i40e/i40evf: Use length to determine if descriptor is done

index 9742beb..68936b6 100644
@@ -1757,6 +1757,7 @@ static bool i40e_add_rx_frag(struct i40e_ring *rx_ring,
  * i40e_fetch_rx_buffer - Allocate skb and populate it
  * @rx_ring: rx descriptor ring to transact packets on
  * @rx_desc: descriptor containing info written by hardware
+ * @size: size of buffer to add to skb
  *
  * 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

@@ -1766,13 +1767,9 @@ static bool i40e_add_rx_frag(struct i40e_ring *rx_ring,
 static inline
 struct sk_buff *i40e_fetch_rx_buffer(struct i40e_ring *rx_ring,
 				     union i40e_rx_desc *rx_desc,
-				     struct sk_buff *skb)
+				     struct sk_buff *skb,
+				     unsigned int size)
 {
-	u64 local_status_error_len =
-		le64_to_cpu(rx_desc->wb.qword1.status_error_len);
-	unsigned int size =
-		(local_status_error_len & I40E_RXD_QW1_LENGTH_PBUF_MASK) >>
-		I40E_RXD_QW1_LENGTH_PBUF_SHIFT;
 	struct i40e_rx_buffer *rx_buffer;
 	struct page *page;
 

@@ -1890,6 +1887,7 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget)
 
 	while (likely(total_rx_packets < budget)) {
 		union i40e_rx_desc *rx_desc;
+		unsigned int size;
 		u16 vlan_tag;
 		u8 rx_ptype;
 		u64 qword;

@@ -1906,19 +1904,21 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget)
 		/* 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
+		 * hardware wrote DD then the length will be non-zero
 		 */
-		if (!i40e_test_staterr(rx_desc,
-				       BIT(I40E_RX_DESC_STATUS_DD_SHIFT)))
+		qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len);
+		size = (qword & I40E_RXD_QW1_LENGTH_PBUF_MASK) >>
+		       I40E_RXD_QW1_LENGTH_PBUF_SHIFT;
+		if (!size)
 			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.
+		 * any other fields out of the rx_desc until we have
+		 * verified the descriptor has been written back.
 		 */
 		dma_rmb();
 
-		skb = i40e_fetch_rx_buffer(rx_ring, rx_desc, skb);
+		skb = i40e_fetch_rx_buffer(rx_ring, rx_desc, skb, size);
 		if (!skb)
 			break;
 

index f1a99a8..e41eb46 100644
@@ -1116,6 +1116,7 @@ static bool i40e_add_rx_frag(struct i40e_ring *rx_ring,
  * i40evf_fetch_rx_buffer - Allocate skb and populate it
  * @rx_ring: rx descriptor ring to transact packets on
  * @rx_desc: descriptor containing info written by hardware
+ * @size: size of buffer to add to skb
  *
  * 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

@@ -1125,13 +1126,9 @@ static bool i40e_add_rx_frag(struct i40e_ring *rx_ring,
 static inline
 struct sk_buff *i40evf_fetch_rx_buffer(struct i40e_ring *rx_ring,
 				       union i40e_rx_desc *rx_desc,
-				       struct sk_buff *skb)
+				       struct sk_buff *skb,
+				       unsigned int size)
 {
-	u64 local_status_error_len =
-		le64_to_cpu(rx_desc->wb.qword1.status_error_len);
-	unsigned int size =
-		(local_status_error_len & I40E_RXD_QW1_LENGTH_PBUF_MASK) >>
-		I40E_RXD_QW1_LENGTH_PBUF_SHIFT;
 	struct i40e_rx_buffer *rx_buffer;
 	struct page *page;
 

@@ -1244,6 +1241,7 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget)
 
 	while (likely(total_rx_packets < budget)) {
 		union i40e_rx_desc *rx_desc;
+		unsigned int size;
 		u16 vlan_tag;
 		u8 rx_ptype;
 		u64 qword;

@@ -1260,19 +1258,21 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget)
 		/* 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
+		 * hardware wrote DD then the length will be non-zero
 		 */
-		if (!i40e_test_staterr(rx_desc,
-				       BIT(I40E_RX_DESC_STATUS_DD_SHIFT)))
+		qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len);
+		size = (qword & I40E_RXD_QW1_LENGTH_PBUF_MASK) >>
+		       I40E_RXD_QW1_LENGTH_PBUF_SHIFT;
+		if (!size)
 			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.
+		 * any other fields out of the rx_desc until we have
+		 * verified the descriptor has been written back.
 		 */
 		dma_rmb();
 
