predicate does not stop if node is not suitable for current pod

[lowang-bh]

caused by Refactor PredicateFn for allocate and preempt actions #2916
Another issue #3049 is also introduced by it.

What happened:

What you expected to happen:

stop predicating if one filter retrun non-success.

How to reproduce it (as minimally and precisely as possible):

The code login is changed when refact.

Anything else we need to know?:

The following issues are relative.
#3049
#2975

Environment:

• Volcano Version: master branch
• Kubernetes version (use kubectl version):
• Cloud provider or hardware configuration:
• OS (e.g. from /etc/os-release):
• Kernel (e.g. uname -a):
• Install tools:
• Others:

[lowang-bh]

/assign @wangyang0616

[lowang-bh]

/assign