On 9/15/23 23:51, Matthew Wilcox wrote:

On Fri, Sep 15, 2023 at 02:32:44PM -0700, Luis Chamberlain wrote:

quoted

However, an issue is that disabling CONFIG_BUFFER_HEAD in practice is not viable
for many Linux distributions since it also means disabling support for most
filesystems other than btrfs and XFS. So we either support larger order folios
on buffer-heads, or we draw up a solution to enable co-existence. Since at LSFMM
2023 it was decided we would not support larger order folios on buffer-heads,

Um, I didn't agree to that.  If block size is equal to folio size, there
are no problems supporting one buffer head per folio.  In fact, we could
probably go up to 8 buffer heads per folio without any trouble (but I'm
not signing up to do that work).

Entirely correct.
I have a patchset ready for doing just that (ie having one buffer head 
per folio); hope I'll find some time next week to get it cleaned up and 
posted.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                Kernel Storage Architect
hare@suse.de                              +49 911 74053 688
SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg
HRB 36809 (AG Nürnberg), Geschäftsführer: Ivo Totev, Andrew
Myers, Andrew McDonald, Martje Boudien Moerman