-		skb = i40evf_fetch_rx_buffer(rx_ring, rx_desc, skb);
+		skb = i40evf_fetch_rx_buffer(rx_ring, rx_desc, skb, size);
 		if (!skb)
 			break;
 

index 68936b6..d1fc0f0 100644
@@ -1754,9 +1754,35 @@ static bool i40e_add_rx_frag(struct i40e_ring *rx_ring,
 }
 
 /**
+ * i40e_get_rx_buffer - Fetch Rx buffer and synchronize data for use
+ * @rx_ring: rx descriptor ring to transact packets on
+ * @size: size of buffer to add to skb
+ *
+ * This function will pull an Rx buffer from the ring and synchronize it
+ * for use by the CPU.
+ */
+static struct i40e_rx_buffer *i40e_get_rx_buffer(struct i40e_ring *rx_ring,
+						 const unsigned int size)
+{
+	struct i40e_rx_buffer *rx_buffer;
+
+	rx_buffer = &rx_ring->rx_bi[rx_ring->next_to_clean];
+	prefetchw(rx_buffer->page);
+
+	/* we are reusing so sync this buffer for CPU use */
+	dma_sync_single_range_for_cpu(rx_ring->dev,
+				      rx_buffer->dma,
+				      rx_buffer->page_offset,
+				      size,
+				      DMA_FROM_DEVICE);
+
+	return rx_buffer;
+}
+
+/**
  * i40e_fetch_rx_buffer - Allocate skb and populate it
  * @rx_ring: rx descriptor ring to transact packets on
- * @rx_desc: descriptor containing info written by hardware
+ * @rx_buffer: rx buffer to pull data from
  * @size: size of buffer to add to skb
  *
  * This function allocates an skb on the fly, and populates it with the page

@@ -1766,19 +1792,13 @@ static bool i40e_add_rx_frag(struct i40e_ring *rx_ring,
  */
 static inline
 struct sk_buff *i40e_fetch_rx_buffer(struct i40e_ring *rx_ring,
-				     union i40e_rx_desc *rx_desc,
+				     struct i40e_rx_buffer *rx_buffer,
 				     struct sk_buff *skb,
 				     unsigned int size)
 {
-	struct i40e_rx_buffer *rx_buffer;
-	struct page *page;
-
-	rx_buffer = &rx_ring->rx_bi[rx_ring->next_to_clean];
-	page = rx_buffer->page;
-	prefetchw(page);
-
 	if (likely(!skb)) {
-		void *page_addr = page_address(page) + rx_buffer->page_offset;
+		void *page_addr = page_address(rx_buffer->page) +
+				  rx_buffer->page_offset;
 
 		/* prefetch first cache line of first page */
 		prefetch(page_addr);

@@ -1794,21 +1814,8 @@ struct sk_buff *i40e_fetch_rx_buffer(struct i40e_ring *rx_ring,
 			rx_ring->rx_stats.alloc_buff_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);
 	}
 
-	/* we are reusing so sync this buffer for CPU use */
-	dma_sync_single_range_for_cpu(rx_ring->dev,
-				      rx_buffer->dma,
-				      rx_buffer->page_offset,
-				      size,
-				      DMA_FROM_DEVICE);
-
 	/* pull page into skb */
 	if (i40e_add_rx_frag(rx_ring, rx_buffer, size, skb)) {
 		/* hand second half of page back to the ring */

@@ -1886,6 +1893,7 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget)
 	bool failure = false;
 
 	while (likely(total_rx_packets < budget)) {
+		struct i40e_rx_buffer *rx_buffer;
 		union i40e_rx_desc *rx_desc;
 		unsigned int size;
 		u16 vlan_tag;

@@ -1918,7 +1926,9 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget)
 		 */
 		dma_rmb();
 
-		skb = i40e_fetch_rx_buffer(rx_ring, rx_desc, skb, size);
+		rx_buffer = i40e_get_rx_buffer(rx_ring, size);
+
+		skb = i40e_fetch_rx_buffer(rx_ring, rx_buffer, skb, size);
 		if (!skb)
 			break;
 

96 lines not on this page · raw message

index d1fc0f0..d7c4e1e 100644
@@ -1294,6 +1294,8 @@ static bool i40e_alloc_mapped_page(struct i40e_ring *rx_ring,
 	bi->dma = dma;
 	bi->page = page;
 	bi->page_offset = 0;
+
+	/* initialize pagecnt_bias to 1 representing we fully own page */
 	bi->pagecnt_bias = 1;
 
 	return true;

@@ -1622,8 +1624,6 @@ static inline bool i40e_page_is_reusable(struct page *page)
  * the adapter for another receive
  *
  * @rx_buffer: buffer containing the page
- * @page: page address from rx_buffer
- * @truesize: actual size of the buffer in this page
  *
  * If page is reusable, rx_buffer->page_offset is adjusted to point to
  * an unused region in the page.

