#include "enic_compat.h"
#include "rq_enet_desc.h"
#include "enic.h"
+#include <rte_ether.h>
+#include <rte_ip.h>
+#include <rte_tcp.h>
#define RTE_PMD_USE_PREFETCH
CQ_ENET_RQ_DESC_BYTES_WRITTEN_MASK;
}
+/* Find the offset to L5. This is needed by enic TSO implementation.
+ * Return 0 if not a TCP packet or can't figure out the length.
+ */
+static inline uint8_t tso_header_len(struct rte_mbuf *mbuf)
+{
+ struct ether_hdr *eh;
+ struct vlan_hdr *vh;
+ struct ipv4_hdr *ip4;
+ struct ipv6_hdr *ip6;
+ struct tcp_hdr *th;
+ uint8_t hdr_len;
+ uint16_t ether_type;
+
+ /* offset past Ethernet header */
+ eh = rte_pktmbuf_mtod(mbuf, struct ether_hdr *);
+ ether_type = eh->ether_type;
+ hdr_len = sizeof(struct ether_hdr);
+ if (ether_type == rte_cpu_to_be_16(ETHER_TYPE_VLAN)) {
+ vh = rte_pktmbuf_mtod_offset(mbuf, struct vlan_hdr *, hdr_len);
+ ether_type = vh->eth_proto;
+ hdr_len += sizeof(struct vlan_hdr);
+ }
+
+ /* offset past IP header */
+ switch (rte_be_to_cpu_16(ether_type)) {
+ case ETHER_TYPE_IPv4:
+ ip4 = rte_pktmbuf_mtod_offset(mbuf, struct ipv4_hdr *, hdr_len);
+ if (ip4->next_proto_id != IPPROTO_TCP)
+ return 0;
+ hdr_len += (ip4->version_ihl & 0xf) * 4;
+ break;
+ case ETHER_TYPE_IPv6:
+ ip6 = rte_pktmbuf_mtod_offset(mbuf, struct ipv6_hdr *, hdr_len);
+ if (ip6->proto != IPPROTO_TCP)
+ return 0;
+ hdr_len += sizeof(struct ipv6_hdr);
+ break;
+ default:
+ return 0;
+ }
+
+ if ((hdr_len + sizeof(struct tcp_hdr)) > mbuf->pkt_len)
+ return 0;
+
+ /* offset past TCP header */
+ th = rte_pktmbuf_mtod_offset(mbuf, struct tcp_hdr *, hdr_len);
+ hdr_len += (th->data_off >> 4) * 4;
+
+ if (hdr_len > mbuf->pkt_len)
+ return 0;
+
+ return hdr_len;
+}
+
static inline uint8_t
enic_cq_rx_check_err(struct cq_desc *cqd)
{
struct cq_enet_rq_desc *cqrd = (struct cq_enet_rq_desc *)cqd;
uint8_t cqrd_flags = cqrd->flags;
static const uint32_t cq_type_table[128] __rte_cache_aligned = {
- [32] = RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV4,
- [34] = RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV4
- | RTE_PTYPE_L4_UDP,
- [36] = RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV4
- | RTE_PTYPE_L4_TCP,
- [96] = RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV4
- | RTE_PTYPE_L4_FRAG,
- [16] = RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV6,
- [18] = RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV6
- | RTE_PTYPE_L4_UDP,
- [20] = RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV6
