| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-Id: <20180326194619.1202-11-jeffrey.t.kirsher@intel.com>
Date: Mon, 26 Mar 2018 12:46:14 -0700
From: Jeff Kirsher <jeffrey.t.kirsher@...el.com>
To: davem@...emloft.net
Cc: Anirudh Venkataramanan <anirudh.venkataramanan@...el.com>,
netdev@...r.kernel.org, nhorman@...hat.com, sassmann@...hat.com,
jogreene@...hat.com, tbogendoerfer@...e.com, bpoirier@...e.com,
Jeff Kirsher <jeffrey.t.kirsher@...el.com>
Subject: [net-next 10/15] ice: Implement transmit and NAPI support
From: Anirudh Venkataramanan <anirudh.venkataramanan@...el.com>
This patch implements ice_start_xmit (the handler for ndo_start_xmit) and
related functions. ice_start_xmit ultimately calls ice_tx_map, where the
Tx descriptor is built and posted to the hardware by bumping the ring tail.
This patch also implements ice_napi_poll, which is invoked when there's an
interrupt on the VSI's queues. The interrupt can be due to either a
completed Tx or an Rx event. In case of a completed Tx/Rx event, resources
are reclaimed. Additionally, in case of an Rx event, the skb is fetched
and passed up to the network stack.
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@...el.com>
Tested-by: Tony Brelinski <tonyx.brelinski@...el.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@...el.com>
---
drivers/net/ethernet/intel/ice/ice.h | 1 +
drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h | 46 ++
drivers/net/ethernet/intel/ice/ice_main.c | 55 ++
drivers/net/ethernet/intel/ice/ice_txrx.c | 1026 +++++++++++++++++++++++-
drivers/net/ethernet/intel/ice/ice_txrx.h | 45 ++
5 files changed, 1171 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h
index 46cf82212ed8..6f01395a7f34 100644
--- a/drivers/net/ethernet/intel/ice/ice.h
+++ b/drivers/net/ethernet/intel/ice/ice.h
@@ -60,6 +60,7 @@
(((val) << ICE_AQ_VSI_UP_TABLE_UP##i##_S) & \
ICE_AQ_VSI_UP_TABLE_UP##i##_M)
+#define ICE_TX_DESC(R, i) (&(((struct ice_tx_desc *)((R)->desc))[i]))
#define ICE_RX_DESC(R, i) (&(((union ice_32b_rx_flex_desc *)((R)->desc))[i]))
#define ice_for_each_txq(vsi, i) \
diff --git a/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h b/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h
index 3c1aeb74d950..e1221b9839d5 100644
--- a/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h
+++ b/drivers/net/ethernet/intel/ice/ice_lan_tx_rx.h
@@ -131,6 +131,33 @@ enum ice_rx_flg64_bits {
ICE_RXFLG_RSVD = 63
};
+/* for ice_32byte_rx_flex_desc.ptype_flexi_flags0 member */
+#define ICE_RX_FLEX_DESC_PTYPE_M (0x3FF) /* 10-bits */
+
+/* for ice_32byte_rx_flex_desc.pkt_length member */
+#define ICE_RX_FLX_DESC_PKT_LEN_M (0x3FFF) /* 14-bits */
+
+enum ice_rx_flex_desc_status_error_0_bits {
+ /* Note: These are predefined bit offsets */
+ ICE_RX_FLEX_DESC_STATUS0_DD_S = 0,
+ ICE_RX_FLEX_DESC_STATUS0_EOF_S,
+ ICE_RX_FLEX_DESC_STATUS0_HBO_S,
+ ICE_RX_FLEX_DESC_STATUS0_L3L4P_S,
+ ICE_RX_FLEX_DESC_STATUS0_XSUM_IPE_S,
+ ICE_RX_FLEX_DESC_STATUS0_XSUM_L4E_S,
+ ICE_RX_FLEX_DESC_STATUS0_XSUM_EIPE_S,
+ ICE_RX_FLEX_DESC_STATUS0_XSUM_EUDPE_S,
+ ICE_RX_FLEX_DESC_STATUS0_LPBK_S,
+ ICE_RX_FLEX_DESC_STATUS0_IPV6EXADD_S,
+ ICE_RX_FLEX_DESC_STATUS0_RXE_S,
+ ICE_RX_FLEX_DESC_STATUS0_CRCP_S,
+ ICE_RX_FLEX_DESC_STATUS0_RSS_VALID_S,
+ ICE_RX_FLEX_DESC_STATUS0_L2TAG1P_S,
+ ICE_RX_FLEX_DESC_STATUS0_XTRMD0_VALID_S,
+ ICE_RX_FLEX_DESC_STATUS0_XTRMD1_VALID_S,
+ ICE_RX_FLEX_DESC_STATUS0_LAST /* this entry must be last!!! */
+};
+
#define ICE_RXQ_CTX_SIZE_DWORDS 8
#define ICE_RXQ_CTX_SZ (ICE_RXQ_CTX_SIZE_DWORDS * sizeof(u32))
@@ -201,6 +228,25 @@ struct ice_tx_desc {
__le64 cmd_type_offset_bsz;
};
+enum ice_tx_desc_dtype_value {
+ ICE_TX_DESC_DTYPE_DATA = 0x0,
+ ICE_TX_DESC_DTYPE_CTX = 0x1,
+ /* DESC_DONE - HW has completed write-back of descriptor */
+ ICE_TX_DESC_DTYPE_DESC_DONE = 0xF,
+};
+
+#define ICE_TXD_QW1_CMD_S 4
+#define ICE_TXD_QW1_CMD_M (0xFFFUL << ICE_TXD_QW1_CMD_S)
+
+enum ice_tx_desc_cmd_bits {
+ ICE_TX_DESC_CMD_EOP = 0x0001,
+ ICE_TX_DESC_CMD_RS = 0x0002,
+};
+
+#define ICE_TXD_QW1_OFFSET_S 16
+#define ICE_TXD_QW1_TX_BUF_SZ_S 34
+#define ICE_TXD_QW1_L2TAG1_S 48
+
#define ICE_LAN_TXQ_MAX_QGRPS 127
#define ICE_LAN_TXQ_MAX_QDIS 1023
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 58e9eb40f64a..dd07bf862afe 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -1258,6 +1258,23 @@ static int ice_vsi_alloc_arrays(struct ice_vsi *vsi, bool alloc_qvectors)
return -ENOMEM;
}
+/**
+ * ice_msix_clean_rings - MSIX mode Interrupt Handler
+ * @irq: interrupt number
+ * @data: pointer to a q_vector
+ */
+static irqreturn_t ice_msix_clean_rings(int __always_unused irq, void *data)
+{
+ struct ice_q_vector *q_vector = (struct ice_q_vector *)data;
+
+ if (!q_vector->tx.ring && !q_vector->rx.ring)
+ return IRQ_HANDLED;
+
+ napi_schedule(&q_vector->napi);
+
+ return IRQ_HANDLED;
+}
+
/**
* ice_vsi_alloc - Allocates the next available struct vsi in the PF
* @pf: board private structure
@@ -1298,6 +1315,8 @@ static struct ice_vsi *ice_vsi_alloc(struct ice_pf *pf, enum ice_vsi_type type)
if (ice_vsi_alloc_arrays(vsi, true))
goto err_rings;
+ /* Setup default MSIX irq handler for VSI */
+ vsi->irq_handler = ice_msix_clean_rings;
break;
default:
dev_warn(&pf->pdev->dev, "Unknown VSI type %d\n", vsi->type);
@@ -1741,6 +1760,9 @@ static int ice_vsi_alloc_q_vector(struct ice_vsi *vsi, int v_idx)
