Re: [PATCH 2/5] buffer: Add kernel-doc for block_dirty_folio()

[Matthew Wilcox]

On Mon, Jan 08, 2024 at 01:35:10PM +0000, Matthew Wilcox wrote:
> On Mon, Jan 08, 2024 at 02:31:17PM +0100, Pankaj Raghav (Samsung) wrote:
> > > + * If the folio has buffers, the uptodate buffers are set dirty, to
> > > + * preserve dirty-state coherency between the folio and the buffers.
> > > + * It the folio does not have buffers then when they are later attached
> > 
> > s/It the folio/If the folio
> > > + * they will all be set dirty.
> > Is it better to rephrase it slightly as follows:
> > 
> > If the folio does not have buffers, they will all be set dirty when they
> > are later attached.
> 
> Yes, I like that better.

Actually, how about:

 * If the folio has buffers, the uptodate buffers are set dirty, to
 * preserve dirty-state coherency between the folio and the buffers.
 * Buffers added to a dirty folio are created dirty.

I considered deleting the sentence entirely as it's not actually related
to what the function does; it's just a note about how the buffer cache
behaves.  That said, information about how buffer heds work is scant
enough that I don't want to delete it.