Commit 126cdfe1 authored by Maciej Fijalkowski's avatar Maciej Fijalkowski Committed by Daniel Borkmann

ice: xsk: Improve AF_XDP ZC Tx and use batching API

Apply the logic that was done for regular XDP from commit 9610bd98
("ice: optimize XDP_TX workloads") to the ZC side of the driver. On top
of that, introduce batching to Tx that is inspired by i40e's
implementation with adjustments to the cleaning logic - take into the
account NAPI budget in ice_clean_xdp_irq_zc().

Separating the stats structs onto separate cache lines seemed to improve
the performance.
Signed-off-by: default avatarMaciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: default avatarDaniel Borkmann <daniel@iogearbox.net>
Reviewed-by: default avatarAlexander Lobakin <alexandr.lobakin@intel.com>
Link: https://lore.kernel.org/bpf/20220125160446.78976-8-maciej.fijalkowski@intel.com
parent 86e3f78c
...@@ -1462,7 +1462,7 @@ int ice_napi_poll(struct napi_struct *napi, int budget) ...@@ -1462,7 +1462,7 @@ int ice_napi_poll(struct napi_struct *napi, int budget)
bool wd; bool wd;
if (tx_ring->xsk_pool) if (tx_ring->xsk_pool)
wd = ice_clean_tx_irq_zc(tx_ring, budget); wd = ice_xmit_zc(tx_ring, ICE_DESC_UNUSED(tx_ring), budget);
else if (ice_ring_is_xdp(tx_ring)) else if (ice_ring_is_xdp(tx_ring))
wd = true; wd = true;
else else
......
...@@ -322,17 +322,17 @@ struct ice_tx_ring { ...@@ -322,17 +322,17 @@ struct ice_tx_ring {
u16 count; /* Number of descriptors */ u16 count; /* Number of descriptors */
u16 q_index; /* Queue number of ring */ u16 q_index; /* Queue number of ring */
/* stats structs */ /* stats structs */
struct ice_q_stats stats;
struct u64_stats_sync syncp;
struct ice_txq_stats tx_stats; struct ice_txq_stats tx_stats;
/* CL3 - 3rd cacheline starts here */ /* CL3 - 3rd cacheline starts here */
struct ice_q_stats stats;
struct u64_stats_sync syncp;
struct rcu_head rcu; /* to avoid race on free */ struct rcu_head rcu; /* to avoid race on free */
DECLARE_BITMAP(xps_state, ICE_TX_NBITS); /* XPS Config State */ DECLARE_BITMAP(xps_state, ICE_TX_NBITS); /* XPS Config State */
struct ice_channel *ch; struct ice_channel *ch;
struct ice_ptp_tx *tx_tstamps; struct ice_ptp_tx *tx_tstamps;
spinlock_t tx_lock; spinlock_t tx_lock;
u32 txq_teid; /* Added Tx queue TEID */ u32 txq_teid; /* Added Tx queue TEID */
/* CL4 - 4th cacheline starts here */
#define ICE_TX_FLAGS_RING_XDP BIT(0) #define ICE_TX_FLAGS_RING_XDP BIT(0)
u8 flags; u8 flags;
u8 dcb_tc; /* Traffic class of ring */ u8 dcb_tc; /* Traffic class of ring */
......