- | RTE_PTYPE_L4_TCP,
- [80] = RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV6
- | RTE_PTYPE_L4_FRAG,
+ [0x00] = RTE_PTYPE_UNKNOWN,
+ [0x20] = RTE_PTYPE_L3_IPV4_EXT_UNKNOWN | RTE_PTYPE_L4_NONFRAG,
+ [0x22] = RTE_PTYPE_L3_IPV4_EXT_UNKNOWN | RTE_PTYPE_L4_UDP,
+ [0x24] = RTE_PTYPE_L3_IPV4_EXT_UNKNOWN | RTE_PTYPE_L4_TCP,
+ [0x60] = RTE_PTYPE_L3_IPV4_EXT_UNKNOWN | RTE_PTYPE_L4_FRAG,
+ [0x62] = RTE_PTYPE_L3_IPV4_EXT_UNKNOWN | RTE_PTYPE_L4_UDP,
+ [0x64] = RTE_PTYPE_L3_IPV4_EXT_UNKNOWN | RTE_PTYPE_L4_TCP,
+ [0x10] = RTE_PTYPE_L3_IPV6_EXT_UNKNOWN | RTE_PTYPE_L4_NONFRAG,
+ [0x12] = RTE_PTYPE_L3_IPV6_EXT_UNKNOWN | RTE_PTYPE_L4_UDP,
+ [0x14] = RTE_PTYPE_L3_IPV6_EXT_UNKNOWN | RTE_PTYPE_L4_TCP,
+ [0x50] = RTE_PTYPE_L3_IPV6_EXT_UNKNOWN | RTE_PTYPE_L4_FRAG,
+ [0x52] = RTE_PTYPE_L3_IPV6_EXT_UNKNOWN | RTE_PTYPE_L4_UDP,
+ [0x54] = RTE_PTYPE_L3_IPV6_EXT_UNKNOWN | RTE_PTYPE_L4_TCP,
/* All others reserved */
};
cqrd_flags &= CQ_ENET_RQ_DESC_FLAGS_IPV4_FRAGMENT
enic_cq_rx_to_pkt_flags(struct cq_desc *cqd, struct rte_mbuf *mbuf)
{
struct cq_enet_rq_desc *cqrd = (struct cq_enet_rq_desc *)cqd;
- uint16_t ciflags, bwflags, pkt_flags = 0;
+ uint16_t ciflags, bwflags, pkt_flags = 0, vlan_tci;
ciflags = enic_cq_rx_desc_ciflags(cqrd);
bwflags = enic_cq_rx_desc_bwflags(cqrd);
+ vlan_tci = enic_cq_rx_desc_vlan(cqrd);
mbuf->ol_flags = 0;
if (unlikely(!enic_cq_rx_desc_eop(ciflags)))
goto mbuf_flags_done;
- /* VLAN stripping */
+ /* VLAN STRIPPED flag. The L2 packet type updated here also */
if (bwflags & CQ_ENET_RQ_DESC_FLAGS_VLAN_STRIPPED) {
pkt_flags |= PKT_RX_VLAN_PKT | PKT_RX_VLAN_STRIPPED;
- mbuf->vlan_tci = enic_cq_rx_desc_vlan(cqrd);
+ mbuf->packet_type |= RTE_PTYPE_L2_ETHER;
} else {
- mbuf->vlan_tci = 0;
+ if (vlan_tci != 0)
+ mbuf->packet_type |= RTE_PTYPE_L2_ETHER_VLAN;
+ else
+ mbuf->packet_type |= RTE_PTYPE_L2_ETHER;
}
-
- /* RSS flag */
- if (enic_cq_rx_desc_rss_type(cqrd)) {
+ mbuf->vlan_tci = vlan_tci;
+
+ if ((cqd->type_color & CQ_DESC_TYPE_MASK) == CQ_DESC_TYPE_CLASSIFIER) {
+ struct cq_enet_rq_clsf_desc *clsf_cqd;
+ uint16_t filter_id;
+ clsf_cqd = (struct cq_enet_rq_clsf_desc *)cqd;
+ filter_id = clsf_cqd->filter_id;
+ if (filter_id) {
+ pkt_flags |= PKT_RX_FDIR;
+ if (filter_id != ENIC_MAGIC_FILTER_ID) {
+ mbuf->hash.fdir.hi = clsf_cqd->filter_id;
+ pkt_flags |= PKT_RX_FDIR_ID;
+ }
+ }
+ } else if (enic_cq_rx_desc_rss_type(cqrd)) {
+ /* RSS flag */
pkt_flags |= PKT_RX_RSS_HASH;
