mksquashfs can read uid/gid/file type from fakeroot(1) database

With this patch I can easily pack chroots built in a fakeroot environment
without running mksquashfs under fakeroot. Instead I can use
`-fakerootdb .fakedata` to get ownership/permissions/device nodes right.

It's possible to convert fakeroot database into mksquashfs' pseudo files
definitions but using the fakeroot database directly is more efficient
(fakeroot uses device/inode number to identify the object) and requires
less code.

Author: asheplyakov

squashfs-tools/Makefile

@@ -153,6 +153,7 @@ INSTALL_DIR = /usr/local/bin
 
 MKSQUASHFS_OBJS = mksquashfs.o read_fs.o action.o swap.o pseudo.o compressor.o \
 	sort.o progressbar.o info.o restore.o process_fragments.o \
+	fakerootdb.o \
 	caches-queues-lists.o reader.o
 
 UNSQUASHFS_OBJS = unsquashfs.o unsquash-1.o unsquash-2.o unsquash-3.o \

squashfs-tools/fakerootdb.c

@@ -0,0 +1,89 @@
+#include <stdlib.h>
+#include <string.h>
+#include <errno.h>
+#include "fakerootdb.h"
+
+#define FAKEROOT_ENTRY_FIELDS	7
+#define FAKEROOT_ENTRY_FMT "dev=%lx,ino=%lu,mode=%o,uid=%u,gid=%u,nlink=%lu,rdev=%lu"
+
+static int compare_by_dev_ino(const void *px, const void *py)
+{
+	struct stat const* const x = px;
+	struct stat const* const y = py;
+
+	if (x->st_dev < y->st_dev)
+		return -1;
+	else if (x->st_dev > y->st_dev)
+		return 1;
+	else if (x->st_ino < y->st_ino)
+		return -1;
+	else if (x->st_ino > y->st_ino)
+		return 1;
+	else
+		return 0;
+}
+
+int fakeroot_read_db(FILE *fakedata, struct fakerootdb *db)
+{
+	struct stat elt, *d;
+	int n;
+	if (!db)
+		return -EINVAL;
+	if (db->db) {
+		free(db->db);
+		db->db = NULL;
+		db->count = 0;
+	}
+	while (!feof(fakedata)) {
+		if (ferror(fakedata))
+			return -EIO;
+		memset(&elt, 0, sizeof(elt));
+		n = fscanf(fakedata,
+			   FAKEROOT_ENTRY_FMT "\n",
+			   &elt.st_dev,
+			   &elt.st_ino,
+			   &elt.st_mode,
+			   &elt.st_uid,
+			   &elt.st_gid,
+			   &elt.st_nlink,
+			   &elt.st_rdev);
+		if (n != FAKEROOT_ENTRY_FIELDS)
+			return -EINVAL;
+
+		/* skip uid = gid = 0 entries, unless they are device nodes.
+		 * fakeroot assumes uid = gid = 0 by default */
+		if (elt.st_uid == 0 && elt.st_gid == 0 && elt.st_rdev == 0)
+			continue;
+
+		d = realloc(db->db, (db->count + 1)*sizeof(elt));
+		if (!d)
+			return -ENOMEM;
+		memcpy(&d[db->count], &elt, sizeof(elt));
+		db->db = d;
+		db->count += 1;
+	}
+	qsort(db->db, db->count, sizeof(elt), compare_by_dev_ino);
+	return 0;
+}
+
+void fakeroot_override_stat(struct stat *st, const struct fakerootdb *db)
+{
+	struct stat key;
+	struct stat const* o = NULL;
+	if (!db|| !db->db || db->count == 0)
+		return;
+	memset(&key, 0, sizeof(key));
+	key.st_dev = st->st_dev;
+	key.st_ino = st->st_ino;
+	o = bsearch(&key, db->db, db->count, sizeof(key), compare_by_dev_ino);
+	if (o) {
+		st->st_mode = o->st_mode;
+		st->st_uid = o->st_uid;
+		st->st_gid = o->st_gid;
+		st->st_rdev = o->st_rdev;
+	} else {
+		/* fakeroot sets uid=gid=0 if the object is not in the DB */
+		st->st_uid = 0;
+		st->st_gid = 0;
+	}
+}

squashfs-tools/fakerootdb.h

@@ -0,0 +1,16 @@
+#ifndef SQUASHFS_FAKEROOTDB_H
+#define SQUASHFS_FAKEROOTDB_H
+#include <stdio.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+
+struct fakerootdb {
+	struct stat *db;
+	size_t count;
+};
+
+int fakeroot_read_db(FILE *fakedata, struct fakerootdb *db);
+
+void fakeroot_override_stat(struct stat *st, const struct fakerootdb *fakerootdb);
+
+#endif /* SQUASHFS_FAKEROOTDB_H */

squashfs-tools/mksquashfs.c

@@ -72,6 +72,7 @@
 #include "restore.h"
 #include "process_fragments.h"
 #include "fnmatch_compat.h"
+#include "fakerootdb.h"
 
 int delete = FALSE;
 int quiet = FALSE;
@@ -275,6 +276,9 @@ int no_hardlinks = FALSE;
 int one_file_system = FALSE;
 dev_t *source_dev;
 dev_t cur_dev;
+/* Override owner/group/permissions/file type from fakeroot(1) database */
+struct fakerootdb fakerootdb;
+char const *fakerootdb_filename = NULL;
 
 static char *read_from_disk(long long start, unsigned int avail_bytes);
 static void add_old_root_entry(char *name, squashfs_inode inode,
@@ -308,6 +312,18 @@ static void check_usable_phys_mem(int total_mem);
 static void print_summary();
 void write_destination(int fd, long long byte, long long bytes, void *buff);
 
