iov_iter cleanups and fixes; there are followups, but this is what had sat in -next this cycle. IMO the macro forest in there became much thinner and easier to follow... The following changes since commit 6efb943b8616ec53a5e444193dccf1af9ad627b5: Linux 5.13-rc1 (2021-05-09 14:17:44 -0700) 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 6852df1266995c35b8621a95dcb7f91ca11ea409: csum_and_copy_to_pipe_iter(): leave handling of csum_state to caller (2021-06-10 11:45:25 -0400) ---------------------------------------------------------------- Al Viro (36): ntfs_copy_from_user_iter(): don't bother with copying iov_iter generic_perform_write()/iomap_write_actor(): saner logics for short copy fuse_fill_write_pages(): don't bother with iov_iter_single_seg_count() teach copy_page_to_iter() to handle compound pages copy_page_to_iter(): fix ITER_DISCARD case [xarray] iov_iter_fault_in_readable() should do nothing in xarray case iov_iter_advance(): use consistent semantics for move past the end iov_iter: switch ..._full() variants of primitives to use of iov_iter_revert() iov_iter: reorder handling of flavours in primitives iov_iter_advance(): don't modify ->iov_offset for ITER_DISCARD iov_iter: separate direction from flavour iov_iter: optimize iov_iter_advance() for iovec and kvec sanitize iov_iter_fault_in_readable() iov_iter_alignment(): don't bother with iterate_all_kinds() iov_iter_gap_alignment(): get rid of iterate_all_kinds() get rid of iterate_all_kinds() in iov_iter_get_pages()/iov_iter_get_pages_alloc() iov_iter_npages(): don't bother with iterate_all_kinds() [xarray] iov_iter_npages(): just use DIV_ROUND_UP() iov_iter: replace iov_iter_copy_from_user_atomic() with iterator-advancing variant csum_and_copy_to_iter(): massage into form closer to csum_and_copy_from_iter() iterate_and_advance(): get rid of magic in case when n is 0 iov_iter: massage iterate_iovec and iterate_kvec to logics similar to iterate_bvec iov_iter: unify iterate_iovec and iterate_kvec iterate_bvec(): expand bvec.h macro forest, massage a bit iov_iter: teach iterate_{bvec,xarray}() about possible short copies iov_iter: get rid of separate bvec and xarray callbacks iov_iter: make the amount already copied available to iterator callbacks iov_iter: make iterator callbacks use base and len instead of iovec pull handling of ->iov_offset into iterate_{iovec,bvec,xarray} iterate_xarray(): only of the first iteration we might get offset != 0 copy_page_to_iter(): don't bother with kmap_atomic() for bvec/kvec cases copy_page_from_iter(): don't need kmap_atomic() for kvec/bvec cases iov_iter: clean csum_and_copy_...() primitives up a bit pipe_zero(): we don't need no stinkin' kmap_atomic()... clean up copy_mc_pipe_to_iter() csum_and_copy_to_pipe_iter(): leave handling of csum_state to caller David Howells (1): iov_iter: Remove iov_iter_for_each_range() Documentation/filesystems/porting.rst | 9 + fs/btrfs/file.c | 23 +- fs/fuse/file.c | 4 +- fs/iomap/buffered-io.c | 35 +- fs/ntfs/file.c | 33 +- include/linux/uio.h | 66 +- include/net/checksum.h | 14 +- lib/iov_iter.c | 1231 +++++++++++++++------------------ mm/filemap.c | 36 +- 9 files changed, 643 insertions(+), 808 deletions(-)