[PATCH 2/4] [fuse] fix deadlock between fuse_put_super() and request_end()

[Miklos Szeredi]

A deadlock was possible, when the last reference to the superblock was
held due to a background request containing a file reference.

Releasing the file would release the vfsmount which in turn would
release the superblock.  Since sbput_sem is held during the fput() and
fuse_put_super() tries to acquire this same semaphore, a deadlock
results.

The solution is to move the fput() outside the region protected by
sbput_sem.

Signed-off-by: Miklos Szeredi <miklos@xxxxxxxxxx>

---

 fs/fuse/dev.c |   13 +++++++++++--
 1 files changed, 11 insertions(+), 2 deletions(-)

cdea55d00688ed9a494345b131521a2cae36d42a
diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c
index 4967bd4..104a62d 100644
--- a/fs/fuse/dev.c
+++ b/fs/fuse/dev.c
@@ -128,12 +128,16 @@ void fuse_put_request(struct fuse_conn *
 	}
 }
 
+/*
+ * Called with sbput_sem held for read (request_end) or write
+ * (fuse_put_super).  By the time fuse_put_super() is finished, all
+ * inodes belonging to background requests must be released, so the
+ * iputs have to be done within the locked region.
+ */
 void fuse_release_background(struct fuse_conn *fc, struct fuse_req *req)
 {
 	iput(req->inode);
 	iput(req->inode2);
-	if (req->file)
-		fput(req->file);
 	spin_lock(&fc->lock);
 	list_del(&req->bg_entry);
 	if (fc->num_background == FUSE_MAX_BACKGROUND) {
@@ -178,6 +182,11 @@ static void request_end(struct fuse_conn
 		if (fc->mounted)
 			fuse_release_background(fc, req);
 		up_read(&fc->sbput_sem);
+
+		/* fput must go outside sbput_sem, otherwise it can deadlock */
+		if (req->file)
+			fput(req->file);
+
 		if (end)
 			end(fc, req);
 		else
-- 
1.2.4

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/