...@@ -679,134 +679,212 @@ int ice_clean_rx_irq_zc(struct ice_rx_ring *rx_ring, int budget) ...@@ -679,134 +679,212 @@ int ice_clean_rx_irq_zc(struct ice_rx_ring *rx_ring, int budget)
} }
/** /**
* ice_xmit_zc - Completes AF_XDP entries, and cleans XDP entries * ice_clean_xdp_tx_buf - Free and unmap XDP Tx buffer
* @xdp_ring: XDP Tx ring * @xdp_ring: XDP Tx ring
* @budget: max number of frames to xmit * @tx_buf: Tx buffer to clean
*/
static void
ice_clean_xdp_tx_buf(struct ice_tx_ring *xdp_ring, struct ice_tx_buf *tx_buf)
{
xdp_return_frame((struct xdp_frame *)tx_buf->raw_buf);
dma_unmap_single(xdp_ring->dev, dma_unmap_addr(tx_buf, dma),
dma_unmap_len(tx_buf, len), DMA_TO_DEVICE);
dma_unmap_len_set(tx_buf, len, 0);
}
/**
* ice_clean_xdp_irq_zc - Reclaim resources after transmit completes on XDP ring
* @xdp_ring: XDP ring to clean
* @napi_budget: amount of descriptors that NAPI allows us to clean
* *
* Returns true if cleanup/transmission is done. * Returns count of cleaned descriptors
*/ */
static bool ice_xmit_zc(struct ice_tx_ring *xdp_ring, int budget) static u16 ice_clean_xdp_irq_zc(struct ice_tx_ring *xdp_ring, int napi_budget)
{ {
struct ice_tx_desc *tx_desc = NULL; u16 tx_thresh = ICE_RING_QUARTER(xdp_ring);
bool work_done = true; int budget = napi_budget / tx_thresh;
struct xdp_desc desc; u16 ntc = xdp_ring->next_to_clean;
dma_addr_t dma; u16 next_dd = xdp_ring->next_dd;
u16 cleared_dds = 0;
while (likely(budget-- > 0)) { do {
struct ice_tx_desc *next_dd_desc;
u16 desc_cnt = xdp_ring->count;
struct ice_tx_buf *tx_buf; struct ice_tx_buf *tx_buf;
u32 xsk_frames;
u16 i;
if (unlikely(!ICE_DESC_UNUSED(xdp_ring))) { next_dd_desc = ICE_TX_DESC(xdp_ring, next_dd);
xdp_ring->tx_stats.tx_busy++; if (!(next_dd_desc->cmd_type_offset_bsz &
work_done = false; cpu_to_le64(ICE_TX_DESC_DTYPE_DESC_DONE)))
break;
}
tx_buf = &xdp_ring->tx_buf[xdp_ring->next_to_use];
if (!xsk_tx_peek_desc(xdp_ring->xsk_pool, &desc))
break; break;
dma = xsk_buff_raw_get_dma(xdp_ring->xsk_pool, desc.addr); cleared_dds++;
xsk_buff_raw_dma_sync_for_device(xdp_ring->xsk_pool, dma, xsk_frames = 0;
desc.len);
tx_buf->bytecount = desc.len; for (i = 0; i < tx_thresh; i++) {
tx_buf = &xdp_ring->tx_buf[ntc];
tx_desc = ICE_TX_DESC(xdp_ring, xdp_ring->next_to_use); if (tx_buf->raw_buf) {
tx_desc->buf_addr = cpu_to_le64(dma); ice_clean_xdp_tx_buf(xdp_ring, tx_buf);
tx_desc->cmd_type_offset_bsz = tx_buf->raw_buf = NULL;
ice_build_ctob(ICE_TXD_LAST_DESC_CMD, 0, desc.len, 0); } else {
xsk_frames++;
}
xdp_ring->next_to_use++; ntc++;
if (xdp_ring->next_to_use == xdp_ring->count) if (ntc >= xdp_ring->count)
xdp_ring->next_to_use = 0; ntc = 0;
} }
if (xsk_frames)
xsk_tx_completed(xdp_ring->xsk_pool, xsk_frames);
next_dd_desc->cmd_type_offset_bsz = 0;
next_dd = next_dd + tx_thresh;
if (next_dd >= desc_cnt)
next_dd = tx_thresh - 1;
} while (budget--);
if (tx_desc) { xdp_ring->next_to_clean = ntc;
ice_xdp_ring_update_tail(xdp_ring); xdp_ring->next_dd = next_dd;
xsk_tx_release(xdp_ring->xsk_pool);
}
return budget > 0 && work_done; return cleared_dds * tx_thresh;
} }
/** /**
