Unshare all shared extents if the user calls fallocate with the new
unshare mode flag set, so that we can guarantee that a subsequent
write will not ENOSPC.
Signed-off-by: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
[hch: pass inode instead of file to xfs_reflink_dirty_range,
use iomap infrastructure for copy up]
Signed-off-by: Christoph Hellwig <hch@xxxxxx>
---
fs/xfs/xfs_file.c | 10 ++
fs/xfs/xfs_reflink.c | 237 ++++++++++++++++++++++++++++++++++++++++++++++++++
fs/xfs/xfs_reflink.h | 2
3 files changed, 247 insertions(+), 2 deletions(-)
diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index 450bf2b..f3e5cb1 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -845,7 +845,7 @@ xfs_file_write_iter(
#define XFS_FALLOC_FL_SUPPORTED \
(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE | \
FALLOC_FL_COLLAPSE_RANGE | FALLOC_FL_ZERO_RANGE | \
- FALLOC_FL_INSERT_RANGE)
+ FALLOC_FL_INSERT_RANGE | FALLOC_FL_UNSHARE_RANGE)
STATIC long
xfs_file_fallocate(
@@ -935,9 +935,15 @@ xfs_file_fallocate(
if (mode & FALLOC_FL_ZERO_RANGE)
error = xfs_zero_file_space(ip, offset, len);
- else
+ else {
+ if (mode & FALLOC_FL_UNSHARE_RANGE) {
+ error = xfs_reflink_unshare(ip, offset, len);
+ if (error)
+ goto out_unlock;
+ }
error = xfs_alloc_file_space(ip, offset, len,
XFS_BMAPI_PREALLOC);
+ }
if (error)
goto out_unlock;
}
diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c
index 77ac810..065e836 100644
--- a/fs/xfs/xfs_reflink.c
+++ b/fs/xfs/xfs_reflink.c
@@ -1472,3 +1472,240 @@ xfs_reflink_remap_range(
trace_xfs_reflink_remap_range_error(dest, error, _RET_IP_);
return error;
}
+
+/*
+ * The user wants to preemptively CoW all shared blocks in this file,
+ * which enables us to turn off the reflink flag. Iterate all
+ * extents which are not prealloc/delalloc to see which ranges are
+ * mentioned in the refcount tree, then read those blocks into the
+ * pagecache, dirty them, fsync them back out, and then we can update
+ * the inode flag. What happens if we run out of memory? :)
+ */
+STATIC int
+xfs_reflink_dirty_extents(
+ struct xfs_inode *ip,
+ xfs_fileoff_t fbno,
+ xfs_filblks_t end,
+ xfs_off_t isize)
+{
+ struct xfs_mount *mp = ip->i_mount;
+ xfs_agnumber_t agno;
+ xfs_agblock_t agbno;
+ xfs_extlen_t aglen;
+ xfs_agblock_t rbno;
+ xfs_extlen_t rlen;
+ xfs_off_t fpos;
+ xfs_off_t flen;
+ struct xfs_bmbt_irec map[2];
+ int nmaps;
+ int error;
+
+ while (end - fbno > 0) {
+ nmaps = 1;
+ /*
+ * Look for extents in the file. Skip holes, delalloc, or
+ * unwritten extents; they can't be reflinked.
+ */
+ error = xfs_bmapi_read(ip, fbno, end - fbno, map, &nmaps, 0);
+ if (error)
+ goto out;
+ if (nmaps == 0)
+ break;
+ if (map[0].br_startblock == HOLESTARTBLOCK ||
+ map[0].br_startblock == DELAYSTARTBLOCK ||
+ ISUNWRITTEN(&map[0]))
+ goto next;
+
+ map[1] = map[0];
+ while (map[1].br_blockcount) {
+ agno = XFS_FSB_TO_AGNO(mp, map[1].br_startblock);
+ agbno = XFS_FSB_TO_AGBNO(mp, map[1].br_startblock);
+ aglen = map[1].br_blockcount;
+
+ error = xfs_reflink_find_shared(mp, agno, agbno, aglen,
+ &rbno, &rlen, true);
+ if (error)
+ goto out;
+ if (rlen == 0)
+ goto skip_copy;
+
+ /* Dirty the pages */
+ xfs_iunlock(ip, XFS_ILOCK_EXCL);
+ fpos = XFS_FSB_TO_B(mp, map[1].br_startoff +
+ (rbno - agbno));
+ flen = XFS_FSB_TO_B(mp, rlen);
+ if (fpos + flen > isize)
+ flen = isize - fpos;
+ error = iomap_file_dirty(VFS_I(ip), fpos, flen,
+ &xfs_iomap_ops);
+ xfs_ilock(ip, XFS_ILOCK_EXCL);
+ if (error)
+ goto out;
+skip_copy:
+ map[1].br_blockcount -= (rbno - agbno + rlen);
+ map[1].br_startoff += (rbno - agbno + rlen);
+ map[1].br_startblock += (rbno - agbno + rlen);
+ }
+
+next:
+ fbno = map[0].br_startoff + map[0].br_blockcount;
+ }
+out:
+ return error;
+}
+
+/* Iterate the extents; if there are no reflinked blocks, clear the flag. */
+STATIC int
+xfs_reflink_try_clear_inode_flag(
+ struct xfs_inode *ip,
+ xfs_off_t old_isize)
+{
+ struct xfs_mount *mp = ip->i_mount;
+ struct xfs_trans *tp;
+ xfs_fileoff_t fbno;
+ xfs_filblks_t end;
+ xfs_agnumber_t agno;
+ xfs_agblock_t agbno;
+ xfs_extlen_t aglen;
+ xfs_agblock_t rbno;
+ xfs_extlen_t rlen;
+ struct xfs_bmbt_irec map[2];
+ int nmaps;
+ int error = 0;
+
+ /* Start a rolling transaction to remove the mappings */
+ error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, 0, 0, 0, &tp);
+ if (error)
+ return error;
+
+ xfs_ilock(ip, XFS_ILOCK_EXCL);
+ xfs_trans_ijoin(tp, ip, 0);
+
+ if (old_isize != i_size_read(VFS_I(ip)))
+ goto cancel;
+ if (!(ip->i_d.di_flags2 & XFS_DIFLAG2_REFLINK))
+ goto cancel;
+
+ fbno = 0;
+ end = XFS_B_TO_FSB(mp, old_isize);
+ while (end - fbno > 0) {
+ nmaps = 1;
+ /*
+ * Look for extents in the file. Skip holes, delalloc, or
+ * unwritten extents; they can't be reflinked.
