Hi all, Here's v2 of the cached writeback mapping invalidation series. Note that this post is a bit premature as I'm still testing this code. Christoph pointed out that he has some patches shifting code around in the same area so I'm posting what I have under test at this point to hopefully work out any potential conflicts. The changes in v2 are basically to beef up the invalidation logic a bit in xfs_map_blocks() via a new (old) xfs_imap_valid() helper and the introduction of patch 5, which attempts to add similar mapping validation in xfs_iomap_write_allocate() to replace the old truncate detection logic. Thoughts, reviews, flames appreciated. Brian fstest: https://marc.info/?l=fstests&m=154756141122233&w=2 v2: - Refactor validation logic into xfs_imap_valid() helper. - Revalidate seqno after the lock cycle in xfs_map_blocks(). - Update *seq in xfs_iomap_write_allocate() regardless of fork type. - Add patch 5 for seqno revalidation on xfs_iomap_write_allocate() lock cycles. v1: https://marc.info/?l=linux-xfs&m=154721212321112&w=2 Brian Foster (5): xfs: eof trim writeback mapping as soon as it is cached xfs: update fork seq counter on data fork changes xfs: validate writeback mapping using data fork seq counter xfs: remove superfluous writeback mapping eof trimming xfs: revalidate imap properly before writeback delalloc conversion fs/xfs/libxfs/xfs_bmap.c | 11 ---- fs/xfs/libxfs/xfs_bmap.h | 1 - fs/xfs/libxfs/xfs_iext_tree.c | 13 ++-- fs/xfs/libxfs/xfs_inode_fork.h | 2 +- fs/xfs/xfs_aops.c | 71 ++++++++++++++-------- fs/xfs/xfs_iomap.c | 105 ++++++++++++++------------------- 6 files changed, 98 insertions(+), 105 deletions(-) -- 2.17.2
