Introduce mb bit in xdp_buff/xdp_frame

RFC In review Lorenzo Bianconi 4 participants 26 messages Jul 23, 2020 – Jul 28, 2020

Subsystems (23)

Networking Network drivers BPF Intel Ethernet +19 more subsystems

lore

Share & tools

lore

Lorenzo Bianconi

author

2020-07-23 11:42 UTC #0 lore raw

cover Introduce mb bit in xdp_buff/xdp_frame

Introduce multi-buffer bit (mb) in xdp_frame/xdp_buffer to specify if shared_info area has been properly initialized for non-linear xdp buffers.
Initialize mb to 0 for all xdp drivers

Lorenzo Bianconi (22):
xdp: introduce mb in xdp_buff/xdp_frame
xdp: initialize xdp_buff mb bit to 0 in netif_receive_generic_xdp
net: virtio_net: initialize mb bit of xdp_buff to 0
net: xen-netfront: initialize mb bit of xdp_buff to 0
net: veth: initialize mb bit of xdp_buff to 0
net: hv_netvsc: initialize mb bit of xdp_buff to 0
net: bnxt: initialize mb bit in xdp_buff to 0
net: dpaa2: initialize mb bit in xdp_buff to 0
net: ti: initialize mb bit in xdp_buff to 0
net: nfp: initialize mb bit in xdp_buff to 0
net: mvpp2: initialize mb bit in xdp_buff to 0
net: sfc: initialize mb bit in xdp_buff to 0
net: qede: initialize mb bit in xdp_buff to 0
net: amazon: ena: initialize mb bit in xdp_buff to 0
net: cavium: thunder: initialize mb bit in xdp_buff to 0
net: socionext: initialize mb bit in xdp_buff to 0
net: tun: initialize mb bit in xdp_buff to 0
net: ixgbe: initialize mb bit in xdp_buff to 0
net: ice: initialize mb bit in xdp_buff to 0
net: i40e: initialize mb bit in xdp_buff to 0
net: mlx5: initialize mb bit in xdp_buff to 0
net: mlx4: initialize mb bit in xdp_buff to 0

 drivers/net/ethernet/amazon/ena/ena_netdev.c        | 1 +
 drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c       | 1 +
 drivers/net/ethernet/cavium/thunder/nicvf_main.c    | 1 +
 drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c    | 1 +
 drivers/net/ethernet/intel/i40e/i40e_txrx.c         | 1 +
 drivers/net/ethernet/intel/ice/ice_txrx.c           | 1 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c       | 1 +
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c   | 1 +
 drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c     | 1 +
 drivers/net/ethernet/mellanox/mlx4/en_rx.c          | 1 +
 drivers/net/ethernet/mellanox/mlx5/core/en_rx.c     | 1 +
 drivers/net/ethernet/netronome/nfp/nfp_net_common.c | 1 +
 drivers/net/ethernet/qlogic/qede/qede_fp.c          | 1 +
 drivers/net/ethernet/sfc/rx.c                       | 1 +
 drivers/net/ethernet/socionext/netsec.c             | 1 +
 drivers/net/ethernet/ti/cpsw.c                      | 1 +
 drivers/net/ethernet/ti/cpsw_new.c                  | 1 +
 drivers/net/hyperv/netvsc_bpf.c                     | 1 +
 drivers/net/tun.c                                   | 2 ++
 drivers/net/veth.c                                  | 1 +
 drivers/net/virtio_net.c                            | 2 ++
 drivers/net/xen-netfront.c                          | 1 +
 include/net/xdp.h                                   | 8 ++++++--
 net/core/dev.c                                      | 1 +
 24 files changed, 31 insertions(+), 2 deletions(-)

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #1 lore raw

RFC 1/22 xdp: introduce mb in xdp_buff/xdp_frame

Introduce multi-buffer bit (mb) in xdp_frame/xdp_buffer to specify if shared_info area has been properly initialized for non-linear xdp buffers

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 include/net/xdp.h | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

include/net/xdp.h +6 -2

index dbe9c60797e1..2ef6935c5646 100644
@@ -72,7 +72,8 @@ struct xdp_buff {
 	void *data_hard_start;
 	struct xdp_rxq_info *rxq;
 	struct xdp_txq_info *txq;
-	u32 frame_sz; /* frame size to deduce data_hard_end/reserved tailroom*/
+	u32 frame_sz:31; /* frame size to deduce data_hard_end/reserved tailroom*/
+	u32 mb:1; /* xdp non-linear buffer */
 };
 
 /* Reserve memory area at end-of data area.