if (cpu_online(v_idx))
cpumask_set_cpu(v_idx, &q_vector->affinity_mask);
+ if (vsi->netdev)
+ netif_napi_add(vsi->netdev, &q_vector->napi, ice_napi_poll,
+ NAPI_POLL_WEIGHT);
/* tie q_vector and vsi together */
vsi->q_vectors[v_idx] = q_vector;
@@ -2914,6 +2936,21 @@ static int ice_vsi_stop_tx_rx_rings(struct ice_vsi *vsi)
return 0;
}
+/**
+ * ice_napi_enable_all - Enable NAPI for all q_vectors in the VSI
+ * @vsi: the VSI being configured
+ */
+static void ice_napi_enable_all(struct ice_vsi *vsi)
+{
+ int q_idx;
+
+ if (!vsi->netdev)
+ return;
+
+ for (q_idx = 0; q_idx < vsi->num_q_vectors; q_idx++)
+ napi_enable(&vsi->q_vectors[q_idx]->napi);
+}
+
/**
* ice_up_complete - Finish the last steps of bringing up a connection
* @vsi: The VSI being configured
@@ -2939,6 +2976,7 @@ static int ice_up_complete(struct ice_vsi *vsi)
return err;
clear_bit(__ICE_DOWN, vsi->state);
+ ice_napi_enable_all(vsi);
ice_vsi_ena_irq(vsi);
if (vsi->port_info &&
@@ -2954,6 +2992,21 @@ static int ice_up_complete(struct ice_vsi *vsi)
return err;
}
+/**
+ * ice_napi_disable_all - Disable NAPI for all q_vectors in the VSI
+ * @vsi: VSI having NAPI disabled
+ */
+static void ice_napi_disable_all(struct ice_vsi *vsi)
+{
+ int q_idx;
+
+ if (!vsi->netdev)
+ return;
+
+ for (q_idx = 0; q_idx < vsi->num_q_vectors; q_idx++)
+ napi_disable(&vsi->q_vectors[q_idx]->napi);
+}
+
/**
* ice_down - Shutdown the connection
* @vsi: The VSI being stopped
@@ -2972,6 +3025,7 @@ static int ice_down(struct ice_vsi *vsi)
ice_vsi_dis_irq(vsi);
err = ice_vsi_stop_tx_rx_rings(vsi);
+ ice_napi_disable_all(vsi);
ice_for_each_txq(vsi, i)
ice_clean_tx_ring(vsi->tx_rings[i]);
@@ -3251,4 +3305,5 @@ static int ice_stop(struct net_device *netdev)
static const struct net_device_ops ice_netdev_ops = {
.ndo_open = ice_open,
.ndo_stop = ice_stop,
+ .ndo_start_xmit = ice_start_xmit,
};
diff --git a/drivers/net/ethernet/intel/ice/ice_txrx.c b/drivers/net/ethernet/intel/ice/ice_txrx.c
index 6190ea30ee01..1ccf8e69b85a 100644
--- a/drivers/net/ethernet/intel/ice/ice_txrx.c
+++ b/drivers/net/ethernet/intel/ice/ice_txrx.c
@@ -7,6 +7,8 @@
#include <linux/mm.h>
#include "ice.h"
+#define ICE_RX_HDR_SIZE 256
+
/**
* ice_unmap_and_free_tx_buf - Release a Tx buffer
* @ring: the ring that owns the buffer
@@ -92,6 +94,129 @@ void ice_free_tx_ring(struct ice_ring *tx_ring)
}
}
+/**
+ * ice_clean_tx_irq - Reclaim resources after transmit completes
+ * @vsi: the VSI we care about
+ * @tx_ring: Tx ring to clean
+ * @napi_budget: Used to determine if we are in netpoll
+ *
+ * Returns true if there's any budget left (e.g. the clean is finished)
+ */
+static bool ice_clean_tx_irq(struct ice_vsi *vsi, struct ice_ring *tx_ring,
+ int napi_budget)
+{
+ unsigned int total_bytes = 0, total_pkts = 0;
+ unsigned int budget = vsi->work_lmt;
+ s16 i = tx_ring->next_to_clean;
+ struct ice_tx_desc *tx_desc;
+ struct ice_tx_buf *tx_buf;
+
+ tx_buf = &tx_ring->tx_buf[i];
+ tx_desc = ICE_TX_DESC(tx_ring, i);
+ i -= tx_ring->count;
+
+ do {
+ struct ice_tx_desc *eop_desc = tx_buf->next_to_watch;
+
+ /* if next_to_watch is not set then there is no work pending */
+ if (!eop_desc)
+ break;
+
+ smp_rmb(); /* prevent any other reads prior to eop_desc */
+
+ /* if the descriptor isn't done, no work yet to do */
+ if (!(eop_desc->cmd_type_offset_bsz &
+ cpu_to_le64(ICE_TX_DESC_DTYPE_DESC_DONE)))
+ break;
+
+ /* clear next_to_watch to prevent false hangs */
+ tx_buf->next_to_watch = NULL;
+
+ /* update the statistics for this packet */
+ total_bytes += tx_buf->bytecount;
+ total_pkts += tx_buf->gso_segs;
+
+ /* free the skb */
+ napi_consume_skb(tx_buf->skb, napi_budget);
+
+ /* unmap skb header data */
+ dma_unmap_single(tx_ring->dev,
+ dma_unmap_addr(tx_buf, dma),
+ dma_unmap_len(tx_buf, len),
+ DMA_TO_DEVICE);
+
+ /* clear tx_buf data */
+ tx_buf->skb = NULL;
+ dma_unmap_len_set(tx_buf, len, 0);
+
+ /* unmap remaining buffers */
+ while (tx_desc != eop_desc) {
+ tx_buf++;
+ tx_desc++;
+ i++;
+ if (unlikely(!i)) {
+ i -= tx_ring->count;
+ tx_buf = tx_ring->tx_buf;
+ tx_desc = ICE_TX_DESC(tx_ring, 0);
+ }
+
+ /* unmap any remaining paged data */
+ if (dma_unmap_len(tx_buf, len)) {
+ dma_unmap_page(tx_ring->dev,
+ dma_unmap_addr(tx_buf, dma),
+ dma_unmap_len(tx_buf, len),
+ DMA_TO_DEVICE);
+ dma_unmap_len_set(tx_buf, len, 0);
+ }
+ }
+
+ /* move us one more past the eop_desc for start of next pkt */
+ tx_buf++;
+ tx_desc++;
+ i++;
+ if (unlikely(!i)) {
+ i -= tx_ring->count;
+ tx_buf = tx_ring->tx_buf;
+ tx_desc = ICE_TX_DESC(tx_ring, 0);
+ }
+
+ prefetch(tx_desc);
+
+ /* update budget accounting */
+ budget--;
+ } while (likely(budget));
+
+ i += tx_ring->count;
+ tx_ring->next_to_clean = i;
+ u64_stats_update_begin(&tx_ring->syncp);
+ tx_ring->stats.bytes += total_bytes;
+ tx_ring->stats.pkts += total_pkts;
+ u64_stats_update_end(&tx_ring->syncp);
+ tx_ring->q_vector->tx.total_bytes += total_bytes;
+ tx_ring->q_vector->tx.total_pkts += total_pkts;
+
+ netdev_tx_completed_queue(txring_txq(tx_ring), total_pkts,
+ total_bytes);
+
+#define TX_WAKE_THRESHOLD ((s16)(DESC_NEEDED * 2))
+ if (unlikely(total_pkts && netif_carrier_ok(tx_ring->netdev) &&
+ (ICE_DESC_UNUSED(tx_ring) >= TX_WAKE_THRESHOLD))) {
+ /* Make sure that anybody stopping the queue after this
+ * sees the new next_to_clean.
