On Thu 15-12-22 21:43:50, Matthew Wilcox (Oracle) wrote: > I was hoping that filesystems would convert from buffer_heads to iomap, > but that's not happening particularly quickly. So the buffer_head > infrastructure needs to be converted from being page-based to being > folio-based. This is the initial patchset that I hope Andrew will take > for 6.3. I have a lot of followup patches, but many of them should go > through individual filesystem trees (ext4, f2fs, etc). They can wait > for 6.4. FWIW I went through all the patches and they look fine to me so feel free to add: Reviewed-by: Jan Kara <jack@xxxxxxx> Honza > > Matthew Wilcox (Oracle) (12): > buffer: Add b_folio as an alias of b_page > buffer: Replace obvious uses of b_page with b_folio > buffer: Use b_folio in touch_buffer() > buffer: Use b_folio in end_buffer_async_read() > buffer: Use b_folio in end_buffer_async_write() > page_io: Remove buffer_head include > buffer: Use b_folio in mark_buffer_dirty() > gfs2: Replace obvious uses of b_page with b_folio > jbd2: Replace obvious uses of b_page with b_folio > nilfs2: Replace obvious uses of b_page with b_folio > reiserfs: Replace obvious uses of b_page with b_folio > mpage: Use b_folio in do_mpage_readpage() > > fs/buffer.c | 54 +++++++++++++++++------------------ > fs/gfs2/glops.c | 2 +- > fs/gfs2/log.c | 2 +- > fs/gfs2/meta_io.c | 2 +- > fs/jbd2/commit.c | 8 ++---- > fs/jbd2/journal.c | 2 +- > fs/mpage.c | 2 +- > fs/nilfs2/btnode.c | 2 +- > fs/nilfs2/btree.c | 2 +- > fs/nilfs2/gcinode.c | 2 +- > fs/nilfs2/mdt.c | 4 +-- > fs/nilfs2/segment.c | 2 +- > fs/reiserfs/journal.c | 4 +-- > fs/reiserfs/tail_conversion.c | 2 +- > include/linux/buffer_head.h | 5 +++- > mm/page_io.c | 1 - > 16 files changed, 47 insertions(+), 49 deletions(-) > > -- > 2.35.1 > -- Jan Kara <jack@xxxxxxxx> SUSE Labs, CR