[PATCH v13 05/23] fsverity: improve flushing performance of fsverity_fill_zerohash

The current version calls flush_dcache_folio(), in memcpy_to_folio(), to
flush whole folio on every digest (which is 128 for 4k). Open code folio
mapping and flushing to copy all digests at once.

Reported-by: Eric Biggers <[email protected]>
Link: https://lore.kernel.org/linux-fsdevel/20260401222717.GH2466@quark/
Signed-off-by: Andrey Albershteyn <[email protected]>
---
 fs/verity/pagecache.c | 24 +++++++++++++++++++++---
 1 file changed, 21 insertions(+), 3 deletions(-)

diff --git a/fs/verity/pagecache.c b/fs/verity/pagecache.c
index 9d82e6b74ba1..270abd652e48 100644
--- a/fs/verity/pagecache.c
+++ b/fs/verity/pagecache.c
@@ -68,14 +68,32 @@ EXPORT_SYMBOL_GPL(generic_readahead_merkle_tree);
 void fsverity_fill_zerohash(struct folio *folio, size_t offset, size_t len,
 			      struct fsverity_info *vi)
 {
-	size_t off = offset;
+	void *vaddr;
+	void *to;
+	size_t chunk;
 
 	WARN_ON_ONCE(!IS_ALIGNED(offset, vi->tree_params.digest_size));
 	WARN_ON_ONCE(!IS_ALIGNED(len, vi->tree_params.digest_size));
+	WARN_ON_ONCE(offset + len > folio_size(folio));
 
-	for (; off < (offset + len); off += vi->tree_params.digest_size)
-		memcpy_to_folio(folio, off, vi->tree_params.zero_digest,
+	do {
+		vaddr = kmap_local_folio(folio, offset);
+		chunk = len;
+		to = vaddr;
+
+		if (folio_test_partial_kmap(folio) &&
+		    chunk > PAGE_SIZE - offset_in_page(offset))
+			chunk = PAGE_SIZE - offset_in_page(offset);
+		for (; to < (vaddr + chunk); to += vi->tree_params.digest_size)
+			memcpy(to, vi->tree_params.zero_digest,
 				vi->tree_params.digest_size);
+		kunmap_local(vaddr);
+
+		offset += chunk;
+		len -= chunk;
+	} while (len > 0);
+
+	flush_dcache_folio(folio);
 }
 EXPORT_SYMBOL_GPL(fsverity_fill_zerohash);
 
-- 
2.54.0