Add support for gnu pax sparse file formats 0.1 and 1.0

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

3 files changed, 231 insertions, 1 deletions

diff --git a/lib/tar/read_header.c b/lib/tar/read_header.c
index 23c23c4..e1c9eaf 100644
--- a/lib/tar/read_header.c
+++ b/lib/tar/read_header.c
@@ -163,6 +163,55 @@ static int pax_read_decimal(const char *str, uint64_t *out)
 	return 0;
 }
 
+static sparse_map_t *read_sparse_map(const char *line)
+{
+	sparse_map_t *last = NULL, *list = NULL, *ent = NULL;
+
+	do {
+		ent = calloc(1, sizeof(*ent));
+		if (ent == NULL)
+			goto fail_errno;
+
+		if (pax_read_decimal(line, &ent->offset))
+			goto fail_format;
+
+		while (isdigit(*line))
+			++line;
+
+		if (*(line++) != ',')
+			goto fail_format;
+
+		if (pax_read_decimal(line, &ent->count))
+			goto fail_format;
+
+		while (isdigit(*line))
+			++line;
+
+		if (last == NULL) {
+			list = last = ent;
+		} else {
+			last->next = ent;
+			last = ent;
+		}
+	} while (*(line++) == ',');
+
+	return list;
+fail_errno:
+	perror("parsing GNU pax sparse file record");
+	goto fail;
+fail_format:
+	fputs("malformed GNU pax sparse file record\n", stderr);
+	goto fail;
+fail:
+	while (list != NULL) {
+		ent = list;
+		list = list->next;
+		free(ent);
+	}
+	free(ent);
+	return NULL;
+}
+
 static int read_pax_header(int fd, uint64_t entsize, unsigned int *set_by_pax,
 			   tar_header_decoded_t *out)
 {
@@ -254,6 +303,19 @@ static int read_pax_header(int fd, uint64_t entsize, unsigned int *set_by_pax,
 				out->sb.st_ctime = field;
 			}
 			*set_by_pax |= PAX_CTIME;
+		} else if (!strncmp(line, "GNU.sparse.name=", 16)) {
+			free(out->name);
+			out->name = strdup(line + 5);
+			if (out->name == NULL)
+				goto fail_errno;
+			*set_by_pax |= PAX_NAME;
+		} else if (!strncmp(line, "GNU.sparse.map=", 15)) {
+			free_sparse_list(out->sparse);
+			sparse_last = NULL;
+
+			out->sparse = read_sparse_map(line + 15);
+			if (out->sparse == NULL)
+				goto fail;
 		} else if (!strncmp(line, "GNU.sparse.size=", 16)) {
 			pax_read_decimal(line + 16, &out->actual_size);
 			*set_by_pax |= PAX_SPARSE_SIZE;
diff --git a/tests/Makemodule.am b/tests/Makemodule.am
index f9232dc..2c8344a 100644
--- a/tests/Makemodule.am
+++ b/tests/Makemodule.am
@@ -55,17 +55,23 @@ test_tar_sparse_gnu1_LDADD = libtar.a libutil.a
 test_tar_sparse_gnu1_CPPFLAGS = $(AM_CPPFLAGS)
 test_tar_sparse_gnu1_CPPFLAGS += -DTESTPATH=$(top_srcdir)/tests/tar
 
+test_tar_sparse_gnu2_SOURCES = tests/tar_sparse_gnu1.c
+test_tar_sparse_gnu2_LDADD = libtar.a libutil.a
+test_tar_sparse_gnu2_CPPFLAGS = $(AM_CPPFLAGS)
+test_tar_sparse_gnu2_CPPFLAGS += -DTESTPATH=$(top_srcdir)/tests/tar
+
 check_PROGRAMS += test_canonicalize_name test_mknode_simple test_mknode_slink
 check_PROGRAMS += test_mknode_reg test_mknode_dir test_gen_inode_table
 check_PROGRAMS += test_add_by_path test_get_path test_fstree_sort
 check_PROGRAMS += test_fstree_from_file test_fstree_init test_fstree_xattr
 check_PROGRAMS += test_tar_ustar test_tar_pax test_tar_gnu test_tar_sparse_gnu
-check_PROGRAMS += test_tar_sparse_gnu1
+check_PROGRAMS += test_tar_sparse_gnu1 test_tar_sparse_gnu2
 
