[PATCH v2] iomap: Support inline data with block size < page size

[Gao Xiang]

Hi Matthew,

On Thu, Jul 29, 2021 at 01:43:38PM +0100, Matthew Wilcox wrote:
> On Thu, Jul 29, 2021 at 05:54:56AM +0200, Andreas Gruenbacher wrote:
> > > -       /* inline data must start page aligned in the file */
> > > -       if (WARN_ON_ONCE(offset_in_page(iomap->offset)))
> > > -               return -EIO;
> > 
> > Maybe add a WARN_ON_ONCE(size > PAGE_SIZE - poff) here?
> 
> Sure!
> 
> > >         if (WARN_ON_ONCE(size > PAGE_SIZE -
> > >                          offset_in_page(iomap->inline_data)))
> > >                 return -EIO;
> > >         if (WARN_ON_ONCE(size > iomap->length))
> > >                 return -EIO;
> > > -       if (WARN_ON_ONCE(page_has_private(page)))
> > > -               return -EIO;
> > > +       if (poff > 0)
> > > +               iomap_page_create(inode, page);
> > >
> > > -       addr = kmap_atomic(page);
> > > +       addr = kmap_atomic(page) + poff;
> > 
> > Maybe kmap_local_page?
> 
> Heh, I do that later when I convert to folios (there is no
> kmap_atomic_folio(), only kmap_local_folio()).  But I can throw that
> in here too.

I don't find any critical point with this patch (and agree with Andreas'
suggestion), maybe some followup folio work could get more input about
this. I'll evaluate them all together later.

Thanks,
Gao Xiang