static int xlgmac_prep_tso(struct sk_buff *skb,
struct xlgmac_pkt_info *pkt_info)
{
int ret;
if (!XLGMAC_GET_REG_BITS(pkt_info->attributes,
TX_PACKET_ATTRIBUTES_TSO_ENABLE_POS,
TX_PACKET_ATTRIBUTES_TSO_ENABLE_LEN))
return 0;
ret = skb_cow_head(skb, 0);
if (ret)
return ret;
pkt_info->header_len = skb_transport_offset(skb) + tcp_hdrlen(skb);
pkt_info->tcp_header_len = tcp_hdrlen(skb);
pkt_info->tcp_payload_len = skb->len - pkt_info->header_len;
pkt_info->mss = skb_shinfo(skb)->gso_size;
XLGMAC_PR("header_len=%u\n", pkt_info->header_len);
XLGMAC_PR("tcp_header_len=%u, tcp_payload_len=%u\n",
pkt_info->tcp_header_len, pkt_info->tcp_payload_len);
XLGMAC_PR("mss=%u\n", pkt_info->mss);
/* Update the number of packets that will ultimately be transmitted
* along with the extra bytes for each extra packet
*/
pkt_info->tx_packets = skb_shinfo(skb)->gso_segs;
pkt_info->tx_bytes += (pkt_info->tx_packets - 1) * pkt_info->header_len;
return 0;
}
static int get_real_size(struct sk_buff *skb, struct net_device *dev,
int *lso_header_size)
{
struct mlx4_en_priv *priv = netdev_priv(dev);
int real_size;
if (skb_is_gso(skb)) {
*lso_header_size = skb_transport_offset(skb) + tcp_hdrlen(skb);
real_size = CTRL_SIZE + skb_shinfo(skb)->nr_frags * DS_SIZE +
ALIGN(*lso_header_size + 4, DS_SIZE);
if (unlikely(*lso_header_size != skb_headlen(skb))) {
/* We add a segment for the skb linear buffer only if
* it contains data */
if (*lso_header_size < skb_headlen(skb))
real_size += DS_SIZE;
else {
if (netif_msg_tx_err(priv))
en_warn(priv, "Non-linear headers\n");
return 0;
}
}
} else {
*lso_header_size = 0;
if (!is_inline(skb, NULL))
real_size = CTRL_SIZE + (skb_shinfo(skb)->nr_frags + 1) * DS_SIZE;
else
real_size = inline_size(skb);
}
return real_size;
}
static inline void enic_queue_wq_skb_tso(struct enic *enic,
struct vnic_wq *wq, struct sk_buff *skb, unsigned int mss,
int vlan_tag_insert, unsigned int vlan_tag)
{
unsigned int head_len = skb_headlen(skb);
unsigned int len_left = skb->len - head_len;
unsigned int hdr_len = skb_transport_offset(skb) + tcp_hdrlen(skb);
int eop = (len_left == 0);
/* Preload TCP csum field with IP pseudo hdr calculated
* with IP length set to zero. HW will later add in length
* to each TCP segment resulting from the TSO.
*/
if (skb->protocol == __constant_htons(ETH_P_IP)) {
ip_hdr(skb)->check = 0;
tcp_hdr(skb)->check = ~csum_tcpudp_magic(ip_hdr(skb)->saddr,
ip_hdr(skb)->daddr, 0, IPPROTO_TCP, 0);
} else if (skb->protocol == __constant_htons(ETH_P_IPV6)) {
tcp_hdr(skb)->check = ~csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
&ipv6_hdr(skb)->daddr, 0, IPPROTO_TCP, 0);
}
/* Queue the main skb fragment */
enic_queue_wq_desc_tso(wq, skb,
pci_map_single(enic->pdev, skb->data,
head_len, PCI_DMA_TODEVICE),
head_len,
mss, hdr_len,
vlan_tag_insert, vlan_tag,
eop);
if (!eop)
enic_queue_wq_skb_cont(enic, wq, skb, len_left);
}
/**
* nfp_net_tx_tso() - Set up Tx descriptor for LSO
* @nn: NFP Net device
* @r_vec: per-ring structure
* @txbuf: Pointer to driver soft TX descriptor
* @txd: Pointer to HW TX descriptor
* @skb: Pointer to SKB
*
* Set up Tx descriptor for LSO, do nothing for non-LSO skbs.
* Return error on packet header greater than maximum supported LSO header size.
*/
static void nfp_net_tx_tso(struct nfp_net *nn, struct nfp_net_r_vector *r_vec,
struct nfp_net_tx_buf *txbuf,
struct nfp_net_tx_desc *txd, struct sk_buff *skb)
{
u32 hdrlen;
u16 mss;
if (!skb_is_gso(skb))
return;
if (!skb->encapsulation)
hdrlen = skb_transport_offset(skb) + tcp_hdrlen(skb);
else
hdrlen = skb_inner_transport_header(skb) - skb->data +
inner_tcp_hdrlen(skb);
txbuf->pkt_cnt = skb_shinfo(skb)->gso_segs;
txbuf->real_len += hdrlen * (txbuf->pkt_cnt - 1);
mss = skb_shinfo(skb)->gso_size & PCIE_DESC_TX_MSS_MASK;
txd->l4_offset = hdrlen;
txd->mss = cpu_to_le16(mss);
txd->flags |= PCIE_DESC_TX_LSO;
u64_stats_update_begin(&r_vec->tx_sync);
r_vec->tx_lso++;
u64_stats_update_end(&r_vec->tx_sync);
}
static unsigned int
tcpoptstrip_mangle_packet(struct sk_buff *skb,
const struct xt_tcpoptstrip_target_info *info,
unsigned int tcphoff, unsigned int minlen)
{
unsigned int optl, i, j;
struct tcphdr *tcph;
u_int16_t n, o;
u_int8_t *opt;
if (!skb_make_writable(skb, skb->len))
return NF_DROP;
tcph = (struct tcphdr *)(skb_network_header(skb) + tcphoff);
opt = (u_int8_t *)tcph;
/*
* Walk through all TCP options - if we find some option to remove,
* set all octets to %TCPOPT_NOP and adjust checksum.
*/
for (i = sizeof(struct tcphdr); i < tcp_hdrlen(skb); i += optl) {
optl = optlen(opt, i);
if (i + optl > tcp_hdrlen(skb))
break;
if (!tcpoptstrip_test_bit(info->strip_bmap, opt[i]))
continue;
for (j = 0; j < optl; ++j) {
o = opt[i+j];
n = TCPOPT_NOP;
if ((i + j) % 2 == 0) {
o <<= 8;
n <<= 8;
}
inet_proto_csum_replace2(&tcph->check, skb, htons(o),
htons(n), 0);
}
memset(opt + i, TCPOPT_NOP, optl);
}
return XT_CONTINUE;
}
static u64 xgene_enet_work_msg(struct sk_buff *skb)
{
struct net_device *ndev = skb->dev;
struct xgene_enet_pdata *pdata = netdev_priv(ndev);
struct iphdr *iph;
u8 l3hlen = 0, l4hlen = 0;
u8 ethhdr, proto = 0, csum_enable = 0;
u64 hopinfo = 0;
u32 hdr_len, mss = 0;
u32 i, len, nr_frags;
ethhdr = xgene_enet_hdr_len(skb->data);
if (unlikely(skb->protocol != htons(ETH_P_IP)) &&
unlikely(skb->protocol != htons(ETH_P_8021Q)))
goto out;
if (unlikely(!(skb->dev->features & NETIF_F_IP_CSUM)))
goto out;
iph = ip_hdr(skb);
if (unlikely(ip_is_fragment(iph)))
goto out;
if (likely(iph->protocol == IPPROTO_TCP)) {
l4hlen = tcp_hdrlen(skb) >> 2;
csum_enable = 1;
proto = TSO_IPPROTO_TCP;
if (ndev->features & NETIF_F_TSO) {
hdr_len = ethhdr + ip_hdrlen(skb) + tcp_hdrlen(skb);
mss = skb_shinfo(skb)->gso_size;
if (skb_is_nonlinear(skb)) {
len = skb_headlen(skb);
nr_frags = skb_shinfo(skb)->nr_frags;
for (i = 0; i < 2 && i < nr_frags; i++)
len += skb_shinfo(skb)->frags[i].size;
/* HW requires header must reside in 3 buffer */
if (unlikely(hdr_len > len)) {
if (skb_linearize(skb))
return 0;
}
}
if (!mss || ((skb->len - hdr_len) <= mss))
goto out;
if (mss != pdata->mss) {
pdata->mss = mss;
pdata->mac_ops->set_mss(pdata);
}
hopinfo |= SET_BIT(ET);
}
} else if (iph->protocol == IPPROTO_UDP) {
/**
* Initializes both "pipeline" and "in" using the data from "tuple", "skb", and the assumption that
* we're translating from 6 to 4.
