From:   David Howells <dhowells@redhat.com>
To:     Jens Axboe <axboe@kernel.dk>, Al Viro <viro@zeniv.linux.org.uk>,
        Christoph Hellwig <hch@infradead.org>
Cc:     David Howells <dhowells@redhat.com>,
        Matthew Wilcox <willy@infradead.org>, Jan Kara <jack@suse.cz>,
        Jeff Layton <jlayton@kernel.org>,
        David Hildenbrand <david@redhat.com>,
        Jason Gunthorpe <jgg@nvidia.com>,
        Logan Gunthorpe <logang@deltatee.com>,
        Hillf Danton <hdanton@sina.com>, linux-fsdevel@vger.kernel.org,
        linux-block@vger.kernel.org, linux-kernel@vger.kernel.org,
        linux-mm@kvack.org
Subject: [PATCH v3 0/5] iov_iter: Adjust styling/location of new splice functions
Date:   Tue, 14 Feb 2023 08:37:05 +0000
Message-Id: <20230214083710.2547248-1-dhowells@redhat.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

Hi Jens, Al, Christoph,

Here are patches to make some changes that Christoph requested[1] to the
new generic file splice functions that I implemented[2].  Apart from one
functional change, they just altering the styling and move one of the
functions to a different file:

 (1) Rename the main functions:

	generic_file_buffered_splice_read() -> filemap_splice_read()
	generic_file_direct_splice_read()   -> direct_splice_read()

 (2) Abstract out the calculation of the location of the head pipe buffer
     into a helper function in linux/pipe_fs_i.h.

 (3) Use init_sync_kiocb() in filemap_splice_read().

     This is where the functional change is.  Some kiocb fields are then
     filled in where they were set to 0 before, including setting ki_flags
     from f_iocb_flags.

 (4) Move filemap_splice_read() to mm/filemap.c.  filemap_get_pages() can
     then be made static again.

 (5) Fix splice-read for a number of filesystems that don't provide a
     ->read_folio() op and for which filemap_get_pages() cannot be used.

I've pushed the patches here also:

	https://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs.git/log/?h=iov-extract-3

I've also updated worked the changes into the commits on my iov-extract
branch if that would be preferable, though that means Jens would need to
update his for-6.3/iov-extract again.

David

Link: https://lore.kernel.org/r/Y+n0n2UE8BQa/OwW@infradead.org/ [1]
Link: https://lore.kernel.org/r/20230207171305.3716974-1-dhowells@redhat.com/ [2]

Changes
=======
ver #3)
 - Fix filesystems/drivers that don't have ->read_folio().

ver #2)
 - Don't attempt to filter IOCB_* flags in filemap_splice_read().

Link: https://lore.kernel.org/r/20230213134619.2198965-1-dhowells@redhat.com/ # v1

David Howells (5):
  splice: Rename new splice functions
  splice: Provide pipe_head_buf() helper
  splice: Use init_sync_kiocb() in filemap_splice_read()
  splice: Move filemap_read_splice() to mm/filemap.c
  shmem, overlayfs, coda, tty, proc, kernfs, random: Fix splice-read

 drivers/char/random.c     |   4 +-
 drivers/tty/tty_io.c      |   4 +-
 fs/coda/file.c            |  36 +++++++++-
 fs/kernfs/file.c          |   2 +-
 fs/overlayfs/file.c       |  36 +++++++++-
 fs/proc/inode.c           |   4 +-
 fs/proc/proc_sysctl.c     |   2 +-
 fs/proc_namespace.c       |   6 +-
 fs/splice.c               | 146 ++------------------------------------
 include/linux/fs.h        |   6 ++
 include/linux/pagemap.h   |   2 -
 include/linux/pipe_fs_i.h |  20 ++++++
 mm/filemap.c              | 136 +++++++++++++++++++++++++++++++++--
 mm/internal.h             |   6 ++
 mm/shmem.c                | 124 +++++++++++++++++++++++++++++++-
 15 files changed, 373 insertions(+), 161 deletions(-)