From: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
To: anthony.l.nguyen@intel.com
Cc: netdev@vger.kernel.org, davem@davemloft.net, kuba@kernel.org,
pabeni@redhat.com, bpf@vger.kernel.org, ast@kernel.org,
daniel@iogearbox.net, andrii@kernel.org,
jesse.brandeburg@intel.com, anthony.l.nguyen@intel.com,
magnus.karlsson@intel.com, jbrouer@redhat.com, toke@redhat.com,
intel-wired-lan@lists.osuosl.org
Subject: Re: [PATCH net-next] ixgbe: add xdp frags support to ndo_xdp_xmit
Date: Wed, 20 Apr 2022 15:09:15 +0200 [thread overview]
Message-ID: <YmAF+wBcluzOGXgJ@lore-desk> (raw)
In-Reply-To: <6de1d7547b60677ad0b0f7ebcbc7ebc76a31dcf7.1649180962.git.lorenzo@kernel.org>
[-- Attachment #1: Type: text/plain, Size: 5034 bytes --]
> Add the capability to map non-linear xdp frames in XDP_TX and ndo_xdp_xmit
> callback.
Hi Tony,
do you have any feedbacks about this patch?
Regards,
Lorenzo
>
> Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
> ---
> drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 99 ++++++++++++-------
> 1 file changed, 63 insertions(+), 36 deletions(-)
>
> diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> index c4a4954aa317..8b84c9b2eecc 100644
> --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> @@ -2344,6 +2344,7 @@ static int ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
> hard_start = page_address(rx_buffer->page) +
> rx_buffer->page_offset - offset;
> xdp_prepare_buff(&xdp, hard_start, offset, size, true);
> + xdp_buff_clear_frags_flag(&xdp);
> #if (PAGE_SIZE > 4096)
> /* At larger PAGE_SIZE, frame_sz depend on len size */
> xdp.frame_sz = ixgbe_rx_frame_truesize(rx_ring, size);
> @@ -8571,57 +8572,83 @@ static u16 ixgbe_select_queue(struct net_device *dev, struct sk_buff *skb,
> int ixgbe_xmit_xdp_ring(struct ixgbe_ring *ring,
> struct xdp_frame *xdpf)
> {
> - struct ixgbe_tx_buffer *tx_buffer;
> - union ixgbe_adv_tx_desc *tx_desc;
> - u32 len, cmd_type;
> - dma_addr_t dma;
> - u16 i;
> -
> - len = xdpf->len;
> + struct skb_shared_info *sinfo = xdp_get_shared_info_from_frame(xdpf);
> + u8 nr_frags = unlikely(xdp_frame_has_frags(xdpf)) ? sinfo->nr_frags : 0;
> + u16 i = 0, index = ring->next_to_use;
> + struct ixgbe_tx_buffer *tx_head = &ring->tx_buffer_info[index];
> + struct ixgbe_tx_buffer *tx_buff = tx_head;
> + union ixgbe_adv_tx_desc *tx_desc = IXGBE_TX_DESC(ring, index);
> + u32 cmd_type, len = xdpf->len;
> + void *data = xdpf->data;
>
> - if (unlikely(!ixgbe_desc_unused(ring)))
> + if (unlikely(ixgbe_desc_unused(ring) < 1 + nr_frags))
> return IXGBE_XDP_CONSUMED;
>
> - dma = dma_map_single(ring->dev, xdpf->data, len, DMA_TO_DEVICE);
> - if (dma_mapping_error(ring->dev, dma))
> - return IXGBE_XDP_CONSUMED;
> + tx_head->bytecount = xdp_get_frame_len(xdpf);
> + tx_head->gso_segs = 1;
> + tx_head->xdpf = xdpf;
>
> - /* record the location of the first descriptor for this packet */
> - tx_buffer = &ring->tx_buffer_info[ring->next_to_use];
> - tx_buffer->bytecount = len;