+static int lstat_with_fakeroot(const char *path, struct stat *stbuf)
+{
+	int err;
+	err = lstat(path, stbuf);
+	if (err < 0)
+		goto out;
+	if (!fakerootdb.db || fakerootdb.count == 0)
+		goto out;
+	fakeroot_override_stat(stbuf, &fakerootdb);
+out:
+	return err;
+}
 
 void prep_exit()
 {
@@ -3350,7 +3366,7 @@ static squashfs_inode scan_single(char *pathname, int progress)
 	 * it to the root directory dir_info structure */
 	dir_ent = create_dir_entry("", NULL, pathname, scan1_opendir("", "", 0));
 
-	if(lstat(pathname, &buf) == -1)
+	if(lstat_with_fakeroot(pathname, &buf) == -1)
 		/* source directory has disappeared? */
 		BAD_ERROR("Cannot stat source directory %s because %s\n",
 						pathname, strerror(errno));
@@ -3392,6 +3408,7 @@ static squashfs_inode scan_encomp(int progress)
 	buf.st_mtime = time(NULL);
 	buf.st_dev = 0;
 	buf.st_ino = 0;
+	fakeroot_override_stat(&buf, &fakerootdb);
 	dir_ent->inode = lookup_inode(&buf);
 	dir_ent->inode->dummy_root_dir = TRUE;
 	dir_ent->dir = root_dir;
@@ -3593,7 +3610,7 @@ static struct dir_info *dir_scan1(char *filename, char *subpath,
 			continue;
 		}
 
-		if(lstat(filename, &buf) == -1) {
+		if(lstat_with_fakeroot(filename, &buf) == -1) {
 			ERROR_START("Cannot stat dir/file %s because %s",
 				filename, strerror(errno));
 			ERROR_EXIT(", ignoring\n");
@@ -4361,7 +4378,7 @@ static struct dir_info *add_source(struct dir_info *sdir, char *source,
 		goto failed_early;
 	}
 
-	res = lstat(file, &buf);
+	res = lstat_with_fakeroot(file, &buf);
 	if (res == -1) {
 		ERROR("Error: Can't stat %s because %s\n", file, strerror(errno));
 		goto failed_early;
@@ -4708,13 +4725,14 @@ static squashfs_inode process_source(int progress)
 		buf.st_uid = getuid();
 		buf.st_gid = getgid();
 		buf.st_mtime = time(NULL);
+		fakeroot_override_stat(&buf, &fakerootdb);
 		entry = create_dir_entry("", NULL, "", new);
 		entry->inode = lookup_inode(&buf);
 		entry->inode->dummy_root_dir = TRUE;
 	} else {
 		char *pathname = absolute ? "/" : ".";
 
-		if(lstat(pathname, &buf) == -1)
+		if(lstat_with_fakeroot(pathname, &buf) == -1)
 			BAD_ERROR("Cannot stat %s because %s\n",
 				pathname, strerror(errno));
 
@@ -6576,6 +6594,12 @@ int main(int argc, char *argv[])
 					exit(1);
 				}
 			}
+		} else if(strcmp(argv[i], "-fakerootdb") == 0) {
+			if(++i == argc) {
+				ERROR("%s: -fakerootdb: missing filename\n", argv[0]);
+				exit(1);
+			}
+			fakerootdb_filename = argv[i];
 		} else if(strcmp(argv[i], "-noI") == 0 ||
 				strcmp(argv[i], "-noInodeCompression") == 0)
 			noI = TRUE;
@@ -6807,6 +6831,24 @@ int main(int argc, char *argv[])
 				strcmp(argv[i], "-log") == 0)
 			i++;
 
+	if (fakerootdb_filename) {
+		int err;
+		FILE *fakedata = NULL;
+		fakedata = fopen(fakerootdb_filename, "r");
+		if (!fakedata) {
+			ERROR("%s: -fakerootdb: failed to open fakeroot database %s\n",
+				argv[0], fakerootdb_filename);
+			EXIT_MKSQUASHFS();
+		}
+		err = fakeroot_read_db(fakedata, &fakerootdb);
+		fclose(fakedata);
+		if (err) {
+			ERROR("%s: -fakerootdb: failed to read fakeroot database %s\n",
+				argv[0], fakerootdb_filename);
+			EXIT_MKSQUASHFS();
+		}
+	}
+
 	if(!delete) {
 	        comp = read_super(fd, &sBlk, destination_file);
 	        if(comp == NULL) {