Re: [PATCH net-next 3/3] net: ethernet: ti: cpsw: add XDP support

[Ivan Khoronzhuk <ivan.khoronzhuk@linaro.org>]

On Fri, May 24, 2019 at 08:49:38PM +0300, grygorii wrote:
>Hi Ivan,
>
>On 23/05/2019 21:20, Ivan Khoronzhuk wrote:
>>Add XDP support based on rx page_pool allocator, one frame per page.
>>Page pool allocator is used with assumption that only one rx_handler
>>is running simultaneously. DMA map/unmap is reused from page pool
>>despite there is no need to map whole page.
>>
>>Due to specific of cpsw, the same TX/RX handler can be used by 2
>>network devices, so special fields in buffer are added to identify
>>an interface the frame is destined to. Thus XDP works for both
>>interfaces, that allows to test xdp redirect between two interfaces
>>easily.
>>
>>XDP prog is common for all channels till appropriate changes are added
>>in XDP infrastructure.
>>
>>Signed-off-by: Ivan Khoronzhuk <ivan.khoronzhuk@linaro.org>
>>---
>>  drivers/net/ethernet/ti/Kconfig        |   1 +
>>  drivers/net/ethernet/ti/cpsw.c         | 555 ++++++++++++++++++++++---
>>  drivers/net/ethernet/ti/cpsw_ethtool.c |  53 +++
>>  drivers/net/ethernet/ti/cpsw_priv.h    |   7 +
>>  4 files changed, 554 insertions(+), 62 deletions(-)
>>
>>diff --git a/drivers/net/ethernet/ti/Kconfig b/drivers/net/ethernet/ti/Kconfig
>>index bd05a977ee7e..3cb8c5214835 100644
>>--- a/drivers/net/ethernet/ti/Kconfig
>>+++ b/drivers/net/ethernet/ti/Kconfig
>>@@ -50,6 +50,7 @@ config TI_CPSW
>>  	depends on ARCH_DAVINCI || ARCH_OMAP2PLUS || COMPILE_TEST
>>  	select TI_DAVINCI_MDIO
>>  	select MFD_SYSCON
>>+	select PAGE_POOL
>>  	select REGMAP
>>  	---help---
>>  	  This driver supports TI's CPSW Ethernet Switch.
>>diff --git a/drivers/net/ethernet/ti/cpsw.c b/drivers/net/ethernet/ti/cpsw.c
>>index 87a600aeee4a..274e6b64ea9e 100644
>>--- a/drivers/net/ethernet/ti/cpsw.c
>>+++ b/drivers/net/ethernet/ti/cpsw.c
>>@@ -31,6 +31,10 @@
>>  #include <linux/if_vlan.h>
>>  #include <linux/kmemleak.h>
>>  #include <linux/sys_soc.h>
>>+#include <net/page_pool.h>
>>+#include <linux/bpf.h>
>>+#include <linux/bpf_trace.h>
>>+#include <linux/filter.h>
>>  #include <linux/pinctrl/consumer.h>
>>  #include <net/pkt_cls.h>
>>@@ -60,6 +64,10 @@ static int descs_pool_size = CPSW_CPDMA_DESCS_POOL_SIZE_DEFAULT;
>>  module_param(descs_pool_size, int, 0444);
>>  MODULE_PARM_DESC(descs_pool_size, "Number of CPDMA CPPI descriptors in pool");
>>+/* The buf includes headroom compatible with both skb and xdpf */
>>+#define CPSW_HEADROOM_NA (max(XDP_PACKET_HEADROOM, NET_SKB_PAD) + NET_IP_ALIGN)
>>+#define CPSW_HEADROOM  ALIGN(CPSW_HEADROOM_NA, sizeof(long))
>>+
>>  #define for_each_slave(priv, func, arg...)				\
>>  	do {								\
>>  		struct cpsw_slave *slave;				\
>>@@ -74,6 +82,8 @@ MODULE_PARM_DESC(descs_pool_size, "Number of CPDMA CPPI descriptors in pool");
>>  				(func)(slave++, ##arg);			\
>>  	} while (0)
>>+#define CPSW_XMETA_OFFSET	ALIGN(sizeof(struct xdp_frame), sizeof(long))
>>+
>>  static int cpsw_ndo_vlan_rx_add_vid(struct net_device *ndev,
>>  				    __be16 proto, u16 vid);
>>@@ -337,24 +347,58 @@ void cpsw_intr_disable(struct cpsw_common *cpsw)
>>  	return;
>>  }
>
>[..]
>
>>+static int cpsw_xdp_tx_frame_mapped(struct cpsw_priv *priv,
>>+				    struct xdp_frame *xdpf, struct page *page)
>>+{
>>+	struct cpsw_common *cpsw = priv->cpsw;
>>+	struct cpsw_meta_xdp *xmeta;
>>+	struct netdev_queue *txq;
>>+	struct cpdma_chan *txch;
>>+	dma_addr_t dma;
>>+	int ret;
>>+
>>+	xmeta = (void *)xdpf + CPSW_XMETA_OFFSET;
>>+	xmeta->ch = 0;
>>+
>>+	txch = cpsw->txv[0].ch;
>>+	dma = (xdpf->data - (void *)xdpf) + page->dma_addr;
>>+	ret = cpdma_chan_submit_mapped(txch, cpsw_xdpf_to_handle(xdpf), dma,
>>+				       xdpf->len,
>>+				       priv->emac_port + cpsw->data.dual_emac);
>>+	if (ret) {
>>+		xdp_return_frame_rx_napi(xdpf);
>>+		goto stop;
>>+	}
>>+
>>+	/* no tx desc - stop sending us tx frames */
>>+	if (unlikely(!cpdma_check_free_tx_desc(txch)))
>>+		goto stop;
>>+
>>+	return ret;
>>+stop:
>>+	txq = netdev_get_tx_queue(priv->ndev, 0);
>>+	netif_tx_stop_queue(txq);
>>+
>>+	/* Barrier, so that stop_queue visible to other cpus */
>>+	smp_mb__after_atomic();
>>+
>>+	if (cpdma_check_free_tx_desc(txch))
>>+		netif_tx_wake_queue(txq);
>>+
>>+	return ret;
>>+}
>>+
>>+static int cpsw_xdp_tx_frame(struct cpsw_priv *priv, struct xdp_frame *xdpf)
>>+{
>>+	struct cpsw_common *cpsw = priv->cpsw;
>>+	struct cpsw_meta_xdp *xmeta;
>>+	struct netdev_queue *txq;
>>+	struct cpdma_chan *txch;
>>+	int ret;
>>+
>>+	xmeta = (void *)xdpf + CPSW_XMETA_OFFSET;
>>+	if (sizeof(*xmeta) > xdpf->headroom)
>>+		return -EINVAL;
>>+
>>+	xmeta->ndev = priv->ndev;
>>+	xmeta->ch = 0;
>>+
>>+	txch = cpsw->txv[0].ch;
>>+	ret = cpdma_chan_submit(txch, cpsw_xdpf_to_handle(xdpf), xdpf->data,
>>+				xdpf->len,
>>+				priv->emac_port + cpsw->data.dual_emac);
>>+	if (ret) {
>>+		xdp_return_frame_rx_napi(xdpf);
>>+		goto stop;
>>+	}
>>+
>>+	/* no tx desc - stop sending us tx frames */
>>+	if (unlikely(!cpdma_check_free_tx_desc(txch)))
>>+		goto stop;
>>+
>>+	return ret;
>>+stop:
>>+	txq = netdev_get_tx_queue(priv->ndev, 0);
>>+	netif_tx_stop_queue(txq);
>>+
>>+	/* Barrier, so that stop_queue visible to other cpus */
>>+	smp_mb__after_atomic();
>>+
>>+	if (cpdma_check_free_tx_desc(txch))
>>+		netif_tx_wake_queue(txq);
>>+
>>+	return ret;
>>+}
>
>Above 2 functions are mostly identical - could you do smth. with it?
... I know it should be, but i hadn't found better way for combining them ....

