Re: [PATCH 9/9] readahead: record mmap read-around states in file_ra_state

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

On Fri, 10 Apr 2009 14:10:06 +0800
Wu Fengguang <fengguang.wu@intel.com> wrote:

> Mmap read-around now shares the same code style and data structure
> with readahead code.
> 
> This also removes do_page_cache_readahead().
> Its last user, mmap read-around, has been changed to call ra_submit().
> 
> The no-readahead-if-congested logic is dumped by the way.
> Users will be pretty sensitive about the slow loading of executables.
> So it's unfavorable to disabled mmap read-around on a congested queue.

Did you verify that the read-congested code ever triggers?

It used to be (and probably still is) the case that
bdi_read_congested() is very very rare, because the read queue is long
and the kernel rarely puts many read requests into it.  You can of
course create this condition with a fake workload with may
threads/processes, but it _is_ fake.

Some real-world workloads (databases?) will of course trigger
bdi_read_congested().  But they're usually doing fixed-sized reads, and
if we're doing _any_ readahead/readaround in that case, readahead is
busted.