Skip to contentSkip to conversation

Fix squashfs failing to load sparse files

PATCH New Campbell Suter 2 participants 2 messages Nov 23, 2020 – Jan 20, 2021 SquashFS file system

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Fix squashfs failing to load sparse files

SquashFS supports sprase blocks in files - that is, if a given block is composed only of zeros, it's not written to the output file to save space and it's on-disk length field is set to zero to indicate that.

Previously the squashfs driver did not recognise that, and would attempt to read and decompress a zero-sized block, which obviously failed.

The following command may be used to create a file for testing:

cat <(dd if=/dev/urandom of=/dev/stdout bs=1M count=1) \
	<(dd if=/dev/zero of=/dev/stdout bs=1M count=1) \
	<(dd if=/dev/urandom of=/dev/stdout bs=1k count=200) >test_file

---

 fs/squashfs/sqfs.c | 54 ++++++++++++++++++++++++++++++----------------
 1 file changed, 35 insertions(+), 19 deletions(-)
fs/squashfs/sqfs.c +35 -19
index 608a2bb454..2516d920f8 100644
@@ -1309,7 +1309,7 @@ int sqfs_read(const char *filename, void *buf, loff_t offset, loff_t len,
 {
 	char *dir = NULL, *fragment_block, *datablock = NULL, *data_buffer = NULL;
 	char *fragment = NULL, *file = NULL, *resolved, *data;
-	u64 start, n_blks, table_size, data_offset, table_offset;
+	u64 start, n_blks, table_size, data_offset, table_offset, sparse_size;
 	int ret, j, i_number, datablk_count = 0;
 	struct squashfs_super_block *sblk = ctxt.sblk;
 	struct squashfs_fragment_block_entry frag_entry;
@@ -1443,28 +1443,43 @@ int sqfs_read(const char *filename, void *buf, loff_t offset, loff_t len,
 		n_blks = DIV_ROUND_UP(table_size + table_offset,
 				      ctxt.cur_dev->blksz);
 
-		data_buffer = malloc_cache_aligned(n_blks * ctxt.cur_dev->blksz);
+		/* Don't load any data for sparse blocks */
+		if (finfo.blk_sizes[j] == 0) {
+			n_blks = 0;
+			table_offset = 0;
+			data_buffer = NULL;
+			data = NULL;
+		} else {
+			data_buffer = malloc_cache_aligned(n_blks * ctxt.cur_dev->blksz);
 
-		if (!data_buffer) {
-			ret = -ENOMEM;
-			goto out;
-		}
+			if (!data_buffer) {
+				ret = -ENOMEM;
+				goto out;
+			}
 
-		ret = sqfs_disk_read(start, n_blks, data_buffer);
-		if (ret < 0) {
-			/*
-			 * Possible causes: too many data blocks or too large
-			 * SquashFS block size. Tip: re-compile the SquashFS
-			 * image with mksquashfs's -b <block_size> option.
-			 */
-			printf("Error: too many data blocks to be read.\n");
-			goto out;
-		}
+			ret = sqfs_disk_read(start, n_blks, data_buffer);
+			if (ret < 0) {
+				/*
+				 * Possible causes: too many data blocks or too large
+				 * SquashFS block size. Tip: re-compile the SquashFS
+				 * image with mksquashfs's -b <block_size> option.
+				 */
+				printf("Error: too many data blocks to be read.\n");
+				goto out;
+			}
 
-		data = data_buffer + table_offset;
+			data = data_buffer + table_offset;
+		}
 
 		/* Load the data */
-		if (SQFS_COMPRESSED_BLOCK(finfo.blk_sizes[j])) {
+		if (finfo.blk_sizes[j] == 0) {
+			/* This is a sparse block */
+			sparse_size = get_unaligned_le32(&sblk->block_size);
+			if ((*actread + sparse_size) > len)
+				sparse_size = len - *actread;
+			memset(buf + *actread, 0, sparse_size);
+			*actread += sparse_size;
+		} else if (SQFS_COMPRESSED_BLOCK(finfo.blk_sizes[j])) {
 			dest_len = get_unaligned_le32(&sblk->block_size);
 			ret = sqfs_decompress(&ctxt, datablock, &dest_len,
 					      data, table_size);
@@ -1483,7 +1498,8 @@ int sqfs_read(const char *filename, void *buf, loff_t offset, loff_t len,
 		}
 
 		data_offset += table_size;
-		free(data_buffer);
+		if (data_buffer)
+			free(data_buffer);
 		data_buffer = NULL;
 		if (*actread >= len)
 			break;
Signature · 1 line
2.28.0
#1 lore raw

PATCH 1/1 Fix squashfs failing to load sparse files

Quoted 7 lines · Campbell Suter

SquashFS supports sprase blocks in files - that is, if a given block is composed only of zeros, it's not written to the output file to save space and it's on-disk length field is set to zero to indicate that.

Previously the squashfs driver did not recognise that, and would attempt to read and decompress a zero-sized block, which obviously failed.

Quoted 7 lines

The following command may be used to create a file for testing:

cat <(dd if=/dev/urandom of=/dev/stdout bs=1M count=1) \
	<(dd if=/dev/zero of=/dev/stdout bs=1M count=1) \
	<(dd if=/dev/urandom of=/dev/stdout bs=1k count=200) >test_file

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/20210120/cf53ff9f/attachment.sig>