Re: [PATCH] mm: consider disabling readahead if there are signs of thrashing

[Roman Gushchin <roman.gushchin@linux.dev>]

Matthew Wilcox <willy@infradead.org> writes:

> On Thu, Jul 10, 2025 at 12:52:32PM -0700, Roman Gushchin wrote:
>> We've noticed in production that under a very heavy memory pressure
>> the readahead behavior becomes unstable causing spikes in memory
>> pressure and CPU contention on zone locks.
>> 
>> The current mmap_miss heuristics considers minor pagefaults as a
>> good reason to decrease mmap_miss and conditionally start async
>> readahead. This creates a vicious cycle: asynchronous readahead
>> loads more pages, which in turn causes more minor pagefaults.
>
> Is the correct response to turn off faultaround, or would we be better
> off scaling it down (eg as low as 64k)?

I think at this point it better to turn it off entirely.

For scaling I wonder if we want to scale it depending on PSI numbers?

>
> I like the signal you're using; I think that makes a lot of sense.

Thanks!