* ice_clean_xdp_tx_buf - Free and unmap XDP Tx buffer * ice_xmit_pkt - produce a single HW Tx descriptor out of AF_XDP descriptor
* @xdp_ring: XDP Tx ring * @xdp_ring: XDP ring to produce the HW Tx descriptor on
* @tx_buf: Tx buffer to clean * @desc: AF_XDP descriptor to pull the DMA address and length from
* @total_bytes: bytes accumulator that will be used for stats update
*/ */
static void static void ice_xmit_pkt(struct ice_tx_ring *xdp_ring, struct xdp_desc *desc,
ice_clean_xdp_tx_buf(struct ice_tx_ring *xdp_ring, struct ice_tx_buf *tx_buf) unsigned int *total_bytes)
{ {
xdp_return_frame((struct xdp_frame *)tx_buf->raw_buf); struct ice_tx_desc *tx_desc;
dma_unmap_single(xdp_ring->dev, dma_unmap_addr(tx_buf, dma), dma_addr_t dma;
dma_unmap_len(tx_buf, len), DMA_TO_DEVICE);
dma_unmap_len_set(tx_buf, len, 0); dma = xsk_buff_raw_get_dma(xdp_ring->xsk_pool, desc->addr);
xsk_buff_raw_dma_sync_for_device(xdp_ring->xsk_pool, dma, desc->len);
tx_desc = ICE_TX_DESC(xdp_ring, xdp_ring->next_to_use++);
tx_desc->buf_addr = cpu_to_le64(dma);
tx_desc->cmd_type_offset_bsz = ice_build_ctob(ICE_TX_DESC_CMD_EOP,
0, desc->len, 0);
*total_bytes += desc->len;
} }
/** /**
* ice_clean_tx_irq_zc - Completes AF_XDP entries, and cleans XDP entries * ice_xmit_pkt_batch - produce a batch of HW Tx descriptors out of AF_XDP descriptors
* @xdp_ring: XDP Tx ring * @xdp_ring: XDP ring to produce the HW Tx descriptors on
* @budget: NAPI budget * @descs: AF_XDP descriptors to pull the DMA addresses and lengths from
* * @total_bytes: bytes accumulator that will be used for stats update
* Returns true if cleanup/tranmission is done.
*/ */
bool ice_clean_tx_irq_zc(struct ice_tx_ring *xdp_ring, int budget) static void ice_xmit_pkt_batch(struct ice_tx_ring *xdp_ring, struct xdp_desc *descs,
unsigned int *total_bytes)
{ {
int total_packets = 0, total_bytes = 0; u16 tx_thresh = ICE_RING_QUARTER(xdp_ring);
s16 ntc = xdp_ring->next_to_clean; u16 ntu = xdp_ring->next_to_use;
struct ice_tx_desc *tx_desc; struct ice_tx_desc *tx_desc;
struct ice_tx_buf *tx_buf; u32 i;
u32 xsk_frames = 0;
bool xmit_done;
tx_desc = ICE_TX_DESC(xdp_ring, ntc); loop_unrolled_for(i = 0; i < PKTS_PER_BATCH; i++) {
tx_buf = &xdp_ring->tx_buf[ntc]; dma_addr_t dma;
ntc -= xdp_ring->count;
do { dma = xsk_buff_raw_get_dma(xdp_ring->xsk_pool, descs[i].addr);
if (!(tx_desc->cmd_type_offset_bsz & xsk_buff_raw_dma_sync_for_device(xdp_ring->xsk_pool, dma, descs[i].len);
cpu_to_le64(ICE_TX_DESC_DTYPE_DESC_DONE)))
break;
total_bytes += tx_buf->bytecount; tx_desc = ICE_TX_DESC(xdp_ring, ntu++);
total_packets++; tx_desc->buf_addr = cpu_to_le64(dma);
tx_desc->cmd_type_offset_bsz = ice_build_ctob(ICE_TX_DESC_CMD_EOP,
0, descs[i].len, 0);
if (tx_buf->raw_buf) { *total_bytes += descs[i].len;
ice_clean_xdp_tx_buf(xdp_ring, tx_buf); }
tx_buf->raw_buf = NULL;
} else {
xsk_frames++;
}
tx_desc->cmd_type_offset_bsz = 0; xdp_ring->next_to_use = ntu;
tx_buf++;
tx_desc++;
ntc++;
if (unlikely(!ntc)) { if (xdp_ring->next_to_use > xdp_ring->next_rs) {
ntc -= xdp_ring->count; tx_desc = ICE_TX_DESC(xdp_ring, xdp_ring->next_rs);
tx_buf = xdp_ring->tx_buf; tx_desc->cmd_type_offset_bsz |=