@@ -96,7 +97,8 @@ struct xdp_frame {
 	u16 len;
 	u16 headroom;
 	u32 metasize:8;
-	u32 frame_sz:24;
+	u32 frame_sz:23;
+	u32 mb:1; /* xdp non-linear frame */
 	/* Lifetime of xdp_rxq_info is limited to NAPI/enqueue time,
 	 * while mem info is valid on remote CPU.
 	 */

@@ -141,6 +143,7 @@ void xdp_convert_frame_to_buff(struct xdp_frame *frame, struct xdp_buff *xdp)
 	xdp->data_end = frame->data + frame->len;
 	xdp->data_meta = frame->data - frame->metasize;
 	xdp->frame_sz = frame->frame_sz;
+	xdp->mb = frame->mb;
 }
 
 static inline

@@ -167,6 +170,7 @@ int xdp_update_frame_from_buff(struct xdp_buff *xdp,
 	xdp_frame->headroom = headroom - sizeof(*xdp_frame);
 	xdp_frame->metasize = metasize;
 	xdp_frame->frame_sz = xdp->frame_sz;
+	xdp_frame->mb = xdp->mb;
 
 	return 0;
 }

Signature · 1 line

2.26.2

Jubran, Samih

2020-07-23 13:51 UTC #2 lore raw

RE: [RFC net-next 01/22] xdp: introduce mb in xdp_buff/xdp_frame

Quoted 34 lines

-----Original Message-----
From: Lorenzo Bianconi <[email protected]>
Sent: Thursday, July 23, 2020 2:42 PM
To: [email protected]
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Jubran, Samih <[email protected]>; [email protected]
Subject: [EXTERNAL] [RFC net-next 01/22] xdp: introduce mb in xdp_buff/xdp_frame
CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you can confirm the sender and know the content is safe.
Introduce multi-buffer bit (mb) in xdp_frame/xdp_buffer to specify if shared_info area has been properly initialized for non-linear xdp buffers
Signed-off-by Lorenzo Bianconi <[email protected]>
---
 include/net/xdp.h | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
include/net/xdp.h index renamed from include/net/xdp.h +0 -0
dbe9c60797e1..2ef6935c5646 100644
include/net/xdp.h +0 -1
@@ -72,7 +72,8 @@ struct xdp_buff {
        void *data_hard_start;
        struct xdp_rxq_info *rxq;
        struct xdp_txq_info *txq;
-       u32 frame_sz; /* frame size to deduce data_hard_end/reserved

Quoted 3 lines

tailroom*/
+ u32 frame_sz:31; /* frame size to deduce data_hard_end/reserved
tailroom*/

It seems strange that you assign a 32 sized field to a 24 sized field. Wouldn't it be better if we used the same size in all places?

Quoted 34 lines

include/net/xdp.h +4 -1

+       u32 mb:1; /* xdp non-linear buffer */
 };
 
 /* Reserve memory area at end-of data area.

@@ -96,7 +97,8 @@ struct xdp_frame {
        u16 len;
        u16 headroom;
        u32 metasize:8;
-       u32 frame_sz:24;
+       u32 frame_sz:23;
+       u32 mb:1; /* xdp non-linear frame */
        /* Lifetime of xdp_rxq_info is limited to NAPI/enqueue time,
         * while mem info is valid on remote CPU.
         */

@@ -141,6 +143,7 @@ void xdp_convert_frame_to_buff(struct xdp_frame
 *frame, struct xdp_buff *xdp)
        xdp->data_end = frame->data + frame->len;
        xdp->data_meta = frame->data - frame->metasize;
        xdp->frame_sz = frame->frame_sz;
+       xdp->mb = frame->mb;
 }

static inline

include/net/xdp.h +1 -0

@@ -167,6 +170,7 @@ int xdp_update_frame_from_buff(struct xdp_buff
 *xdp,
        xdp_frame->headroom = headroom - sizeof(*xdp_frame);
        xdp_frame->metasize = metasize;
        xdp_frame->frame_sz = xdp->frame_sz;
+       xdp_frame->mb = xdp->mb;
 
        return 0;