* "pipeline" defines the sequence of functions that will be executed later and "in" is basically a
* summary of "skb".
*/
static bool init_pipeline_ipv6(struct pipeline *pipeline, struct packet_in *in,
struct nf_conntrack_tuple *tuple, struct sk_buff *skb)
{
struct ipv6hdr *ip6_hdr = ipv6_hdr(skb);
struct hdr_iterator iterator = HDR_ITERATOR_INIT(ip6_hdr);
pipeline->l3_hdr_function = create_ipv4_hdr;
pipeline->create_skb_function = create_skb;
pipeline->l3_post_function = post_ipv4;
in->packet = skb;
in->tuple = tuple;
in->l3_hdr = ip6_hdr;
in->l3_hdr_type = IPPROTO_IPV6;
in->l3_hdr_len = skb_transport_header(skb) - skb_network_header(skb);
in->l3_hdr_basic_len = sizeof(*ip6_hdr);
in->compute_l3_hdr_len = compute_ipv6_hdr_len;
hdr_iterator_last(&iterator);
if (iterator.hdr_type == NEXTHDR_AUTH || iterator.hdr_type == NEXTHDR_ESP) {
// RFC 6146 section 5.1.
log_warning(" Incoming IPv6 packet has an Auth header or an ESP header. Cannot translate; "
"will drop the packet.");
return false;
}
in->l4_hdr_type = iterator.hdr_type;
switch (in->l4_hdr_type) {
case NEXTHDR_TCP:
in->l4_hdr_len = tcp_hdrlen(skb);
pipeline->l4_post_function = post_tcp_ipv4;
pipeline->l4_hdr_and_payload_function = copy_l4_hdr_and_payload;
break;
case NEXTHDR_UDP:
in->l4_hdr_len = sizeof(struct udphdr);
pipeline->l4_hdr_and_payload_function = copy_l4_hdr_and_payload;
pipeline->l4_post_function = post_udp_ipv4;
break;
case NEXTHDR_ICMP:
in->l4_hdr_len = sizeof(struct icmp6hdr);
pipeline->l4_hdr_and_payload_function = create_icmp4_hdr_and_payload;
pipeline->l4_post_function = post_icmp4;
break;
default:
log_warning(" Unsupported l4 protocol (%d). Cannot translate.", in->l4_hdr_type);
return false;
}
in->payload = iterator.data + in->l4_hdr_len;
in->payload_len = be16_to_cpu(ip6_hdr->payload_len) //
- (in->l3_hdr_len - sizeof(*ip6_hdr)) //
- in->l4_hdr_len;
return true;
}
/*
* Return length of individual segments of a gso packet,
* including all headers (MAC, IP, TCP/UDP)
*/
static unsigned int skb_gso_seglen(const struct sk_buff *skb)
{
unsigned int hdr_len = skb_transport_header(skb) - skb_mac_header(skb);
const struct skb_shared_info *shinfo = skb_shinfo(skb);
if (likely(shinfo->gso_type & (SKB_GSO_TCPV4 | SKB_GSO_TCPV6)))
hdr_len += tcp_hdrlen(skb);
else
hdr_len += sizeof(struct udphdr);
return hdr_len + shinfo->gso_size;
}
static bool tcphdr_ok(struct sk_buff *skb)
{
int th_ofs = skb_transport_offset(skb);
int tcp_len;
if (unlikely(!pskb_may_pull(skb, th_ofs + sizeof(struct tcphdr))))
return false;
tcp_len = tcp_hdrlen(skb);
if (unlikely(tcp_len < sizeof(struct tcphdr) ||
skb->len < th_ofs + tcp_len))
return false;
return true;
}
static struct sk_buff *tls_sw_fallback(struct sock *sk, struct sk_buff *skb)
{
int tcp_payload_offset = skb_transport_offset(skb) + tcp_hdrlen(skb);
struct tls_context *tls_ctx = tls_get_ctx(sk);
struct tls_offload_context_tx *ctx = tls_offload_ctx_tx(tls_ctx);
int payload_len = skb->len - tcp_payload_offset;
struct scatterlist *sg_in, sg_out[3];
struct sk_buff *nskb = NULL;
int sg_in_max_elements;
int resync_sgs = 0;
s32 sync_size = 0;
u64 rcd_sn;
/* worst case is:
* MAX_SKB_FRAGS in tls_record_info
* MAX_SKB_FRAGS + 1 in SKB head and frags.
*/
sg_in_max_elements = 2 * MAX_SKB_FRAGS + 1;
if (!payload_len)
return skb;
sg_in = kmalloc_array(sg_in_max_elements, sizeof(*sg_in), GFP_ATOMIC);
if (!sg_in)
goto free_orig;
sg_init_table(sg_in, sg_in_max_elements);
sg_init_table(sg_out, ARRAY_SIZE(sg_out));
if (fill_sg_in(sg_in, skb, ctx, &rcd_sn, &sync_size, &resync_sgs)) {
/* bypass packets before kernel TLS socket option was set */
if (sync_size < 0 && payload_len <= -sync_size)
nskb = skb_get(skb);
goto put_sg;
}
nskb = tls_enc_skb(tls_ctx, sg_out, sg_in, skb, sync_size, rcd_sn);
put_sg:
while (resync_sgs)
put_page(sg_page(&sg_in[--resync_sgs]));
kfree(sg_in);
free_orig:
kfree_skb(skb);
return nskb;
}
/**
* Initializes both "pipeline" and "in" using the data from "tuple", "skb", and the assumption that
* we're translating from 4 to 6.
* "pipeline" defines the sequence of functions that will be executed later and "in" is basically a
* summary of "skb".
*/
static bool init_pipeline_ipv4(struct pipeline *pipeline, struct packet_in *in,
struct nf_conntrack_tuple *tuple, struct sk_buff *skb)
{
struct iphdr *ip4_hdr = ip_hdr(skb);
pipeline->l3_hdr_function = create_ipv6_hdr;
pipeline->create_skb_function = create_skb;
pipeline->l3_post_function = post_ipv6;
in->packet = skb;
in->tuple = tuple;
in->l3_hdr = ip4_hdr;
in->l3_hdr_type = IPPROTO_IP;
in->l3_hdr_len = skb_transport_header(skb) - skb_network_header(skb);
in->l3_hdr_basic_len = sizeof(*ip4_hdr);
in->compute_l3_hdr_len = compute_ipv4_hdr_len;
in->l4_hdr_type = ip4_hdr->protocol;
switch (in->l4_hdr_type) {
case IPPROTO_TCP:
in->l4_hdr_len = tcp_hdrlen(skb);
pipeline->l4_hdr_and_payload_function = copy_l4_hdr_and_payload;
pipeline->l4_post_function = post_tcp_ipv6;
break;
case IPPROTO_UDP:
in->l4_hdr_len = sizeof(struct udphdr);
pipeline->l4_hdr_and_payload_function = copy_l4_hdr_and_payload;
pipeline->l4_post_function = post_udp_ipv6;
break;
case IPPROTO_ICMP:
in->l4_hdr_len = sizeof(struct icmphdr);
pipeline->l4_hdr_and_payload_function = create_icmp6_hdr_and_payload;
pipeline->l4_post_function = post_icmp6;
break;
default:
log_warning(" Unsupported l4 protocol (%d). Cannot translate.", in->l4_hdr_type);
return false;
}
in->payload = skb_transport_header(skb) + in->l4_hdr_len;
in->payload_len = be16_to_cpu(ip4_hdr->tot_len) - in->l3_hdr_len - in->l4_hdr_len;
return true;
}
static int enic_get_skb_header(struct sk_buff *skb, void **iphdr,
void **tcph, u64 *hdr_flags, void *priv)
{
struct cq_enet_rq_desc *cq_desc = priv;
unsigned int ip_len;
struct iphdr *iph;
u8 type, color, eop, sop, ingress_port, vlan_stripped;
u8 fcoe, fcoe_sof, fcoe_fc_crc_ok, fcoe_enc_error, fcoe_eof;
u8 tcp_udp_csum_ok, udp, tcp, ipv4_csum_ok;
u8 ipv6, ipv4, ipv4_fragment, fcs_ok, rss_type, csum_not_calc;
u8 packet_error;
u16 q_number, completed_index, bytes_written, vlan, checksum;
u32 rss_hash;
cq_enet_rq_desc_dec(cq_desc,
&type, &color, &q_number, &completed_index,
&ingress_port, &fcoe, &eop, &sop, &rss_type,
&csum_not_calc, &rss_hash, &bytes_written,
&packet_error, &vlan_stripped, &vlan, &checksum,
&fcoe_sof, &fcoe_fc_crc_ok, &fcoe_enc_error,
&fcoe_eof, &tcp_udp_csum_ok, &udp, &tcp,
&ipv4_csum_ok, &ipv6, &ipv4, &ipv4_fragment,
&fcs_ok);
if (!(ipv4 && tcp && !ipv4_fragment))
