xfs: Introduce XFS_IOC_ZERO_RANGE

XFS_IOC_ZERO_RANGE is the equivalent of an atomic XFS_IOC_UNRESVSP/
XFS_IOC_RESVSP call pair. It enabled ranges of written data to be
turned into zeroes without requiring IO or having to free and
reallocate the extents in the range given as would occur if we had
to punch and then preallocate them separately.  This enables
applications to zero parts of files very quickly without changing
the layout of the files in any way.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>

6 files changed, 27 insertions, 9 deletions

diff --git a/fs/xfs/linux-2.6/xfs_ioctl.c b/fs/xfs/linux-2.6/xfs_ioctl.c
index 3b9e626f7cd1..03aa908a9cb9 100644
--- a/fs/xfs/linux-2.6/xfs_ioctl.c
+++ b/fs/xfs/linux-2.6/xfs_ioctl.c
@@ -1301,7 +1301,8 @@ xfs_file_ioctl(
         case XFS_IOC_ALLOCSP64:
         case XFS_IOC_FREESP64:
         case XFS_IOC_RESVSP64:
-        case XFS_IOC_UNRESVSP64: {
+        case XFS_IOC_UNRESVSP64:
+        case XFS_IOC_ZERO_RANGE: {
                 xfs_flock64_t           bf;
 
                 if (copy_from_user(&bf, arg, sizeof(bf)))
diff --git a/fs/xfs/linux-2.6/xfs_ioctl32.c b/fs/xfs/linux-2.6/xfs_ioctl32.c
index 6c83f7f62dc9..464bcc76e980 100644
--- a/fs/xfs/linux-2.6/xfs_ioctl32.c
+++ b/fs/xfs/linux-2.6/xfs_ioctl32.c
@@ -574,6 +574,7 @@ xfs_file_compat_ioctl(
         case XFS_IOC_FSGEOMETRY_V1:
         case XFS_IOC_FSGROWFSDATA:
         case XFS_IOC_FSGROWFSRT:
+        case XFS_IOC_ZERO_RANGE:
                 return xfs_file_ioctl(filp, cmd, p);
 #else
         case XFS_IOC_ALLOCSP_32:
diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c
index f90dadd5a968..5e33b7862d41 100644
--- a/fs/xfs/xfs_bmap.c
+++ b/fs/xfs/xfs_bmap.c
@@ -4744,8 +4744,12 @@ xfs_bmapi(
                  * Check if writing previously allocated but
                  * unwritten extents.
                  */
-                if (wr && mval->br_state == XFS_EXT_UNWRITTEN &&
-                    ((flags & (XFS_BMAPI_PREALLOC|XFS_BMAPI_DELAY)) == 0)) {
+                if (wr &&
+                    ((mval->br_state == XFS_EXT_UNWRITTEN &&
+                      ((flags & (XFS_BMAPI_PREALLOC|XFS_BMAPI_DELAY)) == 0)) ||
+                     (mval->br_state == XFS_EXT_NORM &&
+                      ((flags & (XFS_BMAPI_PREALLOC|XFS_BMAPI_CONVERT)) ==
+                                (XFS_BMAPI_PREALLOC|XFS_BMAPI_CONVERT))))) {
                         /*
                          * Modify (by adding) the state flag, if writing.
                          */
@@ -4757,7 +4761,9 @@ xfs_bmapi(
                                         *firstblock;
                                 cur->bc_private.b.flist = flist;
                         }
-                        mval->br_state = XFS_EXT_NORM;
+                        mval->br_state = (mval->br_state == XFS_EXT_UNWRITTEN)
+                                                ? XFS_EXT_NORM
+                                                : XFS_EXT_UNWRITTEN;
                         error = xfs_bmap_add_extent(ip, lastx, &cur, mval,
                                 firstblock, flist, &tmp_logflags,
                                 whichfork, (flags & XFS_BMAPI_RSVBLOCKS));
diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h
index b13569a6179b..71ec9b6ecdfc 100644
--- a/fs/xfs/xfs_bmap.h
+++ b/fs/xfs/xfs_bmap.h
@@ -74,9 +74,12 @@ typedef	struct xfs_bmap_free
 #define XFS_BMAPI_IGSTATE       0x080   /* Ignore state - */
                                         /* combine contig. space */
 #define XFS_BMAPI_CONTIG        0x100   /* must allocate only one extent */
-#define XFS_BMAPI_CONVERT       0x200   /* unwritten extent conversion - */
-                                        /* need write cache flushing and no */
-                                        /* additional allocation alignments */
+/*
+ * unwritten extent conversion - this needs write cache flushing and no additional
+ * allocation alignments. When specified with XFS_BMAPI_PREALLOC it converts
+ * from written to unwritten, otherwise convert from unwritten to written.
+ */
+#define XFS_BMAPI_CONVERT       0x200
 
 #define XFS_BMAPI_FLAGS \
         { XFS_BMAPI_WRITE,      "WRITE" }, \
diff --git a/fs/xfs/xfs_fs.h b/fs/xfs/xfs_fs.h
index 87c2e9d02288..10f6093671b0 100644
--- a/fs/xfs/xfs_fs.h
+++ b/fs/xfs/xfs_fs.h
@@ -448,6 +448,7 @@ typedef struct xfs_handle {
 /*      XFS_IOC_SETBIOSIZE ---- deprecated 46      */
 /*      XFS_IOC_GETBIOSIZE ---- deprecated 47      */
 #define XFS_IOC_GETBMAPX        _IOWR('X', 56, struct getbmap)
+#define XFS_IOC_ZERO_RANGE      _IOW ('X', 57, struct xfs_flock64)
 
 /*
  * ioctl commands that replace IRIX syssgi()'s
diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c
index 4c7c7bfb2b2f..dc6e4fb8bbc9 100644
--- a/fs/xfs/xfs_vnodeops.c
+++ b/fs/xfs/xfs_vnodeops.c
@@ -2272,7 +2272,7 @@ xfs_alloc_file_space(
         count = len;
         imapp = &imaps[0];
         nimaps = 1;
-        bmapi_flag = XFS_BMAPI_WRITE | (alloc_type ? XFS_BMAPI_PREALLOC : 0);
+        bmapi_flag = XFS_BMAPI_WRITE | alloc_type;
         startoffset_fsb = XFS_B_TO_FSBT(mp, offset);
         allocatesize_fsb = XFS_B_TO_FSB(mp, count);
 
@@ -2711,6 +2711,7 @@ xfs_change_file_space(
         xfs_off_t       llen;
         xfs_trans_t     *tp;
         struct iattr    iattr;
+        int             prealloc_type;
 
         if (!S_ISREG(ip->i_d.di_mode))
                 return XFS_ERROR(EINVAL);
@@ -2753,12 +2754,17 @@ xfs_change_file_space(
          * size to be changed.
          */
         setprealloc = clrprealloc = 0;
+        prealloc_type = XFS_BMAPI_PREALLOC;
 
         switch (cmd) {
+        case XFS_IOC_ZERO_RANGE:
+                prealloc_type |= XFS_BMAPI_CONVERT;
+                xfs_tosspages(ip, startoffset, startoffset + bf->l_len, 0);
+                /* FALLTHRU */
         case XFS_IOC_RESVSP:
         case XFS_IOC_RESVSP64:
                 error = xfs_alloc_file_space(ip, startoffset, bf->l_len,
-                                                                1, attr_flags);
+                                                prealloc_type, attr_flags);
                 if (error)
                         return error;
                 setprealloc = 1;