The first vfs pile, with deep apologies for being very late in this window.
Assorted cleanups and fixes, plus a large preparatory part of iov_iter work.
There's a lot more of that, but it'll probably go into the next merge window
- it *does* shape up nicely, removes a lot of boilerplate, gets rid of
locking inconsistencie between aio_write and splice_write and I hope to get
Kent's direct-io rewrite merged into the same queue, but some of the stuff
after this point is having (mostly trivial) conflicts with the things already
merged into mainline and with some I want more testing.
This one passes LTP and xfstests without regressions, in addition to usual
beating. BTW, readahead02 in ltp syscalls testsuite has started giving
failures since "mm/readahead.c: fix readahead failure for memoryless NUMA
nodes and limit readahead pages" - might be a false positive, might be
a real regression...
The conflicts are trivial (variable added in mainline next to one removed
in this series, a function removed in mainline next to several removed in
this, new object added in mm/Makefile, a function call in fs/ceph/file.c
replaced with several lines equivalent to it next to an if that was added
after that call in mainline). I've pushed a resolved variant into
vfs.git#proposed-merge, but that's mostly for verification purposes.
There will be another pile, with bits and pieces from other folks; I want
to deal with the bulk of that mess first...
The branch itself is in the usual place -
git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs.git for-linus
Shortlog:
Al Viro (61):
smarter propagate_mnt()
reduce m_start() cost...
usbip: don't open-code sockfd_lookup/sockfd_put
vhost: don't open-code sockfd_put()
switch nbd to sockfd_lookup/sockfd_put
ncpfs: switch to sockfd_lookup()/sockfd_put()
don't bother with {get,put}_write_access() on non-regular files
get rid of DEBUG_WRITECOUNT
fold __get_file_write_access() into its only caller
mark struct file that had write access grabbed by open()
tidy do_dentry_open() up a bit
get_write_access() is inlined, exporting it is pointless
namei.c: move EXPORT_SYMBOL to corresponding definitions
get rid of files_defer_init()
lustre: generic_readlink() is just fine there, TYVM...
new helper: readlink_copy()
ntfs: don't put NULL into ->i_op/->i_fop
get rid of pointless checks for NULL ->i_op
cifs: ->rename() without ->lookup() makes no sense
mn10300: kmap_atomic() returns void *, not unsigned long...
fuse/dev: use atomic maps
pipe: kill ->map() and ->unmap()
switch ->is_partially_uptodate() to saner arguments
callers of iov_copy_from_user_atomic() don't need pagecache_disable()
do_shmem_file_read(): call file_read_actor() directly
introduce copy_page_to_iter, kill loop over iovec in generic_file_aio_read()
cifs_iovec_read(): resubmit shouldn't restart the loop
switch pipe_read() to copy_page_to_iter()
switch vmsplice_to_user() to copy_page_to_iter()
cifs_iovec_read: keep iov_iter between the calls of cifs_readdata_to_iov()
fold cifs_iovec_read() into its (only) caller
read_code(): go through vfs_read() instead of calling the method directly
untangling process_vm_..., part 1
untangling process_vm_..., part 2
untangling process_vm_..., part 3
untangling process_vm_..., part 4
process_vm_access: switch to iov_iter
process_vm_access: switch to copy_page_to_iter/iov_iter_copy_from_user
process_vm_access: take get_user_pages/put_pages one level up
process_vm_rw_pages(): pass accurate amount of bytes
process_vm_access: don't bother with returning the amounts of bytes copied