}
--
2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #3 lore raw

RFC 2/22 xdp: initialize xdp_buff mb bit to 0 in netif_receive_generic_xdp

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 net/core/dev.c | 1 +
 1 file changed, 1 insertion(+)

net/core/dev.c +1 -0

index 316349f6cea5..eb3e54df198c 100644
@@ -4641,6 +4641,7 @@ static u32 netif_receive_generic_xdp(struct sk_buff *skb,
 	/* SKB "head" area always have tailroom for skb_shared_info */
 	xdp->frame_sz  = (void *)skb_end_pointer(skb) - xdp->data_hard_start;
 	xdp->frame_sz += SKB_DATA_ALIGN(sizeof(struct skb_shared_info));
+	xdp->mb = 0;
 
 	orig_data_end = xdp->data_end;
 	orig_data = xdp->data;

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #4 lore raw

RFC 3/22 net: virtio_net: initialize mb bit of xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/virtio_net.c | 2 ++
 1 file changed, 2 insertions(+)

drivers/net/virtio_net.c +2 -0

index ba38765dc490..100fa9b60026 100644
@@ -690,6 +690,7 @@ static struct sk_buff *receive_small(struct net_device *dev,
 		xdp.data_meta = xdp.data;
 		xdp.rxq = &rq->xdp_rxq;
 		xdp.frame_sz = buflen;
+		xdp.mb = 0;
 		orig_data = xdp.data;
 		act = bpf_prog_run_xdp(xdp_prog, &xdp);
 		stats->xdp_packets++;

@@ -860,6 +861,7 @@ static struct sk_buff *receive_mergeable(struct net_device *dev,
 		xdp.data_meta = xdp.data;
 		xdp.rxq = &rq->xdp_rxq;
 		xdp.frame_sz = frame_sz - vi->hdr_len;
+		xdp.mb = 0;
 
 		act = bpf_prog_run_xdp(xdp_prog, &xdp);
 		stats->xdp_packets++;

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #5 lore raw

RFC 4/22 net: xen-netfront: initialize mb bit of xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/xen-netfront.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/xen-netfront.c +1 -0

index ed995df19247..ce862b08bb3b 100644
@@ -868,6 +868,7 @@ static u32 xennet_run_xdp(struct netfront_queue *queue, struct page *pdata,
 	xdp->data_end = xdp->data + len;
 	xdp->rxq = &queue->xdp_rxq;
 	xdp->frame_sz = XEN_PAGE_SIZE - XDP_PACKET_HEADROOM;
+	xdp->mb = 0;
 
 	act = bpf_prog_run_xdp(prog, xdp);
 	switch (act) {

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #6 lore raw

RFC 5/22 net: veth: initialize mb bit of xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/veth.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/veth.c +1 -0

index b594f03eeddb..463865c513b2 100644
@@ -711,6 +711,7 @@ static struct sk_buff *veth_xdp_rcv_skb(struct veth_rq *rq,
 	/* SKB "head" area always have tailroom for skb_shared_info */
 	xdp.frame_sz = (void *)skb_end_pointer(skb) - xdp.data_hard_start;
 	xdp.frame_sz += SKB_DATA_ALIGN(sizeof(struct skb_shared_info));
+	xdp.mb = 0;
 
 	orig_data = xdp.data;
 	orig_data_end = xdp.data_end;

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #7 lore raw

RFC 6/22 net: hv_netvsc: initialize mb bit of xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/hyperv/netvsc_bpf.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/hyperv/netvsc_bpf.c +1 -0

index 8e4141552423..e26c60413128 100644
@@ -50,6 +50,7 @@ u32 netvsc_run_xdp(struct net_device *ndev, struct netvsc_channel *nvchan,
 	xdp->data_end = xdp->data + len;
 	xdp->rxq = &nvchan->xdp_rxq;
 	xdp->frame_sz = PAGE_SIZE;
+	xdp->mb = 0;
 
 	memcpy(xdp->data, data, len);

