Changes since v4 [1]: * Kill the DEFINE_FSDAX_AOPS macro and just open code new address_space_operations instances for each fs (Matthew, Jan, Dave, Christoph) * Rename routines that had a 'dma_' prefix with 'dax_layout_' and merge the dax-layout-break into xfs_break_layouts() (Dave, Christoph) * Rework the implementation to have the fsdax core find the pages, but leave the responsibility of waiting on those pages to the filesystem (Dave). * Drop the nfit_test infrastructure for testing this mechanism, I plan to investigate better mechanisms for injecting arbitrary put_page() delays for dax pages relative to an extent unmap operation. The dm_delay target does not do what I want since it operates at whole device level. A better test interface would be a mechanism to delay I/O completion based on whether a bio referenced a given LBA. Not changed since v4: * This implementation still relies on RCU for synchronizing get_user_pages() and get_user_pages_fast() against dax_layout_busy_page(). We could perform the operation with just barriers if we knew at get_user_pages() time that the pages were flagged for truncation. However, dax_layout_busy_page() does not have the information to flag that a page is actually going to be truncated, only that it *might* be truncated. [1]: https://lists.01.org/pipermail/linux-nvdimm/2017-December/013704.html ---- Background: get_user_pages() in the filesystem pins file backed memory pages for access by devices performing dma. However, it only pins the memory pages not the page-to-file offset association. If a file is truncated the pages are mapped out of the file and dma may continue indefinitely into a page that is owned by a device driver. This breaks coherency of the file vs dma, but the assumption is that if userspace wants the file-space truncated it does not matter what data is inbound from the device, it is not relevant anymore. The only expectation is that dma can safely continue while the filesystem reallocates the block(s). Problem: This expectation that dma can safely continue while the filesystem changes the block map is broken by dax. With dax the target dma page *is* the filesystem block. The model of leaving the page pinned for dma, but truncating the file block out of the file, means that the filesytem is free to reallocate a block under active dma to another file and now the expected data-incoherency situation has turned into active data-corruption. Solution: Defer all filesystem operations (fallocate(), truncate()) on a dax mode file while any page/block in the file is under active dma. This solution assumes that dma is transient. Cases where dma operations are known to not be transient, like RDMA, have been explicitly disabled via commits like 5f1d43de5416 "IB/core: disable memory registration of filesystem-dax vmas". The dax_layout_busy_page() routine is called by filesystems with a lock held against mm faults (i_mmap_lock) to find pinned / busy dax pages. The process of looking up a busy page invalidates all mappings to trigger any subsequent get_user_pages() to block on i_mmap_lock. The filesystem continues to call dax_layout_busy_page() until it finally returns no more active pages. This approach assumes that the page pinning is transient, if that assumption is violated the system would have likely hung from the uncompleted I/O. --- Dan Williams (11): dax: store pfns in the radix xfs, dax: introduce xfs_dax_aops ext4, dax: introduce ext4_dax_aops ext2, dax: introduce ext2_dax_aops fs, dax: use page->mapping to warn if truncate collides with a busy page mm, dax: enable filesystems to trigger dev_pagemap ->page_free callbacks mm, dev_pagemap: introduce CONFIG_DEV_PAGEMAP_OPS wait_bit: introduce {wait_on,wake_up}_atomic_one mm, fs, dax: handle layout changes to pinned dax mappings xfs: prepare xfs_break_layouts() for another layout type xfs, dax: introduce xfs_break_dax_layouts() drivers/dax/super.c | 96 +++++++++++++++-- drivers/nvdimm/pmem.c | 3 - fs/Kconfig | 1 fs/dax.c | 259 +++++++++++++++++++++++++++++++++++++--------- fs/ext2/ext2.h | 1 fs/ext2/inode.c | 28 ++++- fs/ext2/namei.c | 18 --- fs/ext2/super.c | 6 + fs/ext4/inode.c | 11 ++ fs/ext4/super.c | 6 + fs/xfs/xfs_aops.c | 7 + fs/xfs/xfs_aops.h | 1 fs/xfs/xfs_file.c | 94 ++++++++++++++++- fs/xfs/xfs_inode.h | 9 ++ fs/xfs/xfs_ioctl.c | 9 +- fs/xfs/xfs_iops.c | 17 ++- fs/xfs/xfs_pnfs.c | 8 + fs/xfs/xfs_pnfs.h | 4 - fs/xfs/xfs_super.c | 20 ++-- include/linux/dax.h | 45 +++++++- include/linux/memremap.h | 28 ++--- include/linux/mm.h | 61 ++++++++--- include/linux/wait_bit.h | 13 ++ kernel/memremap.c | 30 +++++ kernel/sched/wait_bit.c | 59 +++++++++- mm/Kconfig | 5 + mm/gup.c | 5 + mm/hmm.c | 13 -- mm/swap.c | 3 - 29 files changed, 663 insertions(+), 197 deletions(-)