tx_desc = ICE_TX_DESC(xdp_ring, 0); cpu_to_le64(ICE_TX_DESC_CMD_RS << ICE_TXD_QW1_CMD_S);
} xdp_ring->next_rs += tx_thresh;
}
}
prefetch(tx_desc); /**
* ice_fill_tx_hw_ring - produce the number of Tx descriptors onto ring
* @xdp_ring: XDP ring to produce the HW Tx descriptors on
* @descs: AF_XDP descriptors to pull the DMA addresses and lengths from
* @nb_pkts: count of packets to be send
* @total_bytes: bytes accumulator that will be used for stats update
*/
static void ice_fill_tx_hw_ring(struct ice_tx_ring *xdp_ring, struct xdp_desc *descs,
u32 nb_pkts, unsigned int *total_bytes)
{
u16 tx_thresh = ICE_RING_QUARTER(xdp_ring);
u32 batched, leftover, i;
batched = ALIGN_DOWN(nb_pkts, PKTS_PER_BATCH);
leftover = nb_pkts & (PKTS_PER_BATCH - 1);
for (i = 0; i < batched; i += PKTS_PER_BATCH)
ice_xmit_pkt_batch(xdp_ring, &descs[i], total_bytes);
for (; i < batched + leftover; i++)
ice_xmit_pkt(xdp_ring, &descs[i], total_bytes);
if (xdp_ring->next_to_use > xdp_ring->next_rs) {
struct ice_tx_desc *tx_desc;
tx_desc = ICE_TX_DESC(xdp_ring, xdp_ring->next_rs);
tx_desc->cmd_type_offset_bsz |=
cpu_to_le64(ICE_TX_DESC_CMD_RS << ICE_TXD_QW1_CMD_S);
xdp_ring->next_rs += tx_thresh;
}
}
} while (likely(--budget)); /**
* ice_xmit_zc - take entries from XSK Tx ring and place them onto HW Tx ring
* @xdp_ring: XDP ring to produce the HW Tx descriptors on
* @budget: number of free descriptors on HW Tx ring that can be used
* @napi_budget: amount of descriptors that NAPI allows us to clean
*
* Returns true if there is no more work that needs to be done, false otherwise
*/
bool ice_xmit_zc(struct ice_tx_ring *xdp_ring, u32 budget, int napi_budget)
{
struct xdp_desc *descs = xdp_ring->xsk_pool->tx_descs;
u16 tx_thresh = ICE_RING_QUARTER(xdp_ring);
u32 nb_pkts, nb_processed = 0;
unsigned int total_bytes = 0;
if (budget < tx_thresh)
budget += ice_clean_xdp_irq_zc(xdp_ring, napi_budget);
nb_pkts = xsk_tx_peek_release_desc_batch(xdp_ring->xsk_pool, budget);
if (!nb_pkts)
return true;
if (xdp_ring->next_to_use + nb_pkts >= xdp_ring->count) {
struct ice_tx_desc *tx_desc;
nb_processed = xdp_ring->count - xdp_ring->next_to_use;
ice_fill_tx_hw_ring(xdp_ring, descs, nb_processed, &total_bytes);
tx_desc = ICE_TX_DESC(xdp_ring, xdp_ring->next_rs);
tx_desc->cmd_type_offset_bsz |=
cpu_to_le64(ICE_TX_DESC_CMD_RS << ICE_TXD_QW1_CMD_S);
xdp_ring->next_rs = tx_thresh - 1;
xdp_ring->next_to_use = 0;
}
ntc += xdp_ring->count; ice_fill_tx_hw_ring(xdp_ring, &descs[nb_processed], nb_pkts - nb_processed,
xdp_ring->next_to_clean = ntc; &total_bytes);
if (xsk_frames) ice_xdp_ring_update_tail(xdp_ring);
xsk_tx_completed(xdp_ring->xsk_pool, xsk_frames); ice_update_tx_ring_stats(xdp_ring, nb_pkts, total_bytes);
if (xsk_uses_need_wakeup(xdp_ring->xsk_pool)) if (xsk_uses_need_wakeup(xdp_ring->xsk_pool))
xsk_set_tx_need_wakeup(xdp_ring->xsk_pool); xsk_set_tx_need_wakeup(xdp_ring->xsk_pool);
ice_update_tx_ring_stats(xdp_ring, total_packets, total_bytes); return nb_pkts < budget;
xmit_done = ice_xmit_zc(xdp_ring, ICE_DFLT_IRQ_WORK);
return budget > 0 && xmit_done;
} }
/** /**
......
