Re: [PATCH 15/17] mm/filemap: Don't relock the page after calling readpage

[Christoph Hellwig <hch@xxxxxx>]

On Mon, Nov 02, 2020 at 06:43:10PM +0000, Matthew Wilcox (Oracle) wrote:
> We don't need to get the page lock again; we just need to wait for
> the I/O to finish, so use wait_on_page_locked_killable() like the
> other callers of ->readpage.

As that isn't entirely obvious, what about adding a comment next to
the wait_on_page_locked_killable call to document it?

>  
> +	error = wait_on_page_locked_killable(page);
>  	if (error)
>  		return error;
> +	if (PageUptodate(page))
> +		return 0;
> +	if (!page->mapping)	/* page truncated */
> +		return AOP_TRUNCATED_PAGE;
> +	shrink_readahead_size_eio(&file->f_ra);
> +	return -EIO;

You might notice a theme here, but I hate having the fast path exit
early without a good reason, so I'd be much happier with:

	if (!PageUptodate(page)) {
		if (!page->mapping)	/* page truncated */
			return AOP_TRUNCATED_PAGE;
		shrink_readahead_size_eio(&file->f_ra);
		return -EIO;
	}
	return 0;