dump: Don't unfreeze tasks on dump failure with --no-resume-on-error.

Make it possible to kill or leave stopped tasks if a dump failed
after stopping the tree.

Signed-off-by: Michał Mirosław <[email protected]>

Documentation/criu.txt

@@ -213,6 +213,9 @@ In other words, do not use it unless really needed.
 *-s*, *--leave-stopped*::
     Leave tasks in stopped state after checkpoint, instead of killing.
 
+*--no-resume-on-error*::
+    Leave tasks in stopped state even if checkpoint completed unsuccessfully.
+
 *--external* __type__**[**__id__**]:**__value__::
     Dump an instance of an external resource. The generic syntax is
     'type' of resource, followed by resource 'id' (enclosed in literal

criu/config.c

@@ -412,6 +412,7 @@ void init_opts(void)
 
 	/* Default options */
 	opts.final_state = TASK_DEAD;
+	opts.resume_on_dump_error = true;
 	INIT_LIST_HEAD(&opts.ext_mounts);
 	INIT_LIST_HEAD(&opts.inherit_fds);
 	INIT_LIST_HEAD(&opts.external);
@@ -622,6 +623,7 @@ int parse_options(int argc, char **argv, bool *usage_error, bool *has_exec_cmd,
 		{ "tree", required_argument, 0, 't' },
 		{ "leave-stopped", no_argument, 0, 's' },
 		{ "leave-running", no_argument, 0, 'R' },
+		BOOL_OPT("resume-on-error", &opts.resume_on_dump_error),
 		BOOL_OPT("restore-detached", &opts.restore_detach),
 		BOOL_OPT("restore-sibling", &opts.restore_sibling),
 		BOOL_OPT("daemon", &opts.restore_detach),

criu/cr-dump.c

@@ -2070,7 +2070,9 @@ static int cr_dump_finish(int ret)
 	 *    consistency of the FS and other resources, we simply
 	 *    start rollback procedure and cleanup everything.
 	 */
-	if (ret || post_dump_ret || opts.final_state == TASK_ALIVE) {
+	if (opts.resume_on_dump_error && (ret || post_dump_ret))
+		opts.final_state = TASK_ALIVE;
+	if (opts.final_state == TASK_ALIVE) {
 		unsuspend_lsm();
 		network_unlock();
 		delete_link_remaps();
@@ -2082,7 +2084,7 @@ static int cr_dump_finish(int ret)
 
 	if (arch_set_thread_regs(root_item, true) < 0)
 		return -1;
-	pstree_switch_state(root_item, (ret || post_dump_ret) ? TASK_ALIVE : opts.final_state);
+	pstree_switch_state(root_item, opts.final_state);
 	timing_stop(TIME_FROZEN);
 	free_pstree(root_item);
 	seccomp_free_entries();

criu/cr-service.c

@@ -483,6 +483,9 @@ static int setup_opts_from_req(int sk, CriuOpts *req)
 	if (req->has_leave_stopped && req->leave_stopped)
 		opts.final_state = TASK_STOPPED;
 
+	if (req->has_resume_on_dump_error)
+		opts.resume_on_dump_error = req->resume_on_dump_error;
+
 	if (!req->has_pid) {
 		req->has_pid = true;
 		req->pid = ids.pid;

criu/crtools.c

@@ -394,6 +394,8 @@ int main(int argc, char *argv[], char *envp[])
 	       "  -d|--restore-detached detach after restore\n"
 	       "  -S|--restore-sibling  restore root task as sibling\n"
 	       "  -s|--leave-stopped    leave tasks in stopped state after checkpoint\n"
+	       "     --no-resume-on-error\n"
+	       "                        don't resume tasks on dump failure if they were stopped\n"
 	       "  -R|--leave-running    leave tasks in running state after checkpoint\n"
 	       "  -D|--images-dir DIR   directory for image files\n"
 	       "     --pidfile FILE     write root task, service or page-server pid to FILE\n"

criu/include/cr_options.h

@@ -131,6 +131,7 @@ struct cr_options {
 		bool daemon_mode;
 	};
 	int restore_sibling;
+	int resume_on_dump_error;
 	bool ext_unix_sk;
 	int shell_job;
 	int handle_file_locks;

images/rpc.proto

@@ -145,6 +145,7 @@ message criu_opts {
 	optional bool			leave_stopped		= 69;
 	optional bool			display_stats		= 70;
 	optional bool			log_to_stderr		= 71;
+	optional bool			resume_on_dump_error	= 72 [default = true];
 /*	optional bool			check_mounts		= 128;	*/
 }