Skip to contentSkip to conversation

fs/squashfs: Fix Coverity Scan defects

PATCH New Joao Marcos Costa 2 participants 3 messages Aug 19, 2020 – Aug 25, 2020 SquashFS file system

lore
Share & tools
lore
#0 lore raw

cover fs/squashfs: Fix Coverity Scan defects

Hello,

The following patch fixes most of the defects pointed by Coverity Scan. It depends on the previous patch series adding new decompression algorithms. Some of the defects may be out of SquashFS scope, because they come from a Linux's macro, but sanity checks were added as a possible solution for these tainted values. Please do not hesitate to report any remaining issues.

Best regards,
Joao

Joao Marcos Costa (1):
fs/squashfs: Fix Coverity Scan defects

 fs/squashfs/sqfs.c       | 40 +++++++++++++++++++++++++++++-----------
 fs/squashfs/sqfs_dir.c   |  3 +++
 fs/squashfs/sqfs_inode.c |  5 ++++-
 3 files changed, 36 insertions(+), 12 deletions(-)
Signature · 1 line
2.17.1
#1 lore raw

PATCH 1/1 fs/squashfs: Fix Coverity Scan defects

Fix defects such as uninitialized variables and untrusted pointer operations. Most part of the tainted variables and the related defects actually comes from Linux's macro get_unaligned_le**, extensively used in SquashFS code. Add sanity checks for those variables.

---
 fs/squashfs/sqfs.c       | 40 +++++++++++++++++++++++++++++-----------
 fs/squashfs/sqfs_dir.c   |  3 +++
 fs/squashfs/sqfs_inode.c |  5 ++++-
 3 files changed, 36 insertions(+), 12 deletions(-)
fs/squashfs/sqfs.c +29 -11
index 9bd7b98d88..f67f7c4a40 100644
@@ -154,6 +154,11 @@ static int sqfs_frag_lookup(u32 inode_fragment_index,
 	header = get_unaligned_le16(metadata_buffer + table_offset);
 	metadata = metadata_buffer + table_offset + SQFS_HEADER_SIZE;
 
+	if (!metadata) {
+		ret = -ENOMEM;
+		goto free_buffer;
+	}
+
 	entries = malloc(SQFS_METADATA_BLOCK_SIZE);
 	if (!entries) {
 		ret = -ENOMEM;
@@ -272,8 +277,8 @@ static int sqfs_join(char **strings, char *dest, int start, int end,
  */
 static int sqfs_tokenize(char **tokens, int count, const char *str)
 {
+	int i, j, ret = 0;
 	char *aux, *strc;
-	int i, j;
 
 	strc = strdup(str);
 	if (!strc)
@@ -282,8 +287,8 @@ static int sqfs_tokenize(char **tokens, int count, const char *str)
 	if (!strcmp(strc, "/")) {
 		tokens[0] = strdup(strc);
 		if (!tokens[0]) {
-			free(strc);
-			return -ENOMEM;
+			ret = -ENOMEM;
+			goto free_strc;
 		}
 	} else {
 		for (j = 0; j < count; j++) {
@@ -292,15 +297,16 @@ static int sqfs_tokenize(char **tokens, int count, const char *str)
 			if (!tokens[j]) {
 				for (i = 0; i < j; i++)
 					free(tokens[i]);
-				free(strc);
-				return -ENOMEM;
+				ret = -ENOMEM;
+				goto free_strc;
 			}
 		}
 	}
 
+free_strc:
 	free(strc);
 
-	return 0;
+	return ret;
 }
 
 /*
@@ -428,9 +434,9 @@ static int sqfs_search_dir(struct squashfs_dir_stream *dirs, char **token_list,
 {
 	struct squashfs_super_block *sblk = ctxt.sblk;
 	char *path, *target, **sym_tokens, *res, *rem;
+	struct squashfs_ldir_inode *ldir = NULL;
 	int j, ret, new_inode_number, offset;
 	struct squashfs_symlink_inode *sym;
-	struct squashfs_ldir_inode *ldir;
 	struct squashfs_dir_inode *dir;
 	struct fs_dir_stream *dirsp;
 	struct fs_dirent *dent;
@@ -630,7 +636,7 @@ static int sqfs_read_inode_table(unsigned char **inode_table)
 	int j, ret = 0, metablks_count;
 	unsigned char *src_table, *itb;
 	u32 src_len, dest_offset = 0;
-	unsigned long dest_len;
+	unsigned long dest_len = 0;
 	bool compressed;
 
 	table_size = get_unaligned_le64(&sblk->directory_table_start) -
@@ -685,6 +691,7 @@ static int sqfs_read_inode_table(unsigned char **inode_table)
 				goto free_itb;
 			}
 
+			dest_offset += dest_len;
 		} else {
 			memcpy(*inode_table + (j * SQFS_METADATA_BLOCK_SIZE),
 			       src_table, src_len);
@@ -694,7 +701,7 @@ static int sqfs_read_inode_table(unsigned char **inode_table)
 		 * Offsets to the decompression destination, to the metadata
 		 * buffer 'itb' and to the decompression source, respectively.
 		 */
-		dest_offset += dest_len;
+
 		table_offset += src_len + SQFS_HEADER_SIZE;
 		src_table += src_len + SQFS_HEADER_SIZE;
 	}
@@ -712,7 +719,7 @@ static int sqfs_read_directory_table(unsigned char **dir_table, u32 **pos_list)
 	int j, ret = 0, metablks_count = -1;
 	unsigned char *src_table, *dtb;
 	u32 src_len, dest_offset = 0;
-	unsigned long dest_len;
+	unsigned long dest_len = 0;
 	bool compressed;
 
