sched_ext: fix uninitialized ret on alloc_percpu() failure

Smatch reported:

  kernel/sched/ext.c:5332 scx_alloc_and_add_sched() warn: passing zero to 'ERR_PTR'

In scx_alloc_and_add_sched(), the alloc_percpu() failure path jumps to
err_free_gdsqs without initializing @ret. That can lead to returning
ERR_PTR(0), which violates the ERR_PTR() convention and confuses
callers.

Set @ret to -ENOMEM before jumping to the error path when
alloc_percpu() fails.

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/r/202512141601.yAXDAeA9-lkp@intel.com/
Reported-by: Dan Carpenter <error27@gmail.com>
Fixes: c201ea1 ("sched_ext: Move event_stats_cpu into scx_sched")
Signed-off-by: Liang Jie <liangjie@lixiang.com>
Reviewed-by: Emil Tsalapatis <emil@etsalapatis.com>
Reviewed-by: Andrea Righi <arighi@nvidia.com>
Signed-off-by: Tejun Heo <tj@kernel.org>

Author: Liang Jie

kernel/sched/ext.c

@@ -4783,8 +4783,10 @@ static struct scx_sched *scx_alloc_and_add_sched(struct sched_ext_ops *ops)
 	}
 
 	sch->pcpu = alloc_percpu(struct scx_sched_pcpu);
-	if (!sch->pcpu)
+	if (!sch->pcpu) {
+		ret = -ENOMEM;
 		goto err_free_gdsqs;
+	}
 
 	sch->helper = kthread_run_worker(0, "sched_ext_helper");
 	if (IS_ERR(sch->helper)) {