ixgbe: improve slow-path perf with vector scattered Rx

Provide a wrapper routine to enable receive of scattered packets with a
vector driver. This improves the performance of the slow-path RX.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
This commit is contained in:
Bruce Richardson 2014-09-11 14:15:47 +01:00 committed by Thomas Monjalon
parent 1f22652ca8
commit cf4b4708a8
3 changed files with 186 additions and 9 deletions

View File

@ -3476,12 +3476,20 @@ ixgbe_dev_rx_init(struct rte_eth_dev *dev)
if ((dev->data->dev_conf.rxmode.max_rx_pkt_len +
2 * IXGBE_VLAN_TAG_SIZE) > buf_size){
dev->data->scattered_rx = 1;
#ifdef RTE_IXGBE_INC_VECTOR
dev->rx_pkt_burst = ixgbe_recv_scattered_pkts_vec;
#else
dev->rx_pkt_burst = ixgbe_recv_scattered_pkts;
#endif
}
}
if (dev->data->dev_conf.rxmode.enable_scatter) {
#ifdef RTE_IXGBE_INC_VECTOR
dev->rx_pkt_burst = ixgbe_recv_scattered_pkts_vec;
#else
dev->rx_pkt_burst = ixgbe_recv_scattered_pkts;
#endif
dev->data->scattered_rx = 1;
}
@ -3969,12 +3977,20 @@ ixgbevf_dev_rx_init(struct rte_eth_dev *dev)
if ((dev->data->dev_conf.rxmode.max_rx_pkt_len +
2 * IXGBE_VLAN_TAG_SIZE) > buf_size) {
dev->data->scattered_rx = 1;
#ifdef RTE_IXGBE_INC_VECTOR
dev->rx_pkt_burst = ixgbe_recv_scattered_pkts_vec;
#else
dev->rx_pkt_burst = ixgbe_recv_scattered_pkts;
#endif
}
}
if (dev->data->dev_conf.rxmode.enable_scatter) {
#ifdef RTE_IXGBE_INC_VECTOR
dev->rx_pkt_burst = ixgbe_recv_scattered_pkts_vec;
#else
dev->rx_pkt_burst = ixgbe_recv_scattered_pkts;
#endif
dev->data->scattered_rx = 1;
}

View File

@ -244,8 +244,12 @@ struct ixgbe_txq_ops {
IXGBE_ADVTXD_DCMD_EOP)
#ifdef RTE_IXGBE_INC_VECTOR
uint16_t ixgbe_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts);
uint16_t ixgbe_xmit_pkts_vec(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts);
uint16_t ixgbe_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
uint16_t nb_pkts);
uint16_t ixgbe_recv_scattered_pkts_vec(void *rx_queue,
struct rte_mbuf **rx_pkts, uint16_t nb_pkts);
uint16_t ixgbe_xmit_pkts_vec(void *tx_queue, struct rte_mbuf **tx_pkts,
uint16_t nb_pkts);
int ixgbe_txq_vec_setup(struct igb_tx_queue *txq);
int ixgbe_rxq_vec_setup(struct igb_rx_queue *rxq);
int ixgbe_rx_vec_condition_check(struct rte_eth_dev *dev);

