softlockup with backtrace indicating TLB flush activity after flood of "netif_tcp_helper_alloc_u" messages from third party Solarflare onload driver

Issue

• After a flood of the following message from third part Solarflare onload driver, the system starts experiencing softlockup

kernel: [onload] oo_get_nics: ERROR: No Solarflare network interfaces are active/UP. Please check your config with ip addr or ifconfig
kernel: oo:PROG[pid]: netif_tcp_helper_alloc_u: ci_resource_alloc failed (-19)
kernel: oo:PROG[pid]: netif_tcp_helper_alloc_u: This error can occur if no Solarflare network interfaces are active/UP. Please check your config with ip addr or ifconfig

The softlockup may indicate kernel executing flush_tlb_others_ipi

[ffff8804b12e9dd8] _spin_lock at ffffffff814ddade
[ffff8804b12e9de0] flush_tlb_others_ipi at ffffffff810466a9
[ffff8804b12e9e30] native_flush_tlb_others at ffffffff810467e6
[ffff8804b12e9e60] flush_tlb_mm at ffffffff810469bc