+ */
+ smp_mb();
+ if (__netif_subqueue_stopped(tx_ring->netdev,
+ tx_ring->q_index) &&
+ !test_bit(__ICE_DOWN, vsi->state)) {
+ netif_wake_subqueue(tx_ring->netdev,
+ tx_ring->q_index);
+ ++tx_ring->tx_stats.restart_q;
+ }
+ }
+
+ return !!budget;
+}
+
/**
* ice_setup_tx_ring - Allocate the Tx descriptors
* @tx_ring: the tx ring to set up
@@ -274,13 +399,17 @@ static bool ice_alloc_mapped_page(struct ice_ring *rx_ring,
dma_addr_t dma;
/* since we are recycling buffers we should seldom need to alloc */
- if (likely(page))
+ if (likely(page)) {
+ rx_ring->rx_stats.page_reuse_count++;
return true;
+ }
/* alloc new page for storage */
page = alloc_page(GFP_ATOMIC | __GFP_NOWARN);
- if (unlikely(!page))
+ if (unlikely(!page)) {
+ rx_ring->rx_stats.alloc_page_failed++;
return false;
+ }
/* map page for use */
dma = dma_map_page(rx_ring->dev, page, 0, PAGE_SIZE, DMA_FROM_DEVICE);
@@ -290,6 +419,7 @@ static bool ice_alloc_mapped_page(struct ice_ring *rx_ring,
*/
if (dma_mapping_error(rx_ring->dev, dma)) {
__free_pages(page, 0);
+ rx_ring->rx_stats.alloc_page_failed++;
return false;
}
@@ -359,3 +489,895 @@ bool ice_alloc_rx_bufs(struct ice_ring *rx_ring, u16 cleaned_count)
*/
return true;
}
+
+/**
+ * ice_page_is_reserved - check if reuse is possible
+ * @page: page struct to check
+ */
+static bool ice_page_is_reserved(struct page *page)
+{
+ return (page_to_nid(page) != numa_mem_id()) || page_is_pfmemalloc(page);
+}
+
+/**
+ * ice_add_rx_frag - Add contents of Rx buffer to sk_buff
+ * @rx_buf: buffer containing page to add
+ * @rx_desc: descriptor containing length of buffer written by hardware
+ * @skb: sk_buf to place the data into
+ *
+ * This function will add the data contained in rx_buf->page to the skb.
+ * This is done either through a direct copy if the data in the buffer is
+ * less than the skb header size, otherwise it will just attach the page as
+ * a frag to the skb.
+ *
+ * The function will then update the page offset if necessary and return
+ * true if the buffer can be reused by the adapter.
+ */
+static bool ice_add_rx_frag(struct ice_rx_buf *rx_buf,
+ union ice_32b_rx_flex_desc *rx_desc,
+ struct sk_buff *skb)
+{
+#if (PAGE_SIZE < 8192)
+ unsigned int truesize = ICE_RXBUF_2048;
+#else
+ unsigned int last_offset = PAGE_SIZE - ICE_RXBUF_2048;
+ unsigned int truesize;
+#endif /* PAGE_SIZE < 8192) */
+
+ struct page *page;
+ unsigned int size;
+
+ size = le16_to_cpu(rx_desc->wb.pkt_len) &
+ ICE_RX_FLX_DESC_PKT_LEN_M;
+
+ page = rx_buf->page;
+
+#if (PAGE_SIZE >= 8192)
+ truesize = ALIGN(size, L1_CACHE_BYTES);
+#endif /* PAGE_SIZE >= 8192) */
+
+ /* will the data fit in the skb we allocated? if so, just
+ * copy it as it is pretty small anyway
+ */
+ if (size <= ICE_RX_HDR_SIZE && !skb_is_nonlinear(skb)) {
+ unsigned char *va = page_address(page) + rx_buf->page_offset;
+
+ memcpy(__skb_put(skb, size), va, ALIGN(size, sizeof(long)));
+
+ /* page is not reserved, we can reuse buffer as-is */
+ if (likely(!ice_page_is_reserved(page)))
+ return true;
+
+ /* this page cannot be reused so discard it */
+ __free_pages(page, 0);
+ return false;
+ }
+
+ skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, page,
+ rx_buf->page_offset, size, truesize);
+
+ /* avoid re-using remote pages */
+ if (unlikely(ice_page_is_reserved(page)))
+ return false;
+
+#if (PAGE_SIZE < 8192)
+ /* if we are only owner of page we can reuse it */
+ if (unlikely(page_count(page) != 1))
+ return false;
+
+ /* flip page offset to other buffer */
+ rx_buf->page_offset ^= truesize;
+#else
+ /* move offset up to the next cache line */
+ rx_buf->page_offset += truesize;
+
+ if (rx_buf->page_offset > last_offset)
+ return false;
+#endif /* PAGE_SIZE < 8192) */
+
+ /* Even if we own the page, we are not allowed to use atomic_set()
+ * This would break get_page_unless_zero() users.
