Merge tag 'fsverity-for-linus' of git://git.kernel.org/pub/scm/fs/fsc…

…rypt/fscrypt

Pull fsverity updates from Eric Biggers:

 - Optimize fs-verity sequential read performance by implementing
   readahead of Merkle tree pages. This allows the Merkle tree to be
   read in larger chunks.

 - Optimize FS_IOC_ENABLE_VERITY performance in the uncached case by
   implementing readahead of data pages.

 - Allocate the hash requests from a mempool in order to eliminate the
   possibility of allocation failures during I/O.

* tag 'fsverity-for-linus' of git://git.kernel.org/pub/scm/fs/fscrypt/fscrypt:
  fs-verity: use u64_to_user_ptr()
  fs-verity: use mempool for hash requests
  fs-verity: implement readahead of Merkle tree pages
  fs-verity: implement readahead for FS_IOC_ENABLE_VERITY

fs/ext4/verity.c

@@ -342,12 +342,55 @@ static int ext4_get_verity_descriptor(struct inode *inode, void *buf,
 	return desc_size;
 }
 
+/*
+ * Prefetch some pages from the file's Merkle tree.
+ *
+ * This is basically a stripped-down version of __do_page_cache_readahead()
+ * which works on pages past i_size.
+ */
+static void ext4_merkle_tree_readahead(struct address_space *mapping,
+				       pgoff_t start_index, unsigned long count)
+{
+	LIST_HEAD(pages);
+	unsigned int nr_pages = 0;
+	struct page *page;
+	pgoff_t index;
+	struct blk_plug plug;
+
+	for (index = start_index; index < start_index + count; index++) {
+		page = xa_load(&mapping->i_pages, index);
+		if (!page || xa_is_value(page)) {
+			page = __page_cache_alloc(readahead_gfp_mask(mapping));
+			if (!page)
+				break;
+			page->index = index;
+			list_add(&page->lru, &pages);
+			nr_pages++;
+		}
+	}
+	blk_start_plug(&plug);
+	ext4_mpage_readpages(mapping, &pages, NULL, nr_pages, true);
+	blk_finish_plug(&plug);
+}
+
 static struct page *ext4_read_merkle_tree_page(struct inode *inode,
-					       pgoff_t index)
+					       pgoff_t index,
+					       unsigned long num_ra_pages)
 {
+	struct page *page;
+
 	index += ext4_verity_metadata_pos(inode) >> PAGE_SHIFT;
 
-	return read_mapping_page(inode->i_mapping, index, NULL);
+	page = find_get_page_flags(inode->i_mapping, index, FGP_ACCESSED);
+	if (!page || !PageUptodate(page)) {
+		if (page)
+			put_page(page);
+		else if (num_ra_pages > 1)
+			ext4_merkle_tree_readahead(inode->i_mapping, index,
+						   num_ra_pages);
+		page = read_mapping_page(inode->i_mapping, index, NULL);
+	}
+	return page;
 }
 
 static int ext4_write_merkle_tree_block(struct inode *inode, const void *buf,

fs/f2fs/data.c

@@ -1881,7 +1881,7 @@ static int f2fs_read_single_page(struct inode *inode, struct page *page,
  * use ->readpage() or do the necessary surgery to decouple ->readpages()
  * from read-ahead.
  */
-static int f2fs_mpage_readpages(struct address_space *mapping,
+int f2fs_mpage_readpages(struct address_space *mapping,
 			struct list_head *pages, struct page *page,
 			unsigned nr_pages, bool is_readahead)
 {

fs/f2fs/f2fs.h

@@ -3229,6 +3229,9 @@ int f2fs_reserve_new_block(struct dnode_of_data *dn);
 int f2fs_get_block(struct dnode_of_data *dn, pgoff_t index);
 int f2fs_preallocate_blocks(struct kiocb *iocb, struct iov_iter *from);
 int f2fs_reserve_block(struct dnode_of_data *dn, pgoff_t index);
+int f2fs_mpage_readpages(struct address_space *mapping,
+			struct list_head *pages, struct page *page,
+			unsigned nr_pages, bool is_readahead);
 struct page *f2fs_get_read_data_page(struct inode *inode, pgoff_t index,
 			int op_flags, bool for_write);
 struct page *f2fs_find_data_page(struct inode *inode, pgoff_t index);

fs/f2fs/verity.c

@@ -222,12 +222,55 @@ static int f2fs_get_verity_descriptor(struct inode *inode, void *buf,
 	return size;
 }
 