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #8 lore raw

RFC 7/22 net: bnxt: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c +1 -0

index 5e3b4a3b69ea..615d95cf4629 100644
@@ -139,6 +139,7 @@ bool bnxt_rx_xdp(struct bnxt *bp, struct bnxt_rx_ring_info *rxr, u16 cons,
 	xdp.data_end = *data_ptr + *len;
 	xdp.rxq = &rxr->xdp_rxq;
 	xdp.frame_sz = PAGE_SIZE; /* BNXT_RX_PAGE_MODE(bp) when XDP enabled */
+	xdp.mb = 0;
 	orig_data = xdp.data;
 
 	rcu_read_lock();

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #9 lore raw

RFC 8/22 net: dpaa2: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c +1 -0

index c1bea9132f50..fc37099675c8 100644
@@ -361,6 +361,7 @@ static u32 run_xdp(struct dpaa2_eth_priv *priv,
 
 	xdp.frame_sz = DPAA2_ETH_RX_BUF_RAW_SIZE -
 		(dpaa2_fd_get_offset(fd) - XDP_PACKET_HEADROOM);
+	xdp.mb = 0;
 
 	xdp_act = bpf_prog_run_xdp(xdp_prog, &xdp);

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #10 lore raw

RFC 9/22 net: ti: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/ethernet/ti/cpsw.c     | 1 +
 drivers/net/ethernet/ti/cpsw_new.c | 1 +
 2 files changed, 2 insertions(+)

drivers/net/ethernet/ti/cpsw.c +1 -0

index 9b17bbbe102f..53a55c540adc 100644
@@ -407,6 +407,7 @@ static void cpsw_rx_handler(void *token, int len, int status)
 		xdp.data_hard_start = pa;
 		xdp.rxq = &priv->xdp_rxq[ch];
 		xdp.frame_sz = PAGE_SIZE;
+		xdp.mb = 0;
 
 		port = priv->emac_port + cpsw->data.dual_emac;
 		ret = cpsw_run_xdp(priv, ch, &xdp, page, port);

drivers/net/ethernet/ti/cpsw_new.c +1 -0

index 1247d35d42ef..703d079fd479 100644
@@ -349,6 +349,7 @@ static void cpsw_rx_handler(void *token, int len, int status)
 		xdp.data_hard_start = pa;
 		xdp.rxq = &priv->xdp_rxq[ch];
 		xdp.frame_sz = PAGE_SIZE;
+		xdp.mb = 0;
 
 		ret = cpsw_run_xdp(priv, ch, &xdp, page, priv->emac_port);
 		if (ret != CPSW_XDP_PASS)

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #11 lore raw

RFC 10/22 net: nfp: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/ethernet/netronome/nfp/nfp_net_common.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/ethernet/netronome/nfp/nfp_net_common.c +1 -0

index 44608873d3d9..c6998c34e552 100644
@@ -1824,6 +1824,7 @@ static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget)
 	true_bufsz = xdp_prog ? PAGE_SIZE : dp->fl_bufsz;
 	xdp.frame_sz = PAGE_SIZE - NFP_NET_RX_BUF_HEADROOM;
 	xdp.rxq = &rx_ring->xdp_rxq;
+	xdp.mb = 0;
 	tx_ring = r_vec->xdp_ring;
 
 	while (pkts_polled < budget) {

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #12 lore raw

RFC 11/22 net: mvpp2: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c +1 -0

index 6a3f356640a0..61b1688527ea 100644
@@ -3475,6 +3475,7 @@ static int mvpp2_rx(struct mvpp2_port *port, struct napi_struct *napi,
 			xdp.data = data + MVPP2_MH_SIZE + MVPP2_SKB_HEADROOM;
 			xdp.data_end = xdp.data + rx_bytes;
 			xdp.frame_sz = PAGE_SIZE;
+			xdp.mb = 0;
 
 			if (bm_pool->pkt_size == MVPP2_BM_SHORT_PKT_SIZE)
 				xdp.rxq = &rxq->xdp_rxq_short;