View File

@ -164,12 +164,11 @@ desc_to_olflags_v(__m128i descs[4], struct rte_mbuf **rx_pkts)
* numbers of DD bit
* - don't support ol_flags for rss and csum err
*/
uint16_t
ixgbe_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
uint16_t nb_pkts)
static inline uint16_t
_recv_raw_pkts_vec(struct igb_rx_queue *rxq, struct rte_mbuf **rx_pkts,
uint16_t nb_pkts, uint8_t *split_packet)
{
volatile union ixgbe_adv_rx_desc *rxdp;
struct igb_rx_queue *rxq = rx_queue;
struct igb_rx_entry *sw_ring;
uint16_t nb_pkts_recd;
int pos;
@ -182,7 +181,7 @@ ixgbe_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
-rxq->crc_len, /* sub crc on data_len */
0 /* ignore pkt_type field */
);
__m128i dd_check;
__m128i dd_check, eop_check;
if (unlikely(nb_pkts < RTE_IXGBE_VPMD_RX_BURST))
return 0;
@ -207,6 +206,9 @@ ixgbe_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
/* 4 packets DD mask */
dd_check = _mm_set_epi64x(0x0000000100000001LL, 0x0000000100000001LL);
/* 4 packets EOP mask */
eop_check = _mm_set_epi64x(0x0000000200000002LL, 0x0000000200000002LL);
/* mask to shuffle from desc. to mbuf */
shuf_msk = _mm_set_epi8(
7, 6, 5, 4, /* octet 4~7, 32bits rss */
@ -218,7 +220,6 @@ ixgbe_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
0xFF, 0xFF /* skip pkt_type field */
);
/* Cache is empty -> need to scan the buffer rings, but first move
* the next 'n' mbufs into the cache */
sw_ring = &rxq->sw_ring[rxq->rx_tail];
@ -227,6 +228,7 @@ ixgbe_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
* A. load 4 packet in one loop
* B. copy 4 mbuf point from swring to rx_pkts
* C. calc the number of DD bits among the 4 packets
* [C*. extract the end-of-packet bit, if requested]
* D. fill info. from desc to mbuf
*/
for (pos = 0, nb_pkts_recd = 0; pos < RTE_IXGBE_VPMD_RX_BURST;
@ -237,6 +239,13 @@ ixgbe_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
__m128i zero, staterr, sterr_tmp1, sterr_tmp2;
__m128i mbp1, mbp2; /* two mbuf pointer in one XMM reg. */
if (split_packet) {
rte_prefetch0(&rx_pkts[pos]->cacheline1);
rte_prefetch0(&rx_pkts[pos + 1]->cacheline1);
rte_prefetch0(&rx_pkts[pos + 2]->cacheline1);
rte_prefetch0(&rx_pkts[pos + 3]->cacheline1);
}
/* B.1 load 1 mbuf point */
mbp1 = _mm_loadu_si128((__m128i *)&sw_ring[pos]);
@ -295,7 +304,34 @@ ixgbe_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
pkt_mb2 = _mm_add_epi16(pkt_mb2, crc_adjust);
pkt_mb1 = _mm_add_epi16(pkt_mb1, crc_adjust);
/* C.3 calc avaialbe number of desc */
/* C* extract and record EOP bit */
if (split_packet) {
__m128i eop_shuf_mask = _mm_set_epi8(
0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF,
0xFF, 0xFF, 0xFF, 0xFF,
0x04, 0x0C, 0x00, 0x08
);
/* and with mask to extract bits, flipping 1-0 */
__m128i eop_bits = _mm_andnot_si128(staterr, eop_check);
/* the staterr values are not in order, as the count
* count of dd bits doesn't care. However, for end of
* packet tracking, we do care, so shuffle. This also
* compresses the 32-bit values to 8-bit */
eop_bits = _mm_shuffle_epi8(eop_bits, eop_shuf_mask);
/* store the resulting 32-bit value */
*(int *)split_packet = _mm_cvtsi128_si32(eop_bits);
split_packet += RTE_IXGBE_DESCS_PER_LOOP;
/* zero-out next pointers */
rx_pkts[pos]->next = NULL;
rx_pkts[pos + 1]->next = NULL;
rx_pkts[pos + 2]->next = NULL;
rx_pkts[pos + 3]->next = NULL;
}
/* C.3 calc available number of desc */
staterr = _mm_and_si128(staterr, dd_check);
