Add general purpose flags field to file_info_t

Simplifies some task if we can just add a flag that a file has a framgent
or that it has already been detected as a duplicate.

Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>

6 files changed, 23 insertions, 18 deletions

diff --git a/include/fstree.h b/include/fstree.h
index 23634fc..687b693 100644
--- a/include/fstree.h
+++ b/include/fstree.h
@@ -22,6 +22,14 @@ typedef struct dir_info_t dir_info_t;
 typedef struct fstree_t fstree_t;
 typedef struct tree_xattr_t tree_xattr_t;
 
+enum {
+	FILE_FLAG_HAS_FRAGMENT = 0x01,
+
+	FILE_FLAG_FRAGMENT_IS_DUPLICATE = 0x02,
+
+	FILE_FLAG_BLOCKS_ARE_DUPLICATE = 0x04,
+};
+
 /* Encapsulates a set of key-value pairs attached to a tree_node_t */
 struct tree_xattr_t {
 	/* Number of key-value pairs */
@@ -80,6 +88,9 @@ struct file_info_t {
 
 	uint32_t fragment_chksum;
 
+	/* combination of FILE_FLAG_* flags */
+	uint32_t flags;
+
 	/* Stores data about each full data block. */
 	struct {
 		uint32_t chksum;
diff --git a/lib/fstree/mknode.c b/lib/fstree/mknode.c
index 18602e4..d343145 100644
--- a/lib/fstree/mknode.c
+++ b/lib/fstree/mknode.c
@@ -60,8 +60,6 @@ tree_node_t *fstree_mknode(fstree_t *fs, tree_node_t *parent, const char *name,
 	case S_IFREG:
 		n->data.file = (file_info_t *)n->payload;
 		n->data.file->size = sb->st_size;
-		n->data.file->fragment = 0xFFFFFFFF;
-		n->data.file->fragment_offset = 0xFFFFFFFF;
 		if (extra == NULL)
 			break;
 
diff --git a/lib/sqfs/data_reader.c b/lib/sqfs/data_reader.c
index 5facaa1..a7d605e 100644
--- a/lib/sqfs/data_reader.c
+++ b/lib/sqfs/data_reader.c
@@ -205,8 +205,7 @@ int data_reader_dump_file(data_reader_t *data, file_info_t *fi, int outfd,
 	size_t fragsz = fi->size % data->block_size;
 	size_t count = fi->size / data->block_size;
 
-	if (fragsz != 0 && (fi->fragment == 0xFFFFFFFF ||
-			    fi->fragment_offset == 0xFFFFFFFF)) {
+	if (fragsz != 0 && !(fi->flags & FILE_FLAG_HAS_FRAGMENT)) {
 		fragsz = 0;
 		++count;
 	}
diff --git a/lib/sqfs/data_writer.c b/lib/sqfs/data_writer.c
index 496d636..e9a6193 100644
--- a/lib/sqfs/data_writer.c
+++ b/lib/sqfs/data_writer.c
@@ -138,10 +138,10 @@ static file_info_t *fragment_by_chksum(file_info_t *fi, uint32_t chksum,
 			break;
 		}
 
-		if (it->fragment == 0xFFFFFFFF)
+		if (!(it->flags & FILE_FLAG_HAS_FRAGMENT))
 			continue;
 
-		if (it->fragment_offset == 0xFFFFFFFF)
+		if (it->flags & FILE_FLAG_FRAGMENT_IS_DUPLICATE)
 			continue;
 
 		if ((it->size % block_size) != frag_size)
@@ -158,11 +158,8 @@ static size_t get_block_count(file_info_t *fi, size_t block_size)
 {
 	size_t count = fi->size / block_size;
 
-	if ((fi->size % block_size != 0) &&
-	    (fi->fragment == 0xFFFFFFFF ||
-	     fi->fragment_offset == 0xFFFFFFFF)) {
+	if ((fi->size % block_size) && !(fi->flags & FILE_FLAG_HAS_FRAGMENT))
 		++count;
-	}
 
 	while (count > 0 && fi->blocks[count - 1].size == 0)
 		--count;
@@ -201,21 +198,15 @@ static uint64_t find_equal_blocks(file_info_t *file, file_info_t *list,
 			break;
 	}
 
-	location = 0;
-
 	for (it = list; it != NULL; it = it->next) {
 		if (it == file) {
 			it = NULL;
 			break;
 		}
 
-		/* XXX: list assumed to be processed in order, prevent us from
-		   re-checking files we know are duplicates */
-		if (it->startblock < location)
+		if (it->flags & FILE_FLAG_BLOCKS_ARE_DUPLICATE)
 			continue;
 
-		location = it->startblock;
-
 		cmp_blk_count = get_block_count(it, block_size);
 		if (cmp_blk_count == 0)
 			continue;
@@ -279,6 +270,8 @@ static int flush_data_block(data_writer_t *data, size_t size,
 	chksum = update_crc32(0, data->block, size);
 
 	if (size < data->super->block_size && !(flags & DW_DONT_FRAGMENT)) {
+		fi->flags |= FILE_FLAG_HAS_FRAGMENT;
+
 		ref = fragment_by_chksum(fi, chksum, size, list,
 					 data->super->block_size);
 
@@ -286,6 +279,7 @@ static int flush_data_block(data_writer_t *data, size_t size,
 			fi->fragment_chksum = ref->fragment_chksum;
 			fi->fragment_offset = ref->fragment_offset;
 			fi->fragment = ref->fragment;
+			fi->flags |= FILE_FLAG_FRAGMENT_IS_DUPLICATE;
 			return 0;
 		}
 
@@ -344,6 +338,7 @@ static int end_file(data_writer_t *data, file_info_t *fi,
 
 	if (ref > 0) {
 		fi->startblock = ref;
+		fi->flags |= FILE_FLAG_BLOCKS_ARE_DUPLICATE;
 
 		if (lseek(data->outfd, start, SEEK_SET) == (off_t)-1)
 			goto fail_seek;
diff --git a/lib/sqfs/tree_node_from_inode.c b/lib/sqfs/tree_node_from_inode.c
index 034218c..84ebbb2 100644
--- a/lib/sqfs/tree_node_from_inode.c
+++ b/lib/sqfs/tree_node_from_inode.c
@@ -50,6 +50,8 @@ static void copy_block_sizes(sqfs_inode_generic_t *inode, tree_node_t *out,
 		if (out->data.file->fragment == 0xFFFFFFFF ||
 		    out->data.file->fragment_offset == 0xFFFFFFFF) {
 			++block_count;
+		} else {
+			out->data.file->flags |= FILE_FLAG_HAS_FRAGMENT;
 		}
 	}
 
diff --git a/lib/sqfs/write_inode.c b/lib/sqfs/write_inode.c
index 726206e..efb6a7d 100644
--- a/lib/sqfs/write_inode.c
+++ b/lib/sqfs/write_inode.c
@@ -83,7 +83,7 @@ static int write_file_blocks(fstree_t *fs, file_info_t *fi, meta_writer_t *im)
 	uint32_t bs;
 
 	if ((fi->size % fs->block_size) != 0 &&
-	    (fi->fragment == 0xFFFFFFFF || fi->fragment_offset == 0xFFFFFFFF)) {
+	    !(fi->flags & FILE_FLAG_HAS_FRAGMENT)) {
 		++count;
 	}