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #13 lore raw

RFC 12/22 net: sfc: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/ethernet/sfc/rx.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/ethernet/sfc/rx.c +1 -0

index 59a43d586967..8fd6023995d4 100644
@@ -301,6 +301,7 @@ static bool efx_do_xdp(struct efx_nic *efx, struct efx_channel *channel,
 	xdp.data_end = xdp.data + rx_buf->len;
 	xdp.rxq = &rx_queue->xdp_rxq_info;
 	xdp.frame_sz = efx->rx_page_buf_step;
+	xdp.mb = 0;
 
 	xdp_act = bpf_prog_run_xdp(xdp_prog, &xdp);
 	rcu_read_unlock();

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #14 lore raw

RFC 13/22 net: qede: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/ethernet/qlogic/qede/qede_fp.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/ethernet/qlogic/qede/qede_fp.c +1 -0

index a2494bf85007..14a54094ca08 100644
@@ -1096,6 +1096,7 @@ static bool qede_rx_xdp(struct qede_dev *edev,
 	xdp.data_end = xdp.data + *len;
 	xdp.rxq = &rxq->xdp_rxq;
 	xdp.frame_sz = rxq->rx_buf_seg_size; /* PAGE_SIZE when XDP enabled */
+	xdp.mb = 0;
 
 	/* Queues always have a full reset currently, so for the time
 	 * being until there's atomic program replace just mark read

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #15 lore raw

RFC 14/22 net: amazon: ena: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/ethernet/amazon/ena/ena_netdev.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/ethernet/amazon/ena/ena_netdev.c +1 -0

index 6478c1e0d137..12f7ee1d1f0a 100644
@@ -1613,6 +1613,7 @@ static int ena_clean_rx_irq(struct ena_ring *rx_ring, struct napi_struct *napi,
 	res_budget = budget;
 	xdp.rxq = &rx_ring->xdp_rxq;
 	xdp.frame_sz = ENA_PAGE_SIZE;
+	xdp.mb = 0;
 
 	do {
 		xdp_verdict = XDP_PASS;

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #16 lore raw

RFC 15/22 net: cavium: thunder: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/ethernet/cavium/thunder/nicvf_main.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/ethernet/cavium/thunder/nicvf_main.c +1 -0

index 2ba0ce115e63..bd2592ddfd6d 100644
@@ -553,6 +553,7 @@ static inline bool nicvf_xdp_rx(struct nicvf *nic, struct bpf_prog *prog,
 	xdp.data_end = xdp.data + len;
 	xdp.rxq = &rq->xdp_rxq;
 	xdp.frame_sz = RCV_FRAG_LEN + XDP_PACKET_HEADROOM;
+	xdp.mb = 0;
 	orig_data = xdp.data;
 
 	rcu_read_lock();

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #17 lore raw

RFC 16/22 net: socionext: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/ethernet/socionext/netsec.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/ethernet/socionext/netsec.c +1 -0

index 0f366cc50b74..0c98391224bd 100644
@@ -947,6 +947,7 @@ static int netsec_process_rx(struct netsec_priv *priv, int budget)
 
 	xdp.rxq = &dring->xdp_rxq;
 	xdp.frame_sz = PAGE_SIZE;
+	xdp.mb = 0;
 
 	rcu_read_lock();
 	xdp_prog = READ_ONCE(priv->xdp_prog);

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #18 lore raw

RFC 17/22 net: tun: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/tun.c | 2 ++
 1 file changed, 2 insertions(+)

drivers/net/tun.c +2 -0

index 7adeb91bd368..7007fc344460 100644
@@ -1674,6 +1674,7 @@ static struct sk_buff *tun_build_skb(struct tun_struct *tun,
 		xdp.data_end = xdp.data + len;
 		xdp.rxq = &tfile->xdp_rxq;
 		xdp.frame_sz = buflen;
+		xdp.mb = 0;
 
 		act = bpf_prog_run_xdp(xdp_prog, &xdp);
 		if (act == XDP_REDIRECT || act == XDP_TX) {

@@ -2421,6 +2422,7 @@ static int tun_xdp_one(struct tun_struct *tun,
 		xdp_set_data_meta_invalid(xdp);
 		xdp->rxq = &tfile->xdp_rxq;
 		xdp->frame_sz = buflen;
+		xdp->mb = 0;
 
 		act = bpf_prog_run_xdp(xdp_prog, xdp);
 		err = tun_xdp_act(tun, xdp_prog, xdp, act);

