[0/7] Convert UDF to folios

Message

Matthew Wilcox April 17, 2024, 3:04 p.m. UTC

I'm not making any attempt here to support large folios.  This is just to
remove uses of the page-based APIs.  Most of these places are for inline
data or symlinks, so it wouldn't be appropriate to use large folios
(unless we want to support bs>PS, which seems to be permitted by UDF,
although not widely supported).

Matthew Wilcox (Oracle) (7):
  udf: Convert udf_symlink_filler() to use a folio
  udf: Convert udf_write_begin() to use a folio
  udf: Convert udf_expand_file_adinicb() to use a folio
  udf: Convert udf_adinicb_readpage() to udf_adinicb_read_folio()
  udf: Convert udf_symlink_getattr() to use a folio
  udf: Convert udf_page_mkwrite() to use a folio
  udf: Use a folio in udf_write_end()

 fs/udf/file.c    | 20 +++++++--------
 fs/udf/inode.c   | 65 ++++++++++++++++++++++++------------------------
 fs/udf/symlink.c | 34 +++++++++----------------
 3 files changed, 54 insertions(+), 65 deletions(-)

Comments

Jan Kara April 18, 2024, 10:53 a.m. UTC | #1

On Wed 17-04-24 16:04:06, Matthew Wilcox (Oracle) wrote:
> I'm not making any attempt here to support large folios.  This is just to
> remove uses of the page-based APIs.  Most of these places are for inline
> data or symlinks, so it wouldn't be appropriate to use large folios
> (unless we want to support bs>PS, which seems to be permitted by UDF,
> although not widely supported).

Thanks for the conversion. Overall it looks good and I can fixup the minor
stuff I've found on commit, just I'd like to get a confirmation regarding
the flush_dcache_folio() thing...

								Honza

> Matthew Wilcox (Oracle) (7):
>   udf: Convert udf_symlink_filler() to use a folio
>   udf: Convert udf_write_begin() to use a folio
>   udf: Convert udf_expand_file_adinicb() to use a folio
>   udf: Convert udf_adinicb_readpage() to udf_adinicb_read_folio()
>   udf: Convert udf_symlink_getattr() to use a folio
>   udf: Convert udf_page_mkwrite() to use a folio
>   udf: Use a folio in udf_write_end()
> 
>  fs/udf/file.c    | 20 +++++++--------
>  fs/udf/inode.c   | 65 ++++++++++++++++++++++++------------------------
>  fs/udf/symlink.c | 34 +++++++++----------------
>  3 files changed, 54 insertions(+), 65 deletions(-)
> 
> -- 
> 2.43.0
>

Jan Kara April 23, 2024, 1:44 p.m. UTC | #2

On Thu 18-04-24 12:53:23, Jan Kara wrote:
> On Wed 17-04-24 16:04:06, Matthew Wilcox (Oracle) wrote:
> > I'm not making any attempt here to support large folios.  This is just to
> > remove uses of the page-based APIs.  Most of these places are for inline
> > data or symlinks, so it wouldn't be appropriate to use large folios
> > (unless we want to support bs>PS, which seems to be permitted by UDF,
> > although not widely supported).
> 
> Thanks for the conversion. Overall it looks good and I can fixup the minor
> stuff I've found on commit, just I'd like to get a confirmation regarding
> the flush_dcache_folio() thing...

OK, I've merged the patches into my tree.

								Honza