 	/* DIRECTORY TABLE */
@@ -781,6 +788,8 @@ static int sqfs_read_directory_table(unsigned char **dir_table, u32 **pos_list)
 				dest_offset += dest_len;
 				break;
 			}
+
+			dest_offset += dest_len;
 		} else {
 			memcpy(*dir_table + (j * SQFS_METADATA_BLOCK_SIZE),
 			       src_table, src_len);
@@ -790,7 +799,6 @@ static int sqfs_read_directory_table(unsigned char **dir_table, u32 **pos_list)
 		 * Offsets to the decompression destination, to the metadata
 		 * buffer 'dtb' and to the decompression source, respectively.
 		 */
-		dest_offset += dest_len;
 		table_offset += src_len + SQFS_HEADER_SIZE;
 		src_table += src_len + SQFS_HEADER_SIZE;
 	}
@@ -1138,6 +1146,9 @@ static int sqfs_get_regfile_info(struct squashfs_reg_inode *reg,
 	finfo->start = get_unaligned_le32(&reg->start_block);
 	finfo->frag = SQFS_IS_FRAGMENTED(get_unaligned_le32(&reg->fragment));
 
+	if (finfo->size < 1 || finfo->offset < 0 || finfo->start < 0)
+		return -EINVAL;
+
 	if (finfo->frag) {
 		datablk_count = finfo->size / le32_to_cpu(blksz);
 		ret = sqfs_frag_lookup(get_unaligned_le32(&reg->fragment),
@@ -1145,6 +1156,8 @@ static int sqfs_get_regfile_info(struct squashfs_reg_inode *reg,
 		if (ret < 0)
 			return -EINVAL;
 		finfo->comp = true;
+		if (fentry->size < 1 || fentry->start < 0)
+			return -EINVAL;
 	} else {
 		datablk_count = DIV_ROUND_UP(finfo->size, le32_to_cpu(blksz));
 	}
@@ -1168,6 +1181,9 @@ static int sqfs_get_lregfile_info(struct squashfs_lreg_inode *lreg,
 	finfo->start = get_unaligned_le64(&lreg->start_block);
 	finfo->frag = SQFS_IS_FRAGMENTED(get_unaligned_le32(&lreg->fragment));
 
+	if (finfo->size < 1 || finfo->offset < 0 || finfo->start < 0)
+		return -EINVAL;
+
 	if (finfo->frag) {
 		datablk_count = finfo->size / le32_to_cpu(blksz);
 		ret = sqfs_frag_lookup(get_unaligned_le32(&lreg->fragment),
@@ -1175,6 +1191,8 @@ static int sqfs_get_lregfile_info(struct squashfs_lreg_inode *lreg,
 		if (ret < 0)
 			return -EINVAL;
 		finfo->comp = true;
+		if (fentry->size < 1 || fentry->start < 0)
+			return -EINVAL;
 	} else {
 		datablk_count = DIV_ROUND_UP(finfo->size, le32_to_cpu(blksz));
 	}
fs/squashfs/sqfs_dir.c +3 -0
index 5f7660aeae..00d2891e7d 100644
@@ -53,6 +53,9 @@ int sqfs_dir_offset(void *dir_i, u32 *m_list, int m_count)
 		return -EINVAL;
 	}
 
+	if (offset < 0)
+		return -EINVAL;
+
 	for (j = 0; j < m_count; j++) {
 		if (m_list[j] == start_block)
 			return (++j * SQFS_METADATA_BLOCK_SIZE) + offset;
fs/squashfs/sqfs_inode.c +4 -1
index b037a9b2ac..1387779a85 100644
@@ -138,11 +138,14 @@ void *sqfs_find_inode(void *inode_table, int inode_number, __le32 inode_count,
 int sqfs_read_metablock(unsigned char *file_mapping, int offset,
 			bool *compressed, u32 *data_size)
 {
-	unsigned char *data;
+	const unsigned char *data;
 	u16 header;
 
 	data = file_mapping + offset;
 	header = get_unaligned((u16 *)data);
+	if (!header || !data)
+		return -EINVAL;
+
 	*compressed = SQFS_COMPRESSED_METADATA(header);
 	*data_size = SQFS_METADATA_SIZE(header);
 
Signature · 1 line
2.17.1
#2 lore raw

PATCH 1/1 fs/squashfs: Fix Coverity Scan defects

Quoted 6 lines · Joao Marcos Costa

Fix defects such as uninitialized variables and untrusted pointer operations. Most part of the tainted variables and the related defects actually comes from Linux's macro get_unaligned_le**, extensively used in SquashFS code. Add sanity checks for those variables.

Applied to u-boot/master, thanks!

Signature · 8 lines
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 659 bytes
Desc: not available
URL: <https://lists.denx.de/pipermail/u-boot/attachments/20200825/e8f825eb/attachment.sig>