+ */
+ get_page(rx_buf->page);
+
+ return true;
+}
+
+/**
+ * ice_reuse_rx_page - page flip buffer and store it back on the ring
+ * @rx_ring: rx descriptor ring to store buffers on
+ * @old_buf: donor buffer to have page reused
+ *
+ * Synchronizes page for reuse by the adapter
+ */
+static void ice_reuse_rx_page(struct ice_ring *rx_ring,
+ struct ice_rx_buf *old_buf)
+{
+ u16 nta = rx_ring->next_to_alloc;
+ struct ice_rx_buf *new_buf;
+
+ new_buf = &rx_ring->rx_buf[nta];
+
+ /* update, and store next to alloc */
+ nta++;
+ rx_ring->next_to_alloc = (nta < rx_ring->count) ? nta : 0;
+
+ /* transfer page from old buffer to new buffer */
+ *new_buf = *old_buf;
+}
+
+/**
+ * ice_fetch_rx_buf - Allocate skb and populate it
+ * @rx_ring: rx descriptor ring to transact packets on
+ * @rx_desc: descriptor containing info written by hardware
+ *
+ * This function allocates an skb on the fly, and populates it with the page
+ * data from the current receive descriptor, taking care to set up the skb
+ * correctly, as well as handling calling the page recycle function if
+ * necessary.
+ */
+static struct sk_buff *ice_fetch_rx_buf(struct ice_ring *rx_ring,
+ union ice_32b_rx_flex_desc *rx_desc)
+{
+ struct ice_rx_buf *rx_buf;
+ struct sk_buff *skb;
+ struct page *page;
+
+ rx_buf = &rx_ring->rx_buf[rx_ring->next_to_clean];
+ page = rx_buf->page;
+ prefetchw(page);
+
+ skb = rx_buf->skb;
+
+ if (likely(!skb)) {
+ u8 *page_addr = page_address(page) + rx_buf->page_offset;
+
+ /* prefetch first cache line of first page */
+ prefetch(page_addr);
+#if L1_CACHE_BYTES < 128
+ prefetch((void *)(page_addr + L1_CACHE_BYTES));
+#endif /* L1_CACHE_BYTES */
+
+ /* allocate a skb to store the frags */
+ skb = __napi_alloc_skb(&rx_ring->q_vector->napi,
+ ICE_RX_HDR_SIZE,
+ GFP_ATOMIC | __GFP_NOWARN);
+ if (unlikely(!skb)) {
+ rx_ring->rx_stats.alloc_buf_failed++;
+ return NULL;
+ }
+
+ /* we will be copying header into skb->data in
+ * pskb_may_pull so it is in our interest to prefetch
+ * it now to avoid a possible cache miss
+ */
+ prefetchw(skb->data);
+
+ skb_record_rx_queue(skb, rx_ring->q_index);
+ } else {
+ /* we are reusing so sync this buffer for CPU use */
+ dma_sync_single_range_for_cpu(rx_ring->dev, rx_buf->dma,
+ rx_buf->page_offset,
+ ICE_RXBUF_2048,
+ DMA_FROM_DEVICE);
+
+ rx_buf->skb = NULL;
+ }
+
+ /* pull page into skb */
+ if (ice_add_rx_frag(rx_buf, rx_desc, skb)) {
+ /* hand second half of page back to the ring */
+ ice_reuse_rx_page(rx_ring, rx_buf);
+ rx_ring->rx_stats.page_reuse_count++;
+ } else {
+ /* we are not reusing the buffer so unmap it */
+ dma_unmap_page(rx_ring->dev, rx_buf->dma, PAGE_SIZE,
+ DMA_FROM_DEVICE);
+ }
+
+ /* clear contents of buffer_info */
+ rx_buf->page = NULL;
+
+ return skb;
+}
+
+/**
+ * ice_pull_tail - ice specific version of skb_pull_tail
+ * @skb: pointer to current skb being adjusted
+ *
+ * This function is an ice specific version of __pskb_pull_tail. The
+ * main difference between this version and the original function is that
+ * this function can make several assumptions about the state of things
+ * that allow for significant optimizations versus the standard function.
+ * As a result we can do things like drop a frag and maintain an accurate
+ * truesize for the skb.
+ */
+static void ice_pull_tail(struct sk_buff *skb)
+{
+ struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[0];
+ unsigned int pull_len;
+ unsigned char *va;
+
+ /* it is valid to use page_address instead of kmap since we are
+ * working with pages allocated out of the lomem pool per
+ * alloc_page(GFP_ATOMIC)
+ */
+ va = skb_frag_address(frag);
+
+ /* we need the header to contain the greater of either ETH_HLEN or
+ * 60 bytes if the skb->len is less than 60 for skb_pad.
+ */
+ pull_len = eth_get_headlen(va, ICE_RX_HDR_SIZE);
+
+ /* align pull length to size of long to optimize memcpy performance */
+ skb_copy_to_linear_data(skb, va, ALIGN(pull_len, sizeof(long)));
+
+ /* update all of the pointers */
+ skb_frag_size_sub(frag, pull_len);
+ frag->page_offset += pull_len;
+ skb->data_len -= pull_len;
+ skb->tail += pull_len;
+}
+
+/**
+ * ice_cleanup_headers - Correct empty headers
+ * @skb: pointer to current skb being fixed
+ *
+ * Also address the case where we are pulling data in on pages only
+ * and as such no data is present in the skb header.
+ *
+ * In addition if skb is not at least 60 bytes we need to pad it so that
+ * it is large enough to qualify as a valid Ethernet frame.
+ *
+ * Returns true if an error was encountered and skb was freed.
+ */
+static bool ice_cleanup_headers(struct sk_buff *skb)
+{
+ /* place header in linear portion of buffer */
+ if (skb_is_nonlinear(skb))
+ ice_pull_tail(skb);
+
+ /* if eth_skb_pad returns an error the skb was freed */
+ if (eth_skb_pad(skb))
+ return true;
+
+ return false;
+}
+
+/**
+ * ice_test_staterr - tests bits in Rx descriptor status and error fields
+ * @rx_desc: pointer to receive descriptor (in le64 format)
+ * @stat_err_bits: value to mask
+ *
+ * This function does some fast chicanery in order to return the
+ * value of the mask which is really only used for boolean tests.
+ * The status_error_len doesn't need to be shifted because it begins
+ * at offset zero.
+ */
+static bool ice_test_staterr(union ice_32b_rx_flex_desc *rx_desc,
+ const u16 stat_err_bits)
+{
+ return !!(rx_desc->wb.status_error0 &
+ cpu_to_le16(stat_err_bits));
+}
+
+/**
+ * ice_is_non_eop - process handling of non-EOP buffers
+ * @rx_ring: Rx ring being processed
+ * @rx_desc: Rx descriptor for current buffer
+ * @skb: Current socket buffer containing buffer in progress
+ *
+ * This function updates next to clean. If the buffer is an EOP buffer
+ * this function exits returning false, otherwise it will place the
+ * sk_buff in the next buffer to be chained and return true indicating
+ * that this is in fact a non-EOP buffer.