mbuf->hash.rss = enic_cq_rx_desc_rss_hash(cqrd);
}
/* checksum flags */
- if (!enic_cq_rx_desc_csum_not_calc(cqrd) &&
- (mbuf->packet_type & RTE_PTYPE_L3_IPV4)) {
- if (unlikely(!enic_cq_rx_desc_ipv4_csum_ok(cqrd)))
- pkt_flags |= PKT_RX_IP_CKSUM_BAD;
- if (mbuf->packet_type & (RTE_PTYPE_L4_UDP | RTE_PTYPE_L4_TCP)) {
- if (unlikely(!enic_cq_rx_desc_tcp_udp_csum_ok(cqrd)))
- pkt_flags |= PKT_RX_L4_CKSUM_BAD;
+ if (mbuf->packet_type & RTE_PTYPE_L3_IPV4) {
+ if (enic_cq_rx_desc_csum_not_calc(cqrd))
+ pkt_flags |= (PKT_RX_IP_CKSUM_UNKNOWN &
+ PKT_RX_L4_CKSUM_UNKNOWN);
+ else {
+ uint32_t l4_flags;
+ l4_flags = mbuf->packet_type & RTE_PTYPE_L4_MASK;
+
+ if (enic_cq_rx_desc_ipv4_csum_ok(cqrd))
+ pkt_flags |= PKT_RX_IP_CKSUM_GOOD;
+ else
+ pkt_flags |= PKT_RX_IP_CKSUM_BAD;
+
+ if (l4_flags & (RTE_PTYPE_L4_UDP | RTE_PTYPE_L4_TCP)) {
+ if (enic_cq_rx_desc_tcp_udp_csum_ok(cqrd))
+ pkt_flags |= PKT_RX_L4_CKSUM_GOOD;
+ else
+ pkt_flags |= PKT_RX_L4_CKSUM_BAD;
+ }
}
}
mbuf->ol_flags = pkt_flags;
}
+/* dummy receive function to replace actual function in
+ * order to do safe reconfiguration operations.
+ */
+uint16_t
+enic_dummy_recv_pkts(__rte_unused void *rx_queue,
+ __rte_unused struct rte_mbuf **rx_pkts,
+ __rte_unused uint16_t nb_pkts)
+{
+ return 0;
+}
+
uint16_t
enic_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
uint16_t nb_pkts)
{
- struct vnic_rq *rq = rx_queue;
- struct enic *enic = vnic_dev_priv(rq->vdev);
- unsigned int rx_id;
+ struct vnic_rq *sop_rq = rx_queue;
+ struct vnic_rq *data_rq;
+ struct vnic_rq *rq;
+ struct enic *enic = vnic_dev_priv(sop_rq->vdev);
+ uint16_t cq_idx;
+ uint16_t rq_idx;
+ uint16_t rq_num;
struct rte_mbuf *nmb, *rxmb;
- uint16_t nb_rx = 0, nb_err = 0;
- uint16_t nb_hold;
+ uint16_t nb_rx = 0;
struct vnic_cq *cq;
volatile struct cq_desc *cqd_ptr;
uint8_t color;
+ uint16_t seg_length;
+ struct rte_mbuf *first_seg = sop_rq->pkt_first_seg;
+ struct rte_mbuf *last_seg = sop_rq->pkt_last_seg;
- cq = &enic->cq[enic_cq_rq(enic, rq->index)];
- rx_id = cq->to_clean; /* index of cqd, rqd, mbuf_table */
- cqd_ptr = (struct cq_desc *)(cq->ring.descs) + rx_id;
+ cq = &enic->cq[enic_cq_rq(enic, sop_rq->index)];
+ cq_idx = cq->to_clean; /* index of cqd, rqd, mbuf_table */
+ cqd_ptr = (struct cq_desc *)(cq->ring.descs) + cq_idx;
- nb_hold = rq->rx_nb_hold; /* mbufs held by software */
+ data_rq = &enic->rq[sop_rq->data_queue_idx];
while (nb_rx < nb_pkts) {