...@@ -6,19 +6,37 @@ ...@@ -6,19 +6,37 @@
#include "ice_txrx.h" #include "ice_txrx.h"
#include "ice.h" #include "ice.h"
#define PKTS_PER_BATCH 8
#ifdef __clang__
#define loop_unrolled_for _Pragma("clang loop unroll_count(8)") for
#elif __GNUC__ >= 4
#define loop_unrolled_for _Pragma("GCC unroll 8") for
#else
#define loop_unrolled_for for
#endif
struct ice_vsi; struct ice_vsi;
#ifdef CONFIG_XDP_SOCKETS #ifdef CONFIG_XDP_SOCKETS
int ice_xsk_pool_setup(struct ice_vsi *vsi, struct xsk_buff_pool *pool, int ice_xsk_pool_setup(struct ice_vsi *vsi, struct xsk_buff_pool *pool,
u16 qid); u16 qid);
int ice_clean_rx_irq_zc(struct ice_rx_ring *rx_ring, int budget); int ice_clean_rx_irq_zc(struct ice_rx_ring *rx_ring, int budget);
bool ice_clean_tx_irq_zc(struct ice_tx_ring *xdp_ring, int budget);
int ice_xsk_wakeup(struct net_device *netdev, u32 queue_id, u32 flags); int ice_xsk_wakeup(struct net_device *netdev, u32 queue_id, u32 flags);
bool ice_alloc_rx_bufs_zc(struct ice_rx_ring *rx_ring, u16 count); bool ice_alloc_rx_bufs_zc(struct ice_rx_ring *rx_ring, u16 count);
bool ice_xsk_any_rx_ring_ena(struct ice_vsi *vsi); bool ice_xsk_any_rx_ring_ena(struct ice_vsi *vsi);
void ice_xsk_clean_rx_ring(struct ice_rx_ring *rx_ring); void ice_xsk_clean_rx_ring(struct ice_rx_ring *rx_ring);
void ice_xsk_clean_xdp_ring(struct ice_tx_ring *xdp_ring); void ice_xsk_clean_xdp_ring(struct ice_tx_ring *xdp_ring);
bool ice_xmit_zc(struct ice_tx_ring *xdp_ring, u32 budget, int napi_budget);
#else #else
static inline bool
ice_xmit_zc(struct ice_tx_ring __always_unused *xdp_ring,
u32 __always_unused budget,
int __always_unused napi_budget)
{
return false;
}
static inline int static inline int
ice_xsk_pool_setup(struct ice_vsi __always_unused *vsi, ice_xsk_pool_setup(struct ice_vsi __always_unused *vsi,
struct xsk_buff_pool __always_unused *pool, struct xsk_buff_pool __always_unused *pool,
...@@ -34,13 +52,6 @@ ice_clean_rx_irq_zc(struct ice_rx_ring __always_unused *rx_ring, ...@@ -34,13 +52,6 @@ ice_clean_rx_irq_zc(struct ice_rx_ring __always_unused *rx_ring,
return 0; return 0;
} }
static inline bool
ice_clean_tx_irq_zc(struct ice_tx_ring __always_unused *xdp_ring,
int __always_unused budget)
{
return false;
}
static inline bool static inline bool
ice_alloc_rx_bufs_zc(struct ice_rx_ring __always_unused *rx_ring, ice_alloc_rx_bufs_zc(struct ice_rx_ring __always_unused *rx_ring,
u16 __always_unused count) u16 __always_unused count)
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment