Re: [PATCH v20 03/32] splice: Make direct_read_splice() limit to eof where appropriate

[David Howells]

Okay.  Let's go with that.  So I have to put the handling in vfs_splice_read():

	long vfs_splice_read(struct file *in, loff_t *ppos,
			     struct pipe_inode_info *pipe, size_t len,
			     unsigned int flags)
	{
	...
		if (unlikely(!in->f_op->splice_read))
			return warn_unsupported(in, "read");
		/*
		 * O_DIRECT and DAX don't deal with the pagecache, so we
		 * allocate a buffer, copy into it and splice that into the pipe.
		 */
		if ((in->f_flags & O_DIRECT) || IS_DAX(in->f_mapping->host))
			return copy_splice_read(in, ppos, pipe, len, flags);
		return in->f_op->splice_read(in, ppos, pipe, len, flags);
	}

which leaves very little in generic_file_splice_read:

	ssize_t generic_file_splice_read(struct file *in, loff_t *ppos,
					 struct pipe_inode_info *pipe, size_t len,
					 unsigned int flags)
	{
		if (unlikely(*ppos >= in->f_mapping->host->i_sb->s_maxbytes))
			return 0;
		if (unlikely(!len))
			return 0;
		return filemap_splice_read(in, ppos, pipe, len, flags);
	}

so I wonder if the tests in generic_file_splice_read() can be folded into
vfs_splice_read(), pointers to generic_file_splice_read() be replaced with
pointers to filemap_splice_read() and generic_file_splice_read() just be
removed.

I suspect we can't quite do this because of the *ppos check - but I wonder if
that's actually necessary since filemap_splice_read() checks against
i_size... or if the check can be moved there if we definitely want to do it.

Certainly, the zero-length check can be done in vfs_splice_read().

David