return -1;
skb_reset_network_header(skb);
iph = ip_hdr(skb);
ip_len = ip_hdrlen(skb);
skb_set_transport_header(skb, ip_len);
if (ntohs(iph->tot_len) < ip_len + tcp_hdrlen(skb))
return -1;
*hdr_flags = LRO_IPV4 | LRO_TCP;
*tcph = tcp_hdr(skb);
*iphdr = iph;
return 0;
}
static int get_real_size(const struct sk_buff *skb,
const struct skb_shared_info *shinfo,
struct net_device *dev,
int *lso_header_size,
bool *inline_ok,
void **pfrag)
{
struct mlx4_en_priv *priv = netdev_priv(dev);
int real_size;
if (shinfo->gso_size) {
*inline_ok = false;
if (skb->encapsulation)
*lso_header_size = (skb_inner_transport_header(skb) - skb->data) + inner_tcp_hdrlen(skb);
else
*lso_header_size = skb_transport_offset(skb) + tcp_hdrlen(skb);
real_size = CTRL_SIZE + shinfo->nr_frags * DS_SIZE +
ALIGN(*lso_header_size + 4, DS_SIZE);
if (unlikely(*lso_header_size != skb_headlen(skb))) {
/* We add a segment for the skb linear buffer only if
* it contains data */
if (*lso_header_size < skb_headlen(skb))
real_size += DS_SIZE;
else {
if (netif_msg_tx_err(priv))
en_warn(priv, "Non-linear headers\n");
return 0;
}
}
} else {
*lso_header_size = 0;
*inline_ok = is_inline(priv->prof->inline_thold, skb,
shinfo, pfrag);
if (*inline_ok)
real_size = inline_size(skb);
else
real_size = CTRL_SIZE +
(shinfo->nr_frags + 1) * DS_SIZE;
}
return real_size;
}
void tso_start(struct sk_buff *skb, struct tso_t *tso)
{
int hdr_len = skb_transport_offset(skb) + tcp_hdrlen(skb);
tso->ip_id = ntohs(ip_hdr(skb)->id);
tso->tcp_seq = ntohl(tcp_hdr(skb)->seq);
tso->next_frag_idx = 0;
/* Build first data */
tso->size = skb_headlen(skb) - hdr_len;
tso->data = skb->data + hdr_len;
if ((tso->size == 0) &&
(tso->next_frag_idx < skb_shinfo(skb)->nr_frags)) {
skb_frag_t *frag = &skb_shinfo(skb)->frags[tso->next_frag_idx];
/* Move to next segment */
tso->size = frag->size;
tso->data = page_address(frag->page.p) + frag->page_offset;
tso->next_frag_idx++;
}
}
/* If an incoming SYN or SYNACK frame contains a payload and/or FIN,
* queue this additional data / FIN.
*/
void tcp_fastopen_add_skb(struct sock *sk, struct sk_buff *skb)
{
struct tcp_sock *tp = tcp_sk(sk);
if (TCP_SKB_CB(skb)->end_seq == tp->rcv_nxt)
return;
skb = skb_clone(skb, GFP_ATOMIC);
if (!skb)
return;
skb_dst_drop(skb);
/* segs_in has been initialized to 1 in tcp_create_openreq_child().
* Hence, reset segs_in to 0 before calling tcp_segs_in()
* to avoid double counting. Also, tcp_segs_in() expects
* skb->len to include the tcp_hdrlen. Hence, it should
* be called before __skb_pull().
*/
tp->segs_in = 0;
tcp_segs_in(tp, skb);
__skb_pull(skb, tcp_hdrlen(skb));
sk_forced_mem_schedule(sk, skb->truesize);
skb_set_owner_r(skb, sk);
TCP_SKB_CB(skb)->seq++;
TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_SYN;
tp->rcv_nxt = TCP_SKB_CB(skb)->end_seq;
__skb_queue_tail(&sk->sk_receive_queue, skb);
tp->syn_data_acked = 1;
/* u64_stats_update_begin(&tp->syncp) not needed here,
* as we certainly are not changing upper 32bit value (0)
*/
tp->bytes_received = skb->len;
if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN)
tcp_fin(sk);
}
void tso_build_hdr(struct sk_buff *skb, char *hdr, struct tso_t *tso,
int size, bool is_last)
{
struct iphdr *iph;
struct tcphdr *tcph;
int hdr_len = skb_transport_offset(skb) + tcp_hdrlen(skb);
int mac_hdr_len = skb_network_offset(skb);
memcpy(hdr, skb->data, hdr_len);
iph = (struct iphdr *)(hdr + mac_hdr_len);
iph->id = htons(tso->ip_id);
iph->tot_len = htons(size + hdr_len - mac_hdr_len);
tcph = (struct tcphdr *)(hdr + skb_transport_offset(skb));
tcph->seq = htonl(tso->tcp_seq);
tso->ip_id++;
if (!is_last) {
/* Clear all special flags for not last packet */
tcph->psh = 0;
tcph->fin = 0;
tcph->rst = 0;
}
}
static int iwl_pcie_gen2_build_amsdu(struct iwl_trans *trans,
struct sk_buff *skb,
struct iwl_tfh_tfd *tfd, int start_len,
u8 hdr_len, struct iwl_device_cmd *dev_cmd)
{
#ifdef CONFIG_INET
struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
struct iwl_tx_cmd_gen2 *tx_cmd = (void *)dev_cmd->payload;
struct ieee80211_hdr *hdr = (void *)skb->data;
unsigned int snap_ip_tcp_hdrlen, ip_hdrlen, total_len, hdr_room;
unsigned int mss = skb_shinfo(skb)->gso_size;
u16 length, iv_len, amsdu_pad;
u8 *start_hdr;
struct iwl_tso_hdr_page *hdr_page;
struct page **page_ptr;
struct tso_t tso;
/* if the packet is protected, then it must be CCMP or GCMP */
iv_len = ieee80211_has_protected(hdr->frame_control) ?
IEEE80211_CCMP_HDR_LEN : 0;
trace_iwlwifi_dev_tx(trans->dev, skb, tfd, sizeof(*tfd),
&dev_cmd->hdr, start_len, 0);
ip_hdrlen = skb_transport_header(skb) - skb_network_header(skb);
snap_ip_tcp_hdrlen = 8 + ip_hdrlen + tcp_hdrlen(skb);
total_len = skb->len - snap_ip_tcp_hdrlen - hdr_len - iv_len;
amsdu_pad = 0;
/* total amount of header we may need for this A-MSDU */
hdr_room = DIV_ROUND_UP(total_len, mss) *
(3 + snap_ip_tcp_hdrlen + sizeof(struct ethhdr)) + iv_len;
/* Our device supports 9 segments at most, it will fit in 1 page */
hdr_page = get_page_hdr(trans, hdr_room);
if (!hdr_page)
return -ENOMEM;
get_page(hdr_page->page);
start_hdr = hdr_page->pos;
page_ptr = (void *)((u8 *)skb->cb + trans_pcie->page_offs);
*page_ptr = hdr_page->page;
memcpy(hdr_page->pos, skb->data + hdr_len, iv_len);
hdr_page->pos += iv_len;
/*
* Pull the ieee80211 header + IV to be able to use TSO core,
* we will restore it for the tx_status flow.
*/
skb_pull(skb, hdr_len + iv_len);
/*
* Remove the length of all the headers that we don't actually
* have in the MPDU by themselves, but that we duplicate into
* all the different MSDUs inside the A-MSDU.
*/
le16_add_cpu(&tx_cmd->len, -snap_ip_tcp_hdrlen);
tso_start(skb, &tso);
while (total_len) {
/* this is the data left for this subframe */
unsigned int data_left = min_t(unsigned int, mss, total_len);
struct sk_buff *csum_skb = NULL;
unsigned int tb_len;
dma_addr_t tb_phys;
u8 *subf_hdrs_start = hdr_page->pos;
total_len -= data_left;
memset(hdr_page->pos, 0, amsdu_pad);
hdr_page->pos += amsdu_pad;
amsdu_pad = (4 - (sizeof(struct ethhdr) + snap_ip_tcp_hdrlen +
data_left)) & 0x3;
ether_addr_copy(hdr_page->pos, ieee80211_get_DA(hdr));
hdr_page->pos += ETH_ALEN;
ether_addr_copy(hdr_page->pos, ieee80211_get_SA(hdr));
hdr_page->pos += ETH_ALEN;
length = snap_ip_tcp_hdrlen + data_left;
*((__be16 *)hdr_page->pos) = cpu_to_be16(length);
hdr_page->pos += sizeof(length);
/*
* This will copy the SNAP as well which will be considered
* as MAC header.
*/
tso_build_hdr(skb, hdr_page->pos, &tso, data_left, !total_len);
hdr_page->pos += snap_ip_tcp_hdrlen;
tb_len = hdr_page->pos - start_hdr;
tb_phys = dma_map_single(trans->dev, start_hdr,
tb_len, DMA_TO_DEVICE);
if (unlikely(dma_mapping_error(trans->dev, tb_phys))) {
dev_kfree_skb(csum_skb);
goto out_err;
}
iwl_pcie_gen2_set_tb(trans, tfd, tb_phys, tb_len);
trace_iwlwifi_dev_tx_tb(trans->dev, skb, start_hdr, tb_len);
/* add this subframe's headers' length to the tx_cmd */
le16_add_cpu(&tx_cmd->len, hdr_page->pos - subf_hdrs_start);
/* prepare the start_hdr for the next subframe */
start_hdr = hdr_page->pos;
/* put the payload */
while (data_left) {
tb_len = min_t(unsigned int, tso.size, data_left);
tb_phys = dma_map_single(trans->dev, tso.data,
tb_len, DMA_TO_DEVICE);
if (unlikely(dma_mapping_error(trans->dev, tb_phys))) {
dev_kfree_skb(csum_skb);
goto out_err;
}
iwl_pcie_gen2_set_tb(trans, tfd, tb_phys, tb_len);
trace_iwlwifi_dev_tx_tb(trans->dev, skb, tso.data,
tb_len);
data_left -= tb_len;
tso_build_data(skb, &tso, tb_len);
}
}
/* re -add the WiFi header and IV */
skb_push(skb, hdr_len + iv_len);
return 0;
out_err:
#endif
return -EINVAL;
}
static bool tcp_fastopen_create_child(struct sock *sk,
struct sk_buff *skb,
struct dst_entry *dst,
struct request_sock *req)
{
struct tcp_sock *tp;
struct request_sock_queue *queue = &inet_csk(sk)->icsk_accept_queue;
struct sock *child;
u32 end_seq;
req->num_retrans = 0;
req->num_timeout = 0;
req->sk = NULL;
child = inet_csk(sk)->icsk_af_ops->syn_recv_sock(sk, skb, req, NULL);
if (!child)
return false;
spin_lock(&queue->fastopenq->lock);
queue->fastopenq->qlen++;
spin_unlock(&queue->fastopenq->lock);
/* Initialize the child socket. Have to fix some values to take
* into account the child is a Fast Open socket and is created
* only out of the bits carried in the SYN packet.
*/
tp = tcp_sk(child);
tp->fastopen_rsk = req;
tcp_rsk(req)->tfo_listener = true;
/* RFC1323: The window in SYN & SYN/ACK segments is never
* scaled. So correct it appropriately.
*/
tp->snd_wnd = ntohs(tcp_hdr(skb)->window);
/* Activate the retrans timer so that SYNACK can be retransmitted.
* The request socket is not added to the SYN table of the parent
* because it's been added to the accept queue directly.
*/
inet_csk_reset_xmit_timer(child, ICSK_TIME_RETRANS,
TCP_TIMEOUT_INIT, TCP_RTO_MAX);
atomic_set(&req->rsk_refcnt, 1);
/* Add the child socket directly into the accept queue */
inet_csk_reqsk_queue_add(sk, req, child);
/* Now finish processing the fastopen child socket. */
inet_csk(child)->icsk_af_ops->rebuild_header(child);
tcp_init_congestion_control(child);
tcp_mtup_init(child);
tcp_init_metrics(child);
tcp_init_buffer_space(child);
/* Queue the data carried in the SYN packet. We need to first
* bump skb's refcnt because the caller will attempt to free it.
* Note that IPv6 might also have used skb_get() trick
* in tcp_v6_conn_request() to keep this SYN around (treq->pktopts)
* So we need to eventually get a clone of the packet,
* before inserting it in sk_receive_queue.
*
* XXX (TFO) - we honor a zero-payload TFO request for now,
* (any reason not to?) but no need to queue the skb since
* there is no data. How about SYN+FIN?
*/
end_seq = TCP_SKB_CB(skb)->end_seq;
if (end_seq != TCP_SKB_CB(skb)->seq + 1) {
struct sk_buff *skb2;
if (unlikely(skb_shared(skb)))
skb2 = skb_clone(skb, GFP_ATOMIC);
else
skb2 = skb_get(skb);
if (likely(skb2)) {
skb_dst_drop(skb2);
__skb_pull(skb2, tcp_hdrlen(skb));
skb_set_owner_r(skb2, child);
__skb_queue_tail(&child->sk_receive_queue, skb2);
tp->syn_data_acked = 1;
/* u64_stats_update_begin(&tp->syncp) not needed here,
* as we certainly are not changing upper 32bit value (0)
*/
tp->bytes_received = end_seq - TCP_SKB_CB(skb)->seq - 1;
} else {
end_seq = TCP_SKB_CB(skb)->seq + 1;
}
}
tcp_rsk(req)->rcv_nxt = tp->rcv_nxt = end_seq;
sk->sk_data_ready(sk);
bh_unlock_sock(child);
sock_put(child);
WARN_ON(!req->sk);
return true;
}
static int fill_sg_in(struct scatterlist *sg_in,
struct sk_buff *skb,
struct tls_offload_context_tx *ctx,
u64 *rcd_sn,
s32 *sync_size,
int *resync_sgs)
{
int tcp_payload_offset = skb_transport_offset(skb) + tcp_hdrlen(skb);
int payload_len = skb->len - tcp_payload_offset;
u32 tcp_seq = ntohl(tcp_hdr(skb)->seq);
struct tls_record_info *record;
unsigned long flags;
int remaining;
int i;
spin_lock_irqsave(&ctx->lock, flags);
record = tls_get_record(ctx, tcp_seq, rcd_sn);
if (!record) {
spin_unlock_irqrestore(&ctx->lock, flags);
WARN(1, "Record not found for seq %u\n", tcp_seq);
return -EINVAL;
}
*sync_size = tcp_seq - tls_record_start_seq(record);
if (*sync_size < 0) {
int is_start_marker = tls_record_is_start_marker(record);
spin_unlock_irqrestore(&ctx->lock, flags);
/* This should only occur if the relevant record was
* already acked. In that case it should be ok
* to drop the packet and avoid retransmission.
*
* There is a corner case where the packet contains
* both an acked and a non-acked record.
* We currently don't handle that case and rely
* on TCP to retranmit a packet that doesn't contain
* already acked payload.