+ */
+static bool ice_is_non_eop(struct ice_ring *rx_ring,
+ union ice_32b_rx_flex_desc *rx_desc,
+ struct sk_buff *skb)
+{
+ u32 ntc = rx_ring->next_to_clean + 1;
+
+ /* fetch, update, and store next to clean */
+ ntc = (ntc < rx_ring->count) ? ntc : 0;
+ rx_ring->next_to_clean = ntc;
+
+ prefetch(ICE_RX_DESC(rx_ring, ntc));
+
+ /* if we are the last buffer then there is nothing else to do */
+#define ICE_RXD_EOF BIT(ICE_RX_FLEX_DESC_STATUS0_EOF_S)
+ if (likely(ice_test_staterr(rx_desc, ICE_RXD_EOF)))
+ return false;
+
+ /* place skb in next buffer to be received */
+ rx_ring->rx_buf[ntc].skb = skb;
+ rx_ring->rx_stats.non_eop_descs++;
+
+ return true;
+}
+
+/**
+ * ice_receive_skb - Send a completed packet up the stack
+ * @rx_ring: rx ring in play
+ * @skb: packet to send up
+ * @vlan_tag: vlan tag for packet
+ *
+ * This function sends the completed packet (via. skb) up the stack using
+ * gro receive functions (with/without vlan tag)
+ */
+static void ice_receive_skb(struct ice_ring *rx_ring, struct sk_buff *skb,
+ u16 vlan_tag)
+{
+ if ((rx_ring->netdev->features & NETIF_F_HW_VLAN_CTAG_RX) &&
+ (vlan_tag & VLAN_VID_MASK)) {
+ __vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), vlan_tag);
+ }
+ napi_gro_receive(&rx_ring->q_vector->napi, skb);
+}
+
+/**
+ * ice_clean_rx_irq - Clean completed descriptors from Rx ring - bounce buf
+ * @rx_ring: rx descriptor ring to transact packets on
+ * @budget: Total limit on number of packets to process
+ *
+ * This function provides a "bounce buffer" approach to Rx interrupt
+ * processing. The advantage to this is that on systems that have
+ * expensive overhead for IOMMU access this provides a means of avoiding
+ * it by maintaining the mapping of the page to the system.
+ *
+ * Returns amount of work completed
+ */
+static int ice_clean_rx_irq(struct ice_ring *rx_ring, int budget)
+{
+ unsigned int total_rx_bytes = 0, total_rx_pkts = 0;
+ u16 cleaned_count = ICE_DESC_UNUSED(rx_ring);
+ bool failure = false;
+
+ /* start the loop to process RX packets bounded by 'budget' */
+ while (likely(total_rx_pkts < (unsigned int)budget)) {
+ union ice_32b_rx_flex_desc *rx_desc;
+ struct sk_buff *skb;
+ u16 stat_err_bits;
+ u16 vlan_tag = 0;
+
+ /* return some buffers to hardware, one at a time is too slow */
+ if (cleaned_count >= ICE_RX_BUF_WRITE) {
+ failure = failure ||
+ ice_alloc_rx_bufs(rx_ring, cleaned_count);
+ cleaned_count = 0;
+ }
+
+ /* get the RX desc from RX ring based on 'next_to_clean' */
+ rx_desc = ICE_RX_DESC(rx_ring, rx_ring->next_to_clean);
+
+ /* status_error_len will always be zero for unused descriptors
+ * because it's cleared in cleanup, and overlaps with hdr_addr
+ * which is always zero because packet split isn't used, if the
+ * hardware wrote DD then it will be non-zero
+ */
+ stat_err_bits = BIT(ICE_RX_FLEX_DESC_STATUS0_DD_S);
+ if (!ice_test_staterr(rx_desc, stat_err_bits))
+ break;
+
+ /* This memory barrier is needed to keep us from reading
+ * any other fields out of the rx_desc until we know the
+ * DD bit is set.
+ */
+ dma_rmb();
+
+ /* allocate (if needed) and populate skb */
+ skb = ice_fetch_rx_buf(rx_ring, rx_desc);
+ if (!skb)
+ break;
+
+ cleaned_count++;
+
+ /* skip if it is NOP desc */
+ if (ice_is_non_eop(rx_ring, rx_desc, skb))
+ continue;
+
+ stat_err_bits = BIT(ICE_RX_FLEX_DESC_STATUS0_RXE_S);
+ if (unlikely(ice_test_staterr(rx_desc, stat_err_bits))) {
+ dev_kfree_skb_any(skb);
+ continue;
+ }
+
+ stat_err_bits = BIT(ICE_RX_FLEX_DESC_STATUS0_L2TAG1P_S);
+ if (ice_test_staterr(rx_desc, stat_err_bits))
+ vlan_tag = le16_to_cpu(rx_desc->wb.l2tag1);
+
+ /* correct empty headers and pad skb if needed (to make valid
+ * ethernet frame
+ */
+ if (ice_cleanup_headers(skb)) {
+ skb = NULL;
+ continue;
+ }
+
+ /* probably a little skewed due to removing CRC */
+ total_rx_bytes += skb->len;
+
+ /* send completed skb up the stack */
+ ice_receive_skb(rx_ring, skb, vlan_tag);
+
+ /* update budget accounting */
+ total_rx_pkts++;
+ }
+
+ /* update queue and vector specific stats */
+ u64_stats_update_begin(&rx_ring->syncp);
+ rx_ring->stats.pkts += total_rx_pkts;
+ rx_ring->stats.bytes += total_rx_bytes;
+ u64_stats_update_end(&rx_ring->syncp);
+ rx_ring->q_vector->rx.total_pkts += total_rx_pkts;
+ rx_ring->q_vector->rx.total_bytes += total_rx_bytes;
+
+ /* guarantee a trip back through this routine if there was a failure */
+ return failure ? budget : (int)total_rx_pkts;
+}
+
+/**
+ * ice_napi_poll - NAPI polling Rx/Tx cleanup routine
+ * @napi: napi struct with our devices info in it
+ * @budget: amount of work driver is allowed to do this pass, in packets
+ *
+ * This function will clean all queues associated with a q_vector.
+ *
+ * Returns the amount of work done
+ */
+int ice_napi_poll(struct napi_struct *napi, int budget)
+{
+ struct ice_q_vector *q_vector =
+ container_of(napi, struct ice_q_vector, napi);
+ struct ice_vsi *vsi = q_vector->vsi;
+ struct ice_pf *pf = vsi->back;
+ bool clean_complete = true;
+ int budget_per_ring = 0;
+ struct ice_ring *ring;
+ int work_done = 0;
+
+ /* Since the actual Tx work is minimal, we can give the Tx a larger
+ * budget and be more aggressive about cleaning up the Tx descriptors.
+ */
+ ice_for_each_ring(ring, q_vector->tx)
+ if (!ice_clean_tx_irq(vsi, ring, budget))
+ clean_complete = false;
+
+ /* Handle case where we are called by netpoll with a budget of 0 */
+ if (budget <= 0)
+ return budget;
+
+ /* We attempt to distribute budget to each Rx queue fairly, but don't
+ * allow the budget to go below 1 because that would exit polling early.
+ */
+ if (q_vector->num_ring_rx)
+ budget_per_ring = max(budget / q_vector->num_ring_rx, 1);
+
+ ice_for_each_ring(ring, q_vector->rx) {
+ int cleaned;
+
+ cleaned = ice_clean_rx_irq(ring, budget_per_ring);
+ work_done += cleaned;
+ /* if we clean as many as budgeted, we must not be done */
+ if (cleaned >= budget_per_ring)
+ clean_complete = false;
+ }
+
+ /* If work not completed, return budget and polling will return */
+ if (!clean_complete)
+ return budget;
+
+ /* Work is done so exit the polling mode and re-enable the interrupt */
+ napi_complete_done(napi, work_done);
+ if (test_bit(ICE_FLAG_MSIX_ENA, pf->flags))
+ ice_irq_dynamic_ena(&vsi->back->hw, vsi, q_vector);
+ return 0;
+}
+
+/* helper function for building cmd/type/offset */
+static __le64
+build_ctob(u64 td_cmd, u64 td_offset, unsigned int size, u64 td_tag)
+{
+ return cpu_to_le64(ICE_TX_DESC_DTYPE_DATA |
+ (td_cmd << ICE_TXD_QW1_CMD_S) |
+ (td_offset << ICE_TXD_QW1_OFFSET_S) |
+ ((u64)size << ICE_TXD_QW1_TX_BUF_SZ_S) |
+ (td_tag << ICE_TXD_QW1_L2TAG1_S));
+}
+
+/**
+ * __ice_maybe_stop_tx - 2nd level check for tx stop conditions
+ * @tx_ring: the ring to be checked
+ * @size: the size buffer we want to assure is available
+ *
+ * Returns -EBUSY if a stop is needed, else 0
+ */
+static int __ice_maybe_stop_tx(struct ice_ring *tx_ring, unsigned int size)
+{
+ netif_stop_subqueue(tx_ring->netdev, tx_ring->q_index);
+ /* Memory barrier before checking head and tail */
+ smp_mb();
+
+ /* Check again in a case another CPU has just made room available. */
+ if (likely(ICE_DESC_UNUSED(tx_ring) < size))
+ return -EBUSY;
+
+ /* A reprieve! - use start_subqueue because it doesn't call schedule */
+ netif_start_subqueue(tx_ring->netdev, tx_ring->q_index);
+ ++tx_ring->tx_stats.restart_q;
+ return 0;
+}
+
+/**
+ * ice_maybe_stop_tx - 1st level check for tx stop conditions
+ * @tx_ring: the ring to be checked
+ * @size: the size buffer we want to assure is available
+ *
+ * Returns 0 if stop is not needed
+ */
+static int ice_maybe_stop_tx(struct ice_ring *tx_ring, unsigned int size)
+{
+ if (likely(ICE_DESC_UNUSED(tx_ring) >= size))
+ return 0;
+ return __ice_maybe_stop_tx(tx_ring, size);
+}
+
+/**
+ * ice_tx_map - Build the Tx descriptor
+ * @tx_ring: ring to send buffer on
+ * @first: first buffer info buffer to use
+ *
+ * This function loops over the skb data pointed to by *first
+ * and gets a physical address for each memory location and programs
+ * it and the length into the transmit descriptor.
+ */
+static void ice_tx_map(struct ice_ring *tx_ring, struct ice_tx_buf *first)
+{
+ u64 td_offset = 0, td_tag = 0, td_cmd = 0;
+ u16 i = tx_ring->next_to_use;
+ struct skb_frag_struct *frag;
+ unsigned int data_len, size;
+ struct ice_tx_desc *tx_desc;
+ struct ice_tx_buf *tx_buf;
+ struct sk_buff *skb;
+ dma_addr_t dma;
+
+ skb = first->skb;
+
+ data_len = skb->data_len;
+ size = skb_headlen(skb);
+
+ tx_desc = ICE_TX_DESC(tx_ring, i);
+
+ dma = dma_map_single(tx_ring->dev, skb->data, size, DMA_TO_DEVICE);
+
+ tx_buf = first;
+
+ for (frag = &skb_shinfo(skb)->frags[0];; frag++) {
+ unsigned int max_data = ICE_MAX_DATA_PER_TXD_ALIGNED;
+
+ if (dma_mapping_error(tx_ring->dev, dma))
+ goto dma_error;
+
+ /* record length, and DMA address */
+ dma_unmap_len_set(tx_buf, len, size);
+ dma_unmap_addr_set(tx_buf, dma, dma);
+
+ /* align size to end of page */
+ max_data += -dma & (ICE_MAX_READ_REQ_SIZE - 1);
+ tx_desc->buf_addr = cpu_to_le64(dma);
+
+ /* account for data chunks larger than the hardware
+ * can handle
+ */
+ while (unlikely(size > ICE_MAX_DATA_PER_TXD)) {
+ tx_desc->cmd_type_offset_bsz =
+ build_ctob(td_cmd, td_offset, max_data, td_tag);
+
+ tx_desc++;
+ i++;
+
+ if (i == tx_ring->count) {
+ tx_desc = ICE_TX_DESC(tx_ring, 0);
+ i = 0;
+ }
+
+ dma += max_data;
+ size -= max_data;
+
+ max_data = ICE_MAX_DATA_PER_TXD_ALIGNED;
+ tx_desc->buf_addr = cpu_to_le64(dma);
+ }
+
+ if (likely(!data_len))
+ break;
+
+ tx_desc->cmd_type_offset_bsz = build_ctob(td_cmd, td_offset,
+ size, td_tag);
+
+ tx_desc++;
+ i++;
+
+ if (i == tx_ring->count) {
+ tx_desc = ICE_TX_DESC(tx_ring, 0);
+ i = 0;
+ }
+
+ size = skb_frag_size(frag);
+ data_len -= size;
+
+ dma = skb_frag_dma_map(tx_ring->dev, frag, 0, size,
+ DMA_TO_DEVICE);
+
+ tx_buf = &tx_ring->tx_buf[i];
+ }
+
+ /* record bytecount for BQL */
+ netdev_tx_sent_queue(txring_txq(tx_ring), first->bytecount);
+
+ /* record SW timestamp if HW timestamp is not available */
+ skb_tx_timestamp(first->skb);
+
+ i++;
+ if (i == tx_ring->count)
+ i = 0;
+
+ /* write last descriptor with RS and EOP bits */
+ td_cmd |= (u64)(ICE_TX_DESC_CMD_EOP | ICE_TX_DESC_CMD_RS);
+ tx_desc->cmd_type_offset_bsz =
+ build_ctob(td_cmd, td_offset, size, td_tag);
+
+ /* Force memory writes to complete before letting h/w know there
+ * are new descriptors to fetch.
