* [PATCH v3 net-next] net: socionext: get rid of huge dma sync in netsec_alloc_rx_data
@ 2020-01-14 9:24 Lorenzo Bianconi
2020-01-14 10:59 ` Jesper Dangaard Brouer
2020-01-15 2:46 ` David Miller
0 siblings, 2 replies; 3+ messages in thread
From: Lorenzo Bianconi @ 2020-01-14 9:24 UTC (permalink / raw)
To: netdev; +Cc: lorenzo.bianconi, davem, brouer, ilias.apalodimas, kuba
Socionext driver can run on dma coherent and non-coherent devices.
Get rid of huge dma_sync_single_for_device in netsec_alloc_rx_data since
now the driver can let page_pool API to managed needed DMA sync
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
---
Changes since v2:
- fix checkpatch warnings
Changes since v1:
- rely on original frame size for dma sync
---
drivers/net/ethernet/socionext/netsec.c | 43 +++++++++++++++----------
1 file changed, 26 insertions(+), 17 deletions(-)
diff --git a/drivers/net/ethernet/socionext/netsec.c b/drivers/net/ethernet/socionext/netsec.c
index b5a9e947a4a8..6870a6ce76a6 100644
--- a/drivers/net/ethernet/socionext/netsec.c
+++ b/drivers/net/ethernet/socionext/netsec.c
@@ -243,6 +243,7 @@
NET_IP_ALIGN)
#define NETSEC_RX_BUF_NON_DATA (NETSEC_RXBUF_HEADROOM + \
SKB_DATA_ALIGN(sizeof(struct skb_shared_info)))
+#define NETSEC_RX_BUF_SIZE (PAGE_SIZE - NETSEC_RX_BUF_NON_DATA)
#define DESC_SZ sizeof(struct netsec_de)
@@ -719,7 +720,6 @@ static void *netsec_alloc_rx_data(struct netsec_priv *priv,
{
struct netsec_desc_ring *dring = &priv->desc_ring[NETSEC_RING_RX];
- enum dma_data_direction dma_dir;
struct page *page;
page = page_pool_dev_alloc_pages(dring->page_pool);
@@ -734,9 +734,7 @@ static void *netsec_alloc_rx_data(struct netsec_priv *priv,
/* Make sure the incoming payload fits in the page for XDP and non-XDP
* cases and reserve enough space for headroom + skb_shared_info
*/
- *desc_len = PAGE_SIZE - NETSEC_RX_BUF_NON_DATA;
- dma_dir = page_pool_get_dma_dir(dring->page_pool);
- dma_sync_single_for_device(priv->dev, *dma_handle, *desc_len, dma_dir);
+ *desc_len = NETSEC_RX_BUF_SIZE;
return page_address(page);
}
@@ -883,6 +881,8 @@ static u32 netsec_xdp_xmit_back(struct netsec_priv *priv, struct xdp_buff *xdp)
static u32 netsec_run_xdp(struct netsec_priv *priv, struct bpf_prog *prog,
struct xdp_buff *xdp)
{
+ struct netsec_desc_ring *dring = &priv->desc_ring[NETSEC_RING_RX];
+ unsigned int len = xdp->data_end - xdp->data;
u32 ret = NETSEC_XDP_PASS;
int err;
u32 act;
@@ -896,7 +896,9 @@ static u32 netsec_run_xdp(struct netsec_priv *priv, struct bpf_prog *prog,
case XDP_TX:
ret = netsec_xdp_xmit_back(priv, xdp);
if (ret != NETSEC_XDP_TX)
- xdp_return_buff(xdp);
+ __page_pool_put_page(dring->page_pool,
+ virt_to_head_page(xdp->data),
+ len, true);
break;
case XDP_REDIRECT:
err = xdp_do_redirect(priv->ndev, xdp, prog);
@@ -904,7 +906,9 @@ static u32 netsec_run_xdp(struct netsec_priv *priv, struct bpf_prog *prog,
ret = NETSEC_XDP_REDIR;
} else {
ret = NETSEC_XDP_CONSUMED;
- xdp_return_buff(xdp);
+ __page_pool_put_page(dring->page_pool,
+ virt_to_head_page(xdp->data),
+ len, true);
}
break;
default:
@@ -915,7 +919,9 @@ static u32 netsec_run_xdp(struct netsec_priv *priv, struct bpf_prog *prog,
/* fall through -- handle aborts by dropping packet */
case XDP_DROP:
ret = NETSEC_XDP_CONSUMED;
- xdp_return_buff(xdp);
+ __page_pool_put_page(dring->page_pool,
+ virt_to_head_page(xdp->data),
+ len, true);
break;
}
@@ -1014,7 +1020,8 @@ static int netsec_process_rx(struct netsec_priv *priv, int budget)
* cache state. Since we paid the allocation cost if
* building an skb fails try to put the page into cache
*/
- page_pool_recycle_direct(dring->page_pool, page);
+ __page_pool_put_page(dring->page_pool, page,
+ pkt_len, true);
netif_err(priv, drv, priv->ndev,
"rx failed to build skb\n");
break;
@@ -1272,17 +1279,19 @@ static int netsec_setup_rx_dring(struct netsec_priv *priv)
{
struct netsec_desc_ring *dring = &priv->desc_ring[NETSEC_RING_RX];
struct bpf_prog *xdp_prog = READ_ONCE(priv->xdp_prog);
- struct page_pool_params pp_params = { 0 };
+ struct page_pool_params pp_params = {
+ .order = 0,
