Merge branch 'fixes' of git://git.linux-nfs.org/pub/linux/nfs-2.6

* 'fixes' of git://git.linux-nfs.org/pub/linux/nfs-2.6:
  SUNRPC: Fix obvious refcounting bugs in rpc_pipefs.
  RPC: Ensure that we disconnect TCP socket when client requests error out
  NLM/lockd: remove b_done
  NFS: make 2 functions static
  NFS: Release dcache_lock in an error path of nfs_path

11 files changed, 75 insertions, 62 deletions

diff --git a/fs/lockd/svclock.c b/fs/lockd/svclock.c
index baf5ae513481..c9d419703cf3 100644
--- a/fs/lockd/svclock.c
+++ b/fs/lockd/svclock.c
@@ -638,9 +638,6 @@ static void nlmsvc_grant_callback(struct rpc_task *task, void *data)
         if (task->tk_status < 0) {
                 /* RPC error: Re-insert for retransmission */
                 timeout = 10 * HZ;
-        } else if (block->b_done) {
-                /* Block already removed, kill it for real */
-                timeout = 0;
         } else {
                 /* Call was successful, now wait for client callback */
                 timeout = 60 * HZ;
@@ -709,13 +706,10 @@ nlmsvc_retry_blocked(void)
                         break;
                 if (time_after(block->b_when,jiffies))
                         break;
-                dprintk("nlmsvc_retry_blocked(%p, when=%ld, done=%d)\n",
-                        block, block->b_when, block->b_done);
+                dprintk("nlmsvc_retry_blocked(%p, when=%ld)\n",
+                        block, block->b_when);
                 kref_get(&block->b_count);
-                if (block->b_done)
-                        nlmsvc_unlink_block(block);
-                else
-                        nlmsvc_grant_blocked(block);
+                nlmsvc_grant_blocked(block);
                 nlmsvc_release_block(block);
         }
 
diff --git a/fs/nfs/namespace.c b/fs/nfs/namespace.c
index 19b98ca468eb..86b3169c8cac 100644
--- a/fs/nfs/namespace.c
+++ b/fs/nfs/namespace.c
@@ -51,7 +51,7 @@ char *nfs_path(const char *base, const struct dentry *dentry,
                 namelen = dentry->d_name.len;
                 buflen -= namelen + 1;
                 if (buflen < 0)
-                        goto Elong;
+                        goto Elong_unlock;
                 end -= namelen;
                 memcpy(end, dentry->d_name.name, namelen);
                 *--end = '/';
@@ -68,6 +68,8 @@ char *nfs_path(const char *base, const struct dentry *dentry,
         end -= namelen;
         memcpy(end, base, namelen);
         return end;
+Elong_unlock:
+        spin_unlock(&dcache_lock);
 Elong:
         return ERR_PTR(-ENAMETOOLONG);
 }
diff --git a/fs/nfs/read.c b/fs/nfs/read.c
index 52bf634260a1..65c0c5b32351 100644
--- a/fs/nfs/read.c
+++ b/fs/nfs/read.c
@@ -63,7 +63,7 @@ struct nfs_read_data *nfs_readdata_alloc(unsigned int pagecount)
         return p;
 }
 
-void nfs_readdata_free(struct nfs_read_data *p)
+static void nfs_readdata_free(struct nfs_read_data *p)
 {
         if (p && (p->pagevec != &p->page_array[0]))
                 kfree(p->pagevec);
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 86bac6a5008e..50774991f8d5 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -137,7 +137,7 @@ struct nfs_write_data *nfs_writedata_alloc(unsigned int pagecount)
         return p;
 }
 
-void nfs_writedata_free(struct nfs_write_data *p)
+static void nfs_writedata_free(struct nfs_write_data *p)
 {
         if (p && (p->pagevec != &p->page_array[0]))
                 kfree(p->pagevec);
diff --git a/include/linux/lockd/lockd.h b/include/linux/lockd/lockd.h
index aa4fe905bb4d..0d92c468d55a 100644
--- a/include/linux/lockd/lockd.h
+++ b/include/linux/lockd/lockd.h
@@ -123,7 +123,6 @@ struct nlm_block {
         unsigned int            b_id;           /* block id */
         unsigned char           b_queued;       /* re-queued */
         unsigned char           b_granted;      /* VFS granted lock */
-        unsigned char           b_done;         /* callback complete */
         struct nlm_file *       b_file;         /* file in question */
 };
 
diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h
index 55ea853d57bc..247434553ae8 100644
--- a/include/linux/nfs_fs.h
+++ b/include/linux/nfs_fs.h
@@ -476,10 +476,9 @@ static inline int nfs_wb_page(struct inode *inode, struct page* page)
 }
 
