Alas they uglify the error paths during restart, but I can't
count on two hands the number of times I've had to add these and
recompile to find one silly bug or other, so I suggest that it
may be worth having these in the code anyway.

Signed-off-by: Serge E. Hallyn <[email protected]>
---
 checkpoint/restart.c |   50 +++++++++++++++++++++++++++++++++++++++-----------
 1 files changed, 39 insertions(+), 11 deletions(-)

diff --git a/checkpoint/restart.c b/checkpoint/restart.c
index 1085ed5..ebdd5ba 100644
--- a/checkpoint/restart.c
+++ b/checkpoint/restart.c
@@ -662,6 +662,7 @@ static int restore_activate_next(struct ckpt_ctx *ctx)
                rcu_read_unlock();
 
                if (!task) {
+                       ckpt_debug("could not find task %d\n", pid);
                        restore_notify_error(ctx, -ESRCH);
                        return -ESRCH;
                }
@@ -685,6 +686,7 @@ static int wait_task_active(struct ckpt_ctx *ctx)
        ckpt_debug("active %d < %d (ret %d)\n",
                   ctx->active_pid, ctx->nr_pids, ret);
        if (!ret && ckpt_test_ctx_error(ctx)) {
+               ckpt_debug("wait_event_interruptible returned %d\n", ret);
                force_sig(SIGKILL, current);
                ret = -EBUSY;
        }
@@ -827,20 +829,25 @@ static int do_restore_task(void)
 
        /* wait for our turn, do the restore, and tell next task in line */
        ret = wait_task_active(ctx);
-       if (ret < 0)
+       if (ret < 0) {
+               ckpt_debug("wait_task_active returned %d\n", ret);
                goto out;
+       }
 
        ckpt_debug_log_running(ctx);
 
        zombie = restore_task(ctx);
        if (zombie < 0) {
+               ckpt_debug("restore_task returned %d\n", ret);
                ret = zombie;
                goto out;
        }
 
        ret = restore_activate_next(ctx);
-       if (ret < 0)
+       if (ret < 0) {
+               ckpt_debug("restore_activate_next returned %d\n", ret);
                goto out;
+       }
 
        /*
         * zombie: we're done here; do_exit() will notice the @ctx on
@@ -855,6 +862,8 @@ static int do_restore_task(void)
 
        restore_task_done(ctx);
        ret = wait_task_sync(ctx);
+       if (ret)
+               ckpt_debug("wait_task_sync returned %d\n", ret);
  out:
        old_ctx = xchg(&current->checkpoint_ctx, NULL);
        if (old_ctx)
@@ -970,12 +979,14 @@ static int wait_all_tasks_finish(struct ckpt_ctx *ctx)
 
        BUG_ON(ctx->active_pid != -1);
        ret = restore_activate_next(ctx);
-       if (ret < 0)
+       if (ret < 0) {
+               ckpt_debug("restore_activate_next returned %d\n", ret);
                return ret;
+       }
 
        ret = wait_for_completion_interruptible(&ctx->complete);
 
-       ckpt_debug("final sync kflags %#lx\n", ctx->kflags);
+       ckpt_debug("final sync kflags %#lx (ret %d)\n", ctx->kflags, ret);
        /*
         * Usually when restart fails, the restarting task will first
         * set @ctx->errno before waking us up. In the rare event that
@@ -1051,18 +1062,24 @@ static int do_restore_coord(struct ckpt_ctx *ctx, pid_t 
pid)
        ckpt_debug_log_running(ctx);
 
        ret = restore_read_header(ctx);
-       if (ret < 0)
+       if (ret < 0) {
+               ckpt_debug("restore_read_header returned %d\n", ret);
                return ret;
+       }
        ret = restore_read_tree(ctx);
-       if (ret < 0)
+       if (ret < 0) {
+               ckpt_debug("restore_read_tree returned %d\n", ret);
                return ret;
+       }
 
        if ((ctx->uflags & RESTART_TASKSELF) && ctx->nr_pids != 1)
                return -EINVAL;
 
        ret = init_restart_ctx(ctx, pid);
-       if (ret < 0)
+       if (ret < 0) {
+               ckpt_debug("init_restart_ctx returned %d\n", ret);
                return ret;
+       }
 
        /*
         * Populate own ->checkpoint_ctx: if an ancestor attempts to
@@ -1094,8 +1111,10 @@ static int do_restore_coord(struct ckpt_ctx *ctx, pid_t 
pid)
        } else {
                /* prepare descendants' t->checkpoint_ctx point to coord */
                ret = prepare_descendants(ctx, ctx->root_task);
-               if (ret < 0)
+               if (ret < 0) {
+                       ckpt_debug("prepare_descendants returned %d", ret);
                        goto out;
+               }
 
                /* tell tasks we're ready for them to dec  ctx->nr_running */
                wake_up_all(&ctx->waitq);
@@ -1108,17 +1127,23 @@ static int do_restore_coord(struct ckpt_ctx *ctx, pid_t 
pid)
 
                /* wait for all other tasks to complete do_restore_task() */
                ret = wait_all_tasks_finish(ctx);
-               if (ret < 0)
+               if (ret < 0) {
+                       ckpt_debug("wait_all_tasks_finish returned %d", ret);
                        goto out;
+               }
        }
 
        ret = deferqueue_run(ctx->deferqueue);  /* run deferred work */
-       if (ret < 0)
+       if (ret < 0) {
+               ckpt_debug("deferqueue_run returned %d\n", ret);
                goto out;
+       }
 
        ret = restore_read_tail(ctx);
-       if (ret < 0)
+       if (ret < 0) {
+               ckpt_debug("restore_read_tail returned %d\n", ret);
                goto out;
+       }
 
        if (ctx->uflags & RESTART_FROZEN) {
                ret = cgroup_freezer_make_frozen(ctx->root_task);
@@ -1202,6 +1227,9 @@ long do_restart(struct ckpt_ctx *ctx, pid_t pid, unsigned 
long flags)
        else
                ret = do_restore_task();
 
+       if (ret < 0)
+               ckpt_debug("ret %ld\n", ret);
+
        /* restart(2) isn't idempotent: should not be auto-restarted */
        if (ret == -ERESTARTSYS || ret == -ERESTARTNOINTR ||
            ret == -ERESTARTNOHAND || ret == -ERESTART_RESTARTBLOCK)
-- 
1.6.1

_______________________________________________
Containers mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/containers

_______________________________________________
Devel mailing list
[email protected]
https://openvz.org/mailman/listinfo/devel

Reply via email to