We should always ensure memory allocation for rx nbufs
should always happen from phy address 0x50000000 and above
else drop the nbuf and try for a new nbuf.
Change-Id: I675a0f1289e04f720949ad9b6917bbb733270a78
CRs-Fixed: 2003174
Add NULL termination to the list of rx frames
before passing them to the high level OS.
Change-Id: Ibc51f14b8b9ef2a8dff7147f4072910e11540093
CRs-Fixed: 1116939
Added a separate loop to dequeue the rx pkts and relinquish
REO and rx-Refill ring faster. Additionl rx processing is
moved to a different per vdev loop.
CRs-Fixed: 1098453
Change-Id: I43b1fa302cf7e7c77f6114ea491306c0b724e365
The MC BC packets cause a NULL REO queue descriptor exception,
and hence are directed to the ring that is configured to receive
this exception from REO. As of now, this error is directed to
WBM release ring. Add handling to pass these packets to stack
instead of dropping them, when received on the WBM release ring.
Note that, the final fix may be route these to REO exception ring
from which the driver may need to pass them to stack.
CRs-Fixed: 1098153
Change-Id: If544911e7882a9f5e7dbffa5ef705618e9fcbbda
