Merge tag 'xfs-for-linus-4.8-rc1' of git://git.kernel.org/pub/scm/lin…

…ux/kernel/git/dgc/linux-xfs

Pull xfs updates from Dave Chinner:
 "The major addition is the new iomap based block mapping
  infrastructure.  We've been kicking this about locally for years, but
  there are other filesystems want to use it too (e.g. gfs2).  Now it
  is fully working, reviewed and ready for merge and be used by other
  filesystems.

  There are a lot of other fixes and cleanups in the tree, but those are
  XFS internal things and none are of the scale or visibility of the
  iomap changes.  See below for details.

  I am likely to send another pull request next week - we're just about
  ready to merge some new functionality (on disk block->owner reverse
  mapping infrastructure), but that's a huge chunk of code (74 files
  changed, 7283 insertions(+), 1114 deletions(-)) so I'm keeping that
  separate to all the "normal" pull request changes so they don't get
  lost in the noise.

  Summary of changes in this update:
   - generic iomap based IO path infrastructure
   - generic iomap based fiemap implementation
   - xfs iomap based Io path implementation
   - buffer error handling fixes
   - tracking of in flight buffer IO for unmount serialisation
   - direct IO and DAX io path separation and simplification
   - shortform directory format definition changes for wider platform
     compatibility
   - various buffer cache fixes
   - cleanups in preparation for rmap merge
   - error injection cleanups and fixes
   - log item format buffer memory allocation restructuring to prevent
     rare OOM reclaim deadlocks
   - sparse inode chunks are now fully supported"

* tag 'xfs-for-linus-4.8-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/dgc/linux-xfs: (53 commits)
  xfs: remove EXPERIMENTAL tag from sparse inode feature
  xfs: bufferhead chains are invalid after end_page_writeback
  xfs: allocate log vector buffers outside CIL context lock
  libxfs: directory node splitting does not have an extra block
  xfs: remove dax code from object file when disabled
  xfs: skip dirty pages in ->releasepage()
  xfs: remove __arch_pack
  xfs: kill xfs_dir2_inou_t
  xfs: kill xfs_dir2_sf_off_t
  xfs: split direct I/O and DAX path
  xfs: direct calls in the direct I/O path
  xfs: stop using generic_file_read_iter for direct I/O
  xfs: split xfs_file_read_iter into buffered and direct I/O helpers
  xfs: remove s_maxbytes enforcement in xfs_file_read_iter
  xfs: kill ioflags
  xfs: don't pass ioflags around in the ioctl path
  xfs: track and serialize in-flight async buffers against unmount
  xfs: exclude never-released buffers from buftarg I/O accounting
  xfs: don't reset b_retries to 0 on every failure
  xfs: remove extraneous buffer flag changes
  ...

fs/Kconfig

@@ -10,6 +10,9 @@ config DCACHE_WORD_ACCESS
 
 if BLOCK
 
+config FS_IOMAP
+	bool
+
 source "fs/ext2/Kconfig"
 source "fs/ext4/Kconfig"
 source "fs/jbd2/Kconfig"

fs/Makefile

@@ -49,6 +49,7 @@ obj-$(CONFIG_COREDUMP)		+= coredump.o
 obj-$(CONFIG_SYSCTL)		+= drop_caches.o
 
 obj-$(CONFIG_FHANDLE)		+= fhandle.o
+obj-$(CONFIG_FS_IOMAP)		+= iomap.o
 
 obj-y				+= quota/
 

fs/buffer.c

@@ -21,6 +21,7 @@
 #include <linux/kernel.h>
 #include <linux/syscalls.h>
 #include <linux/fs.h>
+#include <linux/iomap.h>
 #include <linux/mm.h>
 #include <linux/percpu.h>
 #include <linux/slab.h>
@@ -1892,8 +1893,62 @@ void page_zero_new_buffers(struct page *page, unsigned from, unsigned to)
 }
 EXPORT_SYMBOL(page_zero_new_buffers);
 
