From: Ira Weiny <ira.weiny@xxxxxxxxx> At LSF/MM'19 [1] [2] we discussed applications that overestimate memory consumption due to their inability to detect whether the kernel will instantiate page cache for a file, and cases where a global dax enable via a mount option is too coarse. The following patch series enables selecting the use of DAX on individual files and/or directories on xfs, and lays some groundwork to do so in ext4. In this scheme the dax mount option can be omitted to allow the per-file property to take effect. The insight at LSF/MM was to separate the per-mount or per-file "physical" capability switch from an "effective" attribute for the file. At LSF/MM we discussed the difficulties of switching the mode of a file with active mappings / page cache. Rather than solve those races the decision was to just limit mode flips to 0-length files. Finally, the physical DAX flag inheritance is maintained from previous work on XFS but should be added for other file systems for consistence. [1] https://lwn.net/Articles/787973/ [2] https://lwn.net/Articles/787233/ To: linux-kernel@xxxxxxxxxxxxxxx Cc: Alexander Viro <viro@xxxxxxxxxxxxxxxxxx> Cc: "Darrick J. Wong" <darrick.wong@xxxxxxxxxx> Cc: Dan Williams <dan.j.williams@xxxxxxxxx> Cc: Dave Chinner <david@xxxxxxxxxxxxx> Cc: Christoph Hellwig <hch@xxxxxx> Cc: "Theodore Y. Ts'o" <tytso@xxxxxxx> Cc: Jan Kara <jack@xxxxxxx> Cc: linux-ext4@xxxxxxxxxxxxxxx Cc: linux-xfs@xxxxxxxxxxxxxxx Cc: linux-fsdevel@xxxxxxxxxxxxxxx Ira Weiny (5): fs/stat: Define DAX statx attribute fs/xfs: Isolate the physical DAX flag from effective fs/xfs: Separate functionality of xfs_inode_supports_dax() fs/xfs: Clean up DAX support check fs/xfs: Allow toggle of physical DAX flag fs/stat.c | 3 +++ fs/xfs/xfs_ioctl.c | 32 ++++++++++++++------------------ fs/xfs/xfs_iops.c | 36 ++++++++++++++++++++++++++++++------ fs/xfs/xfs_iops.h | 2 ++ include/uapi/linux/stat.h | 1 + 5 files changed, 50 insertions(+), 24 deletions(-) -- 2.20.1
