wireguard-linux/fs/orangefs/super.c, branch jd/unified-crypt-queue

Merge tag 'for-linus-5.2-ofs1' of git://git.kernel.org/pub/scm/linux/kernel/git/hubcap/linux

2019-05-09T16:37:25Z

Pull orangefs updates from Mike Marshall: "This includes one fix and our "Orangefs through the pagecache" patch series which greatly improves our small IO performance and helps us pass more xfstests than before. Fix: - orangefs: truncate before updating size Pagecache series: - all the rest" * tag 'for-linus-5.2-ofs1' of git://git.kernel.org/pub/scm/linux/kernel/git/hubcap/linux: (23 commits) orangefs: truncate before updating size orangefs: copy Orangefs-sized blocks into the pagecache if possible. orangefs: pass slot index back to readpage. orangefs: remember count when reading. orangefs: add orangefs_revalidate_mapping orangefs: implement writepages orangefs: write range tracking orangefs: avoid fsync service operation on flush orangefs: skip inode writeout if nothing to write orangefs: move do_readv_writev to direct_IO orangefs: do not return successful read when the client-core disappeared orangefs: implement writepage orangefs: migrate to generic_file_read_iter orangefs: service ops done for writeback are not killable orangefs: remove orangefs_readpages orangefs: reorganize setattr functions to track attribute changes orangefs: let setattr write to cached inode orangefs: set up and use backing_dev_info orangefs: hold i_lock during inode_getattr orangefs: update attributes rather than relying on server ...

orangefs: reorganize setattr functions to track attribute changes

2019-05-03T18:32:38Z

OrangeFS accepts a mask indicating which attributes were changed. The kernel must not set any bits except those that were actually changed. The kernel must set the uid/gid of the request to the actual uid/gid responsible for the change. Code path for notify_change initiated setattrs is orangefs_setattr(dentry, iattr) -> __orangefs_setattr(inode, iattr) In kernel changes are initiated by calling __orangefs_setattr. Code path for writeback is orangefs_write_inode -> orangefs_inode_setattr attr_valid and attr_uid and attr_gid change together under i_lock. I_DIRTY changes separately. __orangefs_setattr lock if needs to be cleaned first, unlock and retry set attr_valid copy data in unlock mark_inode_dirty orangefs_inode_setattr lock copy attributes out unlock clear getattr_time # __writeback_single_inode clears dirty orangefs_inode_getattr # possible to get here with attr_valid set and not dirty lock if getattr_time ok or attr_valid set, unlock and return unlock do server operation # another thread may getattr or setattr, so check for that lock if getattr_time ok or attr_valid, unlock and return else, copy in update getattr_time unlock Signed-off-by: Martin Brandenburg Signed-off-by: Mike Marshall

orangefs: let setattr write to cached inode

2019-05-03T18:32:38Z

This is a fairly big change, but ultimately it's not a lot of code. Implement write_inode and then avoid the call to orangefs_inode_setattr within orangefs_setattr. Signed-off-by: Martin Brandenburg Signed-off-by: Mike Marshall

orangefs: set up and use backing_dev_info

2019-05-03T18:32:38Z

Signed-off-by: Martin Brandenburg Signed-off-by: Mike Marshall

orangefs: implement xattr cache

2019-05-03T18:32:37Z

This uses the same timeout as the getattr cache. This substantially increases performance when writing files with smaller buffer sizes. When writing, the size is (often) changed, which causes a call to notify_change which calls security_inode_need_killpriv which needs a getxattr. Caching it reduces traffic to the server. Signed-off-by: Martin Brandenburg Signed-off-by: Mike Marshall

orangefs: make use of ->free_inode()

2019-05-02T02:43:27Z

Acked-by: Mike Marshall Signed-off-by: Al Viro

orangefs: formatting cleanups

2018-06-01T18:51:30Z

Signed-off-by: Mike Marshall

orangefs: revamp block sizes

2018-06-01T18:48:31Z

Now the superblock block size is PAGE_SIZE. The inode block size is PAGE_SIZE for directories and symlinks, but is the server-reported block size for regular files. The block size in the OrangeFS private inode is now deleted. Stat now reports PAGE_SIZE for directories and symlinks and the server-reported block size for regular files. The user-space visible change is that the block size for directores and symlinks and the superblock is now PAGE_SIZE rather than the size of the client-core shared memory buffers, which was typically four megabytes. Reported-by: Becky Ligon Signed-off-by: Martin Brandenburg Cc: hubcap@omnibond.com Cc: walt@omnibond.com Signed-off-by: Mike Marshall

orangefs_kill_sb(): deal with allocation failures

2018-04-16T03:49:12Z

orangefs_fill_sb() might've failed to allocate ORANGEFS_SB(s); don't oops in that case. Cc: stable@kernel.org Signed-off-by: Al Viro

Merge tag 'for-linus-4.16' of git://git.kernel.org/pub/scm/linux/kernel/git/hubcap/linux

2018-02-08T20:20:41Z

Pull orangefs updates from Mike Marshall: "Mostly cleanups, but three bug fixes: - don't pass garbage return codes back up the call chain (Mike Marshall) - fix stale inode test (Martin Brandenburg) - fix off-by-one errors (Xiongfeng Wang) Also add Martin as a reviewer in the Maintainers file" * tag 'for-linus-4.16' of git://git.kernel.org/pub/scm/linux/kernel/git/hubcap/linux: orangefs: reverse sense of is-inode-stale test in d_revalidate orangefs: simplify orangefs_inode_is_stale Orangefs: don't propogate whacky error codes orangefs: use correct string length orangefs: make orangefs_make_bad_inode static orangefs: remove ORANGEFS_KERNEL_DEBUG orangefs: remove gossip_ldebug and gossip_lerr orangefs: make orangefs_client_debug_init static MAINTAINERS: update orangefs list and add myself as reviewer