Commit Graph

38 Commits

Author SHA1 Message Date
Al Viro 6d4c1a30b3 orangefs: fix do_readv_writev() handling of error halfway through
Error should only be returned if nothing had been read/written.
Otherwise we need to report a short read/write instead.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-03-25 22:30:54 -04:00
Al Viro 177f8fc491 orangefs: sanitize ->llseek()
a) open files can't have NULL inodes
b) it's SEEK_END, not ORANGEFS_SEEK_END; no need to get cute.
c) make_bad_inode() on lseek()?

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-03-25 22:30:54 -04:00
Al Viro b8a99a8f9f orangefs: saner calling conventions for getting a slot
just have it return the slot number or -E... - the caller checks
the sign anyway

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-03-25 22:30:54 -04:00
Al Viro bf6bf606e5 orangefs_copy_{to,from}_bufmap(): don't pass bufmap pointer
it's always __orangefs_bufmap

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-03-25 22:30:54 -04:00
Martin Brandenburg e2f7f0d798 orangefs: use new orangefs_inode_getattr to get size in write and llseek
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-03-23 17:36:14 -04:00
Martin Brandenburg 3c9cf98d7b orangefs: rename orangefs_inode_getattr to orangefs_inode_old_getattr
This is motivated by orangefs_inode_old_getattr's habit of writing over
live inodes.

Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-03-23 17:36:14 -04:00
Mike Marshall 162ada7764 Orangefs: improve the POSIXness of interrupted writes...
Don't return EINTR on interrupted writes if some data has already
been written.

Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-03-09 13:12:37 -05:00
Mike Marshall 9d9e7ba9ee Orangefs: improve gossip statements
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-03-03 13:46:48 -05:00
Martin Brandenburg 9c2bcf288e orangefs: remove vestigial async io code
I have verified that there is nothing in the userspace daemon version we
are implementing this protocol against that ever looks at this field.

Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-02-24 17:07:50 -05:00
Al Viro 897c5df6cf orangefs: get rid of op->done
shouldn't be needed now

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-02-19 13:45:55 -05:00
Al Viro 78699e29fd orangefs: delay freeing slot until cancel completes
Make cancels reuse the aborted read/write op, to make sure they do not
fail on lack of memory.

Don't issue a cancel unless the daemon has seen our read/write, has not
replied and isn't being shut down.

If cancel *is* issued, don't wait for it to complete; stash the slot
in there and just have it freed when cancel is finally replied to or
purged (and delay dropping the reference until then, obviously).

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-02-19 13:45:53 -05:00
Al Viro 1357d06d49 get rid of bufmap argument of orangefs_bufmap_put()
it's always equal to __orangefs_bufmap and the latter can't change
until we are done

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-02-12 15:05:33 -05:00
Al Viro c0eae8cd77 orangefs: get rid of handle_io_error()
the second caller never needs to cancel, actually

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-02-12 15:05:32 -05:00
Al Viro 7b9761af86 orangefs: wait_for_direct_io(): restore the position in iter when restarting
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-02-12 15:05:32 -05:00
Al Viro e17be9fd4d orangefs: avoid freeing a slot twice in wait_for_direct_io()
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-02-12 15:05:32 -05:00
Mike Marshall 6ebcc3fcda Orangefs: added a couple of WARN_ONs, perhaps just temporarily.
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-02-04 16:28:31 -05:00
Martin Brandenburg 99109822f5 orangefs: Fix revalidate.
Previously, it would update a live inode. This was fixed, but it did not
ever check that the inode attributes in the dcache are correct. This
checks all inode attributes and rejects any that are not correct, which
causes a lookup and thus a new getattr.

Perhaps inode_operations->permission should replace or augment some of
this.

There is no actual caching, and this does a rather excessive amount of
network operations back to the filesystem server.

Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-01-28 15:08:40 -05:00
Al Viro b0bc3a7b62 orangefs: move handle_io_error() to file.c
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-01-23 15:20:11 -05:00
Al Viro 2a9e5c2260 orangefs: don't reinvent completion.h...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-01-23 15:20:11 -05:00
Al Viro ed42fe0593 orangefs: hopefully saner op refcounting and locking
* create with refcount 1
* make op_release() decrement and free if zero (i.e. old put_op()
  has become that).
* mark when submitter has given up waiting; from that point nobody
  else can move between the lists, change state, etc.
* have daemon read/write_iter grab a reference when picking op
  and *always* give it up in the end
* don't put into hash until we know it's been successfully passed to
  daemon

* move op->lock _lower_ than htab_in_progress_lock (and make sure
  to take it in purge_inprogress_ops())

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2016-01-23 13:03:12 -05:00
Mike Marshall 575e946125 Orangefs: change pvfs2 filenames to orangefs
Also changed references within source files that referred to
header files whose names had changed.

Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-12-04 12:56:14 -05:00
Yi Liu 8bb8aefd5a OrangeFS: Change almost all instances of the string PVFS2 to OrangeFS.
OrangeFS was formerly known as PVFS2 and retains the name in many places.

I leave the device /dev/pvfs2-req since this affects userspace.

I leave the filesystem type pvfs2 since this affects userspace. Further
the OrangeFS sysint library reads fstab for an entry of type pvfs2
independently of kernel mounts.

I leave extended attribute keys user.pvfs2 and system.pvfs2 as the
sysint library understands these.

I leave references to userspace binaries still named pvfs2.

I leave the filenames.

Signed-off-by: Yi Liu <yi9@clemson.edu>
[martin@omnibond.com: clairify above constraints and merge]
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-12-03 14:45:44 -05:00
Al Viro 555fa0fa61 fs: out of bounds on stack in iov_iter_advance
On Wed, Nov 11, 2015 at 10:19:48AM +0000, Al Viro wrote:

> I'll cook the minimal fixup for API change after I get some sleep and
> send it your way, unless somebody gets there first...

This should do it - switches ->ioctl() to pvfs2_inode_[gs]etxattr() and
converts xattr_handler ->[gs]et() to new API.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-11-16 11:27:24 -05:00
Mike Marshall 3f1b6947dc Orangefs: set pos after generic_write_checks
if we are appending, generic_write_checks would have updated
pos to the end of the file...

Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-11-13 13:05:11 -05:00
Al Viro 74f68fce2a orangefs: make pvfs2_inode_read() take iov_iter
... and make the only caller use page-backed iov_iter,
getting rid of kmap/kunmap *and* of the bug with
attempted use of iovec-backed copy_page_to_iter()
on a kernel pointer.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-11-13 11:36:31 -05:00
Al Viro 0071ed1ec6 orangefs: make do_readv_writev() take iov_iter
no need to build a copy of what the caller already has;
what's more, we want the one given to caller properly
advanced *and* we shouldn't depend upon it being an
iovec-backed one.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-11-13 11:35:28 -05:00
Al Viro dc4067f671 orangefs: don't bother with splitting iovecs
copy_page_{to,from}_iter() advances it just fine *and* it has no
problem with partially consumed segments.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-11-13 11:23:02 -05:00
Al Viro 3c2fcfcb68 orangefs: make wait_for_direct_io() take iov_iter
incidentally, insane or compromised server returning *more* than
requested on read should not oops the kernel - initialize the
iov_iter for read according to the iovec we've got.  That's why
pvfs_bufmap_copy_to_iovec() needed a separate size argument - we
shouldn't abuse iov_iter_count(iter) for passing that.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-11-13 11:21:37 -05:00
Al Viro a5c126a522 orangefs: make precopy_buffers() take iov_iter
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-11-13 11:18:50 -05:00
Al Viro 5f0e3c953f orangefs: make postcopy_buffers() take iov_iter
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-11-13 11:11:55 -05:00
Al Viro 5c278228bb orangefs: explicitly pass the size to pvfs_bufmap_copy_to_iovec()
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-11-13 10:25:01 -05:00
Mike Marshall 548049495c Orangefs: fix some checkpatch.pl complaints that had creeped in.
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-10-05 13:44:24 -04:00
Martin Brandenburg 353908035f Orangefs: Use readonly mmap since writepage is not implemented.
Previously the code silently failed to update the disk. Now it will not
allow writable and shared mmaps.

Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-10-03 11:44:39 -04:00
Mike Marshall f957ae2dec Orangefs: choose return codes from among the expected ones.
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-10-03 11:44:34 -04:00
Mike Marshall 4d1c44043b Orangefs: use iov_iter interface
replace opencoded pvfs_bufmap_copy_to_kernel_iovec,
pvfs_bufmap_copy_to_user_iovec, pvfs_bufmap_copy_iovec_from_kernel,
and pvfs_bufmap_copy_iovec_from_user with pvfs_bufmap_copy_to_iovec
and pvfs_bufmap_copy_from_iovec, which both use the iov_iter
interface.

Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-10-03 11:44:00 -04:00
Mike Marshall eeaa3d448c Orangefs: address problems found by static checker
Don't check for negative rc from boolean.

  Don't pointlessly initialize variables, it short-circuits
  gcc's uninitialized variable warnings. And max_new_nr_segs
  can never be zero, so don't check for it.

  Preserve original kstrdup pointer for freeing later.

  Don't check for negative value in unsigned variable.

Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-10-03 11:40:03 -04:00
Mike Marshall 84d02150de Orangefs: sooth most sparse complaints
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-10-03 11:40:02 -04:00
Mike Marshall 5db11c21a9 Orangefs: kernel client part 2
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
2015-10-03 11:39:54 -04:00