>
>>+
>>+static int cpsw_run_xdp(struct cpsw_priv *priv, struct cpsw_vector *rxv,
>>+			struct xdp_buff *xdp, struct page *page)
>>+{
>>+	struct net_device *ndev = priv->ndev;
>>+	struct xdp_frame *xdpf;
>>+	struct bpf_prog *prog;
>>+	int ret = 1;
>>+	u32 act;
>>+
>>+	rcu_read_lock();
>>+
>>+	prog = READ_ONCE(priv->xdp_prog);
>>+	if (!prog) {
>>+		ret = 0;
>>+		goto out;
>>+	}
>>+
>>+	act = bpf_prog_run_xdp(prog, xdp);
>>+	switch (act) {
>>+	case XDP_PASS:
>>+		ret = 0;
>>+		break;
>>+	case XDP_TX:
>>+		xdpf = convert_to_xdp_frame(xdp);
>>+		if (unlikely(!xdpf))
>>+			xdp_return_buff(xdp);
>>+		else
>>+			cpsw_xdp_tx_frame_mapped(priv, xdpf, page);
>>+		break;
>>+	case XDP_REDIRECT:
>>+		if (xdp_do_redirect(ndev, xdp, prog))
>>+			xdp_return_buff(xdp);
>>+		else
>>+			ret = 2;
>
>could we avoid using consts as return values?
>may be some informative defines/enum?
Ok, for all "const" cases.

-- 
Regards,
Ivan Khoronzhuk