volatile struct rq_enet_desc *rqd_ptr;
dma_addr_t dma_addr;
struct cq_desc cqd;
uint8_t packet_error;
+ uint16_t ciflags;
/* Check for pkts available */
color = (cqd_ptr->type_color >> CQ_DESC_COLOR_SHIFT)
if (color == cq->last_color)
break;
- /* Get the cq descriptor and rq pointer */
+ /* Get the cq descriptor and extract rq info from it */
cqd = *cqd_ptr;
- rqd_ptr = (struct rq_enet_desc *)(rq->ring.descs) + rx_id;
+ rq_num = cqd.q_number & CQ_DESC_Q_NUM_MASK;
+ rq_idx = cqd.completed_index & CQ_DESC_COMP_NDX_MASK;
+
+ rq = &enic->rq[rq_num];
+ rqd_ptr = ((struct rq_enet_desc *)rq->ring.descs) + rq_idx;
/* allocate a new mbuf */
nmb = rte_mbuf_raw_alloc(rq->mp);
packet_error = enic_cq_rx_check_err(&cqd);
/* Get the mbuf to return and replace with one just allocated */
- rxmb = rq->mbuf_ring[rx_id];
- rq->mbuf_ring[rx_id] = nmb;
+ rxmb = rq->mbuf_ring[rq_idx];
+ rq->mbuf_ring[rq_idx] = nmb;
/* Increment cqd, rqd, mbuf_table index */
- rx_id++;
- if (unlikely(rx_id == rq->ring.desc_count)) {
- rx_id = 0;
+ cq_idx++;
+ if (unlikely(cq_idx == cq->ring.desc_count)) {
+ cq_idx = 0;
cq->last_color = cq->last_color ? 0 : 1;
}
/* Prefetch next mbuf & desc while processing current one */
- cqd_ptr = (struct cq_desc *)(cq->ring.descs) + rx_id;
+ cqd_ptr = (struct cq_desc *)(cq->ring.descs) + cq_idx;
rte_enic_prefetch(cqd_ptr);
- rte_enic_prefetch(rq->mbuf_ring[rx_id]);
- rte_enic_prefetch((struct rq_enet_desc *)(rq->ring.descs)
- + rx_id);
+
+ ciflags = enic_cq_rx_desc_ciflags(
+ (struct cq_enet_rq_desc *)&cqd);
/* Push descriptor for newly allocated mbuf */
- dma_addr = (dma_addr_t)(nmb->buf_physaddr
- + RTE_PKTMBUF_HEADROOM);
- rqd_ptr->address = rte_cpu_to_le_64(dma_addr);
- rqd_ptr->length_type = cpu_to_le16(nmb->buf_len
- - RTE_PKTMBUF_HEADROOM);
+ nmb->data_off = RTE_PKTMBUF_HEADROOM;
+ dma_addr = (dma_addr_t)(nmb->buf_physaddr +
+ RTE_PKTMBUF_HEADROOM);
+ rq_enet_desc_enc(rqd_ptr, dma_addr,
+ (rq->is_sop ? RQ_ENET_TYPE_ONLY_SOP
+ : RQ_ENET_TYPE_NOT_SOP),
+ nmb->buf_len - RTE_PKTMBUF_HEADROOM);
+
+ /* Fill in the rest of the mbuf */
+ seg_length = enic_cq_rx_desc_n_bytes(&cqd);
+
+ if (rq->is_sop) {
+ first_seg = rxmb;
+ first_seg->pkt_len = seg_length;
+ } else {
+ first_seg->pkt_len = (uint16_t)(first_seg->pkt_len
+ + seg_length);
+ first_seg->nb_segs++;
+ last_seg->next = rxmb;
+ }
+
+ rxmb->port = enic->port_id;
+ rxmb->data_len = seg_length;
+
+ rq->rx_nb_hold++;
+
+ if (!(enic_cq_rx_desc_eop(ciflags))) {