*/
if (!is_start_marker)
*sync_size = 0;
return -EINVAL;
}
remaining = *sync_size;
for (i = 0; remaining > 0; i++) {
skb_frag_t *frag = &record->frags[i];
__skb_frag_ref(frag);
sg_set_page(sg_in + i, skb_frag_page(frag),
skb_frag_size(frag), frag->page_offset);
remaining -= skb_frag_size(frag);
if (remaining < 0)
sg_in[i].length += remaining;
}
*resync_sgs = i;
spin_unlock_irqrestore(&ctx->lock, flags);
if (skb_to_sgvec(skb, &sg_in[i], tcp_payload_offset, payload_len) < 0)
return -EINVAL;
return 0;
}
static inline int
dma_xmit(struct sk_buff *skb, struct net_device *dev, END_DEVICE *ei_local, int gmac_no)
{
struct netdev_queue *txq;
dma_addr_t frag_addr;
u32 frag_size, nr_desc;
u32 txd_info3, txd_info4;
#if defined (CONFIG_RAETH_SG_DMA_TX)
u32 i, nr_frags;
const skb_frag_t *tx_frag;
const struct skb_shared_info *shinfo;
#else
#define nr_frags 0
#endif
#if defined (CONFIG_RA_HW_NAT) || defined (CONFIG_RA_HW_NAT_MODULE)
if (ra_sw_nat_hook_tx != NULL) {
#if defined (CONFIG_RA_HW_NAT_WIFI) || defined (CONFIG_RA_HW_NAT_PCI)
if (IS_DPORT_PPE_VALID(skb))
gmac_no = PSE_PORT_PPE;
else
#endif
if (ra_sw_nat_hook_tx(skb, gmac_no) == 0) {
dev_kfree_skb(skb);
return NETDEV_TX_OK;
}
}
#endif
txd_info3 = TX3_QDMA_SWC;
if (gmac_no != PSE_PORT_PPE) {
u32 QID = M2Q_table[(skb->mark & 0x3f)];
if (QID < 8 && M2Q_wan_lan) {
#if defined (CONFIG_PSEUDO_SUPPORT)
if (gmac_no == PSE_PORT_GMAC2)
QID += 8;
#elif defined (CONFIG_RAETH_HW_VLAN_TX)
if ((skb_vlan_tag_get(skb) & VLAN_VID_MASK) > 1)
QID += 8;
#endif
}
txd_info3 |= TX3_QDMA_QID(QID);
}
txd_info4 = TX4_DMA_FPORT(gmac_no);
#if defined (CONFIG_RAETH_CHECKSUM_OFFLOAD)
if (skb->ip_summed == CHECKSUM_PARTIAL)
txd_info4 |= TX4_DMA_TUI_CO(7);
#endif
#if defined (CONFIG_RAETH_HW_VLAN_TX)
if (skb_vlan_tag_present(skb))
txd_info4 |= (0x10000 | skb_vlan_tag_get(skb));
#endif
#if defined (CONFIG_RAETH_SG_DMA_TX)
shinfo = skb_shinfo(skb);
#endif
#if defined (CONFIG_RAETH_TSO)
/* fill MSS info in tcp checksum field */
if (shinfo->gso_size) {
u32 hdr_len;
if (!(shinfo->gso_type & (SKB_GSO_TCPV4|SKB_GSO_TCPV6))) {
dev_kfree_skb(skb);
return NETDEV_TX_OK;
}
if (skb_header_cloned(skb)) {
if (pskb_expand_head(skb, 0, 0, GFP_ATOMIC)) {
dev_kfree_skb(skb);
return NETDEV_TX_OK;
}
}
hdr_len = (skb_transport_offset(skb) + tcp_hdrlen(skb));
if (hdr_len >= skb->len) {
dev_kfree_skb(skb);
return NETDEV_TX_OK;
}
tcp_hdr(skb)->check = htons(shinfo->gso_size);
txd_info4 |= TX4_DMA_TSO;
}
#endif
nr_desc = DIV_ROUND_UP(skb_headlen(skb), TXD_MAX_SEG_SIZE);
#if defined (CONFIG_RAETH_SG_DMA_TX)
nr_frags = (u32)shinfo->nr_frags;
for (i = 0; i < nr_frags; i++) {
tx_frag = &shinfo->frags[i];
nr_desc += DIV_ROUND_UP(skb_frag_size(tx_frag), TXD_MAX_SEG_SIZE);
}
#endif
txq = netdev_get_tx_queue(dev, 0);
/* flush main skb part before spin_lock() */
frag_size = (u32)skb_headlen(skb);
frag_addr = dma_map_single(NULL, skb->data, frag_size, DMA_TO_DEVICE);
/* protect TX ring access (from eth2/eth3 queues) */
spin_lock(&ei_local->page_lock);
/* check nr_desc+2 free descriptors (2 need to prevent head/tail overlap) */
if (ei_local->txd_pool_free_num < (nr_desc+2)) {
spin_unlock(&ei_local->page_lock);
netif_tx_stop_queue(txq);
#if defined (CONFIG_RAETH_DEBUG)
if (net_ratelimit())
printk("%s: QDMA TX pool is run out! (GMAC: %d)\n", RAETH_DEV_NAME, gmac_no);
#endif
return NETDEV_TX_BUSY;
}
qdma_write_skb_fragment(ei_local, frag_addr, frag_size,
txd_info3, txd_info4, skb, nr_frags == 0);
#if defined (CONFIG_RAETH_SG_DMA_TX)
for (i = 0; i < nr_frags; i++) {
tx_frag = &shinfo->frags[i];
frag_size = skb_frag_size(tx_frag);
frag_addr = skb_frag_dma_map(NULL, tx_frag, 0, frag_size, DMA_TO_DEVICE);
qdma_write_skb_fragment(ei_local, frag_addr, frag_size,
txd_info3, txd_info4, skb, i == nr_frags - 1);
}
#endif
#if defined (CONFIG_RAETH_BQL)
netdev_tx_sent_queue(txq, skb->len);
#endif
#if !defined (CONFIG_RAETH_BQL) || !defined (CONFIG_SMP)
/* smp_mb() already inlined in netdev_tx_sent_queue */
wmb();
#endif
/* kick the QDMA TX */
sysRegWrite(QTX_CTX_PTR, (u32)get_txd_ptr_phy(ei_local, ei_local->txd_last_idx));
spin_unlock(&ei_local->page_lock);
return NETDEV_TX_OK;
}
static int xenvif_tx_submit(struct xenvif *vif)
{
struct gnttab_map_grant_ref *gop_map = vif->tx_map_ops;
struct gnttab_copy *gop_copy = vif->tx_copy_ops;
struct sk_buff *skb;
int work_done = 0;
while ((skb = __skb_dequeue(&vif->tx_queue)) != NULL) {
struct xen_netif_tx_request *txp;
u16 pending_idx;
unsigned data_len;
pending_idx = XENVIF_TX_CB(skb)->pending_idx;
txp = &vif->pending_tx_info[pending_idx].req;
/* Check the remap error code. */
if (unlikely(xenvif_tx_check_gop(vif, skb, &gop_map, &gop_copy))) {
skb_shinfo(skb)->nr_frags = 0;
kfree_skb(skb);
continue;
}
data_len = skb->len;
callback_param(vif, pending_idx).ctx = NULL;
if (data_len < txp->size) {
/* Append the packet payload as a fragment. */
txp->offset += data_len;
txp->size -= data_len;
} else {
/* Schedule a response immediately. */
xenvif_idx_release(vif, pending_idx,
XEN_NETIF_RSP_OKAY);
}
if (txp->flags & XEN_NETTXF_csum_blank)
skb->ip_summed = CHECKSUM_PARTIAL;
else if (txp->flags & XEN_NETTXF_data_validated)
skb->ip_summed = CHECKSUM_UNNECESSARY;
xenvif_fill_frags(vif, skb);
if (unlikely(skb_has_frag_list(skb))) {
if (xenvif_handle_frag_list(vif, skb)) {
if (net_ratelimit())
netdev_err(vif->dev,
"Not enough memory to consolidate frag_list!\n");
skb_shinfo(skb)->tx_flags |= SKBTX_DEV_ZEROCOPY;
kfree_skb(skb);
continue;
}
}
if (skb_is_nonlinear(skb) && skb_headlen(skb) < PKT_PROT_LEN) {
int target = min_t(int, skb->len, PKT_PROT_LEN);
__pskb_pull_tail(skb, target - skb_headlen(skb));
}
skb->dev = vif->dev;
skb->protocol = eth_type_trans(skb, skb->dev);
skb_reset_network_header(skb);
if (checksum_setup(vif, skb)) {
netdev_dbg(vif->dev,
"Can't setup checksum in net_tx_action\n");
/* We have to set this flag to trigger the callback */
if (skb_shinfo(skb)->destructor_arg)
skb_shinfo(skb)->tx_flags |= SKBTX_DEV_ZEROCOPY;
kfree_skb(skb);
continue;
}
skb_probe_transport_header(skb, 0);
/* If the packet is GSO then we will have just set up the
* transport header offset in checksum_setup so it's now
* straightforward to calculate gso_segs.
*/
if (skb_is_gso(skb)) {
int mss = skb_shinfo(skb)->gso_size;
int hdrlen = skb_transport_header(skb) -
skb_mac_header(skb) +
tcp_hdrlen(skb);
skb_shinfo(skb)->gso_segs =
DIV_ROUND_UP(skb->len - hdrlen, mss);
}
vif->dev->stats.rx_bytes += skb->len;
vif->dev->stats.rx_packets++;
work_done++;
/* Set this flag right before netif_receive_skb, otherwise
* someone might think this packet already left netback, and
* do a skb_copy_ubufs while we are still in control of the
* skb. E.g. the __pskb_pull_tail earlier can do such thing.
