System shows high load in KVM when isolcpus in use with most CPUs

Issue

• The CPUs not in isolcpus list had huge runqueues and the majority of tasks were handling function tdp_mmu_zap_root_work.

PID: 1165     TASK: ffff888121a6c700  CPU: 0    COMMAND: "kworker/u193:5"
 #0 [fffffe000000bd10] machine_kexec at ffffffff8106c237
 #1 [fffffe000000bd68] __crash_kexec at ffffffff811c3c9a
 #2 [fffffe000000be28] panic at ffffffff81ad7dc6
 #3 [fffffe000000beb0] unknown_nmi_error.cold at ffffffff81ad1453
 #4 [fffffe000000bec8] exc_nmi at ffffffff81b26c2d
 #5 [fffffe000000bef0] end_repeat_nmi at ffffffff81c0156a
    [exception RIP: native_queued_spin_lock_slowpath+0x42]
    RIP: ffffffff81165192  RSP: ffffc9000f4b3da0  RFLAGS: 00000202
    RAX: 0000000000c40101  RBX: ffffc9002cd21000  RCX: 0000000000000000
    RDX: ffffc9002cd21004  RSI: 0000000000000000  RDI: ffffc9002cd21004
    RBP: ffffc9002cd21004   R8: 000000000000001b   R9: 0000000008000000
    R10: 0000000000000000  R11: 0000000000000000  R12: 0000000000000000
    R13: 0000000000000001  R14: 0000000000000004  R15: ffff888121a6c700
    ORIG_RAX: ffffffffffffffff  CS: 0010  SS: 0018
--- <NMI exception stack> ---
 #6 [ffffc9000f4b3da0] native_queued_spin_lock_slowpath at ffffffff81165192
 #7 [ffffc9000f4b3da0] queued_read_lock_slowpath at ffffffff81165744
 #8 [ffffc9000f4b3db8] __cond_resched_rwlock_read at ffffffff8112c74f
 #9 [ffffc9000f4b3dc8] __tdp_mmu_zap_root at ffffffffc1213221 [kvm]
#10 [ffffc9000f4b3e70] tdp_mmu_zap_root_work at ffffffffc1213fa4 [kvm]
#11 [ffffc9000f4b3e90] process_one_work at ffffffff81118365
#12 [ffffc9000f4b3ed8] worker_thread at ffffffff81118930
#13 [ffffc9000f4b3f18] kthread at ffffffff8111f956
#14 [ffffc9000f4b3f50] ret_from_fork at ffffffff81002d9f