On Tue, 2011-12-13 at 11:34 +0800, Miao Xie wrote:
> On thu, 8 Dec 2011 10:04:34 -0800, Kamal Mostafa wrote:
> >  /**
> > - * writeback_inodes_sb_if_idle	-	start writeback if none underway
> > + * try_to_writeback_inodes_sb_nr	-	start writeback if none underway
> >   * @sb: the superblock
> >   * @nr: the number of pages to write
> >   *
> >   * Invoke writeback_inodes_sb if no writeback is currently underway.
> >   * Returns 1 if writeback was started, 0 if not.
> >   */
> > -int writeback_inodes_sb_nr_if_idle(struct super_block *sb,
> > +int try_to_writeback_inodes_sb_nr(struct super_block *sb,
> >  				   unsigned long nr,
> >  				   enum wb_reason reason)
> >  {
> >  	if (!writeback_in_progress(sb->s_bdi)) {
> >  		down_read(&sb->s_umount);
> > -		writeback_inodes_sb_nr(sb, nr, reason);
> > +		if (nr == 0)
> > +			writeback_inodes_sb(sb, reason);
> > +		else
> > +			writeback_inodes_sb_nr(sb, nr, reason);
> >  		up_read(&sb->s_umount);
> >  		return 1;
> >  	} else
> >  		return 0;
> 
> The comment said "Returns 1 if writeback was started", so if writeback_in_progress()
> return true, I think this function also should return 1.

My interpretation of that comment is that it will return 1 only if this
call results in a new writeback being started (not if a writeback was
already in progress).

This patch [4/7] intentionally does not introduce any functional
changes.

> BTW: Does anyone know when this patchset will be merged into the main tree?

I also eagerly await the merge of this patch set.

 -Kamal