@@ -1646,14 +1646,13 @@ static inline bool i40e_page_is_reusable(struct page *page)
  *
  * In either case, if the page is reusable its refcount is increased.
  **/
-static bool i40e_can_reuse_rx_page(struct i40e_rx_buffer *rx_buffer,
-				   struct page *page,
-				   const unsigned int truesize)
+static bool i40e_can_reuse_rx_page(struct i40e_rx_buffer *rx_buffer)
 {
 #if (PAGE_SIZE >= 8192)
 	unsigned int last_offset = PAGE_SIZE - I40E_RXBUFFER_2048;
 #endif
-	unsigned int pagecnt_bias = rx_buffer->pagecnt_bias--;
+	unsigned int pagecnt_bias = rx_buffer->pagecnt_bias;
+	struct page *page = rx_buffer->page;
 
 	/* Is any reuse possible? */
 	if (unlikely(!i40e_page_is_reusable(page)))

@@ -1661,15 +1660,9 @@ static bool i40e_can_reuse_rx_page(struct i40e_rx_buffer *rx_buffer,
 
 #if (PAGE_SIZE < 8192)
 	/* if we are only owner of page we can reuse it */
-	if (unlikely(page_count(page) != pagecnt_bias))
+	if (unlikely((page_count(page) - pagecnt_bias) > 1))
 		return false;
-
-	/* flip page offset to other buffer */
-	rx_buffer->page_offset ^= truesize;
 #else
-	/* move offset up to the next cache line */
-	rx_buffer->page_offset += truesize;
-
 	if (rx_buffer->page_offset > last_offset)
 		return false;
 #endif

@@ -1678,10 +1671,11 @@ static bool i40e_can_reuse_rx_page(struct i40e_rx_buffer *rx_buffer,
 	 * the pagecnt_bias and page count so that we fully restock the
 	 * number of references the driver holds.
 	 */
-	if (unlikely(pagecnt_bias == 1)) {
+	if (unlikely(!pagecnt_bias)) {
 		page_ref_add(page, USHRT_MAX);
 		rx_buffer->pagecnt_bias = USHRT_MAX;
 	}
+
 	return true;
 }
 

@@ -1689,8 +1683,8 @@ static bool i40e_can_reuse_rx_page(struct i40e_rx_buffer *rx_buffer,
  * i40e_add_rx_frag - Add contents of Rx buffer to sk_buff
  * @rx_ring: rx descriptor ring to transact packets on
  * @rx_buffer: buffer containing page to add
- * @size: packet length from rx_desc
  * @skb: sk_buff to place the data into
+ * @size: packet length from rx_desc
  *
  * This function will add the data contained in rx_buffer->page to the skb.
  * This is done either through a direct copy if the data in the buffer is

@@ -1700,10 +1694,10 @@ static bool i40e_can_reuse_rx_page(struct i40e_rx_buffer *rx_buffer,
  * The function will then update the page offset if necessary and return
  * true if the buffer can be reused by the adapter.
  **/
-static bool i40e_add_rx_frag(struct i40e_ring *rx_ring,
+static void i40e_add_rx_frag(struct i40e_ring *rx_ring,
 			     struct i40e_rx_buffer *rx_buffer,
-			     unsigned int size,
-			     struct sk_buff *skb)
+			     struct sk_buff *skb,
+			     unsigned int size)
 {
 	struct page *page = rx_buffer->page;
 	unsigned char *va = page_address(page) + rx_buffer->page_offset;

@@ -1723,12 +1717,11 @@ static bool i40e_add_rx_frag(struct i40e_ring *rx_ring,
 	if (size <= I40E_RX_HDR_SIZE) {
 		memcpy(__skb_put(skb, size), va, ALIGN(size, sizeof(long)));
 