*/
if (skb_shinfo(skb)->destructor_arg) {
skb_shinfo(skb)->tx_flags |= SKBTX_DEV_ZEROCOPY;
vif->tx_zerocopy_sent++;
}
netif_receive_skb(skb);
}
static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_sq *sq, struct sk_buff *skb)
{
struct mlx5_wq_cyc *wq = &sq->wq;
u16 pi = sq->pc & wq->sz_m1;
struct mlx5e_tx_wqe *wqe = mlx5_wq_cyc_get_wqe(wq, pi);
struct mlx5e_tx_wqe_info *wi = &sq->wqe_info[pi];
struct mlx5_wqe_ctrl_seg *cseg = &wqe->ctrl;
struct mlx5_wqe_eth_seg *eseg = &wqe->eth;
struct mlx5_wqe_data_seg *dseg;
unsigned char *skb_data = skb->data;
unsigned int skb_len = skb->len;
u8 opcode = MLX5_OPCODE_SEND;
dma_addr_t dma_addr = 0;
unsigned int num_bytes;
bool bf = false;
u16 headlen;
u16 ds_cnt = sizeof(*wqe) / MLX5_SEND_WQE_DS;
u16 ihs;
int i;
memset(wqe, 0, sizeof(*wqe));
if (likely(skb->ip_summed == CHECKSUM_PARTIAL)) {
eseg->cs_flags = MLX5_ETH_WQE_L3_CSUM;
if (skb->encapsulation) {
eseg->cs_flags |= MLX5_ETH_WQE_L3_INNER_CSUM |
MLX5_ETH_WQE_L4_INNER_CSUM;
sq->stats.csum_partial_inner++;
} else {
eseg->cs_flags |= MLX5_ETH_WQE_L4_CSUM;
sq->stats.csum_partial++;
}
} else
sq->stats.csum_none++;
if (sq->cc != sq->prev_cc) {
sq->prev_cc = sq->cc;
sq->bf_budget = (sq->cc == sq->pc) ? MLX5E_SQ_BF_BUDGET : 0;
}
if (skb_is_gso(skb)) {
eseg->mss = cpu_to_be16(skb_shinfo(skb)->gso_size);
opcode = MLX5_OPCODE_LSO;
if (skb->encapsulation) {
ihs = skb_inner_transport_header(skb) - skb->data +
inner_tcp_hdrlen(skb);
sq->stats.tso_inner_packets++;
sq->stats.tso_inner_bytes += skb->len - ihs;
} else {
ihs = skb_transport_offset(skb) + tcp_hdrlen(skb);
sq->stats.tso_packets++;
sq->stats.tso_bytes += skb->len - ihs;
}
num_bytes = skb->len + (skb_shinfo(skb)->gso_segs - 1) * ihs;
} else {
bf = sq->bf_budget &&
!skb->xmit_more &&
!skb_shinfo(skb)->nr_frags;
ihs = mlx5e_get_inline_hdr_size(sq, skb, bf);
num_bytes = max_t(unsigned int, skb->len, ETH_ZLEN);
}
wi->num_bytes = num_bytes;
if (skb_vlan_tag_present(skb)) {
mlx5e_insert_vlan(eseg->inline_hdr_start, skb, ihs, &skb_data,
&skb_len);
ihs += VLAN_HLEN;
} else {
memcpy(eseg->inline_hdr_start, skb_data, ihs);
mlx5e_tx_skb_pull_inline(&skb_data, &skb_len, ihs);
}
eseg->inline_hdr_sz = cpu_to_be16(ihs);
ds_cnt += DIV_ROUND_UP(ihs - sizeof(eseg->inline_hdr_start),
MLX5_SEND_WQE_DS);
dseg = (struct mlx5_wqe_data_seg *)cseg + ds_cnt;
wi->num_dma = 0;
headlen = skb_len - skb->data_len;
if (headlen) {
dma_addr = dma_map_single(sq->pdev, skb_data, headlen,
DMA_TO_DEVICE);
if (unlikely(dma_mapping_error(sq->pdev, dma_addr)))
goto dma_unmap_wqe_err;
dseg->addr = cpu_to_be64(dma_addr);
dseg->lkey = sq->mkey_be;
dseg->byte_count = cpu_to_be32(headlen);
mlx5e_dma_push(sq, dma_addr, headlen, MLX5E_DMA_MAP_SINGLE);
wi->num_dma++;
dseg++;
}
for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[i];
int fsz = skb_frag_size(frag);
dma_addr = skb_frag_dma_map(sq->pdev, frag, 0, fsz,
DMA_TO_DEVICE);
if (unlikely(dma_mapping_error(sq->pdev, dma_addr)))
goto dma_unmap_wqe_err;
dseg->addr = cpu_to_be64(dma_addr);
dseg->lkey = sq->mkey_be;
dseg->byte_count = cpu_to_be32(fsz);
mlx5e_dma_push(sq, dma_addr, fsz, MLX5E_DMA_MAP_PAGE);
wi->num_dma++;
dseg++;
}
ds_cnt += wi->num_dma;
cseg->opmod_idx_opcode = cpu_to_be32((sq->pc << 8) | opcode);
cseg->qpn_ds = cpu_to_be32((sq->sqn << 8) | ds_cnt);
sq->skb[pi] = skb;
wi->num_wqebbs = DIV_ROUND_UP(ds_cnt, MLX5_SEND_WQEBB_NUM_DS);
sq->pc += wi->num_wqebbs;
if (unlikely(MLX5E_TX_HW_STAMP(sq->channel->priv, skb)))
skb_shinfo(skb)->tx_flags |= SKBTX_IN_PROGRESS;
netdev_tx_sent_queue(sq->txq, wi->num_bytes);
if (unlikely(!mlx5e_sq_has_room_for(sq, MLX5E_SQ_STOP_ROOM))) {
netif_tx_stop_queue(sq->txq);
sq->stats.queue_stopped++;
}
if (!skb->xmit_more || netif_xmit_stopped(sq->txq)) {
int bf_sz = 0;
if (bf && sq->uar_bf_map)
bf_sz = wi->num_wqebbs << 3;
cseg->fm_ce_se = MLX5_WQE_CTRL_CQ_UPDATE;
mlx5e_tx_notify_hw(sq, &wqe->ctrl, bf_sz);
}
sq->bf_budget = bf ? sq->bf_budget - 1 : 0;
/* fill sq edge with nops to avoid wqe wrap around */
while ((sq->pc & wq->sz_m1) > sq->edge)
mlx5e_send_nop(sq, false);
sq->stats.packets++;
sq->stats.bytes += num_bytes;
return NETDEV_TX_OK;
dma_unmap_wqe_err:
sq->stats.queue_dropped++;
mlx5e_dma_unmap_wqe_err(sq, wi->num_dma);
dev_kfree_skb_any(skb);
return NETDEV_TX_OK;
}
static inline int
dma_xmit(struct sk_buff* skb, struct net_device *dev, END_DEVICE *ei_local, int gmac_no)
{
struct netdev_queue *txq;
dma_addr_t frag_addr;
u32 frag_size, nr_desc;
u32 next_idx, desc_odd = 0;
u32 txd_info2 = 0, txd_info4;
#if defined (CONFIG_RAETH_SG_DMA_TX)
u32 i, nr_frags;
const skb_frag_t *tx_frag;
const struct skb_shared_info *shinfo;
#else
#define nr_frags 0
#endif
#if defined (CONFIG_RA_HW_NAT) || defined (CONFIG_RA_HW_NAT_MODULE)
if (ra_sw_nat_hook_tx != NULL) {
#if defined (CONFIG_RA_HW_NAT_WIFI) || defined (CONFIG_RA_HW_NAT_PCI)
if (IS_DPORT_PPE_VALID(skb))
gmac_no = PSE_PORT_PPE;
else
#endif
if (ra_sw_nat_hook_tx(skb, gmac_no) == 0) {
dev_kfree_skb(skb);
return NETDEV_TX_OK;
}
}
#endif
#if !defined (RAETH_HW_PADPKT)
if (skb->len < ei_local->min_pkt_len) {
if (skb_padto(skb, ei_local->min_pkt_len)) {
#if defined (CONFIG_RAETH_DEBUG)
if (net_ratelimit())
printk(KERN_ERR "%s: skb_padto failed\n", RAETH_DEV_NAME);
#endif
return NETDEV_TX_OK;
}
skb_put(skb, ei_local->min_pkt_len - skb->len);
}
#endif
#if defined (CONFIG_RALINK_MT7620)
if (gmac_no == PSE_PORT_PPE)
txd_info4 = TX4_DMA_FP_BMAP(0x80); /* P7 */
else
#if defined (CONFIG_RAETH_HAS_PORT5) && !defined (CONFIG_RAETH_HAS_PORT4) && !defined (CONFIG_RAETH_ESW)
txd_info4 = TX4_DMA_FP_BMAP(0x20); /* P5 */
#elif defined (CONFIG_RAETH_HAS_PORT4) && !defined (CONFIG_RAETH_HAS_PORT5) && !defined (CONFIG_RAETH_ESW)
txd_info4 = TX4_DMA_FP_BMAP(0x10); /* P4 */
#else
txd_info4 = 0; /* routing by DA */
#endif
#elif defined (CONFIG_RALINK_MT7621)
