[Linux-cachefs] hung task on page invalidate

[David Howells]

Milosz Tanski <milosz at adfin.com> wrote:

> In my test cluster I started seeing an issue when the fscache get
> stuck waiting on pending writes when doing page invalidate. The
> problem happens because it looks like the page never leaves the
> cookie->store page tree.
> 
> I've been reading the different code paths in fscache/page.c to
> understand why this could be the case, but on first look it looks like
> it does the correct thing. Do you have any ideas what could be causing
> this / where to look for the smoking gun.
> 
> I've only begun to see this recently, I think our work load on the
> test cluster changed a bit and it's been showing up more.

The most likely answer is that there's a race whereby a page is added to the
store tree, but the write-out operation isn't rescheduled.

David