+ last_seg = rxmb;
+ continue;
+ }
+
+ /* cq rx flags are only valid if eop bit is set */
+ first_seg->packet_type = enic_cq_rx_flags_to_pkt_type(&cqd);
+ enic_cq_rx_to_pkt_flags(&cqd, first_seg);
- /* Drop incoming bad packet */
if (unlikely(packet_error)) {
- rte_pktmbuf_free(rxmb);
- nb_err++;
+ rte_pktmbuf_free(first_seg);
+ rte_atomic64_inc(&enic->soft_stats.rx_packet_errors);
continue;
}
- /* Fill in the rest of the mbuf */
- rxmb->data_off = RTE_PKTMBUF_HEADROOM;
- rxmb->nb_segs = 1;
- rxmb->next = NULL;
- rxmb->port = enic->port_id;
- rxmb->pkt_len = enic_cq_rx_desc_n_bytes(&cqd);
- rxmb->packet_type = enic_cq_rx_flags_to_pkt_type(&cqd);
- enic_cq_rx_to_pkt_flags(&cqd, rxmb);
- rxmb->data_len = rxmb->pkt_len;
/* prefetch mbuf data for caller */
- rte_packet_prefetch(RTE_PTR_ADD(rxmb->buf_addr,
+ rte_packet_prefetch(RTE_PTR_ADD(first_seg->buf_addr,
RTE_PKTMBUF_HEADROOM));
/* store the mbuf address into the next entry of the array */
- rx_pkts[nb_rx++] = rxmb;
+ rx_pkts[nb_rx++] = first_seg;
}
- nb_hold += nb_rx + nb_err;
- cq->to_clean = rx_id;
+ sop_rq->pkt_first_seg = first_seg;
+ sop_rq->pkt_last_seg = last_seg;
+
+ cq->to_clean = cq_idx;
+
+ if ((sop_rq->rx_nb_hold + data_rq->rx_nb_hold) >
+ sop_rq->rx_free_thresh) {
+ if (data_rq->in_use) {
+ data_rq->posted_index =
+ enic_ring_add(data_rq->ring.desc_count,
+ data_rq->posted_index,
+ data_rq->rx_nb_hold);
+ data_rq->rx_nb_hold = 0;
+ }
+ sop_rq->posted_index = enic_ring_add(sop_rq->ring.desc_count,
+ sop_rq->posted_index,
+ sop_rq->rx_nb_hold);
+ sop_rq->rx_nb_hold = 0;
- if (nb_hold > rq->rx_free_thresh) {
- rq->posted_index = enic_ring_add(rq->ring.desc_count,
- rq->posted_index, nb_hold);
- nb_hold = 0;
rte_mb();
- iowrite32(rq->posted_index, &rq->ctrl->posted_index);
+ if (data_rq->in_use)
+ iowrite32_relaxed(data_rq->posted_index,
+ &data_rq->ctrl->posted_index);
+ rte_compiler_barrier();
+ iowrite32_relaxed(sop_rq->posted_index,
+ &sop_rq->ctrl->posted_index);
}
- rq->rx_nb_hold = nb_hold;
return nb_rx;
}
pool = ((struct rte_mbuf *)buf->mb)->pool;
for (i = 0; i < nb_to_free; i++) {
buf = &wq->bufs[tail_idx];
- m = (struct rte_mbuf *)(buf->mb);
+ m = rte_pktmbuf_prefree_seg((struct rte_mbuf *)(buf->mb));
+ buf->mb = NULL;
+
+ if (unlikely(m == NULL)) {
+ tail_idx = enic_ring_incr(desc_count, tail_idx);
+ continue;
+ }
+
if (likely(m->pool == pool)) {
- ENIC_ASSERT(nb_free < ENIC_MAX_WQ_DESCS);
+ RTE_ASSERT(nb_free < ENIC_MAX_WQ_DESCS);
