On Thu, Aug 12, 2021 at 12:57:41PM +0100, David Howells wrote:

Hi Willy, Trond,

Here's a change to make reads from the swapfile use async DIO rather than
readpage(), as requested by Willy.

Whilst trying to make this work, I found that NFS's support for swapfiles
seems to have been non-functional since Aug 2019 (I think), so the first
patch fixes that.  Question is: do we actually *want* to keep this
functionality, given that it seems that no one's tested it with an upstream
kernel in the last couple of years?

Independ of the NFS use using the direct I/O code for swap seems like
the right thing to do in generlal.  e.g. for XFS a lookup in the extent
btree will be more efficient than the weird swap extent map.