+ /* internal DMA mapping in page_pool */
+ .flags = PP_FLAG_DMA_MAP | PP_FLAG_DMA_SYNC_DEV,
+ .pool_size = DESC_NUM,
+ .nid = NUMA_NO_NODE,
+ .dev = priv->dev,
+ .dma_dir = xdp_prog ? DMA_BIDIRECTIONAL : DMA_FROM_DEVICE,
+ .offset = NETSEC_RXBUF_HEADROOM,
+ .max_len = NETSEC_RX_BUF_SIZE,
+ };
int i, err;
- pp_params.order = 0;
- /* internal DMA mapping in page_pool */
- pp_params.flags = PP_FLAG_DMA_MAP;
- pp_params.pool_size = DESC_NUM;
- pp_params.nid = NUMA_NO_NODE;
- pp_params.dev = priv->dev;
- pp_params.dma_dir = xdp_prog ? DMA_BIDIRECTIONAL : DMA_FROM_DEVICE;
-
dring->page_pool = page_pool_create(&pp_params);
if (IS_ERR(dring->page_pool)) {
err = PTR_ERR(dring->page_pool);
--
2.21.1
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH v3 net-next] net: socionext: get rid of huge dma sync in netsec_alloc_rx_data
2020-01-14 9:24 [PATCH v3 net-next] net: socionext: get rid of huge dma sync in netsec_alloc_rx_data Lorenzo Bianconi
@ 2020-01-14 10:59 ` Jesper Dangaard Brouer
2020-01-15 2:46 ` David Miller
1 sibling, 0 replies; 3+ messages in thread
From: Jesper Dangaard Brouer @ 2020-01-14 10:59 UTC (permalink / raw)
To: Lorenzo Bianconi
Cc: netdev, lorenzo.bianconi, davem, ilias.apalodimas, kuba, brouer
On Tue, 14 Jan 2020 10:24:19 +0100
Lorenzo Bianconi <lorenzo@kernel.org> wrote:
> Socionext driver can run on dma coherent and non-coherent devices.
> Get rid of huge dma_sync_single_for_device in netsec_alloc_rx_data since
> now the driver can let page_pool API to managed needed DMA sync
>
> Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
> Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
> ---
> Changes since v2:
> - fix checkpatch warnings
>
> Changes since v1:
> - rely on original frame size for dma sync
> ---
> drivers/net/ethernet/socionext/netsec.c | 43 +++++++++++++++----------
> 1 file changed, 26 insertions(+), 17 deletions(-)
>
> diff --git a/drivers/net/ethernet/socionext/netsec.c b/drivers/net/ethernet/socionext/netsec.c
> index b5a9e947a4a8..6870a6ce76a6 100644
> --- a/drivers/net/ethernet/socionext/netsec.c
> +++ b/drivers/net/ethernet/socionext/netsec.c
[...]
> @@ -883,6 +881,8 @@ static u32 netsec_xdp_xmit_back(struct netsec_priv *priv, struct xdp_buff *xdp)
> static u32 netsec_run_xdp(struct netsec_priv *priv, struct bpf_prog *prog,
> struct xdp_buff *xdp)
> {
> + struct netsec_desc_ring *dring = &priv->desc_ring[NETSEC_RING_RX];
> + unsigned int len = xdp->data_end - xdp->data;
This is correct because you calc len before BPF-prog can change these,
and because we can currently only shrink tail (data_end).
Acked-by: Jesper Dangaard Brouer <brouer@redhat.com>
--
Best regards,
Jesper Dangaard Brouer
MSc.CS, Principal Kernel Engineer at Red Hat
LinkedIn: http://www.linkedin.com/in/brouer
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH v3 net-next] net: socionext: get rid of huge dma sync in netsec_alloc_rx_data
2020-01-14 9:24 [PATCH v3 net-next] net: socionext: get rid of huge dma sync in netsec_alloc_rx_data Lorenzo Bianconi
2020-01-14 10:59 ` Jesper Dangaard Brouer
@ 2020-01-15 2:46 ` David Miller
1 sibling, 0 replies; 3+ messages in thread
From: David Miller @ 2020-01-15 2:46 UTC (permalink / raw)
To: lorenzo; +Cc: netdev, lorenzo.bianconi, brouer, ilias.apalodimas, kuba
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Tue, 14 Jan 2020 10:24:19 +0100
> Socionext driver can run on dma coherent and non-coherent devices.
> Get rid of huge dma_sync_single_for_device in netsec_alloc_rx_data since
> now the driver can let page_pool API to managed needed DMA sync
>
> Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
> Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Applied, thanks.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2020-01-15 2:47 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-01-14 9:24 [PATCH v3 net-next] net: socionext: get rid of huge dma sync in netsec_alloc_rx_data Lorenzo Bianconi
2020-01-14 10:59 ` Jesper Dangaard Brouer
2020-01-15 2:46 ` David Miller
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).