+ *
+ * We also use this memory barrier to make certain all of the
+ * status bits have been updated before next_to_watch is written.
+ */
+ wmb();
+
+ /* set next_to_watch value indicating a packet is present */
+ first->next_to_watch = tx_desc;
+
+ tx_ring->next_to_use = i;
+
+ ice_maybe_stop_tx(tx_ring, DESC_NEEDED);
+
+ /* notify HW of packet */
+ if (netif_xmit_stopped(txring_txq(tx_ring)) || !skb->xmit_more) {
+ writel(i, tx_ring->tail);
+
+ /* we need this if more than one processor can write to our tail
+ * at a time, it synchronizes IO on IA64/Altix systems
+ */
+ mmiowb();
+ }
+
+ return;
+
+dma_error:
+ /* clear dma mappings for failed tx_buf map */
+ for (;;) {
+ tx_buf = &tx_ring->tx_buf[i];
+ ice_unmap_and_free_tx_buf(tx_ring, tx_buf);
+ if (tx_buf == first)
+ break;
+ if (i == 0)
+ i = tx_ring->count;
+ i--;
+ }
+
+ tx_ring->next_to_use = i;
+}
+
+/**
+ * ice_txd_use_count - estimate the number of descriptors needed for Tx
+ * @size: transmit request size in bytes
+ *
+ * Due to hardware alignment restrictions (4K alignment), we need to
+ * assume that we can have no more than 12K of data per descriptor, even
+ * though each descriptor can take up to 16K - 1 bytes of aligned memory.
+ * Thus, we need to divide by 12K. But division is slow! Instead,
+ * we decompose the operation into shifts and one relatively cheap
+ * multiply operation.
+ *
+ * To divide by 12K, we first divide by 4K, then divide by 3:
+ * To divide by 4K, shift right by 12 bits
+ * To divide by 3, multiply by 85, then divide by 256
+ * (Divide by 256 is done by shifting right by 8 bits)
+ * Finally, we add one to round up. Because 256 isn't an exact multiple of
+ * 3, we'll underestimate near each multiple of 12K. This is actually more
+ * accurate as we have 4K - 1 of wiggle room that we can fit into the last
+ * segment. For our purposes this is accurate out to 1M which is orders of
+ * magnitude greater than our largest possible GSO size.
+ *
+ * This would then be implemented as:
+ * return (((size >> 12) * 85) >> 8) + 1;
+ *
+ * Since multiplication and division are commutative, we can reorder
+ * operations into:
+ * return ((size * 85) >> 20) + 1;
+ */
+static unsigned int ice_txd_use_count(unsigned int size)
+{
+ return ((size * 85) >> 20) + 1;
+}
+
+/**
+ * ice_xmit_desc_count - calculate number of tx descriptors needed
+ * @skb: send buffer
+ *
+ * Returns number of data descriptors needed for this skb.
+ */
+static unsigned int ice_xmit_desc_count(struct sk_buff *skb)
+{
+ const struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[0];
+ unsigned int nr_frags = skb_shinfo(skb)->nr_frags;
+ unsigned int count = 0, size = skb_headlen(skb);
+
+ for (;;) {
+ count += ice_txd_use_count(size);
+
+ if (!nr_frags--)
+ break;
+
+ size = skb_frag_size(frag++);
+ }
+
+ return count;
+}
+
+/**
+ * __ice_chk_linearize - Check if there are more than 8 buffers per packet
+ * @skb: send buffer
+ *
+ * Note: This HW can't DMA more than 8 buffers to build a packet on the wire
+ * and so we need to figure out the cases where we need to linearize the skb.
+ *
+ * For TSO we need to count the TSO header and segment payload separately.
+ * As such we need to check cases where we have 7 fragments or more as we
+ * can potentially require 9 DMA transactions, 1 for the TSO header, 1 for
+ * the segment payload in the first descriptor, and another 7 for the
+ * fragments.
+ */
+static bool __ice_chk_linearize(struct sk_buff *skb)
+{
+ const struct skb_frag_struct *frag, *stale;
+ int nr_frags, sum;
+
+ /* no need to check if number of frags is less than 7 */
+ nr_frags = skb_shinfo(skb)->nr_frags;
+ if (nr_frags < (ICE_MAX_BUF_TXD - 1))
+ return false;
+
+ /* We need to walk through the list and validate that each group
+ * of 6 fragments totals at least gso_size.
+ */
+ nr_frags -= ICE_MAX_BUF_TXD - 2;
+ frag = &skb_shinfo(skb)->frags[0];
+
+ /* Initialize size to the negative value of gso_size minus 1. We
+ * use this as the worst case scenerio in which the frag ahead
+ * of us only provides one byte which is why we are limited to 6
+ * descriptors for a single transmit as the header and previous
+ * fragment are already consuming 2 descriptors.
+ */
+ sum = 1 - skb_shinfo(skb)->gso_size;
+
+ /* Add size of frags 0 through 4 to create our initial sum */
+ sum += skb_frag_size(frag++);
+ sum += skb_frag_size(frag++);
+ sum += skb_frag_size(frag++);
+ sum += skb_frag_size(frag++);
+ sum += skb_frag_size(frag++);
+
+ /* Walk through fragments adding latest fragment, testing it, and
+ * then removing stale fragments from the sum.
+ */
+ stale = &skb_shinfo(skb)->frags[0];
+ for (;;) {
+ sum += skb_frag_size(frag++);
+
+ /* if sum is negative we failed to make sufficient progress */
+ if (sum < 0)
+ return true;
+
+ if (!nr_frags--)
+ break;
+
+ sum -= skb_frag_size(stale++);
+ }
+
+ return false;
+}
+
+/**
+ * ice_chk_linearize - Check if there are more than 8 fragments per packet
+ * @skb: send buffer
+ * @count: number of buffers used
+ *
+ * Note: Our HW can't scatter-gather more than 8 fragments to build
+ * a packet on the wire and so we need to figure out the cases where we
+ * need to linearize the skb.