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #19 lore raw

RFC 18/22 net: ixgbe: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c     | 1 +
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 1 +
 2 files changed, 2 insertions(+)

drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +1 -0

index 4d898ff21a46..327265ae3b68 100644
@@ -2303,6 +2303,7 @@ static int ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
 #if (PAGE_SIZE < 8192)
 	xdp.frame_sz = ixgbe_rx_frame_truesize(rx_ring, 0);
 #endif
+	xdp.mb = 0;
 
 	while (likely(total_rx_packets < budget)) {
 		union ixgbe_adv_rx_desc *rx_desc;

drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c +1 -0

index 6e9a397db583..134e8f1a0b77 100644
@@ -1133,6 +1133,7 @@ static int ixgbevf_clean_rx_irq(struct ixgbevf_q_vector *q_vector,
 	struct xdp_buff xdp;
 
 	xdp.rxq = &rx_ring->xdp_rxq;
+	xdp.mb = 0;
 
 	/* Frame size depend on rx_ring setup when PAGE_SIZE=4K */
 #if (PAGE_SIZE < 8192)

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:43 UTC #20 lore raw

RFC 19/22 net: ice: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/ethernet/intel/ice/ice_txrx.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/ethernet/intel/ice/ice_txrx.c +1 -0

index abdb137c8bb7..90453b3ffa2c 100644
@@ -1098,6 +1098,7 @@ int ice_clean_rx_irq(struct ice_ring *rx_ring, int budget)
 #if (PAGE_SIZE < 8192)
 	xdp.frame_sz = ice_rx_frame_truesize(rx_ring, 0);
 #endif
+	xdp.mb = 0;
 
 	/* start the loop to process Rx packets bounded by 'budget' */
 	while (likely(total_rx_pkts < (unsigned int)budget)) {

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:44 UTC #21 lore raw

RFC 20/22 net: i40e: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/ethernet/intel/i40e/i40e_txrx.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/ethernet/intel/i40e/i40e_txrx.c +1 -0

index 3e5c566ceb01..7d5a8afa8a4c 100644
@@ -2324,6 +2324,7 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget)
 	xdp.frame_sz = i40e_rx_frame_truesize(rx_ring, 0);
 #endif
 	xdp.rxq = &rx_ring->xdp_rxq;
+	xdp.mb = 0;
 
 	while (likely(total_rx_packets < (unsigned int)budget)) {
 		struct i40e_rx_buffer *rx_buffer;

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:44 UTC #22 lore raw

RFC 21/22 net: mlx5: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/ethernet/mellanox/mlx5/core/en_rx.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +1 -0

index 74860f3827b1..bfaa03134505 100644
@@ -1104,6 +1104,7 @@ static void mlx5e_fill_xdp_buff(struct mlx5e_rq *rq, void *va, u16 headroom,
 	xdp->data_end = xdp->data + len;
 	xdp->rxq = &rq->xdp_rxq;
 	xdp->frame_sz = rq->buff.frame0_sz;
+	xdp->mb = 0;
 }
 
 struct sk_buff *

Signature · 1 line

2.26.2

Lorenzo Bianconi

author

2020-07-23 11:44 UTC #23 lore raw

RFC 22/22 net: mlx4: initialize mb bit in xdp_buff to 0

In reply to Lorenzo Bianconi

Initialize multi-buffer bit (mb) to 0 in xdp_buff data structure. This is a preliminary patch to enable xdp multi-buffer support.

Signed-off-by Lorenzo Bianconi <[email protected]>

---
 drivers/net/ethernet/mellanox/mlx4/en_rx.c | 1 +
 1 file changed, 1 insertion(+)

drivers/net/ethernet/mellanox/mlx4/en_rx.c +1 -0

index 8a10285b0e10..0d42e81a7ad8 100644
@@ -684,6 +684,7 @@ int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int bud
 	xdp_prog = rcu_dereference(ring->xdp_prog);
 	xdp.rxq = &ring->xdp_rxq;
 	xdp.frame_sz = priv->frag_info[0].frag_stride;
+	xdp.mb = 0;
 	doorbell_pending = 0;
 