+/*
+ * Prefetch some pages from the file's Merkle tree.
+ *
+ * This is basically a stripped-down version of __do_page_cache_readahead()
+ * which works on pages past i_size.
+ */
+static void f2fs_merkle_tree_readahead(struct address_space *mapping,
+				       pgoff_t start_index, unsigned long count)
+{
+	LIST_HEAD(pages);
+	unsigned int nr_pages = 0;
+	struct page *page;
+	pgoff_t index;
+	struct blk_plug plug;
+
+	for (index = start_index; index < start_index + count; index++) {
+		page = xa_load(&mapping->i_pages, index);
+		if (!page || xa_is_value(page)) {
+			page = __page_cache_alloc(readahead_gfp_mask(mapping));
+			if (!page)
+				break;
+			page->index = index;
+			list_add(&page->lru, &pages);
+			nr_pages++;
+		}
+	}
+	blk_start_plug(&plug);
+	f2fs_mpage_readpages(mapping, &pages, NULL, nr_pages, true);
+	blk_finish_plug(&plug);
+}
+
 static struct page *f2fs_read_merkle_tree_page(struct inode *inode,
-					       pgoff_t index)
+					       pgoff_t index,
+					       unsigned long num_ra_pages)
 {
+	struct page *page;
+
 	index += f2fs_verity_metadata_pos(inode) >> PAGE_SHIFT;
 
-	return read_mapping_page(inode->i_mapping, index, NULL);
+	page = find_get_page_flags(inode->i_mapping, index, FGP_ACCESSED);
+	if (!page || !PageUptodate(page)) {
+		if (page)
+			put_page(page);
+		else if (num_ra_pages > 1)
+			f2fs_merkle_tree_readahead(inode->i_mapping, index,
+						   num_ra_pages);
+		page = read_mapping_page(inode->i_mapping, index, NULL);
+	}
+	return page;
 }
 
 static int f2fs_write_merkle_tree_block(struct inode *inode, const void *buf,

fs/verity/enable.c

@@ -8,18 +8,48 @@
 #include "fsverity_private.h"
 
 #include <crypto/hash.h>
+#include <linux/backing-dev.h>
 #include <linux/mount.h>
 #include <linux/pagemap.h>
 #include <linux/sched/signal.h>
 #include <linux/uaccess.h>
 
-static int build_merkle_tree_level(struct inode *inode, unsigned int level,
+/*
+ * Read a file data page for Merkle tree construction.  Do aggressive readahead,
+ * since we're sequentially reading the entire file.
+ */
+static struct page *read_file_data_page(struct file *filp, pgoff_t index,
+					struct file_ra_state *ra,
+					unsigned long remaining_pages)
+{
+	struct page *page;
+
+	page = find_get_page_flags(filp->f_mapping, index, FGP_ACCESSED);
+	if (!page || !PageUptodate(page)) {
+		if (page)
+			put_page(page);
+		else
+			page_cache_sync_readahead(filp->f_mapping, ra, filp,
+						  index, remaining_pages);
+		page = read_mapping_page(filp->f_mapping, index, NULL);
+		if (IS_ERR(page))
+			return page;
+	}
+	if (PageReadahead(page))
+		page_cache_async_readahead(filp->f_mapping, ra, filp, page,
+					   index, remaining_pages);
+	return page;
+}
+
+static int build_merkle_tree_level(struct file *filp, unsigned int level,
 				   u64 num_blocks_to_hash,
 				   const struct merkle_tree_params *params,
 				   u8 *pending_hashes,
 				   struct ahash_request *req)
 {
+	struct inode *inode = file_inode(filp);
 	const struct fsverity_operations *vops = inode->i_sb->s_vop;
+	struct file_ra_state ra = { 0 };
 	unsigned int pending_size = 0;
 	u64 dst_block_num;
 	u64 i;
@@ -36,6 +66,8 @@ static int build_merkle_tree_level(struct inode *inode, unsigned int level,
 		dst_block_num = 0; /* unused */
 	}
 
+	file_ra_state_init(&ra, filp->f_mapping);
+
 	for (i = 0; i < num_blocks_to_hash; i++) {
 		struct page *src_page;
 
@@ -45,7 +77,8 @@ static int build_merkle_tree_level(struct inode *inode, unsigned int level,
 
 		if (level == 0) {
 			/* Leaf: hashing a data block */
-			src_page = read_mapping_page(inode->i_mapping, i, NULL);
+			src_page = read_file_data_page(filp, i, &ra,
+						       num_blocks_to_hash - i);
 			if (IS_ERR(src_page)) {
 				err = PTR_ERR(src_page);
 				fsverity_err(inode,
@@ -54,9 +87,14 @@ static int build_merkle_tree_level(struct inode *inode, unsigned int level,
 				return err;
 			}
 		} else {
+			unsigned long num_ra_pages =
+				min_t(unsigned long, num_blocks_to_hash - i,
+				      inode->i_sb->s_bdi->io_pages);
+
 			/* Non-leaf: hashing hash block from level below */
 			src_page = vops->read_merkle_tree_page(inode,
-					params->level_start[level - 1] + i);
+					params->level_start[level - 1] + i,
+					num_ra_pages);
 			if (IS_ERR(src_page)) {
 				err = PTR_ERR(src_page);
 				fsverity_err(inode,
@@ -103,17 +141,18 @@ static int build_merkle_tree_level(struct inode *inode, unsigned int level,
 }
 
 /*
- * Build the Merkle tree for the given inode using the given parameters, and
+ * Build the Merkle tree for the given file using the given parameters, and
  * return the root hash in @root_hash.
  *
  * The tree is written to a filesystem-specific location as determined by the
  * ->write_merkle_tree_block() method.  However, the blocks that comprise the
  * tree are the same for all filesystems.
  */
-static int build_merkle_tree(struct inode *inode,
+static int build_merkle_tree(struct file *filp,
 			     const struct merkle_tree_params *params,
 			     u8 *root_hash)
 {
+	struct inode *inode = file_inode(filp);
 	u8 *pending_hashes;
 	struct ahash_request *req;
 	u64 blocks;
@@ -126,9 +165,11 @@ static int build_merkle_tree(struct inode *inode,
 		return 0;
 	}
 
+	/* This allocation never fails, since it's mempool-backed. */
+	req = fsverity_alloc_hash_request(params->hash_alg, GFP_KERNEL);
+
 	pending_hashes = kmalloc(params->block_size, GFP_KERNEL);
-	req = ahash_request_alloc(params->hash_alg->tfm, GFP_KERNEL);
-	if (!pending_hashes || !req)
+	if (!pending_hashes)
 		goto out;
 
 	/*
@@ -139,7 +180,7 @@ static int build_merkle_tree(struct inode *inode,
 	blocks = (inode->i_size + params->block_size - 1) >>
 		 params->log_blocksize;
 	for (level = 0; level <= params->num_levels; level++) {
-		err = build_merkle_tree_level(inode, level, blocks, params,
+		err = build_merkle_tree_level(filp, level, blocks, params,
 					      pending_hashes, req);
 		if (err)
 			goto out;
@@ -150,7 +191,7 @@ static int build_merkle_tree(struct inode *inode,
 	err = 0;
 out:
 	kfree(pending_hashes);
-	ahash_request_free(req);
+	fsverity_free_hash_request(params->hash_alg, req);
 	return err;
 }
 
@@ -175,8 +216,7 @@ static int enable_verity(struct file *filp,
 
 	/* Get the salt if the user provided one */
 	if (arg->salt_size &&
-	    copy_from_user(desc->salt,
-			   (const u8 __user *)(uintptr_t)arg->salt_ptr,
+	    copy_from_user(desc->salt, u64_to_user_ptr(arg->salt_ptr),
 			   arg->salt_size)) {
 		err = -EFAULT;
 		goto out;
@@ -185,8 +225,7 @@ static int enable_verity(struct file *filp,
 
 	/* Get the signature if the user provided one */
 	if (arg->sig_size &&
-	    copy_from_user(desc->signature,
-			   (const u8 __user *)(uintptr_t)arg->sig_ptr,
+	    copy_from_user(desc->signature, u64_to_user_ptr(arg->sig_ptr),
 			   arg->sig_size)) {
 		err = -EFAULT;
 		goto out;
@@ -227,7 +266,7 @@ static int enable_verity(struct file *filp,
 	 */
 	pr_debug("Building Merkle tree...\n");
 	BUILD_BUG_ON(sizeof(desc->root_hash) < FS_VERITY_MAX_DIGEST_SIZE);
-	err = build_merkle_tree(inode, &params, desc->root_hash);
+	err = build_merkle_tree(filp, &params, desc->root_hash);
 	if (err) {
 		fsverity_err(inode, "Error %d building Merkle tree", err);
 		goto rollback;

fs/verity/fsverity_private.h

@@ -16,6 +16,7 @@
 
 #include <crypto/sha.h>
 #include <linux/fsverity.h>
+#include <linux/mempool.h>
 
 struct ahash_request;
 
@@ -37,11 +38,12 @@ struct fsverity_hash_alg {
 	const char *name;	  /* crypto API name, e.g. sha256 */
 	unsigned int digest_size; /* digest size in bytes, e.g. 32 for SHA-256 */
 	unsigned int block_size;  /* block size in bytes, e.g. 64 for SHA-256 */
+	mempool_t req_pool;	  /* mempool with a preallocated hash request */
 };
 
 /* Merkle tree parameters: hash algorithm, initial hash state, and topology */
 struct merkle_tree_params {
-	const struct fsverity_hash_alg *hash_alg; /* the hash algorithm */
+	struct fsverity_hash_alg *hash_alg; /* the hash algorithm */
 	const u8 *hashstate;		/* initial hash state or NULL */
 	unsigned int digest_size;	/* same as hash_alg->digest_size */
 	unsigned int block_size;	/* size of data and tree blocks */
@@ -50,6 +52,7 @@ struct merkle_tree_params {
 	unsigned int log_arity;		/* log2(hashes_per_block) */
 	unsigned int num_levels;	/* number of levels in Merkle tree */
 	u64 tree_size;			/* Merkle tree size in bytes */
+	unsigned long level0_blocks;	/* number of blocks in tree level 0 */
 
 	/*
 	 * Starting block index for each tree level, ordered from leaf level (0)
@@ -114,14 +117,18 @@ struct fsverity_signed_digest {
 
 extern struct fsverity_hash_alg fsverity_hash_algs[];
 
-const struct fsverity_hash_alg *fsverity_get_hash_alg(const struct inode *inode,
-						      unsigned int num);
-const u8 *fsverity_prepare_hash_state(const struct fsverity_hash_alg *alg,
+struct fsverity_hash_alg *fsverity_get_hash_alg(const struct inode *inode,
+						unsigned int num);
+struct ahash_request *fsverity_alloc_hash_request(struct fsverity_hash_alg *alg,
+						  gfp_t gfp_flags);
+void fsverity_free_hash_request(struct fsverity_hash_alg *alg,
+				struct ahash_request *req);
+const u8 *fsverity_prepare_hash_state(struct fsverity_hash_alg *alg,
 				      const u8 *salt, size_t salt_size);
 int fsverity_hash_page(const struct merkle_tree_params *params,
 		       const struct inode *inode,
 		       struct ahash_request *req, struct page *page, u8 *out);
-int fsverity_hash_buffer(const struct fsverity_hash_alg *alg,
+int fsverity_hash_buffer(struct fsverity_hash_alg *alg,
 			 const void *data, size_t size, u8 *out);
 void __init fsverity_check_hash_algs(void);