> - tx_buffer->gso_segs = 1;
> - tx_buffer->protocol = 0;
> + tx_desc->read.olinfo_status =
> + cpu_to_le32(tx_head->bytecount << IXGBE_ADVTXD_PAYLEN_SHIFT);
>
> - i = ring->next_to_use;
> - tx_desc = IXGBE_TX_DESC(ring, i);
> + for (;;) {
> + dma_addr_t dma;
>
> - dma_unmap_len_set(tx_buffer, len, len);
> - dma_unmap_addr_set(tx_buffer, dma, dma);
> - tx_buffer->xdpf = xdpf;
> + dma = dma_map_single(ring->dev, data, len, DMA_TO_DEVICE);
> + if (dma_mapping_error(ring->dev, dma))
> + goto unmap;
>
> - tx_desc->read.buffer_addr = cpu_to_le64(dma);
> + dma_unmap_len_set(tx_buff, len, len);
> + dma_unmap_addr_set(tx_buff, dma, dma);
> +
> + cmd_type = IXGBE_ADVTXD_DTYP_DATA | IXGBE_ADVTXD_DCMD_DEXT |
> + IXGBE_ADVTXD_DCMD_IFCS | len;
> + tx_desc->read.cmd_type_len = cpu_to_le32(cmd_type);
> + tx_desc->read.buffer_addr = cpu_to_le64(dma);
> + tx_buff->protocol = 0;
> +
> + if (++index == ring->count)
> + index = 0;
> +
> + if (i == nr_frags)
> + break;
> +
> + tx_buff = &ring->tx_buffer_info[index];
> + tx_desc = IXGBE_TX_DESC(ring, index);
> + tx_desc->read.olinfo_status = 0;
>
> + data = skb_frag_address(&sinfo->frags[i]);
> + len = skb_frag_size(&sinfo->frags[i]);
> + i++;
> + }
> /* put descriptor type bits */
> - cmd_type = IXGBE_ADVTXD_DTYP_DATA |
> - IXGBE_ADVTXD_DCMD_DEXT |
> - IXGBE_ADVTXD_DCMD_IFCS;
> - cmd_type |= len | IXGBE_TXD_CMD;
> - tx_desc->read.cmd_type_len = cpu_to_le32(cmd_type);
> - tx_desc->read.olinfo_status =
> - cpu_to_le32(len << IXGBE_ADVTXD_PAYLEN_SHIFT);
> + tx_desc->read.cmd_type_len |= cpu_to_le32(IXGBE_TXD_CMD);
>
> /* Avoid any potential race with xdp_xmit and cleanup */
> smp_wmb();
>
> - /* set next_to_watch value indicating a packet is present */
> - i++;
> - if (i == ring->count)
> - i = 0;
> -
> - tx_buffer->next_to_watch = tx_desc;
> - ring->next_to_use = i;
> + tx_head->next_to_watch = tx_desc;
> + ring->next_to_use = index;
>
> return IXGBE_XDP_TX;
> +
> +unmap:
> + for (;;) {
> + tx_buff = &ring->tx_buffer_info[index];
> + if (dma_unmap_len(tx_buff, len))
> + dma_unmap_page(ring->dev, dma_unmap_addr(tx_buff, dma),
> + dma_unmap_len(tx_buff, len),
> + DMA_TO_DEVICE);
> + dma_unmap_len_set(tx_buff, len, 0);
> + if (tx_buff == tx_head)
> + break;
> +
> + if (!index)
> + index += ring->count;
> + index--;
> + }
> +
> + return IXGBE_XDP_CONSUMED;
> }
>
> netdev_tx_t ixgbe_xmit_frame_ring(struct sk_buff *skb,
> --
> 2.35.1
>
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]
WARNING: multiple messages have this Message-ID (diff)
From: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
To: intel-wired-lan@osuosl.org
Subject: [Intel-wired-lan] [PATCH net-next] ixgbe: add xdp frags support to ndo_xdp_xmit
Date: Wed, 20 Apr 2022 15:09:15 +0200 [thread overview]
Message-ID: <YmAF+wBcluzOGXgJ@lore-desk> (raw)
In-Reply-To: <6de1d7547b60677ad0b0f7ebcbc7ebc76a31dcf7.1649180962.git.lorenzo@kernel.org>
> Add the capability to map non-linear xdp frames in XDP_TX and ndo_xdp_xmit
> callback.