 	/* We assume a 1:1 mapping between CQEs and Rx descriptors, so Rx

Signature · 1 line

2.26.2

Jesper Dangaard Brouer

2020-07-28 14:49 UTC #24 lore raw

In reply to Lorenzo Bianconi

Quoted 4 lines · On Thu, 23 Jul 2020 13:42:12 +0200 Lorenzo Bianconi <[email protected]> wrote:

Introduce multi-buffer bit (mb) in xdp_frame/xdp_buffer to specify if shared_info area has been properly initialized for non-linear xdp buffers.
Initialize mb to 0 for all xdp drivers

It is nice to see that we have some many XDP drivers, but 20 separate patches for drivers is a bit much. Perhaps we can do all the drivers in one patch? What do others think?

Quoted 24 lines

Lorenzo Bianconi (22):
xdp: introduce mb in xdp_buff/xdp_frame
xdp: initialize xdp_buff mb bit to 0 in netif_receive_generic_xdp
net: virtio_net: initialize mb bit of xdp_buff to 0
net: xen-netfront: initialize mb bit of xdp_buff to 0
net: veth: initialize mb bit of xdp_buff to 0
net: hv_netvsc: initialize mb bit of xdp_buff to 0
net: bnxt: initialize mb bit in xdp_buff to 0
net: dpaa2: initialize mb bit in xdp_buff to 0
net: ti: initialize mb bit in xdp_buff to 0
net: nfp: initialize mb bit in xdp_buff to 0
net: mvpp2: initialize mb bit in xdp_buff to 0
net: sfc: initialize mb bit in xdp_buff to 0
net: qede: initialize mb bit in xdp_buff to 0
net: amazon: ena: initialize mb bit in xdp_buff to 0
net: cavium: thunder: initialize mb bit in xdp_buff to 0
net: socionext: initialize mb bit in xdp_buff to 0
net: tun: initialize mb bit in xdp_buff to 0
net: ixgbe: initialize mb bit in xdp_buff to 0
net: ice: initialize mb bit in xdp_buff to 0
net: i40e: initialize mb bit in xdp_buff to 0
net: mlx5: initialize mb bit in xdp_buff to 0
net: mlx4: initialize mb bit in xdp_buff to 0

Quoted 9 lines

 drivers/net/ethernet/amazon/ena/ena_netdev.c        | 1 +
 drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c       | 1 +
 drivers/net/ethernet/cavium/thunder/nicvf_main.c    | 1 +
 drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c    | 1 +
 drivers/net/ethernet/intel/i40e/i40e_txrx.c         | 1 +
 drivers/net/ethernet/intel/ice/ice_txrx.c           | 1 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c       | 1 +
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c   | 1 +
 drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c     | 1 +

I see that mvneta is missing, but maybe it is doing another kind of init of struct xdp_buff?

Quoted 16 lines

 drivers/net/ethernet/mellanox/mlx4/en_rx.c          | 1 +
 drivers/net/ethernet/mellanox/mlx5/core/en_rx.c     | 1 +
 drivers/net/ethernet/netronome/nfp/nfp_net_common.c | 1 +
 drivers/net/ethernet/qlogic/qede/qede_fp.c          | 1 +
 drivers/net/ethernet/sfc/rx.c                       | 1 +
 drivers/net/ethernet/socionext/netsec.c             | 1 +
 drivers/net/ethernet/ti/cpsw.c                      | 1 +
 drivers/net/ethernet/ti/cpsw_new.c                  | 1 +
 drivers/net/hyperv/netvsc_bpf.c                     | 1 +
 drivers/net/tun.c                                   | 2 ++
 drivers/net/veth.c                                  | 1 +
 drivers/net/virtio_net.c                            | 2 ++
 drivers/net/xen-netfront.c                          | 1 +
 include/net/xdp.h                                   | 8 ++++++--
 net/core/dev.c                                      | 1 +
 24 files changed, 31 insertions(+), 2 deletions(-)

Signature · 4 lines

Best regards,
  Jesper Dangaard Brouer
  MSc.CS, Principal Kernel Engineer at Red Hat
  LinkedIn: http://www.linkedin.com/in/brouer

Lorenzo Bianconi

2020-07-28 15:10 UTC #25 lore raw

Quoted 11 lines

Quoted 4 lines · On Thu, 23 Jul 2020 13:42:12 +0200 Lorenzo Bianconi <[email protected]> wrote:
Introduce multi-buffer bit (mb) in xdp_frame/xdp_buffer to specify if shared_info area has been properly initialized for non-linear xdp buffers.
Initialize mb to 0 for all xdp drivers
It is nice to see that we have some many XDP drivers, but 20 separate patches for drivers is a bit much. Perhaps we can do all the drivers in one patch? What do others think?

