gfs2: Do not call iomap_zero_range beyond eof

Since commit eb65540aa9fc ("iomap: warn on zero range of a post-eof
folio"), iomap_zero_range() warns when asked to zero a folio beyond eof.
The warning triggers on the following code path:

  gfs2_fallocate(FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE)
    __gfs2_punch_hole()
      gfs2_block_zero_range()
        iomap_zero_range()

In __gfs2_punch_hole(), gfs2 zeroes out partial folios at the beginning
and at the end of the specified range, whether those folios are beyond
eof or not.  This may add folios to the page cache which are entirely
beyond eof, which isn't of any use.  Avoid that by truncating the range
to zero out at eof.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>

diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c
index 366516b98b3f31c822639e49e18241980c2a1e5b..070905493f8faa1f356d890da367c0b4bcea23e2 100644 (file)
--- a/fs/gfs2/bmap.c
+++ b/fs/gfs2/bmap.c
@@ -1296,10 +1296,12 @@ int gfs2_alloc_extent(struct inode *inode, u64 lblock, u64 *dblock,
  * uses iomap write to perform its actions, which begin their own transactions
  * (iomap_begin, get_folio, etc.)
  */
-static int gfs2_block_zero_range(struct inode *inode, loff_t from,
-                                unsigned int length)
+static int gfs2_block_zero_range(struct inode *inode, loff_t from, loff_t length)
 {
        BUG_ON(current->journal_info);
+       if (from >= inode->i_size)
+               return 0;
+       length = min(length, inode->i_size - from);
        return iomap_zero_range(inode, from, length, NULL, &gfs2_iomap_ops,
                        NULL);
 }