-int __block_write_begin(struct page *page, loff_t pos, unsigned len,
-		get_block_t *get_block)
+static void
+iomap_to_bh(struct inode *inode, sector_t block, struct buffer_head *bh,
+		struct iomap *iomap)
+{
+	loff_t offset = block << inode->i_blkbits;
+
+	bh->b_bdev = iomap->bdev;
+
+	/*
+	 * Block points to offset in file we need to map, iomap contains
+	 * the offset at which the map starts. If the map ends before the
+	 * current block, then do not map the buffer and let the caller
+	 * handle it.
+	 */
+	BUG_ON(offset >= iomap->offset + iomap->length);
+
+	switch (iomap->type) {
+	case IOMAP_HOLE:
+		/*
+		 * If the buffer is not up to date or beyond the current EOF,
+		 * we need to mark it as new to ensure sub-block zeroing is
+		 * executed if necessary.
+		 */
+		if (!buffer_uptodate(bh) ||
+		    (offset >= i_size_read(inode)))
+			set_buffer_new(bh);
+		break;
+	case IOMAP_DELALLOC:
+		if (!buffer_uptodate(bh) ||
+		    (offset >= i_size_read(inode)))
+			set_buffer_new(bh);
+		set_buffer_uptodate(bh);
+		set_buffer_mapped(bh);
+		set_buffer_delay(bh);
+		break;
+	case IOMAP_UNWRITTEN:
+		/*
+		 * For unwritten regions, we always need to ensure that
+		 * sub-block writes cause the regions in the block we are not
+		 * writing to are zeroed. Set the buffer as new to ensure this.
+		 */
+		set_buffer_new(bh);
+		set_buffer_unwritten(bh);
+		/* FALLTHRU */
+	case IOMAP_MAPPED:
+		if (offset >= i_size_read(inode))
+			set_buffer_new(bh);
+		bh->b_blocknr = (iomap->blkno >> (inode->i_blkbits - 9)) +
+				((offset - iomap->offset) >> inode->i_blkbits);
+		set_buffer_mapped(bh);
+		break;
+	}
+}
+
+int __block_write_begin_int(struct page *page, loff_t pos, unsigned len,
+		get_block_t *get_block, struct iomap *iomap)
 {
 	unsigned from = pos & (PAGE_SIZE - 1);
 	unsigned to = from + len;
@@ -1929,9 +1984,14 @@ int __block_write_begin(struct page *page, loff_t pos, unsigned len,
 			clear_buffer_new(bh);
 		if (!buffer_mapped(bh)) {
 			WARN_ON(bh->b_size != blocksize);
-			err = get_block(inode, block, bh, 1);
-			if (err)
-				break;
+			if (get_block) {
+				err = get_block(inode, block, bh, 1);
+				if (err)
+					break;
+			} else {
+				iomap_to_bh(inode, block, bh, iomap);
+			}
+
 			if (buffer_new(bh)) {
 				unmap_underlying_metadata(bh->b_bdev,
 							bh->b_blocknr);
@@ -1972,6 +2032,12 @@ int __block_write_begin(struct page *page, loff_t pos, unsigned len,
 		page_zero_new_buffers(page, from, to);
 	return err;
 }
+
+int __block_write_begin(struct page *page, loff_t pos, unsigned len,
+		get_block_t *get_block)
+{
+	return __block_write_begin_int(page, pos, len, get_block, NULL);
+}
 EXPORT_SYMBOL(__block_write_begin);
 
 static int __block_commit_write(struct inode *inode, struct page *page,

fs/internal.h

@@ -11,6 +11,7 @@
 
 struct super_block;
 struct file_system_type;
+struct iomap;
 struct linux_binprm;
 struct path;
 struct mount;
@@ -39,6 +40,8 @@ static inline int __sync_blockdev(struct block_device *bdev, int wait)
  * buffer.c
  */
 extern void guard_bio_eod(int rw, struct bio *bio);
+extern int __block_write_begin_int(struct page *page, loff_t pos, unsigned len,
+		get_block_t *get_block, struct iomap *iomap);
 
 /*
  * char_dev.c