txd_info4 = TX4_DMA_FPORT(gmac_no);
#else
txd_info4 = (TX4_DMA_QN(3) | TX4_DMA_PN(gmac_no));
#endif
#if defined (CONFIG_RAETH_CHECKSUM_OFFLOAD) && !defined (RAETH_SDMA)
if (skb->ip_summed == CHECKSUM_PARTIAL)
txd_info4 |= TX4_DMA_TUI_CO(7);
#endif
#if defined (CONFIG_RAETH_HW_VLAN_TX)
if (skb_vlan_tag_present(skb)) {
#if defined (RAETH_HW_VLAN4K)
txd_info4 |= (0x10000 | skb_vlan_tag_get(skb));
#else
u32 vlan_tci = skb_vlan_tag_get(skb);
txd_info4 |= (TX4_DMA_INSV | TX4_DMA_VPRI(vlan_tci));
txd_info4 |= (u32)ei_local->vlan_4k_map[(vlan_tci & VLAN_VID_MASK)];
#endif
}
#endif
#if defined (CONFIG_RAETH_SG_DMA_TX)
shinfo = skb_shinfo(skb);
#endif
#if defined (CONFIG_RAETH_TSO)
/* fill MSS info in tcp checksum field */
if (shinfo->gso_size) {
u32 hdr_len;
if (!(shinfo->gso_type & (SKB_GSO_TCPV4|SKB_GSO_TCPV6))) {
dev_kfree_skb(skb);
return NETDEV_TX_OK;
}
if (skb_header_cloned(skb)) {
if (pskb_expand_head(skb, 0, 0, GFP_ATOMIC)) {
dev_kfree_skb(skb);
return NETDEV_TX_OK;
}
}
hdr_len = (skb_transport_offset(skb) + tcp_hdrlen(skb));
if (hdr_len >= skb->len) {
dev_kfree_skb(skb);
return NETDEV_TX_OK;
}
tcp_hdr(skb)->check = htons(shinfo->gso_size);
txd_info4 |= TX4_DMA_TSO;
}
#endif
nr_desc = DIV_ROUND_UP(skb_headlen(skb), TXD_MAX_SEG_SIZE);
#if defined (CONFIG_RAETH_SG_DMA_TX)
nr_frags = (u32)shinfo->nr_frags;
for (i = 0; i < nr_frags; i++) {
tx_frag = &shinfo->frags[i];
nr_desc += DIV_ROUND_UP(skb_frag_size(tx_frag), TXD_MAX_SEG_SIZE);
}
#endif
nr_desc = DIV_ROUND_UP(nr_desc, 2);
txq = netdev_get_tx_queue(dev, 0);
/* flush main skb part before spin_lock() */
frag_size = (u32)skb_headlen(skb);
frag_addr = dma_map_single(NULL, skb->data, frag_size, DMA_TO_DEVICE);
/* protect TX ring access (from eth2/eth3 queues) */
spin_lock(&ei_local->page_lock);
/* check nr_desc+1 free descriptors */
next_idx = (ei_local->txd_last_idx + nr_desc) % NUM_TX_DESC;
if (ei_local->txd_buff[ei_local->txd_last_idx] || ei_local->txd_buff[next_idx]) {
spin_unlock(&ei_local->page_lock);
netif_tx_stop_queue(txq);
#if defined (CONFIG_RAETH_DEBUG)
if (net_ratelimit())
printk("%s: PDMA TX ring is full! (GMAC: %d)\n", RAETH_DEV_NAME, gmac_no);
#endif
return NETDEV_TX_BUSY;
}
pdma_write_skb_fragment(ei_local, frag_addr, frag_size, &desc_odd,
&txd_info2, txd_info4, skb, nr_frags == 0);
#if defined (CONFIG_RAETH_SG_DMA_TX)
for (i = 0; i < nr_frags; i++) {
tx_frag = &shinfo->frags[i];
frag_size = skb_frag_size(tx_frag);
frag_addr = skb_frag_dma_map(NULL, tx_frag, 0, frag_size, DMA_TO_DEVICE);
pdma_write_skb_fragment(ei_local, frag_addr, frag_size, &desc_odd,
&txd_info2, txd_info4, skb, i == nr_frags - 1);
}
#endif
#if defined (CONFIG_RAETH_BQL)
netdev_tx_sent_queue(txq, skb->len);
#endif
#if !defined (CONFIG_RAETH_BQL) || !defined (CONFIG_SMP)
/* smp_mb() already inlined in netdev_tx_sent_queue */
wmb();
#endif
/* kick the DMA TX */
sysRegWrite(TX_CTX_IDX0, cpu_to_le32(ei_local->txd_last_idx));
spin_unlock(&ei_local->page_lock);
return NETDEV_TX_OK;
}
static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_sq *sq, struct sk_buff *skb)
{
struct mlx5_wq_cyc *wq = &sq->wq;
u16 pi = sq->pc & wq->sz_m1;
struct mlx5e_tx_wqe *wqe = mlx5_wq_cyc_get_wqe(wq, pi);
struct mlx5_wqe_ctrl_seg *cseg = &wqe->ctrl;
struct mlx5_wqe_eth_seg *eseg = &wqe->eth;
struct mlx5_wqe_data_seg *dseg;
u8 opcode = MLX5_OPCODE_SEND;
dma_addr_t dma_addr = 0;
bool bf = false;
u16 headlen;
u16 ds_cnt;
u16 ihs;
int i;
memset(wqe, 0, sizeof(*wqe));
if (likely(skb->ip_summed == CHECKSUM_PARTIAL))
eseg->cs_flags = MLX5_ETH_WQE_L3_CSUM | MLX5_ETH_WQE_L4_CSUM;
else
sq->stats.csum_offload_none++;
if (sq->cc != sq->prev_cc) {
sq->prev_cc = sq->cc;
sq->bf_budget = (sq->cc == sq->pc) ? MLX5E_SQ_BF_BUDGET : 0;
}
if (skb_is_gso(skb)) {
u32 payload_len;
eseg->mss = cpu_to_be16(skb_shinfo(skb)->gso_size);
opcode = MLX5_OPCODE_LSO;
ihs = skb_transport_offset(skb) + tcp_hdrlen(skb);
payload_len = skb->len - ihs;
MLX5E_TX_SKB_CB(skb)->num_bytes = skb->len +
(skb_shinfo(skb)->gso_segs - 1) * ihs;
sq->stats.tso_packets++;
sq->stats.tso_bytes += payload_len;
} else {
bf = sq->bf_budget &&
!skb->xmit_more &&
!skb_shinfo(skb)->nr_frags;
ihs = mlx5e_get_inline_hdr_size(sq, skb, bf);
MLX5E_TX_SKB_CB(skb)->num_bytes = max_t(unsigned int, skb->len,
ETH_ZLEN);
}
skb_copy_from_linear_data(skb, eseg->inline_hdr_start, ihs);
skb_pull_inline(skb, ihs);
eseg->inline_hdr_sz = cpu_to_be16(ihs);
ds_cnt = sizeof(*wqe) / MLX5_SEND_WQE_DS;
ds_cnt += DIV_ROUND_UP(ihs - sizeof(eseg->inline_hdr_start),
MLX5_SEND_WQE_DS);
dseg = (struct mlx5_wqe_data_seg *)cseg + ds_cnt;
MLX5E_TX_SKB_CB(skb)->num_dma = 0;
headlen = skb_headlen(skb);
if (headlen) {
dma_addr = dma_map_single(sq->pdev, skb->data, headlen,
DMA_TO_DEVICE);
if (unlikely(dma_mapping_error(sq->pdev, dma_addr)))
goto dma_unmap_wqe_err;
dseg->addr = cpu_to_be64(dma_addr);
dseg->lkey = sq->mkey_be;
dseg->byte_count = cpu_to_be32(headlen);
mlx5e_dma_push(sq, dma_addr, headlen);
MLX5E_TX_SKB_CB(skb)->num_dma++;
dseg++;
}
for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[i];
int fsz = skb_frag_size(frag);
dma_addr = skb_frag_dma_map(sq->pdev, frag, 0, fsz,
DMA_TO_DEVICE);
if (unlikely(dma_mapping_error(sq->pdev, dma_addr)))
goto dma_unmap_wqe_err;
dseg->addr = cpu_to_be64(dma_addr);
dseg->lkey = sq->mkey_be;
dseg->byte_count = cpu_to_be32(fsz);
mlx5e_dma_push(sq, dma_addr, fsz);
MLX5E_TX_SKB_CB(skb)->num_dma++;
dseg++;
}
ds_cnt += MLX5E_TX_SKB_CB(skb)->num_dma;
cseg->opmod_idx_opcode = cpu_to_be32((sq->pc << 8) | opcode);
cseg->qpn_ds = cpu_to_be32((sq->sqn << 8) | ds_cnt);
sq->skb[pi] = skb;
MLX5E_TX_SKB_CB(skb)->num_wqebbs = DIV_ROUND_UP(ds_cnt,
MLX5_SEND_WQEBB_NUM_DS);
sq->pc += MLX5E_TX_SKB_CB(skb)->num_wqebbs;
netdev_tx_sent_queue(sq->txq, MLX5E_TX_SKB_CB(skb)->num_bytes);
if (unlikely(!mlx5e_sq_has_room_for(sq, MLX5E_SQ_STOP_ROOM))) {
