diff --git a/mm/shmem.c b/mm/shmem.c
index 4ae47f54c82247bbcc3bfc9419ec0a1f79474512..c919ed578f0a2909b1d149bb3ed4793a9ec6845b 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -80,6 +80,7 @@
 enum sgp_type {
 	SGP_READ,	/* don't exceed i_size, don't allocate page */
 	SGP_CACHE,	/* don't exceed i_size, may allocate page */
+	SGP_DIRTY,	/* like SGP_CACHE, but set new page dirty */
 	SGP_WRITE,	/* may exceed i_size, may allocate page */
 };
 
@@ -1333,6 +1334,8 @@ static int shmem_getpage(struct inode *inode, unsigned long idx,
 		clear_highpage(filepage);
 		flush_dcache_page(filepage);
 		SetPageUptodate(filepage);
+		if (sgp == SGP_DIRTY)
+			set_page_dirty(filepage);
 	}
 done:
 	*pagep = filepage;
@@ -1518,6 +1521,15 @@ static void do_shmem_file_read(struct file *filp, loff_t *ppos, read_descriptor_
 	struct inode *inode = filp->f_path.dentry->d_inode;
 	struct address_space *mapping = inode->i_mapping;
 	unsigned long index, offset;
+	enum sgp_type sgp = SGP_READ;
+
+	/*
+	 * Might this read be for a stacking filesystem?  Then when reading
+	 * holes of a sparse file, we actually need to allocate those pages,
+	 * and even mark them dirty, so it cannot exceed the max_blocks limit.
+	 */
+	if (segment_eq(get_fs(), KERNEL_DS))
+		sgp = SGP_DIRTY;
 
 	index = *ppos >> PAGE_CACHE_SHIFT;
 	offset = *ppos & ~PAGE_CACHE_MASK;
@@ -1536,7 +1548,7 @@ static void do_shmem_file_read(struct file *filp, loff_t *ppos, read_descriptor_
 				break;
 		}
 
-		desc->error = shmem_getpage(inode, index, &page, SGP_READ, NULL);
+		desc->error = shmem_getpage(inode, index, &page, sgp, NULL);
 		if (desc->error) {
 			if (desc->error == -EINVAL)
 				desc->error = 0;