ON(sizeof(struct iovec) * FUSE_IOCTL_MAX_IOV > PAGE_SIZE);
+ BUILD_BUG_ON(sizeof(struct fuse_ioctl_iovec) * FUSE_IOCTL_MAX_IOV > PAGE_SIZE);
err = -ENOMEM;
pages = kzalloc(sizeof(pages[0]) * FUSE_MAX_PAGES_PER_REQ, GFP_KERNEL);
- iov_page = alloc_page(GFP_KERNEL);
+ iov_page = (struct iovec *) __get_free_page(GFP_KERNEL);
if (!pages || !iov_page)
goto out;
@@ -1707,7 +1841,7 @@ long fuse_do_ioctl(struct file *file, unsigned int cmd, unsigned long arg,
* RETRY from server is not allowed.
*/
if (!(flags & FUSE_IOCTL_UNRESTRICTED)) {
- struct iovec *iov = page_address(iov_page);
+ struct iovec *iov = iov_page;
iov->iov_base = (void __user *)arg;
iov->iov_len = _IOC_SIZE(cmd);
@@ -1788,7 +1922,7 @@ long fuse_do_ioctl(struct file *file, unsigned int cmd, unsigned long arg,
/* did it ask for retry? */
if (outarg.flags & FUSE_IOCTL_RETRY) {
- char *vaddr;
+ void *vaddr;
/* no retry if in restricted mode */
err = -EIO;
@@ -1808,18 +1942,25 @@ long fuse_do_ioctl(struct file *file, unsigned int cmd, unsigned long arg,
in_iovs + out_iovs > FUSE_IOCTL_MAX_IOV)
goto out;
- err = -EIO;
- if ((in_iovs + out_iovs) * sizeof(struct iovec) != transferred)
- goto out;
-
- /* okay, copy in iovs and retry */
vaddr = kmap_atomic(pages[0], KM_USER0);
- memcpy(page_address(iov_page), vaddr, transferred);
+ err = fuse_copy_ioctl_iovec(fc, iov_page, vaddr,
+ transferred, in_iovs + out_iovs,
+ (flags & FUSE_IOCTL_COMPAT) != 0);
kunmap_atomic(vaddr, KM_USER0);
+ if (err)
+ goto out;
- in_iov = page_address(iov_page);
+ in_iov = iov_page;
out_iov = in_iov + in_iovs;
+ err = fuse_verify_ioctl_iov(in_iov, in_iovs);
+ if (err)
+ goto out;
+
+ err = fuse_verify_ioctl_iov(out_iov, out_iovs);
+ if (err)
+ goto out;
+
goto retry;
}
@@ -1831,8 +1972,7 @@ long fuse_do_ioctl(struct file *file, unsigned int cmd, unsigned long arg,
out:
if (req)
fuse_put_request(fc, req);
- if (iov_page)
- __free_page(iov_page);
+ free_page((unsigned long) iov_page);
while (num_pages)
__free_page(pages[--num_pages]);
kfree(pages);
diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h
index 57d4a3a0f102..b788becada76 100644
--- a/fs/fuse/fuse_i.h
+++ b/fs/fuse/fuse_i.h
@@ -21,6 +21,7 @@
#include <linux/rwsem.h>
#include <linux/rbtree.h>
#include <linux/poll.h>
+#include <linux/workqueue.h>
/** Max number of pages that can be used in a single read request */
#define FUSE_MAX_PAGES_PER_REQ 32
@@ -53,6 +54,12 @@ extern struct mutex fuse_mutex;
extern unsigned max_user_bgreq;
extern unsigned max_user_congthresh;
+/* One forget request */
+struct fuse_forget_link {
+ struct fuse_forget_one forget_one;
+ struct fuse_forget_link *next;
+};
+
/** FUSE inode */
struct fuse_inode {
/** Inode data */
@@ -66,7 +73,7 @@ struct fuse_inode {
u64 nlookup;
/** The request used for sending the FORGET message */
- struct fuse_req *forget_req;
+ struct fuse_forget_link *forget;
/** Time in jiffies until the file attributes are valid */
u64 i_time;
@@ -255,15 +262,16 @@ struct fuse_req {
/** Data for asynchronous requests */
union {
- struct fuse_forget_in forget_in;
struct {
- struct fuse_release_in in;
+ union {
+ struct fuse_release_in in;
+ struct work_struct work;
+ };
struct path path;
} release;
struct fuse_init_in init_in;
struct fuse_init_out init_out;
struct cuse_init_in cuse_init_in;
- struct cuse_init_out cuse_init_out;
struct {
struct fuse_read_in in;
u64 attr_ver;
@@ -369,6 +377,13 @@ struct fuse_conn {
/** Pending interrupts */
struct list_head interrupts;
+ /** Queue of pending forgets */
+ struct fuse_forget_link forget_list_head;
+ struct fuse_forget_link *forget_list_tail;
+
+ /** Batching of FORGET requests (positive indicates FORGET batch) */
+ int forget_batch;
+
/** Flag indicating if connection is blocked. This will be
the case before the INIT reply is received, and if there
are too many outstading backgrounds requests */
@@ -543,8 +558,10 @@ int fuse_lookup_name(struct super_block *sb, u64 nodeid, struct qstr *name,
/**
* Send FORGET command
*/
-void fuse_send_forget(struct fuse_conn *fc, struct fuse_req *req,
- u64 nodeid, u64 nlookup);
+void fuse_queue_forget(struct fuse_conn *fc, struct fuse_forget_link *forget,
+ u64 nodeid, u64 nlookup);
+
+struct fuse_forget_link *fuse_alloc_forget(void);