 /*
- * Allocate and free nfs_write_data structures
+ * Allocate nfs_write_data structures
  */
 extern struct nfs_write_data *nfs_writedata_alloc(unsigned int pagecount);
-extern void nfs_writedata_free(struct nfs_write_data *p);
 
 /*
  * linux/fs/nfs/read.c
@@ -491,10 +490,9 @@ extern int  nfs_readpage_result(struct rpc_task *, struct nfs_read_data *);
 extern void nfs_readdata_release(void *data);
 
 /*
- * Allocate and free nfs_read_data structures
+ * Allocate nfs_read_data structures
  */
 extern struct nfs_read_data *nfs_readdata_alloc(unsigned int pagecount);
-extern void nfs_readdata_free(struct nfs_read_data *p);
 
 /*
  * linux/fs/nfs3proc.c
diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h
index e8bbe8118de8..840e47a4ccc5 100644
--- a/include/linux/sunrpc/xprt.h
+++ b/include/linux/sunrpc/xprt.h
@@ -229,7 +229,7 @@ int			xprt_reserve_xprt(struct rpc_task *task);
 int                     xprt_reserve_xprt_cong(struct rpc_task *task);
 int                     xprt_prepare_transmit(struct rpc_task *task);
 void                    xprt_transmit(struct rpc_task *task);
-void                    xprt_abort_transmit(struct rpc_task *task);
+void                    xprt_end_transmit(struct rpc_task *task);
 int                     xprt_adjust_timeout(struct rpc_rqst *req);
 void                    xprt_release_xprt(struct rpc_xprt *xprt, struct rpc_task *task);
 void                    xprt_release_xprt_cong(struct rpc_xprt *xprt, struct rpc_task *task);
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index 4ba271f892c8..d6409e757219 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -921,26 +921,43 @@ call_transmit(struct rpc_task *task)
         task->tk_status = xprt_prepare_transmit(task);
         if (task->tk_status != 0)
                 return;
+        task->tk_action = call_transmit_status;
         /* Encode here so that rpcsec_gss can use correct sequence number. */
         if (rpc_task_need_encode(task)) {
-                task->tk_rqstp->rq_bytes_sent = 0;
+                BUG_ON(task->tk_rqstp->rq_bytes_sent != 0);
                 call_encode(task);
                 /* Did the encode result in an error condition? */
                 if (task->tk_status != 0)
-                        goto out_nosend;
+                        return;
         }
-        task->tk_action = call_transmit_status;
         xprt_transmit(task);
         if (task->tk_status < 0)
                 return;
-        if (!task->tk_msg.rpc_proc->p_decode) {
-                task->tk_action = rpc_exit_task;
-                rpc_wake_up_task(task);
-        }
-        return;
-out_nosend:
-        /* release socket write lock before attempting to handle error */
-        xprt_abort_transmit(task);
+        /*
+         * On success, ensure that we call xprt_end_transmit() before sleeping
+         * in order to allow access to the socket to other RPC requests.
+         */
+        call_transmit_status(task);
+        if (task->tk_msg.rpc_proc->p_decode != NULL)
+                return;
+        task->tk_action = rpc_exit_task;
+        rpc_wake_up_task(task);
+}
+
+/*
+ * 5a.  Handle cleanup after a transmission
+ */
+static void
+call_transmit_status(struct rpc_task *task)
+{
+        task->tk_action = call_status;
+        /*
+         * Special case: if we've been waiting on the socket's write_space()
+         * callback, then don't call xprt_end_transmit().
+         */
+        if (task->tk_status == -EAGAIN)
+                return;
+        xprt_end_transmit(task);
         rpc_task_force_reencode(task);
 }
 
