On Mon, Jun 14, 2021 at 05:55:30PM +0200, Christoph Hellwig wrote:

On Mon, Jun 14, 2021 at 05:53:33PM +0200, Christoph Hellwig wrote:

quoted

On Mon, Jun 14, 2021 at 01:23:22PM +0100, Matthew Wilcox wrote:

quoted

i have a somewhat similar series in the works ...

https://git.infradead.org/users/willy/pagecache.git/commitdiff/1e7e8c2d82666b55690705d5bbe908e31d437edb
https://git.infradead.org/users/willy/pagecache.git/commitdiff/bf767a4969c0bc6735275ff7457a8082eef4c3fd

... the other patches rather depend on the folio work.

Yes, these looks useful to me as well.

And in fact I suspect the code in __set_page_dirty_no_writeback should
really be the default if no ->set_page_dirty is set up.  It is the
same code as the no-mapping case and really makes sense as the default
case..

yes; i was even wondering if that should be conditional on one of the
"no writeback" bits in mapping ...