Hi all, Here's the last dfops refactoring series. The objective is to embed dfops in the transaction. In doing so, the boilerplate xfs_trans_alloc() -> xfs_defer_init() -> xfs_defer_finish() -> xfs_trans_commit() pattern is reduced to a transaction allocation and commit. Patches 1-7 are various preparatory, cleanup and/or bug fix patches. Patch 8 adds an internal dfops field to xfs_trans with some supporting mechanism bits. Patches 9-11 convert the various deferred op using paths over to use the internal transaction dfops instead of the on-stack variant. Patches 12-13 remove a bunch of the boilerplate dfops code that is made superfluous by the previous changes. Finally, patch 14 drops the no longer necessary dfops param to xfs_defer_finish(). I think most of this is relatively straightforward and survives my testing so far. FWIW, I think there are still a couple more related cleanups that could be made on top of this series. For one, I need to optimize on top of Darrick's recently posted xfs_defer_finish() patch to return a clean transaction. Other things to consider might be to do away with support for external dfops and the ->t_dfops pointer indirection, or perhaps even consider going the other direction: allocate dfops from a separate zone to save some memory on non-permanent transactions (note that 16 of 28 transactions use a permanent log res. last I looked, so it may not be worth it atm). I know Christoph also had thoughts around condensing some of the items joined to the dfops to those with the transaction. I have yet to think about that one, but I do have an RFC quality patch laying around that replaces the ->dop_low flag with a transaction flag (->t_flags), eliminating the need for that extra byte in xfs_defer_ops. The one quirk associated with that is the question of whether we want to preserve the behavior where low mode remains active across the series of transactions associated with the traditional (on-stack) dfops or is reset on transaction roll (a la firstblock). I'll post that RFC separately for a more proper discussion.. Brian Brian Foster (14): xfs: pull up dfops from xfs_itruncate_extents() xfs: use ->t_dfops in log recovery intent processing xfs: fix transaction leak on remote attr set/remove failure xfs: make deferred processing safe for embedded dfops xfs: remove unused deferred ops committed field xfs: reset dfops to initial state after finish xfs: pack holes in xfs_defer_ops and xfs_trans xfs: support embedded dfops in transaction xfs: use internal dfops in cow blocks cancel xfs: use internal dfops in attr code xfs: use internal dfops during [b|c]ui recovery xfs: remove all boilerplate defer init/finish code xfs: remove unnecessary dfops init calls in xattr code xfs: drop unnecessary xfs_defer_finish() dfops parameter fs/xfs/libxfs/xfs_alloc_btree.c | 1 + fs/xfs/libxfs/xfs_attr.c | 49 ++++++--------- fs/xfs/libxfs/xfs_attr_leaf.c | 1 + fs/xfs/libxfs/xfs_attr_remote.c | 13 ++-- fs/xfs/libxfs/xfs_bmap.c | 16 +---- fs/xfs/libxfs/xfs_da_btree.c | 1 + fs/xfs/libxfs/xfs_defer.c | 95 +++++++++++++++++++++++------- fs/xfs/libxfs/xfs_defer.h | 9 +-- fs/xfs/libxfs/xfs_dir2_block.c | 1 + fs/xfs/libxfs/xfs_dir2_data.c | 1 + fs/xfs/libxfs/xfs_dir2_leaf.c | 1 + fs/xfs/libxfs/xfs_dir2_node.c | 1 + fs/xfs/libxfs/xfs_dir2_sf.c | 1 + fs/xfs/libxfs/xfs_dquot_buf.c | 1 + fs/xfs/libxfs/xfs_ialloc_btree.c | 1 + fs/xfs/libxfs/xfs_inode_fork.c | 1 + fs/xfs/libxfs/xfs_refcount.c | 10 +--- fs/xfs/libxfs/xfs_refcount_btree.c | 1 + fs/xfs/libxfs/xfs_symlink_remote.c | 1 + fs/xfs/libxfs/xfs_trans_resv.c | 1 + fs/xfs/xfs_aops.c | 1 + fs/xfs/xfs_attr_inactive.c | 1 + fs/xfs/xfs_attr_list.c | 1 + fs/xfs/xfs_bmap_item.c | 21 +++---- fs/xfs/xfs_bmap_util.c | 45 +++----------- fs/xfs/xfs_buf_item.c | 1 + fs/xfs/xfs_dir2_readdir.c | 1 + fs/xfs/xfs_dquot.c | 8 +-- fs/xfs/xfs_dquot_item.c | 1 + fs/xfs/xfs_export.c | 1 + fs/xfs/xfs_extent_busy.c | 1 + fs/xfs/xfs_extfree_item.c | 1 + fs/xfs/xfs_file.c | 1 + fs/xfs/xfs_icache.c | 1 + fs/xfs/xfs_icreate_item.c | 1 + fs/xfs/xfs_inode.c | 87 +++++---------------------- fs/xfs/xfs_inode_item.c | 1 + fs/xfs/xfs_ioctl.c | 1 + fs/xfs/xfs_iomap.c | 26 +------- fs/xfs/xfs_iops.c | 1 + fs/xfs/xfs_log.c | 1 + fs/xfs/xfs_log_cil.c | 1 + fs/xfs/xfs_log_recover.c | 12 +--- fs/xfs/xfs_pnfs.c | 1 + fs/xfs/xfs_qm.c | 1 + fs/xfs/xfs_qm_bhv.c | 1 + fs/xfs/xfs_qm_syscalls.c | 1 + fs/xfs/xfs_quotaops.c | 1 + fs/xfs/xfs_refcount_item.c | 30 +++++----- fs/xfs/xfs_reflink.c | 51 ++++++---------- fs/xfs/xfs_rtalloc.c | 9 +-- fs/xfs/xfs_super.c | 1 + fs/xfs/xfs_symlink.c | 38 +++--------- fs/xfs/xfs_trace.h | 8 +-- fs/xfs/xfs_trans.c | 32 ++++++++-- fs/xfs/xfs_trans.h | 12 +++- fs/xfs/xfs_trans_ail.c | 1 + fs/xfs/xfs_trans_buf.c | 1 + fs/xfs/xfs_trans_dquot.c | 1 + fs/xfs/xfs_trans_inode.c | 1 + 60 files changed, 266 insertions(+), 346 deletions(-) -- 2.17.1 -- To unsubscribe from this list: send the line "unsubscribe linux-xfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