@@ -992,18 +1009,7 @@ call_status(struct rpc_task *task)
 }
 
 /*
- * 6a.  Handle transmission errors.
- */
-static void
-call_transmit_status(struct rpc_task *task)
-{
-        if (task->tk_status != -EAGAIN)
-                rpc_task_force_reencode(task);
-        call_status(task);
-}
-
-/*
- * 6b.  Handle RPC timeout
+ * 6a.  Handle RPC timeout
  *      We do not release the request slot, so we keep using the
  *      same XID for all retransmits.
  */
diff --git a/net/sunrpc/rpc_pipe.c b/net/sunrpc/rpc_pipe.c
index dc6cb93c8830..a3bd2db2e024 100644
--- a/net/sunrpc/rpc_pipe.c
+++ b/net/sunrpc/rpc_pipe.c
@@ -667,10 +667,11 @@ rpc_mkdir(char *path, struct rpc_clnt *rpc_client)
                         RPCAUTH_info, RPCAUTH_EOF);
         if (error)
                 goto err_depopulate;
+        dget(dentry);
 out:
         mutex_unlock(&dir->i_mutex);
         rpc_release_path(&nd);
-        return dget(dentry);
+        return dentry;
 err_depopulate:
         rpc_depopulate(dentry);
         __rpc_rmdir(dir, dentry);
@@ -731,10 +732,11 @@ rpc_mkpipe(char *path, void *private, struct rpc_pipe_ops *ops, int flags)
         rpci->flags = flags;
         rpci->ops = ops;
         inode_dir_notify(dir, DN_CREATE);
+        dget(dentry);
 out:
         mutex_unlock(&dir->i_mutex);
         rpc_release_path(&nd);
-        return dget(dentry);
+        return dentry;
 err_dput:
         dput(dentry);
         dentry = ERR_PTR(-ENOMEM);
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index 313b68d892c6..e8c2bc4977f3 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -707,12 +707,9 @@ out_unlock:
         return err;
 }
 
-void
-xprt_abort_transmit(struct rpc_task *task)
+void xprt_end_transmit(struct rpc_task *task)
 {
-        struct rpc_xprt *xprt = task->tk_xprt;
-
-        xprt_release_write(xprt, task);
+        xprt_release_write(task->tk_xprt, task);
 }
 
 /**
@@ -761,8 +758,6 @@ void xprt_transmit(struct rpc_task *task)
                         task->tk_status = -ENOTCONN;
                 else if (!req->rq_received)
                         rpc_sleep_on(&xprt->pending, task, NULL, xprt_timer);
-
-                xprt->ops->release_xprt(xprt, task);
                 spin_unlock_bh(&xprt->transport_lock);
                 return;
         }
@@ -772,18 +767,8 @@ void xprt_transmit(struct rpc_task *task)
          *       schedq, and being picked up by a parallel run of rpciod().
          */
         task->tk_status = status;
-
-        switch (status) {
-        case -ECONNREFUSED:
+        if (status == -ECONNREFUSED)
                 rpc_sleep_on(&xprt->sending, task, NULL, NULL);
-        case -EAGAIN:
-        case -ENOTCONN:
-                return;
-        default:
-                break;
-        }
-        xprt_release_write(xprt, task);
-        return;
 }
 
 static inline void do_xprt_reserve(struct rpc_task *task)
diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
index ee678ed13b6f..441bd53f5eca 100644
--- a/net/sunrpc/xprtsock.c
+++ b/net/sunrpc/xprtsock.c
@@ -414,6 +414,33 @@ static int xs_tcp_send_request(struct rpc_task *task)
 }
 
 /**
+ * xs_tcp_release_xprt - clean up after a tcp transmission
+ * @xprt: transport
+ * @task: rpc task
+ *
+ * This cleans up if an error causes us to abort the transmission of a request.
+ * In this case, the socket may need to be reset in order to avoid confusing
+ * the server.
+ */
+static void xs_tcp_release_xprt(struct rpc_xprt *xprt, struct rpc_task *task)
+{
+        struct rpc_rqst *req;
+
+        if (task != xprt->snd_task)
+                return;
+        if (task == NULL)
+                goto out_release;
+        req = task->tk_rqstp;
+        if (req->rq_bytes_sent == 0)
+                goto out_release;
+        if (req->rq_bytes_sent == req->rq_snd_buf.len)
+                goto out_release;
+        set_bit(XPRT_CLOSE_WAIT, &task->tk_xprt->state);
+out_release:
+        xprt_release_xprt(xprt, task);
+}
+
+/**
  * xs_close - close a socket
  * @xprt: transport
  *
@@ -1250,7 +1277,7 @@ static struct rpc_xprt_ops xs_udp_ops = {
 
 static struct rpc_xprt_ops xs_tcp_ops = {
         .reserve_xprt           = xprt_reserve_xprt,
-        .release_xprt           = xprt_release_xprt,
+        .release_xprt           = xs_tcp_release_xprt,
         .set_port               = xs_set_port,
         .connect                = xs_connect,
         .buf_alloc              = rpc_malloc,