Expand tx dma maps to allow packets up to the MTU limit to be transmitted.

Don't clear TPB_TX_BUF_EN when taking the interface down, as somehow that
causes transmission to fail when the interface comes up again if any large
packets had been sent.

Handle packets received into multiple rx buffers using the same approach as
ixl(4), with the addition that a receive error can be signaled on any of
the buffers rather than just the last one.  Set the low water mark on the
rx ring so there are always enough buffers to receive the largest packet
supported by the interface, and fix the calculation of the data pointer
in rx buffers so there's actually MCLBYTES space available.

ok dlg@

1 files changed, 46 insertions, 13 deletions

diff --git a/sys/dev/pci/if_aq_pci.c b/sys/dev/pci/if_aq_pci.c
index 90c68ffd6b5..f7cf6339c66 100644
--- a/sys/dev/pci/if_aq_pci.c
+++ b/sys/dev/pci/if_aq_pci.c
@@ -1,4 +1,4 @@
-/* $OpenBSD: if_aq_pci.c,v 1.11 2022/03/20 00:01:33 jmatthew Exp $ */
+/* $OpenBSD: if_aq_pci.c,v 1.12 2022/03/26 06:04:20 jmatthew Exp $ */
 /*	$NetBSD: if_aq.c,v 1.27 2021/06/16 00:21:18 riastradh Exp $	*/
 
 /*
@@ -698,6 +698,10 @@ struct aq_rxring {
 	struct if_rxring	 rx_rxr;
 	uint32_t		 rx_prod;
 	uint32_t		 rx_cons;
+
+	struct mbuf		*rx_m_head;
+	struct mbuf		**rx_m_tail;
+	int			 rx_m_error;
 };
 
 struct aq_txring {
@@ -1119,6 +1123,9 @@ aq_attach(struct device *parent, struct device *self, void *aux)
 		rx->rx_q = i;
 		rx->rx_irq = i * 2;
 		rx->rx_ifiq = ifp->if_iqs[i];
+		rx->rx_m_head = NULL;
+		rx->rx_m_tail = &rx->rx_m_head;
+		rx->rx_m_error = 0;
 		ifp->if_iqs[i]->ifiq_softc = aq;
 		timeout_set(&rx->rx_refill, aq_refill, rx);
 
@@ -2136,7 +2143,7 @@ aq_rx_fill_slots(struct aq_softc *sc, struct aq_rxring *rx, uint nslots)
 		if (m == NULL)
 			break;
 
-		m->m_data += (m->m_ext.ext_size - MCLBYTES);
+		m->m_data += (m->m_ext.ext_size - (MCLBYTES + ETHER_ALIGN));
 		m->m_data += ETHER_ALIGN;
 		m->m_len = m->m_pkthdr.len = MCLBYTES;
 
@@ -2198,7 +2205,7 @@ aq_rxeof(struct aq_softc *sc, struct aq_rxring *rx)
 	uint32_t end, idx;
 	uint16_t pktlen, status;
 	uint32_t rxd_type;
-	struct mbuf *m;
+	struct mbuf *m, *mb;
 	struct mbuf_list ml = MBUF_LIST_INITIALIZER();
 	int rxfree;
 
@@ -2227,13 +2234,20 @@ aq_rxeof(struct aq_softc *sc, struct aq_rxring *rx)
 			break;
 
 		rxfree++;
-		m = as->as_m;
+		mb = as->as_m;
 		as->as_m = NULL;
 
 		pktlen = lemtoh16(&rxd->pkt_len);
 		rxd_type = lemtoh32(&rxd->type);
 		/* rss hash */
 