/**
* Initialize READ or READDIR request
@@ -656,11 +673,6 @@ void fuse_put_request(struct fuse_conn *fc, struct fuse_req *req);
void fuse_request_send(struct fuse_conn *fc, struct fuse_req *req);
/**
- * Send a request with no reply
- */
-void fuse_request_send_noreply(struct fuse_conn *fc, struct fuse_req *req);
-
-/**
* Send a request in the background
*/
void fuse_request_send_background(struct fuse_conn *fc, struct fuse_req *req);
diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
index cfce3ad86a92..cc6ec4b2f0ff 100644
--- a/fs/fuse/inode.c
+++ b/fs/fuse/inode.c
@@ -71,6 +71,11 @@ struct fuse_mount_data {
unsigned blksize;
};
+struct fuse_forget_link *fuse_alloc_forget()
+{
+ return kzalloc(sizeof(struct fuse_forget_link), GFP_KERNEL);
+}
+
static struct inode *fuse_alloc_inode(struct super_block *sb)
{
struct inode *inode;
@@ -90,8 +95,8 @@ static struct inode *fuse_alloc_inode(struct super_block *sb)
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) {
+ fi->forget = fuse_alloc_forget();
+ if (!fi->forget) {
kmem_cache_free(fuse_inode_cachep, inode);
return NULL;
}
@@ -99,27 +104,20 @@ static struct inode *fuse_alloc_inode(struct super_block *sb)
return inode;
}
-static void fuse_destroy_inode(struct inode *inode)
+static void fuse_i_callback(struct rcu_head *head)
{
- 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);
+ struct inode *inode = container_of(head, struct inode, i_rcu);
+ INIT_LIST_HEAD(&inode->i_dentry);
kmem_cache_free(fuse_inode_cachep, inode);
}
-void fuse_send_forget(struct fuse_conn *fc, struct fuse_req *req,
- u64 nodeid, u64 nlookup)
+static void fuse_destroy_inode(struct inode *inode)
{
- struct fuse_forget_in *inarg = &req->misc.forget_in;
- inarg->nlookup = nlookup;
- req->in.h.opcode = FUSE_FORGET;
- req->in.h.nodeid = nodeid;
- req->in.numargs = 1;
- req->in.args[0].size = sizeof(struct fuse_forget_in);
- req->in.args[0].value = inarg;
- fuse_request_send_noreply(fc, req);
+ struct fuse_inode *fi = get_fuse_inode(inode);
+ BUG_ON(!list_empty(&fi->write_files));
+ BUG_ON(!list_empty(&fi->queued_writes));
+ kfree(fi->forget);
+ call_rcu(&inode->i_rcu, fuse_i_callback);
}
static void fuse_evict_inode(struct inode *inode)
@@ -129,8 +127,8 @@ static void fuse_evict_inode(struct inode *inode)
if (inode->i_sb->s_flags & MS_ACTIVE) {
struct fuse_conn *fc = get_fuse_conn(inode);
struct fuse_inode *fi = get_fuse_inode(inode);
- fuse_send_forget(fc, fi->forget_req, fi->nodeid, fi->nlookup);
- fi->forget_req = NULL;
+ fuse_queue_forget(fc, fi->forget, fi->nodeid, fi->nlookup);
+ fi->forget = NULL;
}
}
@@ -534,6 +532,7 @@ void fuse_conn_init(struct fuse_conn *fc)
INIT_LIST_HEAD(&fc->interrupts);
INIT_LIST_HEAD(&fc->bg_queue);
INIT_LIST_HEAD(&fc->entry);
+ fc->forget_list_tail = &fc->forget_list_head;
atomic_set(&fc->num_waiting, 0);
fc->max_background = FUSE_DEFAULT_MAX_BACKGROUND;
fc->congestion_threshold = FUSE_DEFAULT_CONGESTION_THRESHOLD;
@@ -618,10 +617,8 @@ static struct dentry *fuse_get_dentry(struct super_block *sb,
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;
+ if (!IS_ERR(entry) && get_node_id(inode) != FUSE_ROOT_ID)
fuse_invalidate_entry_cache(entry);
- }
return entry;
@@ -640,8 +637,10 @@ static int fuse_encode_fh(struct dentry *dentry, u32 *fh, int *max_len,
u64 nodeid;
u32 generation;
- if (*max_len < len)
+ if (*max_len < len) {
+ *max_len = len;
return 255;
+ }
nodeid = get_fuse_inode(inode)->nodeid;
generation = inode->i_generation;
@@ -720,10 +719,8 @@ static struct dentry *fuse_get_parent(struct dentry *child)
}
parent = d_obtain_alias(inode);
- if (!IS_ERR(parent) && get_node_id(inode) != FUSE_ROOT_ID) {
- parent->d_op = &fuse_dentry_operations;
+ if (!IS_ERR(parent) && get_node_id(inode) != FUSE_ROOT_ID)
fuse_invalidate_entry_cache(parent);
- }
return parent;
}
@@ -873,7 +870,6 @@ static int fuse_bdi_init(struct fuse_conn *fc, struct super_block *sb)
fc->bdi.name = "fuse";
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;
@@ -990,6 +986,8 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent)
iput(root);
goto err_put_conn;
}
+ /* only now - we want root dentry with NULL ->d_op */
+ sb->s_d_op = &fuse_dentry_operations;
init_req = fuse_request_alloc();
if (!init_req)