Mark generic HDLC + PPP as broken.
[sfrench/cifs-2.6.git] / fs / fs-writeback.c
index 0fca82021d7652df2d2fc8e1b1f33204e8712651..06557679ca411d6fbce09a33dddd042f938ecbc2 100644 (file)
@@ -284,7 +284,17 @@ __sync_single_inode(struct inode *inode, struct writeback_control *wbc)
                                 * soon as the queue becomes uncongested.
                                 */
                                inode->i_state |= I_DIRTY_PAGES;
-                               requeue_io(inode);
+                               if (wbc->nr_to_write <= 0) {
+                                       /*
+                                        * slice used up: queue for next turn
+                                        */
+                                       requeue_io(inode);
+                               } else {
+                                       /*
+                                        * somehow blocked: retry later
+                                        */
+                                       redirty_tail(inode);
+                               }
                        } else {
                                /*
                                 * Otherwise fully redirty the inode so that
@@ -334,9 +344,6 @@ __writeback_single_inode(struct inode *inode, struct writeback_control *wbc)
                WARN_ON(inode->i_state & I_WILL_FREE);
 
        if ((wbc->sync_mode != WB_SYNC_ALL) && (inode->i_state & I_SYNC)) {
-               struct address_space *mapping = inode->i_mapping;
-               int ret;
-
                /*
                 * We're skipping this inode because it's locked, and we're not
                 * doing writeback-for-data-integrity.  Move it to s_more_io so
@@ -345,15 +352,7 @@ __writeback_single_inode(struct inode *inode, struct writeback_control *wbc)
                 * completed a full scan of s_io.
                 */
                requeue_io(inode);
-
-               /*
-                * Even if we don't actually write the inode itself here,
-                * we can at least start some of the data writeout..
-                */
-               spin_unlock(&inode_lock);
-               ret = do_writepages(mapping, wbc);
-               spin_lock(&inode_lock);
-               return ret;
+               return 0;
        }
 
        /*
@@ -479,11 +478,13 @@ sync_sb_inodes(struct super_block *sb, struct writeback_control *wbc)
                iput(inode);
                cond_resched();
                spin_lock(&inode_lock);
-               if (wbc->nr_to_write <= 0)
+               if (wbc->nr_to_write <= 0) {
+                       wbc->more_io = 1;
                        break;
+               }
+               if (!list_empty(&sb->s_more_io))
+                       wbc->more_io = 1;
        }
-       if (!list_empty(&sb->s_more_io))
-               wbc->more_io = 1;
        return;         /* Leave any unwritten inodes on s_io */
 }
 
@@ -514,8 +515,7 @@ writeback_inodes(struct writeback_control *wbc)
        might_sleep();
        spin_lock(&sb_lock);
 restart:
-       sb = sb_entry(super_blocks.prev);
-       for (; sb != sb_entry(&super_blocks); sb = sb_entry(sb->s_list.prev)) {
+       list_for_each_entry_reverse(sb, &super_blocks, s_list) {
                if (sb_has_dirty_inodes(sb)) {
                        /* we're making our own get_super here */
                        sb->s_count++;
@@ -580,10 +580,8 @@ static void set_sb_syncing(int val)
 {
        struct super_block *sb;
        spin_lock(&sb_lock);
-       sb = sb_entry(super_blocks.prev);
-       for (; sb != sb_entry(&super_blocks); sb = sb_entry(sb->s_list.prev)) {
+       list_for_each_entry_reverse(sb, &super_blocks, s_list)
                sb->s_syncing = val;
-       }
        spin_unlock(&sb_lock);
 }
 
@@ -657,7 +655,7 @@ int write_inode_now(struct inode *inode, int sync)
        int ret;
        struct writeback_control wbc = {
                .nr_to_write = LONG_MAX,
-               .sync_mode = WB_SYNC_ALL,
+               .sync_mode = sync ? WB_SYNC_ALL : WB_SYNC_NONE,
                .range_start = 0,
                .range_end = LLONG_MAX,
        };
@@ -753,7 +751,7 @@ int generic_osync_inode(struct inode *inode, struct address_space *mapping, int
 EXPORT_SYMBOL(generic_osync_inode);
 
 /**
- * writeback_acquire: attempt to get exclusive writeback access to a device
+ * writeback_acquire - attempt to get exclusive writeback access to a device
  * @bdi: the device's backing_dev_info structure
  *
  * It is a waste of resources to have more than one pdflush thread blocked on
@@ -770,7 +768,7 @@ int writeback_acquire(struct backing_dev_info *bdi)
 }
 
 /**
- * writeback_in_progress: determine whether there is writeback in progress
+ * writeback_in_progress - determine whether there is writeback in progress
  * @bdi: the device's backing_dev_info structure.
  *
  * Determine whether there is writeback in progress against a backing device.
@@ -781,7 +779,7 @@ int writeback_in_progress(struct backing_dev_info *bdi)
 }
 
 /**
- * writeback_release: relinquish exclusive writeback access against a device.
+ * writeback_release - relinquish exclusive writeback access against a device.
  * @bdi: the device's backing_dev_info structure
  */
 void writeback_release(struct backing_dev_info *bdi)