I am fine with it, not sure what is the best approach.

Quoted 35 lines

Quoted 25 lines
Lorenzo Bianconi (22):
xdp: introduce mb in xdp_buff/xdp_frame
xdp: initialize xdp_buff mb bit to 0 in netif_receive_generic_xdp
net: virtio_net: initialize mb bit of xdp_buff to 0
net: xen-netfront: initialize mb bit of xdp_buff to 0
net: veth: initialize mb bit of xdp_buff to 0
net: hv_netvsc: initialize mb bit of xdp_buff to 0
net: bnxt: initialize mb bit in xdp_buff to 0
net: dpaa2: initialize mb bit in xdp_buff to 0
net: ti: initialize mb bit in xdp_buff to 0
net: nfp: initialize mb bit in xdp_buff to 0
net: mvpp2: initialize mb bit in xdp_buff to 0
net: sfc: initialize mb bit in xdp_buff to 0
net: qede: initialize mb bit in xdp_buff to 0
net: amazon: ena: initialize mb bit in xdp_buff to 0
net: cavium: thunder: initialize mb bit in xdp_buff to 0
net: socionext: initialize mb bit in xdp_buff to 0
net: tun: initialize mb bit in xdp_buff to 0
net: ixgbe: initialize mb bit in xdp_buff to 0
net: ice: initialize mb bit in xdp_buff to 0
net: i40e: initialize mb bit in xdp_buff to 0
net: mlx5: initialize mb bit in xdp_buff to 0
net: mlx4: initialize mb bit in xdp_buff to 0
Quoted 9 lines
 drivers/net/ethernet/amazon/ena/ena_netdev.c        | 1 +
 drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c       | 1 +
 drivers/net/ethernet/cavium/thunder/nicvf_main.c    | 1 +
 drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c    | 1 +
 drivers/net/ethernet/intel/i40e/i40e_txrx.c         | 1 +
 drivers/net/ethernet/intel/ice/ice_txrx.c           | 1 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c       | 1 +
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c   | 1 +
 drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c     | 1 +

Quoted 2 lines

I see that mvneta is missing, but maybe it is doing another kind of init of struct xdp_buff?

yes, mvneta has been done in a previous patch.

Regards,
Lorenzo

Quoted 24 lines

Quoted 17 lines

 drivers/net/ethernet/mellanox/mlx4/en_rx.c          | 1 +
 drivers/net/ethernet/mellanox/mlx5/core/en_rx.c     | 1 +
 drivers/net/ethernet/netronome/nfp/nfp_net_common.c | 1 +
 drivers/net/ethernet/qlogic/qede/qede_fp.c          | 1 +
 drivers/net/ethernet/sfc/rx.c                       | 1 +
 drivers/net/ethernet/socionext/netsec.c             | 1 +
 drivers/net/ethernet/ti/cpsw.c                      | 1 +
 drivers/net/ethernet/ti/cpsw_new.c                  | 1 +
 drivers/net/hyperv/netvsc_bpf.c                     | 1 +
 drivers/net/tun.c                                   | 2 ++
 drivers/net/veth.c                                  | 1 +
 drivers/net/virtio_net.c                            | 2 ++
 drivers/net/xen-netfront.c                          | 1 +
 include/net/xdp.h                                   | 8 ++++++--
 net/core/dev.c                                      | 1 +
 24 files changed, 31 insertions(+), 2 deletions(-)

Signature · 4 lines

Best regards,
  Jesper Dangaard Brouer
  MSc.CS, Principal Kernel Engineer at Red Hat
  LinkedIn: http://www.linkedin.com/in/brouer