Make DMA tags per-queue to facilate parallel mappings

Defer mbuf allocation and initialization until after data has already been
received in a cluster

This reduces cpu utilization somewhat, but it only improves the rx path.
Recent changes to TCP appear to make us rate limited by the TX path.

This is the first step in reducing mbuf management overhead for manipulating
clusters.

MFC after: 3 days

1 files changed, 4 insertions, 1 deletions

diff --git a/sys/dev/cxgb/cxgb_adapter.h b/sys/dev/cxgb/cxgb_adapter.h
index 62efb1a..288d620 100644
--- a/sys/dev/cxgb/cxgb_adapter.h
+++ b/sys/dev/cxgb/cxgb_adapter.h
@@ -175,7 +175,9 @@ struct sge_fl {
 	uint64_t	empty;
 	bus_dma_tag_t	desc_tag;
 	bus_dmamap_t	desc_map;
-	struct mtx      fl_locks[8];
+	bus_dma_tag_t   entry_tag;
+	uma_zone_t      zone;
+	int             type;
 };
 
 struct tx_desc;
@@ -201,6 +203,7 @@ struct sge_txq {
 	uint64_t	restarts;
 	bus_dma_tag_t	desc_tag;
 	bus_dmamap_t	desc_map;
+	bus_dma_tag_t   entry_tag;
 	struct mtx      lock;
 };