staterr = _mm_packs_epi32(staterr, zero);
@ -319,6 +355,127 @@ ixgbe_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
return nb_pkts_recd;
}
/*
* vPMD receive routine, now only accept (nb_pkts == RTE_IXGBE_VPMD_RX_BURST)
* in one loop
*
* Notice:
* - nb_pkts < RTE_IXGBE_VPMD_RX_BURST, just return no packet
* - nb_pkts > RTE_IXGBE_VPMD_RX_BURST, only scan RTE_IXGBE_VPMD_RX_BURST
* numbers of DD bit
* - don't support ol_flags for rss and csum err
*/
uint16_t
ixgbe_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
uint16_t nb_pkts)
{
return _recv_raw_pkts_vec(rx_queue, rx_pkts, nb_pkts, NULL);
}
static inline uint16_t
reassemble_packets(struct igb_rx_queue *rxq, struct rte_mbuf **rx_bufs,
uint16_t nb_bufs, uint8_t *split_flags)
{
struct rte_mbuf *pkts[RTE_IXGBE_VPMD_RX_BURST]; /*finished pkts*/
struct rte_mbuf *start = rxq->pkt_first_seg;
struct rte_mbuf *end = rxq->pkt_last_seg;
unsigned pkt_idx = 0, buf_idx = 0;
while (buf_idx < nb_bufs) {
if (end != NULL) {
/* processing a split packet */
end->next = rx_bufs[buf_idx];
rx_bufs[buf_idx]->data_len += rxq->crc_len;
start->nb_segs++;
start->pkt_len += rx_bufs[buf_idx]->data_len;
end = end->next;
if (!split_flags[buf_idx]) {
/* it's the last packet of the set */
start->hash = end->hash;
start->ol_flags = end->ol_flags;
/* we need to strip crc for the whole packet */
start->pkt_len -= rxq->crc_len;
if (end->data_len > rxq->crc_len)
end->data_len -= rxq->crc_len;
else {
/* free up last mbuf */
struct rte_mbuf *secondlast = start;
while (secondlast->next != end)
secondlast = secondlast->next;
secondlast->data_len -= (rxq->crc_len -
end->data_len);
secondlast->next = NULL;
rte_pktmbuf_free_seg(end);
end = secondlast;
}
pkts[pkt_idx++] = start;
start = end = NULL;
}
} else {
/* not processing a split packet */
if (!split_flags[buf_idx]) {
/* not a split packet, save and skip */
pkts[pkt_idx++] = rx_bufs[buf_idx];
continue;
}
end = start = rx_bufs[buf_idx];
rx_bufs[buf_idx]->data_len += rxq->crc_len;
rx_bufs[buf_idx]->pkt_len += rxq->crc_len;
}
buf_idx++;
}
/* save the partial packet for next time */
rxq->pkt_first_seg = start;
rxq->pkt_last_seg = end;
memcpy(rx_bufs, pkts, pkt_idx * (sizeof(*pkts)));
return pkt_idx;
}
/*
* vPMD receive routine that reassembles scattered packets
*
* Notice:
* - don't support ol_flags for rss and csum err
* - now only accept (nb_pkts == RTE_IXGBE_VPMD_RX_BURST)
*/
uint16_t
ixgbe_recv_scattered_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
uint16_t nb_pkts)
{
struct igb_rx_queue *rxq = rx_queue;
uint8_t split_flags[RTE_IXGBE_VPMD_RX_BURST] = {0};
/* get some new buffers */
uint16_t nb_bufs = _recv_raw_pkts_vec(rxq, rx_pkts, nb_pkts,
split_flags);
if (nb_bufs == 0)
return 0;
/* happy day case, full burst + no packets to be joined */
const uint32_t *split_fl32 = (uint32_t *)split_flags;
if (rxq->pkt_first_seg == NULL &&
split_fl32[0] == 0 && split_fl32[1] == 0 &&
split_fl32[2] == 0 && split_fl32[3] == 0)
return nb_bufs;
/* reassemble any packets that need reassembly*/
unsigned i = 0;
if (rxq->pkt_first_seg == NULL) {
/* find the first split flag, and only reassemble then*/
while (!split_flags[i] && i < nb_bufs)
i++;
if (i == nb_bufs)
return nb_bufs;
}
return i + reassemble_packets(rxq, &rx_pkts[i], nb_bufs - i,
&split_flags[i]);
}
static inline void
vtx1(volatile union ixgbe_adv_tx_desc *txdp,
struct rte_mbuf *pkt, uint64_t flags)