[git pull] vfs.git iov_iter

[git pull] vfs.git iov_iter

[Al Viro]

	A couple of iov_iter patches - Christoph's crapectomy (the last
remaining user of iov_for_each() went away with lustre, IIRC) and Eric'c
optimization of sanity checks.

The following changes since commit bfeffd155283772bbe78c6a05dec7c0128ee500c:

  Linux 5.0-rc1 (2019-01-06 17:08:20 -0800)

are available in the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs.git work.iov_iter

for you to fetch changes up to 6daef95b8c914866a46247232a048447fff97279:

  iov_iter: optimize page_copy_sane() (2019-02-26 14:05:20 -0500)

----------------------------------------------------------------
Christoph Hellwig (1):
      uio: remove the unused iov_for_each macro

Eric Dumazet (1):
      iov_iter: optimize page_copy_sane()

 .clang-format       |  1 -
 include/linux/uio.h |  8 --------
 lib/iov_iter.c      | 17 +++++++++++++++--
 3 files changed, 15 insertions(+), 11 deletions(-)

Re: [git pull] vfs.git iov_iter

[Linus Torvalds]

On Sun, Mar 10, 2019 at 5:41 PM Al Viro <viro@zeniv.linux.org.uk> wrote:
>
>         A couple of iov_iter patches - Christoph's crapectomy (the last
> remaining user of iov_for_each() went away with lustre, IIRC) and Eric'c
> optimization of sanity checks.

I certainly heartily approve of making page_copy_sane() be less insane
than it used to be.

That said - has the warning that it has triggered lately?

If that sanity check still shows up in profiles as a performance
problem, maybe the whole thing could at least be made dependent on
some sanity checking option to get rid of it entirely for the common
case?

                   Linus

Re: [git pull] vfs.git iov_iter

[pr-tracker-bot]

The pull request you sent on Mon, 11 Mar 2019 00:41:46 +0000:

> git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs.git work.iov_iter

has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/dbc2fba3fc46084f502aec53183995a632998dcd

Thank you!

-- 
Deet-doot-dot, I am a bot.
https://korg.wiki.kernel.org/userdoc/prtracker

[git pull] vfs.git iov_iter

[Al Viro]

	* bio_{map,copy}_user_iov() series; those are cleanups - fixes from
the same pile went into mainline (and stable) in late September.
	* fs/iomap.c iov_iter-related fixes
	* new primitive - iov_iter_for_each_range(); apply a function to
kernel-mapped segments of an iov_iter.  Usable for kvec and bvec ones,
the latter does kmap()/kunmap() around the callback.  _Not_ usable for
iovec- or pipe-backed iov_iter; the latter is not hard to fix if the need
ever appears, the former is by design.  Another related primitive will
have to wait for the next cycle - it passes page + offset + size instead
of pointer + size, and that one will be usable for everything _except_
kvec.  Unfortunately, that one didn't get exposure in -next yet, so...
	* a bit more lustre iov_iter work, including a use case for
iov_iter_for_each_range() (checksum calculation).
	* vhost/scsi leak fix in failure exit.
	* misc cleanups and detritectomy...

The following changes since commit 1cfd0ddd82232804e03f3023f6a58b50dfef0574:

  bio_copy_user_iov(): don't ignore ->iov_offset (2017-10-10 23:55:14 -0400)

are available in the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs.git work.iov_iter

for you to fetch changes up to cfe057f7db1ff026c8db75469a3f9ba9736e1975:

  iomap_dio_actor(): fix iov_iter bugs (2017-10-11 22:38:02 -0400)

----------------------------------------------------------------
Al Viro (21):
      bio_map_user_iov(): switch to iov_iter_get_pages()/iov_iter_advance()
      ... and with iov_iter_get_pages_alloc() it becomes even simpler
      don't rely upon subsequent bio_add_pc_page() calls failing
      bio_map_user_iov(): move alignment check into the main loop
      bio_map_user_iov(): get rid of the iov_for_each()
      blk_rq_map_user_iov(): move iov_iter_advance() down
      move more stuff down into bio_copy_user_iov()
      bio_copy_from_iter(): get rid of copying iov_iter
      bio_map_user_iov(): get rid of copying iov_iter
      bio_copy_user_iov(): saner bio size calculation
      bio_alloc_map_data(): do bmd->iter setup right there
      kill iov_shorten()
      orangefs: remove detritus from struct orangefs_kiocb_s
      xen: don't open-code iov_iter_kvec()
      lnet_return_rx_credits_locked: don't abuse list_entry
      new primitive: iov_iter_for_each_range()
      fix a page leak in vhost_scsi_iov_to_sgl() error recovery
      vhost/scsi: switch to iov_iter_get_pages()
      lustre: switch struct ksock_conn to iov_iter
      switch ksocknal_lib_recv_...() to use of iov_iter_for_each_range()
      iomap_dio_actor(): fix iov_iter bugs

 block/bio.c                                        | 192 ++++++++-------------
 block/blk-map.c                                    |   7 -
 .../staging/lustre/lnet/klnds/socklnd/socklnd.c    |   4 +-
 .../staging/lustre/lnet/klnds/socklnd/socklnd.h    |   9 +-
 .../staging/lustre/lnet/klnds/socklnd/socklnd_cb.c | 157 ++++-------------
 .../lustre/lnet/klnds/socklnd/socklnd_lib.c        |  99 +++--------
 drivers/staging/lustre/lnet/lnet/lib-move.c        |   2 +-
 drivers/vhost/scsi.c                               |  73 +++-----
 drivers/xen/pvcalls-back.c                         |  16 +-
 fs/iomap.c                                         |  24 ++-
 fs/orangefs/orangefs-kernel.h                      |   6 -
 fs/read_write.c                                    |  21 ---
 include/linux/bio.h                                |   4 +-
 include/linux/uio.h                                |   6 +-
 lib/iov_iter.c                                     |  22 +++
 15 files changed, 202 insertions(+), 440 deletions(-)