netif_tx_stop_queue(sq->txq);
sq->stats.stopped++;
}
if (!skb->xmit_more || netif_xmit_stopped(sq->txq)) {
int bf_sz = 0;
if (bf && sq->uar_bf_map)
bf_sz = MLX5E_TX_SKB_CB(skb)->num_wqebbs << 3;
cseg->fm_ce_se = MLX5_WQE_CTRL_CQ_UPDATE;
mlx5e_tx_notify_hw(sq, wqe, bf_sz);
}
/* fill sq edge with nops to avoid wqe wrap around */
while ((sq->pc & wq->sz_m1) > sq->edge)
mlx5e_send_nop(sq, false);
sq->bf_budget = bf ? sq->bf_budget - 1 : 0;
sq->stats.packets++;
return NETDEV_TX_OK;
dma_unmap_wqe_err:
sq->stats.dropped++;
mlx5e_dma_unmap_wqe_err(sq, skb);
dev_kfree_skb_any(skb);
return NETDEV_TX_OK;
}
static struct sk_buff *tls_enc_skb(struct tls_context *tls_ctx,
struct scatterlist sg_out[3],
struct scatterlist *sg_in,
struct sk_buff *skb,
s32 sync_size, u64 rcd_sn)
{
int tcp_payload_offset = skb_transport_offset(skb) + tcp_hdrlen(skb);
struct tls_offload_context_tx *ctx = tls_offload_ctx_tx(tls_ctx);
int payload_len = skb->len - tcp_payload_offset;
void *buf, *iv, *aad, *dummy_buf;
struct aead_request *aead_req;
struct sk_buff *nskb = NULL;
int buf_len;
aead_req = tls_alloc_aead_request(ctx->aead_send, GFP_ATOMIC);
if (!aead_req)
return NULL;
buf_len = TLS_CIPHER_AES_GCM_128_SALT_SIZE +
TLS_CIPHER_AES_GCM_128_IV_SIZE +
TLS_AAD_SPACE_SIZE +
sync_size +
TLS_CIPHER_AES_GCM_128_TAG_SIZE;
buf = kmalloc(buf_len, GFP_ATOMIC);
if (!buf)
goto free_req;
iv = buf;
memcpy(iv, tls_ctx->crypto_send.aes_gcm_128.salt,
TLS_CIPHER_AES_GCM_128_SALT_SIZE);
aad = buf + TLS_CIPHER_AES_GCM_128_SALT_SIZE +
TLS_CIPHER_AES_GCM_128_IV_SIZE;
dummy_buf = aad + TLS_AAD_SPACE_SIZE;
nskb = alloc_skb(skb_headroom(skb) + skb->len, GFP_ATOMIC);
if (!nskb)
goto free_buf;
skb_reserve(nskb, skb_headroom(skb));
fill_sg_out(sg_out, buf, tls_ctx, nskb, tcp_payload_offset,
payload_len, sync_size, dummy_buf);
if (tls_enc_records(aead_req, ctx->aead_send, sg_in, sg_out, aad, iv,
rcd_sn, sync_size + payload_len) < 0)
goto free_nskb;
complete_skb(nskb, skb, tcp_payload_offset);
/* validate_xmit_skb_list assumes that if the skb wasn't segmented
* nskb->prev will point to the skb itself
*/
nskb->prev = nskb;
free_buf:
kfree(buf);
free_req:
kfree(aead_req);
return nskb;
free_nskb:
kfree_skb(nskb);
nskb = NULL;
goto free_buf;
}
static int xenvif_tx_submit(struct xenvif *vif)
{
struct gnttab_copy *gop = vif->tx_copy_ops;
struct sk_buff *skb;
int work_done = 0;
while ((skb = __skb_dequeue(&vif->tx_queue)) != NULL) {
struct xen_netif_tx_request *txp;
u16 pending_idx;
unsigned data_len;
pending_idx = *((u16 *)skb->data);
txp = &vif->pending_tx_info[pending_idx].req;
/* Check the remap error code. */
if (unlikely(xenvif_tx_check_gop(vif, skb, &gop))) {
netdev_dbg(vif->dev, "netback grant failed.\n");
skb_shinfo(skb)->nr_frags = 0;
kfree_skb(skb);
continue;
}
data_len = skb->len;
memcpy(skb->data,
(void *)(idx_to_kaddr(vif, pending_idx)|txp->offset),
data_len);
if (data_len < txp->size) {
/* Append the packet payload as a fragment. */
txp->offset += data_len;
txp->size -= data_len;
} else {
/* Schedule a response immediately. */
xenvif_idx_release(vif, pending_idx,
XEN_NETIF_RSP_OKAY);
}
if (txp->flags & XEN_NETTXF_csum_blank)
skb->ip_summed = CHECKSUM_PARTIAL;
else if (txp->flags & XEN_NETTXF_data_validated)
skb->ip_summed = CHECKSUM_UNNECESSARY;
xenvif_fill_frags(vif, skb);
if (skb_is_nonlinear(skb) && skb_headlen(skb) < PKT_PROT_LEN) {
int target = min_t(int, skb->len, PKT_PROT_LEN);
__pskb_pull_tail(skb, target - skb_headlen(skb));
}
skb->dev = vif->dev;
skb->protocol = eth_type_trans(skb, skb->dev);
skb_reset_network_header(skb);
if (checksum_setup(vif, skb)) {
netdev_dbg(vif->dev,
"Can't setup checksum in net_tx_action\n");
kfree_skb(skb);
continue;
}
skb_probe_transport_header(skb, 0);
/* If the packet is GSO then we will have just set up the
* transport header offset in checksum_setup so it's now
* straightforward to calculate gso_segs.
*/
if (skb_is_gso(skb)) {
int mss = skb_shinfo(skb)->gso_size;
int hdrlen = skb_transport_header(skb) -
skb_mac_header(skb) +
tcp_hdrlen(skb);
skb_shinfo(skb)->gso_segs =
DIV_ROUND_UP(skb->len - hdrlen, mss);
}
vif->dev->stats.rx_bytes += skb->len;
vif->dev->stats.rx_packets++;
work_done++;
netif_receive_skb(skb);
}
static inline void enic_queue_wq_skb_tso(struct enic *enic,
struct vnic_wq *wq, struct sk_buff *skb, unsigned int mss,
int vlan_tag_insert, unsigned int vlan_tag)
{
unsigned int frag_len_left = skb_headlen(skb);
unsigned int len_left = skb->len - frag_len_left;
unsigned int hdr_len = skb_transport_offset(skb) + tcp_hdrlen(skb);
int eop = (len_left == 0);
unsigned int len;
dma_addr_t dma_addr;
unsigned int offset = 0;
skb_frag_t *frag;
if (skb->protocol == cpu_to_be16(ETH_P_IP)) {
ip_hdr(skb)->check = 0;
tcp_hdr(skb)->check = ~csum_tcpudp_magic(ip_hdr(skb)->saddr,
ip_hdr(skb)->daddr, 0, IPPROTO_TCP, 0);
} else if (skb->protocol == cpu_to_be16(ETH_P_IPV6)) {
tcp_hdr(skb)->check = ~csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
&ipv6_hdr(skb)->daddr, 0, IPPROTO_TCP, 0);
}
while (frag_len_left) {
len = min(frag_len_left, (unsigned int)WQ_ENET_MAX_DESC_LEN);
dma_addr = pci_map_single(enic->pdev, skb->data + offset,
len, PCI_DMA_TODEVICE);
enic_queue_wq_desc_tso(wq, skb,
dma_addr,
len,
mss, hdr_len,
vlan_tag_insert, vlan_tag,
eop && (len == frag_len_left));
frag_len_left -= len;
offset += len;
}
if (eop)
return;
for (frag = skb_shinfo(skb)->frags; len_left; frag++) {
len_left -= frag->size;
frag_len_left = frag->size;
offset = frag->page_offset;
while (frag_len_left) {
len = min(frag_len_left,
(unsigned int)WQ_ENET_MAX_DESC_LEN);
dma_addr = pci_map_page(enic->pdev, frag->page,
offset, len,
PCI_DMA_TODEVICE);
enic_queue_wq_desc_cont(wq, skb,
dma_addr,
len,
(len_left == 0) &&
(len == frag_len_left));
frag_len_left -= len;
offset += len;
}
}
}