[Linux-cachefs] Can the GFP flags to releasepage() be trusted? -- was Re: [PATCH v2 3/8] nfs: Move to using the alternate fallback fscache I/O API

[David Howells]

Trond Myklebust <trondmy at hammerspace.com> wrote:

> To elaborate a bit: we used to have code here that would check whether
> the page had been cleaned but was unstable, and if an argument of
> GFP_KERNEL or above was set, we'd try to call COMMIT to ensure the page
> was synched to disk on the server (and we'd wait for that call to
> complete).
> 
> That code would end up deadlocking in all sorts of horrible ways, so we
> ended up having to pull it.

I don't think that a deadlock should be possible with this.  PG_fscache is now
only being used to indicate that a DIO write to the cache is in progress on
the page.  It will complete and remove the mark at some point.

David