+ */
+ error = xfs_bmapi_read(ip, fbno, end - fbno, map, &nmaps, 0);
+ if (error)
+ goto cancel;
+ if (nmaps == 0)
+ break;
+ if (map[0].br_startblock == HOLESTARTBLOCK ||
+ map[0].br_startblock == DELAYSTARTBLOCK ||
+ ISUNWRITTEN(&map[0]))
+ goto next;
+
+ map[1] = map[0];
+ while (map[1].br_blockcount) {
+ agno = XFS_FSB_TO_AGNO(mp, map[1].br_startblock);
+ agbno = XFS_FSB_TO_AGBNO(mp, map[1].br_startblock);
+ aglen = map[1].br_blockcount;
+
+ error = xfs_reflink_find_shared(mp, agno, agbno, aglen,
+ &rbno, &rlen, false);
+ if (error)
+ goto cancel;
+ /* Is there still a shared block here? */
+ if (rlen > 0) {
+ error = 0;
+ goto cancel;
+ }
+
+ map[1].br_blockcount -= aglen;
+ map[1].br_startoff += aglen;
+ map[1].br_startblock += aglen;
+ }
+
+next:
+ fbno = map[0].br_startoff + map[0].br_blockcount;
+ }
+
+ /*
+ * We didn't find any shared blocks so turn off the reflink flag.
+ * First, get rid of any leftover CoW mappings.
+ */
+ error = xfs_reflink_cancel_cow_blocks(ip, &tp, 0, NULLFILEOFF);
+ if (error)
+ goto cancel;
+
+ /* Clear the inode flag. */
+ trace_xfs_reflink_unset_inode_flag(ip);
+ ip->i_d.di_flags2 &= ~XFS_DIFLAG2_REFLINK;
+ xfs_trans_ijoin(tp, ip, 0);
+ xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+
+ error = xfs_trans_commit(tp);
+ if (error)
+ goto out;
+
+ xfs_iunlock(ip, XFS_ILOCK_EXCL);
+ return 0;
+cancel:
+ xfs_trans_cancel(tp);
+out:
+ xfs_iunlock(ip, XFS_ILOCK_EXCL);
+ return error;
+}
+
+/*
+ * Pre-COW all shared blocks within a given byte range of a file and turn off
+ * the reflink flag if we unshare all of the file's blocks.
+ */
+int
+xfs_reflink_unshare(
+ struct xfs_inode *ip,
+ xfs_off_t offset,
+ xfs_off_t len)
+{
+ struct xfs_mount *mp = ip->i_mount;
+ xfs_fileoff_t fbno;
+ xfs_filblks_t end;
+ xfs_off_t old_isize, isize;
+ int error;
+
+ if (!xfs_is_reflink_inode(ip))
+ return 0;
+
+ trace_xfs_reflink_unshare(ip, offset, len);
+
+ inode_dio_wait(VFS_I(ip));
+
+ /* Try to CoW the selected ranges */
+ xfs_ilock(ip, XFS_ILOCK_EXCL);
+ fbno = XFS_B_TO_FSB(mp, offset);
+ old_isize = isize = i_size_read(VFS_I(ip));
+ end = XFS_B_TO_FSB(mp, offset + len);
+ error = xfs_reflink_dirty_extents(ip, fbno, end, isize);
+ if (error)
+ goto out_unlock;
+ xfs_iunlock(ip, XFS_ILOCK_EXCL);
+
+ /* Wait for the IO to finish */
+ error = filemap_write_and_wait(VFS_I(ip)->i_mapping);
+ if (error)
+ goto out;
+
+ /* Turn off the reflink flag if we unshared the whole file */
+ if (offset == 0 && len == isize) {
+ error = xfs_reflink_try_clear_inode_flag(ip, old_isize);
+ if (error)
+ goto out;
+ }
+
+ return 0;
+
+out_unlock:
+ xfs_iunlock(ip, XFS_ILOCK_EXCL);
+out:
+ trace_xfs_reflink_unshare_error(ip, error, _RET_IP_);
+ return error;
+}
diff --git a/fs/xfs/xfs_reflink.h b/fs/xfs/xfs_reflink.h
index df82b20..ad4fc61 100644
--- a/fs/xfs/xfs_reflink.h
+++ b/fs/xfs/xfs_reflink.h
@@ -48,5 +48,7 @@ extern int xfs_reflink_recover_cow(struct xfs_mount *mp);
extern int xfs_reflink_remap_range(struct xfs_inode *src, xfs_off_t srcoff,
struct xfs_inode *dest, xfs_off_t destoff, xfs_off_t len,
unsigned int flags);
+extern int xfs_reflink_unshare(struct xfs_inode *ip, xfs_off_t offset,
+ xfs_off_t len);
#endif /* __XFS_REFLINK_H */
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
