Re: [PATCH] mm: readahead: fix do_readahead for no readpage(s)

[Andrew Morton <akpm@linux-foundation.org>]

On Tue, 28 Jan 2014 11:14:19 +0000 Mark Rutland <mark.rutland@arm.com> wrote:

> Commit 63d0f0a3c7e1 (mm/readahead.c:do_readhead(): don't check for
> ->readpage) unintentionally made do_readahead return 0 for all valid
> files regardless of whether readahead was supported, rather than the
> expected -EINVAL. This gets forwarded on to userspace, and results in
> sys_readahead appearing to succeed in cases that don't make sense (e.g.
> when called on pipes or sockets). This issue is detected by the LTP
> readahead01 testcase.

How can this be?

: static ssize_t
: do_readahead(struct address_space *mapping, struct file *filp,
: 	     pgoff_t index, unsigned long nr)
: {
: 	if (!mapping || !mapping->a_ops)
: 		return -EINVAL;
: 
: 	return force_page_cache_readahead(mapping, filp, index, nr);
: }

and

: int force_page_cache_readahead(struct address_space *mapping, struct file *filp,
: 		pgoff_t offset, unsigned long nr_to_read)
: {
: 	if (unlikely(!mapping->a_ops->readpage && !mapping->a_ops->readpages))
: 		return -EINVAL;

Clearly, do_readahead() will return -EINVAL if neither ->readpage or
->readpages are implemented.

I can see that the behaviour would change if the address_space
implements only one of ->readpage and ->readpages, but that doesn't
appear to match your description and the new behaviour is correct - we
can now perform readahead for address_spaces which implement
->readpages and not ->readpage (which would be odd and might not work
for other reasons..).