CVE-2026-98617: eth: fbnic: use the Rx queue napi pointer to find the napi vector

In the Linux kernel, the following vulnerability has been resolved:

eth: fbnic: use the Rx queue napi pointer to find the napi vector

The queue management ndos pick the napi vector for an Rx queue with: nv = fbn->napi[idx % fbn->numnapi];

The issue is this is only correct in the cases where there are no standalone Tx vectors. In those cases we were allocating the Tx vectors first and then the Rx so the queues would be pointing to Tx NAPI vectors instead of the Rx ones.

The mapping the ndos want is already recorded. fbnicsetnetifnapi() publishes it with netifqueuesetnapi(), which stores the napi pointer in netdevrxqueue.napi, and fbnicresetnetifnapi() clears it again. Both run under the netdev instance lock that the queue management ndos also hold, so the pointer can be read directly.

Use it and drop the divide. The pointer is NULL exactly while the datapath is down, so fbnicqueuememalloc() can reject that case rather than reaching into freed state: netdevrxqueuerestart() calls it before it tests netifrunning(), and fbnicpmsuspend() leaves netifrunning() true across a PCIe recovery that never completes, so a queue restart can arrive after fbnicstop() has freed the rings and the vectors. fbnicstop() clears the association in fbnicresetnetifqueues() before fbnicfreenapivectors(), so the NULL is always published first. fbnicqueuestart() and fbnicqueuestop() need no check of their own, as netdevrxqueuereconfig() only reaches them once fbnicqueuememalloc() has succeeded under the same instance lock.