 TESTS += test_canonicalize_name test_mknode_simple test_mknode_slink
 TESTS += test_mknode_reg test_mknode_dir test_gen_inode_table
 TESTS += test_add_by_path test_get_path test_fstree_sort test_fstree_from_file
 TESTS += test_fstree_init test_fstree_xattr test_tar_ustar test_tar_pax
 TESTS += test_tar_gnu test_tar_sparse_gnu test_tar_sparse_gnu1
+TESTS += test_tar_sparse_gnu2
 
 EXTRA_DIST += $(top_srcdir)/tests/tar
diff --git a/tests/tar_sparse_gnu2.c b/tests/tar_sparse_gnu2.c
new file mode 100644
index 0000000..edf1fb5
--- /dev/null
+++ b/tests/tar_sparse_gnu2.c
@@ -0,0 +1,162 @@
+/* SPDX-License-Identifier: GPL-3.0-or-later */
+#include "util.h"
+#include "tar.h"
+
+#include <unistd.h>
+#include <assert.h>
+#include <stdlib.h>
+#include <string.h>
+#include <fcntl.h>
+#include <stdio.h>
+
+#define STR(x) #x
+#define STRVALUE(x) STR(x)
+
+#define TEST_PATH STRVALUE(TESTPATH)
+
+static int open_read(const char *path)
+{
+	int fd = open(path, O_RDONLY);
+
+	if (fd < 0) {
+		perror(path);
+		exit(EXIT_FAILURE);
+	}
+
+	return fd;
+}
+
+int main(void)
+{
+	tar_header_decoded_t hdr;
+	sparse_map_t *sparse;
+	int fd;
+
+	assert(chdir(TEST_PATH) == 0);
+
+	fd = open_read("sparse-files/pax-gnu0-1.tar");
+	assert(read_header(fd, &hdr) == 0);
+	assert(hdr.sb.st_mode == (S_IFREG | 0644));
+	assert(hdr.sb.st_uid == 01750);
+	assert(hdr.sb.st_gid == 01750);
+	assert(hdr.sb.st_size == 2097152);
+	assert(hdr.actual_size == 2097152);
+	assert(hdr.record_size == 32768);
+	assert(strcmp(hdr.name, "input.bin") == 0);
+	assert(!hdr.unknown_record);
+
+	sparse = hdr.sparse;
+	assert(sparse != NULL);
+	assert(sparse->offset == 0);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 262144);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 524288);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 786432);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 1048576);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 1310720);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 1572864);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 1835008);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 2097152);
+	assert(sparse->count == 0);
+
+	sparse = sparse->next;
+	assert(sparse == NULL);
+
+	clear_header(&hdr);
+	close(fd);
+
+	fd = open_read("sparse-files/pax-gnu1-0.tar");
+	assert(read_header(fd, &hdr) == 0);
+	assert(hdr.sb.st_mode == (S_IFREG | 0644));
+	assert(hdr.sb.st_uid == 01750);
+	assert(hdr.sb.st_gid == 01750);
+	assert(hdr.sb.st_size == 2097152);
+	assert(hdr.actual_size == 2097152);
+	assert(hdr.record_size == 32768);
+	assert(strcmp(hdr.name, "input.bin") == 0);
+	assert(!hdr.unknown_record);
+
+	sparse = hdr.sparse;
+	assert(sparse != NULL);
+	assert(sparse->offset == 0);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 262144);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 524288);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 786432);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 1048576);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 1310720);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 1572864);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 1835008);
+	assert(sparse->count == 4096);
+
+	sparse = sparse->next;
+	assert(sparse != NULL);
+	assert(sparse->offset == 2097152);
+	assert(sparse->count == 0);
+
+	sparse = sparse->next;
+	assert(sparse == NULL);
+
+	clear_header(&hdr);
+	close(fd);
+
+	return EXIT_SUCCESS;
+}