Hi Tony,
do you have any feedbacks about this patch?
Regards,
Lorenzo
>
> Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
> ---
> drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 99 ++++++++++++-------
> 1 file changed, 63 insertions(+), 36 deletions(-)
>
> diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> index c4a4954aa317..8b84c9b2eecc 100644
> --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> @@ -2344,6 +2344,7 @@ static int ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
> hard_start = page_address(rx_buffer->page) +
> rx_buffer->page_offset - offset;
> xdp_prepare_buff(&xdp, hard_start, offset, size, true);
> + xdp_buff_clear_frags_flag(&xdp);
> #if (PAGE_SIZE > 4096)
> /* At larger PAGE_SIZE, frame_sz depend on len size */
> xdp.frame_sz = ixgbe_rx_frame_truesize(rx_ring, size);
> @@ -8571,57 +8572,83 @@ static u16 ixgbe_select_queue(struct net_device *dev, struct sk_buff *skb,
> int ixgbe_xmit_xdp_ring(struct ixgbe_ring *ring,
> struct xdp_frame *xdpf)
> {
> - struct ixgbe_tx_buffer *tx_buffer;
> - union ixgbe_adv_tx_desc *tx_desc;
> - u32 len, cmd_type;
> - dma_addr_t dma;
> - u16 i;
> -
> - len = xdpf->len;
> + struct skb_shared_info *sinfo = xdp_get_shared_info_from_frame(xdpf);
> + u8 nr_frags = unlikely(xdp_frame_has_frags(xdpf)) ? sinfo->nr_frags : 0;
> + u16 i = 0, index = ring->next_to_use;
> + struct ixgbe_tx_buffer *tx_head = &ring->tx_buffer_info[index];
> + struct ixgbe_tx_buffer *tx_buff = tx_head;
> + union ixgbe_adv_tx_desc *tx_desc = IXGBE_TX_DESC(ring, index);
> + u32 cmd_type, len = xdpf->len;
> + void *data = xdpf->data;
>
> - if (unlikely(!ixgbe_desc_unused(ring)))
> + if (unlikely(ixgbe_desc_unused(ring) < 1 + nr_frags))
> return IXGBE_XDP_CONSUMED;
>
> - dma = dma_map_single(ring->dev, xdpf->data, len, DMA_TO_DEVICE);
> - if (dma_mapping_error(ring->dev, dma))
> - return IXGBE_XDP_CONSUMED;
> + tx_head->bytecount = xdp_get_frame_len(xdpf);
> + tx_head->gso_segs = 1;
> + tx_head->xdpf = xdpf;
>
> - /* record the location of the first descriptor for this packet */
> - tx_buffer = &ring->tx_buffer_info[ring->next_to_use];
> - tx_buffer->bytecount = len;
> - tx_buffer->gso_segs = 1;
> - tx_buffer->protocol = 0;
> + tx_desc->read.olinfo_status =
> + cpu_to_le32(tx_head->bytecount << IXGBE_ADVTXD_PAYLEN_SHIFT);
>
> - i = ring->next_to_use;
> - tx_desc = IXGBE_TX_DESC(ring, i);
> + for (;;) {
> + dma_addr_t dma;
>
> - dma_unmap_len_set(tx_buffer, len, len);
> - dma_unmap_addr_set(tx_buffer, dma, dma);
> - tx_buffer->xdpf = xdpf;
> + dma = dma_map_single(ring->dev, data, len, DMA_TO_DEVICE);
> + if (dma_mapping_error(ring->dev, dma))
> + goto unmap;
>
> - tx_desc->read.buffer_addr = cpu_to_le64(dma);
> + dma_unmap_len_set(tx_buff, len, len);
> + dma_unmap_addr_set(tx_buff, dma, dma);
