netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Simon Horman <simon.horman@netronome.com>
To: Yangchun Fu <yangchun@google.com>
Cc: netdev@vger.kernel.org, Catherine Sullivan <csully@google.com>
Subject: Re: [PATCH net] gve: Fixes DMA synchronization.
Date: Tue, 29 Oct 2019 09:18:52 +0100	[thread overview]
Message-ID: <20191029081851.GA23615@netronome.com> (raw)
In-Reply-To: <20191028182309.73313-1-yangchun@google.com>

Hi Yungchun,

thanks for your patch.

On Mon, Oct 28, 2019 at 11:23:09AM -0700, Yangchun Fu wrote:
> Synces the DMA buffer properly in order for CPU and device to see
> the most up-to-data data.
> 
> Signed-off-by: Yangchun Fu <yangchun@google.com>
> Reviewed-by: Catherine Sullivan <csully@google.com>
> ---
>  drivers/net/ethernet/google/gve/gve_rx.c |  2 ++
>  drivers/net/ethernet/google/gve/gve_tx.c | 26 ++++++++++++++++++++++--
>  2 files changed, 26 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/ethernet/google/gve/gve_rx.c b/drivers/net/ethernet/google/gve/gve_rx.c
> index 59564ac99d2a..edec61dfc868 100644
> --- a/drivers/net/ethernet/google/gve/gve_rx.c
> +++ b/drivers/net/ethernet/google/gve/gve_rx.c
> @@ -289,6 +289,8 @@ static bool gve_rx(struct gve_rx_ring *rx, struct gve_rx_desc *rx_desc,
>  
>  	len = be16_to_cpu(rx_desc->len) - GVE_RX_PAD;
>  	page_info = &rx->data.page_info[idx];
> +	dma_sync_single_for_cpu(&priv->pdev->dev, rx->data.qpl->page_buses[idx],
> +				PAGE_SIZE, DMA_FROM_DEVICE);
>  
>  	/* gvnic can only receive into registered segments. If the buffer
>  	 * can't be recycled, our only choice is to copy the data out of
> diff --git a/drivers/net/ethernet/google/gve/gve_tx.c b/drivers/net/ethernet/google/gve/gve_tx.c
> index 778b87b5a06c..d8342b7b9764 100644
> --- a/drivers/net/ethernet/google/gve/gve_tx.c
> +++ b/drivers/net/ethernet/google/gve/gve_tx.c
> @@ -390,7 +390,23 @@ static void gve_tx_fill_seg_desc(union gve_tx_desc *seg_desc,
>  	seg_desc->seg.seg_addr = cpu_to_be64(addr);
>  }
>  
> -static int gve_tx_add_skb(struct gve_tx_ring *tx, struct sk_buff *skb)
> +static inline void gve_dma_sync_for_device(struct gve_priv *priv,

It seems that only priv->pdev->dev is used in this function.  Perhaps it
would be better to pass it to this function rather than all of priv.

> +					   dma_addr_t *page_buses,
> +					   u64 iov_offset, u64 iov_len)
> +{
> +	u64 addr;
> +	dma_addr_t dma;
> +
> +	for (addr = iov_offset; addr < iov_offset + iov_len;
> +	     addr += PAGE_SIZE) {
> +		dma = page_buses[addr / PAGE_SIZE];
> +		dma_sync_single_for_device(&priv->pdev->dev, dma, PAGE_SIZE,
> +					   DMA_TO_DEVICE);
> +	}
> +}
> +
> +static int gve_tx_add_skb(struct gve_tx_ring *tx, struct sk_buff *skb,
> +			  struct gve_priv *priv)
>  {
>  	int pad_bytes, hlen, hdr_nfrags, payload_nfrags, l4_hdr_offset;
>  	union gve_tx_desc *pkt_desc, *seg_desc;
> @@ -432,6 +448,9 @@ static int gve_tx_add_skb(struct gve_tx_ring *tx, struct sk_buff *skb)
>  	skb_copy_bits(skb, 0,
>  		      tx->tx_fifo.base + info->iov[hdr_nfrags - 1].iov_offset,
>  		      hlen);
> +	gve_dma_sync_for_device(priv, tx->tx_fifo.qpl->page_buses,
> +				info->iov[hdr_nfrags - 1].iov_offset,
> +				info->iov[hdr_nfrags - 1].iov_len);
>  	copy_offset = hlen;
>  
>  	for (i = payload_iov; i < payload_nfrags + payload_iov; i++) {
> @@ -445,6 +464,9 @@ static int gve_tx_add_skb(struct gve_tx_ring *tx, struct sk_buff *skb)
>  		skb_copy_bits(skb, copy_offset,
>  			      tx->tx_fifo.base + info->iov[i].iov_offset,
>  			      info->iov[i].iov_len);
> +		gve_dma_sync_for_device(priv, tx->tx_fifo.qpl->page_buses,
> +					info->iov[i].iov_offset,
> +					info->iov[i].iov_len);
>  		copy_offset += info->iov[i].iov_len;
>  	}
>  
> @@ -473,7 +495,7 @@ netdev_tx_t gve_tx(struct sk_buff *skb, struct net_device *dev)
>  		gve_tx_put_doorbell(priv, tx->q_resources, tx->req);
>  		return NETDEV_TX_BUSY;
>  	}
> -	nsegs = gve_tx_add_skb(tx, skb);
> +	nsegs = gve_tx_add_skb(tx, skb, priv);
>  
>  	netdev_tx_sent_queue(tx->netdev_txq, skb->len);
>  	skb_tx_timestamp(skb);
> -- 
> 2.24.0.rc0.303.g954a862665-goog
> 

  reply	other threads:[~2019-10-29  8:18 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-10-28 18:23 [PATCH net] gve: Fixes DMA synchronization Yangchun Fu
2019-10-29  8:18 ` Simon Horman [this message]
2019-10-30 21:50   ` Yangchun Fu
2019-10-30  0:41 ` David Miller
2019-10-30 21:51   ` Yangchun Fu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20191029081851.GA23615@netronome.com \
    --to=simon.horman@netronome.com \
    --cc=csully@google.com \
    --cc=netdev@vger.kernel.org \
    --cc=yangchun@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).