From: Salil Mehta <salil.mehta@huawei.com>
To: <davem@davemloft.net>
Cc: <salil.mehta@huawei.com>, <yisen.zhuang@huawei.com>,
<lipeng321@huawei.com>, <mehta.salil@opnsrc.net>,
<netdev@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
<linux-rdma@vger.kernel.org>, <linuxarm@huawei.com>
Subject: [PATCH V2 net-next 4/5] net: hns3: Add skb chain when num of RX buf exceeds MAX_SKB_FRAGS
Date: Thu, 15 Nov 2018 09:29:24 +0000 [thread overview]
Message-ID: <20181115092925.11812-5-salil.mehta@huawei.com> (raw)
In-Reply-To: <20181115092925.11812-1-salil.mehta@huawei.com>
From: Peng Li <lipeng321@huawei.com>
MAX_SKB_FRAGS in protocol stack is defined as:
MAX_SKB_FRAGS is 17 when PAGE_SIZE is 4K. If HW enable GRO, it may
merge small packets and the rx buffer may be more than
MAX_SKB_FRAGS. So driver will add skb chain when RX buffer num.
more than MAX_SKB_FRAGS.
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
---
.../net/ethernet/hisilicon/hns3/hns3_enet.c | 37 ++++++++++++++++++-
.../net/ethernet/hisilicon/hns3/hns3_enet.h | 2 +
2 files changed, 37 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
index 860067898471..7776089b6bc2 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
@@ -2361,6 +2361,9 @@ static void hns3_rx_checksum(struct hns3_enet_ring *ring, struct sk_buff *skb,
static void hns3_rx_skb(struct hns3_enet_ring *ring, struct sk_buff *skb)
{
+ if (skb_has_frag_list(skb))
+ napi_gro_flush(&ring->tqp_vector->napi, false);
+
napi_gro_receive(&ring->tqp_vector->napi, skb);
}
@@ -2417,6 +2420,8 @@ static int hns3_alloc_skb(struct hns3_enet_ring *ring, int length,
prefetchw(skb->data);
ring->pending_buf = 1;
+ ring->frag_num = 0;
+ ring->tail_skb = NULL;
if (length <= HNS3_RX_HEAD_SIZE) {
memcpy(__skb_put(skb, length), va, ALIGN(length, sizeof(long)));
@@ -2435,7 +2440,7 @@ static int hns3_alloc_skb(struct hns3_enet_ring *ring, int length,
ring->pull_len = eth_get_headlen(va, HNS3_RX_HEAD_SIZE);
__skb_put(skb, ring->pull_len);
- hns3_nic_reuse_page(skb, 0, ring, ring->pull_len,
+ hns3_nic_reuse_page(skb, ring->frag_num++, ring, ring->pull_len,
desc_cb);
ring_ptr_move_fw(ring, next_to_clean);
@@ -2446,6 +2451,8 @@ static int hns3_add_frag(struct hns3_enet_ring *ring, struct hns3_desc *desc,
struct sk_buff **out_skb, bool pending)
{
struct sk_buff *skb = *out_skb;
+ struct sk_buff *head_skb = *out_skb;
+ struct sk_buff *new_skb;
struct hns3_desc_cb *desc_cb;
struct hns3_desc *pre_desc;
u32 bd_base_info;
@@ -2470,7 +2477,33 @@ static int hns3_add_frag(struct hns3_enet_ring *ring, struct hns3_desc *desc,
if (!hnae3_get_bit(bd_base_info, HNS3_RXD_VLD_B))
return -ENXIO;
- hns3_nic_reuse_page(skb, ring->pending_buf, ring, 0, desc_cb);
+ if (unlikely(ring->frag_num >= MAX_SKB_FRAGS)) {
+ new_skb = napi_alloc_skb(&ring->tqp_vector->napi,
+ HNS3_RX_HEAD_SIZE);
+ if (unlikely(!new_skb)) {
+ netdev_err(ring->tqp->handle->kinfo.netdev,
+ "alloc rx skb frag fail\n");
+ return -ENXIO;
+ }
+ ring->frag_num = 0;
+
+ if (ring->tail_skb) {
+ ring->tail_skb->next = new_skb;
+ ring->tail_skb = new_skb;
+ } else {
+ skb_shinfo(skb)->frag_list = new_skb;
+ ring->tail_skb = new_skb;
+ }
+ }
+
+ if (ring->tail_skb) {
+ head_skb->truesize += hnae3_buf_size(ring);
+ head_skb->data_len += le16_to_cpu(desc->rx.size);
+ head_skb->len += le16_to_cpu(desc->rx.size);
+ skb = ring->tail_skb;
+ }
+
+ hns3_nic_reuse_page(skb, ring->frag_num++, ring, 0, desc_cb);
ring_ptr_move_fw(ring, next_to_clean);
ring->pending_buf++;
}
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h
index d8c0998127be..8e56b7e44978 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h
@@ -402,6 +402,7 @@ struct hns3_enet_ring {
int next_to_clean;
int pull_len; /* head length for current packet */
+ u32 frag_num;
unsigned char *va; /* first buffer address for current packet */
u32 flag; /* ring attribute */
@@ -412,6 +413,7 @@ struct hns3_enet_ring {
int pending_buf;
struct sk_buff *skb;
+ struct sk_buff *tail_skb;
};
struct hns_queue;
--
2.17.1
next prev parent reply other threads:[~2018-11-15 9:30 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-11-15 9:29 [PATCH V2 net-next 0/5] net: hns3: Add support of hardware GRO to HNS3 Driver Salil Mehta
2018-11-15 9:29 ` [PATCH V2 net-next 1/5] net: hns3: Enable HW GRO for Rev B(=0x21) HNS3 hardware Salil Mehta
2018-11-15 9:29 ` [PATCH V2 net-next 2/5] net: hns3: Add handling of GRO Pkts not fully RX'ed in NAPI poll Salil Mehta
2018-11-15 9:29 ` [PATCH V2 net-next 3/5] net: hns3: Add support for ethtool -K to enable/disable HW GRO Salil Mehta
2018-11-15 9:29 ` Salil Mehta [this message]
2018-11-15 9:29 ` [PATCH V2 net-next 5/5] net: hns3: Adds GRO params to SKB for the stack Salil Mehta
2018-11-15 17:46 ` [PATCH V2 net-next 0/5] net: hns3: Add support of hardware GRO to HNS3 Driver David Miller
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=20181115092925.11812-5-salil.mehta@huawei.com \
--to=salil.mehta@huawei.com \
--cc=davem@davemloft.net \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-rdma@vger.kernel.org \
--cc=linuxarm@huawei.com \
--cc=lipeng321@huawei.com \
--cc=mehta.salil@opnsrc.net \
--cc=netdev@vger.kernel.org \
--cc=yisen.zhuang@huawei.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).