+ */
+static bool ice_chk_linearize(struct sk_buff *skb, unsigned int count)
+{
+ /* Both TSO and single send will work if count is less than 8 */
+ if (likely(count < ICE_MAX_BUF_TXD))
+ return false;
+
+ if (skb_is_gso(skb))
+ return __ice_chk_linearize(skb);
+
+ /* we can support up to 8 data buffers for a single send */
+ return count != ICE_MAX_BUF_TXD;
+}
+
+/**
+ * ice_xmit_frame_ring - Sends buffer on Tx ring
+ * @skb: send buffer
+ * @tx_ring: ring to send buffer on
+ *
+ * Returns NETDEV_TX_OK if sent, else an error code
+ */
+static netdev_tx_t
+ice_xmit_frame_ring(struct sk_buff *skb, struct ice_ring *tx_ring)
+{
+ struct ice_tx_buf *first;
+ unsigned int count;
+
+ count = ice_xmit_desc_count(skb);
+ if (ice_chk_linearize(skb, count)) {
+ if (__skb_linearize(skb))
+ goto out_drop;
+ count = ice_txd_use_count(skb->len);
+ tx_ring->tx_stats.tx_linearize++;
+ }
+
+ /* need: 1 descriptor per page * PAGE_SIZE/ICE_MAX_DATA_PER_TXD,
+ * + 1 desc for skb_head_len/ICE_MAX_DATA_PER_TXD,
+ * + 4 desc gap to avoid the cache line where head is,
+ * + 1 desc for context descriptor,
+ * otherwise try next time
+ */
+ if (ice_maybe_stop_tx(tx_ring, count + 4 + 1)) {
+ tx_ring->tx_stats.tx_busy++;
+ return NETDEV_TX_BUSY;
+ }
+
+ /* record the location of the first descriptor for this packet */
+ first = &tx_ring->tx_buf[tx_ring->next_to_use];
+ first->skb = skb;
+ first->bytecount = max_t(unsigned int, skb->len, ETH_ZLEN);
+ first->gso_segs = 1;
+
+ ice_tx_map(tx_ring, first);
+ return NETDEV_TX_OK;
+
+out_drop:
+ dev_kfree_skb_any(skb);
+ return NETDEV_TX_OK;
+}
+
+/**
+ * ice_start_xmit - Selects the correct VSI and Tx queue to send buffer
+ * @skb: send buffer
+ * @netdev: network interface device structure
+ *
+ * Returns NETDEV_TX_OK if sent, else an error code
+ */
+netdev_tx_t ice_start_xmit(struct sk_buff *skb, struct net_device *netdev)
+{
+ struct ice_netdev_priv *np = netdev_priv(netdev);
+ struct ice_vsi *vsi = np->vsi;
+ struct ice_ring *tx_ring;
+
+ tx_ring = vsi->tx_rings[skb->queue_mapping];
+
+ /* hardware can't handle really short frames, hardware padding works
+ * beyond this point
+ */
+ if (skb_put_padto(skb, ICE_MIN_TX_LEN))
+ return NETDEV_TX_OK;
+
+ return ice_xmit_frame_ring(skb, tx_ring);
+}
diff --git a/drivers/net/ethernet/intel/ice/ice_txrx.h b/drivers/net/ethernet/intel/ice/ice_txrx.h
index b7015cfad2d7..8fa4450514b4 100644
--- a/drivers/net/ethernet/intel/ice/ice_txrx.h
+++ b/drivers/net/ethernet/intel/ice/ice_txrx.h
@@ -7,8 +7,23 @@
#define ICE_DFLT_IRQ_WORK 256
#define ICE_RXBUF_2048 2048
#define ICE_MAX_CHAINED_RX_BUFS 5
+#define ICE_MAX_BUF_TXD 8
+#define ICE_MIN_TX_LEN 17
+
+/* The size limit for a transmit buffer in a descriptor is (16K - 1).
+ * In order to align with the read requests we will align the value to
+ * the nearest 4K which represents our maximum read request size.
+ */
+#define ICE_MAX_READ_REQ_SIZE 4096
+#define ICE_MAX_DATA_PER_TXD (16 * 1024 - 1)
+#define ICE_MAX_DATA_PER_TXD_ALIGNED \
+ (~(ICE_MAX_READ_REQ_SIZE - 1) & ICE_MAX_DATA_PER_TXD)
+
+#define ICE_RX_BUF_WRITE 16 /* Must be power of 2 */
#define ICE_MAX_TXQ_PER_TXQG 128
+/* Tx Descriptors needed, worst case */
+#define DESC_NEEDED (MAX_SKB_FRAGS + 4)
#define ICE_DESC_UNUSED(R) \
((((R)->next_to_clean > (R)->next_to_use) ? 0 : (R)->count) + \
(R)->next_to_clean - (R)->next_to_use - 1)
@@ -30,6 +45,24 @@ struct ice_rx_buf {
unsigned int page_offset;
};
+struct ice_q_stats {
+ u64 pkts;
+ u64 bytes;
+};
+
+struct ice_txq_stats {
+ u64 restart_q;
+ u64 tx_busy;
+ u64 tx_linearize;
+};
+
+struct ice_rxq_stats {
+ u64 non_eop_descs;
+ u64 alloc_page_failed;
+ u64 alloc_buf_failed;
+ u64 page_reuse_count;
+};
+
/* this enum matches hardware bits and is meant to be used by DYN_CTLN
* registers and QINT registers or more generally anywhere in the manual
* mentioning ITR_INDX, ITR_NONE cannot be used as an index 'n' into any
@@ -94,6 +127,15 @@ struct ice_ring {
u16 next_to_clean;
bool ring_active; /* is ring online or not */
+
+ /* stats structs */
+ struct ice_q_stats stats;
+ struct u64_stats_sync syncp;
+ union {
+ struct ice_txq_stats tx_stats;
+ struct ice_rxq_stats rx_stats;
+ };
+
unsigned int size; /* length of descriptor ring in bytes */
dma_addr_t dma; /* physical address of ring */
struct rcu_head rcu; /* to avoid race on free */
@@ -121,10 +163,13 @@ struct ice_ring_container {
for (pos = (head).ring; pos; pos = pos->next)
bool ice_alloc_rx_bufs(struct ice_ring *rxr, u16 cleaned_count);
+netdev_tx_t ice_start_xmit(struct sk_buff *skb, struct net_device *netdev);
void ice_clean_tx_ring(struct ice_ring *tx_ring);
void ice_clean_rx_ring(struct ice_ring *rx_ring);
int ice_setup_tx_ring(struct ice_ring *tx_ring);
int ice_setup_rx_ring(struct ice_ring *rx_ring);
void ice_free_tx_ring(struct ice_ring *tx_ring);
void ice_free_rx_ring(struct ice_ring *rx_ring);
+int ice_napi_poll(struct napi_struct *napi, int budget);
+
#endif /* _ICE_TXRX_H_ */
--
2.14.3
Powered by blists - more mailing lists