+		mb->m_pkthdr.len = 0;
+		mb->m_next = NULL;
+		*rx->rx_m_tail = mb;
+		rx->rx_m_tail = &mb->m_next;
+
+		m = rx->rx_m_head;
+
 #if NVLAN > 0
 		if (rxd_type & (AQ_RXDESC_TYPE_VLAN | AQ_RXDESC_TYPE_VLAN2)) {
 			m->m_pkthdr.ether_vtag = lemtoh16(&rxd->vlan);
@@ -2254,10 +2268,25 @@ aq_rxeof(struct aq_softc *sc, struct aq_rxring *rx)
 		    (rxd_type & AQ_RXDESC_TYPE_DMA_ERR)) {
 			printf("%s:rx: rx error (status %x type %x)\n",
 			    DEVNAME(sc), status, rxd_type);
-			m_freem(m);
+			rx->rx_m_error = 1;
+		}
+
+		if (status & AQ_RXDESC_STATUS_EOP) {
+			mb->m_len = pktlen - m->m_pkthdr.len;
+			m->m_pkthdr.len = pktlen;
+			if (rx->rx_m_error != 0) {
+				ifp->if_ierrors++;
+				m_freem(m);
+			} else {
+				ml_enqueue(&ml, m);
+			}
+
+			rx->rx_m_head = NULL;
+			rx->rx_m_tail = &rx->rx_m_head;
+			rx->rx_m_error = 0;
 		} else {
-			m->m_pkthdr.len = m->m_len = pktlen;
-			ml_enqueue(&ml, m);
+			mb->m_len = MCLBYTES;
+			m->m_pkthdr.len += mb->m_len;
 		}
 
 		idx++;
@@ -2486,7 +2515,7 @@ aq_queue_up(struct aq_softc *sc, struct aq_queues *aq)
 	struct aq_rxring *rx;
 	struct aq_txring *tx;
 	struct aq_slot *as;
-	int i;
+	int i, mtu;
 
 	rx = &aq->q_rx;
 	rx->rx_slots = mallocarray(sizeof(*as), AQ_RXD_NUM, M_DEVBUF,
@@ -2524,11 +2553,11 @@ aq_queue_up(struct aq_softc *sc, struct aq_queues *aq)
 		goto destroy_rx_ring;
 	}
 
+	mtu = sc->sc_arpcom.ac_if.if_hardmtu;
 	for (i = 0; i < AQ_TXD_NUM; i++) {
 		as = &tx->tx_slots[i];
-		if (bus_dmamap_create(sc->sc_dmat, MCLBYTES,
-		    AQ_TX_MAX_SEGMENTS, MCLBYTES, 0,
-		    BUS_DMA_WAITOK | BUS_DMA_ALLOCNOW | BUS_DMA_64BIT,
+		if (bus_dmamap_create(sc->sc_dmat, mtu, AQ_TX_MAX_SEGMENTS,
+		    MCLBYTES, 0, BUS_DMA_WAITOK | BUS_DMA_ALLOCNOW | BUS_DMA_64BIT,
 		    &as->as_map) != 0) {
 			printf("%s: failed to allocated tx dma maps %d\n",
 			    DEVNAME(sc), aq->q_index);
@@ -2576,6 +2605,10 @@ aq_queue_down(struct aq_softc *sc, struct aq_queues *aq)
 	aq_dmamem_free(sc, &tx->tx_mem);
 
 	rx = &aq->q_rx;
+	m_freem(rx->rx_m_head);
+	rx->rx_m_head = NULL;
+	rx->rx_m_tail = &rx->rx_m_head;
+	rx->rx_m_error = 0;
 	aq_rxring_reset(sc, &aq->q_rx, 0);
 	if (rx->rx_slots != NULL) {
 		aq_free_slots(sc, rx->rx_slots, AQ_RXD_NUM, AQ_RXD_NUM);
@@ -2624,7 +2657,8 @@ aq_up(struct aq_softc *sc)
 	for (i = 0; i < sc->sc_nqueues; i++) {
 		struct aq_queues *aq = &sc->sc_queues[i];
 
-		if_rxr_init(&aq->q_rx.rx_rxr, 1, AQ_RXD_NUM - 1);
+		if_rxr_init(&aq->q_rx.rx_rxr, howmany(ifp->if_hardmtu, MCLBYTES),
+		    AQ_RXD_NUM - 1);
 		aq_rx_fill(sc, &aq->q_rx);
 
 		ifq_clr_oactive(aq->q_tx.tx_ifq);
@@ -2649,7 +2683,6 @@ aq_down(struct aq_softc *sc)
 	aq_enable_intr(sc, 1, 0);
 	intr_barrier(sc->sc_ih);
 
-	AQ_WRITE_REG_BIT(sc, TPB_TX_BUF_REG, TPB_TX_BUF_EN, 0);
 	AQ_WRITE_REG_BIT(sc, RPB_RPF_RX_REG, RPB_RPF_RX_BUF_EN, 0);
 	for (i = 0; i < sc->sc_nqueues; i++) {
 		/* queue intr barrier? */