> +
> + cmd_type = IXGBE_ADVTXD_DTYP_DATA | IXGBE_ADVTXD_DCMD_DEXT |
> + IXGBE_ADVTXD_DCMD_IFCS | len;
> + tx_desc->read.cmd_type_len = cpu_to_le32(cmd_type);
> + tx_desc->read.buffer_addr = cpu_to_le64(dma);
> + tx_buff->protocol = 0;
> +
> + if (++index == ring->count)
> + index = 0;
> +
> + if (i == nr_frags)
> + break;
> +
> + tx_buff = &ring->tx_buffer_info[index];
> + tx_desc = IXGBE_TX_DESC(ring, index);
> + tx_desc->read.olinfo_status = 0;
>
> + data = skb_frag_address(&sinfo->frags[i]);
> + len = skb_frag_size(&sinfo->frags[i]);
> + i++;
> + }
> /* put descriptor type bits */
> - cmd_type = IXGBE_ADVTXD_DTYP_DATA |
> - IXGBE_ADVTXD_DCMD_DEXT |
> - IXGBE_ADVTXD_DCMD_IFCS;
> - cmd_type |= len | IXGBE_TXD_CMD;
> - tx_desc->read.cmd_type_len = cpu_to_le32(cmd_type);
> - tx_desc->read.olinfo_status =
> - cpu_to_le32(len << IXGBE_ADVTXD_PAYLEN_SHIFT);
> + tx_desc->read.cmd_type_len |= cpu_to_le32(IXGBE_TXD_CMD);
>
> /* Avoid any potential race with xdp_xmit and cleanup */
> smp_wmb();
>
> - /* set next_to_watch value indicating a packet is present */
> - i++;
> - if (i == ring->count)
> - i = 0;
> -
> - tx_buffer->next_to_watch = tx_desc;
> - ring->next_to_use = i;
> + tx_head->next_to_watch = tx_desc;
> + ring->next_to_use = index;
>
> return IXGBE_XDP_TX;
> +
> +unmap:
> + for (;;) {
> + tx_buff = &ring->tx_buffer_info[index];
> + if (dma_unmap_len(tx_buff, len))
> + dma_unmap_page(ring->dev, dma_unmap_addr(tx_buff, dma),
> + dma_unmap_len(tx_buff, len),
> + DMA_TO_DEVICE);
> + dma_unmap_len_set(tx_buff, len, 0);
> + if (tx_buff == tx_head)
> + break;
> +
> + if (!index)
> + index += ring->count;
> + index--;
> + }
> +
> + return IXGBE_XDP_CONSUMED;
> }
>
> netdev_tx_t ixgbe_xmit_frame_ring(struct sk_buff *skb,
> --
> 2.35.1
>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
URL: <http://lists.osuosl.org/pipermail/intel-wired-lan/attachments/20220420/573289c8/attachment.asc>
next prev parent reply other threads:[~2022-04-20 13:09 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-04-05 17:53 [PATCH net-next] ixgbe: add xdp frags support to ndo_xdp_xmit Lorenzo Bianconi
2022-04-20 13:09 ` Lorenzo Bianconi [this message]
2022-04-20 13:09 ` [Intel-wired-lan] " Lorenzo Bianconi
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=YmAF+wBcluzOGXgJ@lore-desk \
--to=lorenzo.bianconi@redhat.com \
--cc=andrii@kernel.org \
--cc=anthony.l.nguyen@intel.com \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=davem@davemloft.net \
--cc=intel-wired-lan@lists.osuosl.org \
--cc=jbrouer@redhat.com \
--cc=jesse.brandeburg@intel.com \
--cc=kuba@kernel.org \
--cc=magnus.karlsson@intel.com \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=toke@redhat.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.