free[nb_free++] = m;
} else {
rte_mempool_put_bulk(pool, (void *)free, nb_free);
pool = m->pool;
}
tail_idx = enic_ring_incr(desc_count, tail_idx);
- buf->mb = NULL;
}
- rte_mempool_put_bulk(pool, (void **)free, nb_free);
+ if (nb_free > 0)
+ rte_mempool_put_bulk(pool, (void **)free, nb_free);
wq->tail_idx = tail_idx;
wq->ring.desc_avail += nb_to_free;
uint8_t vlan_tag_insert;
uint8_t eop;
uint64_t bus_addr;
+ uint8_t offload_mode;
+ uint16_t header_len;
enic_cleanup_wq(enic, wq);
wq_desc_avail = vnic_wq_desc_avail(wq);
for (index = 0; index < nb_pkts; index++) {
tx_pkt = *tx_pkts++;
+ pkt_len = tx_pkt->pkt_len;
+ data_len = tx_pkt->data_len;
+ ol_flags = tx_pkt->ol_flags;
nb_segs = tx_pkt->nb_segs;
+
+ if (pkt_len > ENIC_TX_MAX_PKT_SIZE) {
+ rte_pktmbuf_free(tx_pkt);
+ rte_atomic64_inc(&enic->soft_stats.tx_oversized);
+ continue;
+ }
+
if (nb_segs > wq_desc_avail) {
if (index > 0)
goto post;
goto done;
}
- pkt_len = tx_pkt->pkt_len;
- data_len = tx_pkt->data_len;
- ol_flags = tx_pkt->ol_flags;
mss = 0;
vlan_id = 0;
vlan_tag_insert = 0;
desc_p = descs + head_idx;
eop = (data_len == pkt_len);
-
- if (ol_flags & ol_flags_mask) {
- if (ol_flags & PKT_TX_VLAN_PKT) {
- vlan_tag_insert = 1;
- vlan_id = tx_pkt->vlan_tci;
+ offload_mode = WQ_ENET_OFFLOAD_MODE_CSUM;
+ header_len = 0;
+
+ if (tx_pkt->tso_segsz) {
+ header_len = tso_header_len(tx_pkt);
+ if (header_len) {
+ offload_mode = WQ_ENET_OFFLOAD_MODE_TSO;
+ mss = tx_pkt->tso_segsz;
}
-
+ }
+ if ((ol_flags & ol_flags_mask) && (header_len == 0)) {
if (ol_flags & PKT_TX_IP_CKSUM)
mss |= ENIC_CALC_IP_CKSUM;
}
}
- wq_enet_desc_enc(&desc_tmp, bus_addr, data_len, mss, 0, 0, eop,
- eop, 0, vlan_tag_insert, vlan_id, 0);
+ if (ol_flags & PKT_TX_VLAN_PKT) {
+ vlan_tag_insert = 1;
+ vlan_id = tx_pkt->vlan_tci;
+ }
+
+ wq_enet_desc_enc(&desc_tmp, bus_addr, data_len, mss, header_len,
+ offload_mode, eop, eop, 0, vlan_tag_insert,
+ vlan_id, 0);
*desc_p = desc_tmp;
buf = &wq->bufs[head_idx];
+ tx_pkt->data_off);
wq_enet_desc_enc((struct wq_enet_desc *)
&desc_tmp, bus_addr, data_len,
- mss, 0, 0, eop, eop, 0,
- vlan_tag_insert, vlan_id, 0);
+ mss, 0, offload_mode, eop, eop,
+ 0, vlan_tag_insert, vlan_id,
+ 0);
*desc_p = desc_tmp;
buf = &wq->bufs[head_idx];
}
post:
rte_wmb();
- iowrite32(head_idx, &wq->ctrl->posted_index);
+ iowrite32_relaxed(head_idx, &wq->ctrl->posted_index);
done:
wq->ring.desc_avail = wq_desc_avail;
wq->head_idx = head_idx;