process_vm_access: tidy up a bit
take iov_iter stuff to mm/iov_iter.c
ocfs2: don't open-code kernel_sendmsg()
lustre: switch to kernel_sendmsg()
constify blk_rq_map_user_iov() and friends
drbd: don't open-code kernel_recvmsg()
ocfs2: don't open-code kernel_recvmsg()
lustre: don't open-code kernel_recvmsg()
kill the 4th argument of __generic_file_aio_write()
kill the 5th argument of generic_file_buffered_write()
btrfs_file_aio_write(): get rid of ppos
generic_file_direct_write(): get rid of ppos argument
export generic_perform_write(), start getting rid of generic_file_buffer_write()
xfs_file_buffered_aio_write(): switch to generic_perform_write()
ceph_aio_write(): switch to generic_perform_write()
ocfs2_file_aio_write(): switch to generic_perform_write()
kill generic_file_buffered_write()
ceph_sync_{,direct_}write: fix an oops on ceph_osdc_new_request() failure
cifs: fix the race in cifs_writev()
missing bits of "splice: fix racy pipe->buffers uses"
David Howells (1):
VFS: Make delayed_free() call free_vfsmnt()
Kent Overstreet (1):
iov_iter: Move iov_iter to uio.h
Diffstat:
Documentation/filesystems/Locking | 2 +-
Documentation/filesystems/vfs.txt | 2 +-
arch/mn10300/include/asm/highmem.h | 4 +-
arch/powerpc/configs/ppc6xx_defconfig | 1 -
arch/powerpc/configs/ps3_defconfig | 1 -
arch/s390/configs/default_defconfig | 1 -
arch/sh/configs/rsk7203_defconfig | 1 -
arch/xtensa/configs/iss_defconfig | 1 -
arch/xtensa/configs/s6105_defconfig | 1 -
block/blk-map.c | 2 +-
drivers/block/drbd/drbd_receiver.c | 12 +-
drivers/block/nbd.c | 48 ++-
drivers/char/virtio_console.c | 4 +-
.../lustre/lnet/klnds/socklnd/socklnd_lib-linux.c | 60 +---
.../lustre/lustre/libcfs/linux/linux-tcpip.c | 24 +-
drivers/staging/lustre/lustre/llite/symlink.c | 23 +-
drivers/staging/usbip/stub_dev.c | 8 +-
drivers/staging/usbip/usbip_common.c | 25 --
drivers/staging/usbip/usbip_common.h | 1 -
drivers/staging/usbip/vhci_hcd.c | 4 +-
drivers/staging/usbip/vhci_sysfs.c | 6 +-
drivers/vhost/net.c | 14 +-
fs/bio.c | 10 +-
fs/block_dev.c | 2 +-
fs/btrfs/file.c | 16 +-
fs/buffer.c | 6 +-
fs/cachefiles/bind.c | 1 -
fs/cachefiles/namei.c | 3 +-
fs/ceph/file.c | 12 +-
fs/cifs/cifsfs.c | 1 -
fs/cifs/file.c | 128 +++-----
fs/exec.c | 2 +-
fs/ext4/file.c | 2 +-
fs/file.c | 11 +-
fs/file_table.c | 43 +--
fs/fuse/dev.c | 14 +-
fs/fuse/file.c | 5 +-
fs/mount.h | 5 +-
fs/namei.c | 67 ++--
fs/namespace.c | 56 ++--
fs/ncpfs/inode.c | 50 +--
fs/ncpfs/ncp_fs_sb.h | 2 -
fs/ntfs/inode.c | 2 -
fs/ocfs2/cluster/tcp.c | 49 +--
fs/ocfs2/file.c | 9 +-
fs/open.c | 68 +---
fs/pipe.c | 133 +-------
fs/pnode.c | 198 ++++++-----
fs/pnode.h | 3 +
fs/proc/namespaces.c | 14 +-
fs/proc/self.c | 2 +-
fs/proc_namespace.c | 1 +
fs/splice.c | 126 ++-----
fs/udf/file.c | 2 +-
fs/xfs/xfs_file.c | 13 +-
fs/xfs/xfs_ioctl.c | 28 +-
include/linux/bio.h | 5 +-
include/linux/blkdev.h | 4 +-
include/linux/buffer_head.h | 4 +-
include/linux/fdtable.h | 2 -
include/linux/fs.h | 97 +-----
include/linux/mount.h | 3 +
include/linux/nbd.h | 3 +-
include/linux/pipe_fs_i.h | 19 --
include/linux/uio.h | 52 +++
kernel/relay.c | 4 +-
kernel/trace/trace.c | 8 +-
lib/Kconfig.debug | 10 -
mm/Makefile | 2 +-
mm/filemap.c | 344 ++++----------------
mm/iov_iter.c | 224 +++++++++++++
mm/process_vm_access.c | 250 +++++---------
mm/shmem.c | 79 ++---
security/integrity/evm/evm_crypto.c | 2 +-
security/integrity/evm/evm_main.c | 2 +-
security/tomoyo/realpath.c | 4 +-
76 files changed, 910 insertions(+), 1537 deletions(-)
create mode 100644 mm/iov_iter.c
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