-		/* page is reusable, we can reuse buffer as-is */
-		if (likely(i40e_page_is_reusable(page)))
-			return true;
-
-		/* this page cannot be reused so discard it */
-		return false;
+		/* page is to be freed, increase pagecnt_bias instead of
+		 * decreasing page count.
+		 */
+		rx_buffer->pagecnt_bias++;
+		return;
 	}
 
 	/* we need the header to contain the greater of either

@@ -1750,7 +1743,12 @@ static bool i40e_add_rx_frag(struct i40e_ring *rx_ring,
 	skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, page,
 			(unsigned long)va & ~PAGE_MASK, size, truesize);
 
-	return i40e_can_reuse_rx_page(rx_buffer, page, truesize);
+	/* page is being used so we must update the page offset */
+#if (PAGE_SIZE < 8192)
+	rx_buffer->page_offset ^= truesize;
+#else
+	rx_buffer->page_offset += truesize;
+#endif
 }
 
 /**

@@ -1776,6 +1774,9 @@ static struct i40e_rx_buffer *i40e_get_rx_buffer(struct i40e_ring *rx_ring,
 				      size,
 				      DMA_FROM_DEVICE);
 
+	/* We have pulled a buffer for use, so decrement pagecnt_bias */
+	rx_buffer->pagecnt_bias--;
+
 	return rx_buffer;
 }
 

@@ -1812,12 +1813,29 @@ struct sk_buff *i40e_fetch_rx_buffer(struct i40e_ring *rx_ring,
 				       GFP_ATOMIC | __GFP_NOWARN);
 		if (unlikely(!skb)) {
 			rx_ring->rx_stats.alloc_buff_failed++;
+			rx_buffer->pagecnt_bias++;
 			return NULL;
 		}
 	}
 
 	/* pull page into skb */
-	if (i40e_add_rx_frag(rx_ring, rx_buffer, size, skb)) {
+	i40e_add_rx_frag(rx_ring, rx_buffer, skb, size);
+
+	return skb;
+}
+
+/**
+ * i40e_put_rx_buffer - Clean up used buffer and either recycle or free
+ * @rx_ring: rx descriptor ring to transact packets on
+ * @rx_buffer: rx buffer to pull data from
+ *
+ * This function will clean up the contents of the rx_buffer.  It will
+ * either recycle the bufer or unmap it and free the associated resources.
+ */
+static void i40e_put_rx_buffer(struct i40e_ring *rx_ring,
+			       struct i40e_rx_buffer *rx_buffer)
+{
+	if (i40e_can_reuse_rx_page(rx_buffer)) {
 		/* hand second half of page back to the ring */
 		i40e_reuse_rx_page(rx_ring, rx_buffer);
 		rx_ring->rx_stats.page_reuse_count++;

@@ -1831,8 +1849,6 @@ struct sk_buff *i40e_fetch_rx_buffer(struct i40e_ring *rx_ring,
 
 	/* clear contents of buffer_info */
 	rx_buffer->page = NULL;
-
-	return skb;
 }
 
 /**

@@ -1932,6 +1948,7 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget)
 		if (!skb)
 			break;
 
+		i40e_put_rx_buffer(rx_ring, rx_buffer);
 		cleaned_count++;
 
 		if (i40e_is_non_eop(rx_ring, rx_desc, skb))

163 lines not on this page · raw message

index d7c4e1e..433309f 100644
@@ -1687,61 +1687,23 @@ static bool i40e_can_reuse_rx_page(struct i40e_rx_buffer *rx_buffer)
  * @size: packet length from rx_desc
  *
  * This function will add the data contained in rx_buffer->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.
+ * 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.
+ * The function will then update the page offset.
  **/
 static void i40e_add_rx_frag(struct i40e_ring *rx_ring,
 			     struct i40e_rx_buffer *rx_buffer,
 			     struct sk_buff *skb,
 			     unsigned int size)
 {
-	struct page *page = rx_buffer->page;
-	unsigned char *va = page_address(page) + rx_buffer->page_offset;
 #if (PAGE_SIZE < 8192)
 	unsigned int truesize = I40E_RXBUFFER_2048;
 #else
-	unsigned int truesize = ALIGN(size, L1_CACHE_BYTES);
+	unsigned int truesize = SKB_DATA_ALIGN(size);
 #endif
-	unsigned int pull_len;
-
-	if (unlikely(skb_is_nonlinear(skb)))
-		goto add_tail_frag;
-
-	/* will the data fit in the skb we allocated? if so, just
-	 * copy it as it is pretty small anyway
-	 */
-	if (size <= I40E_RX_HDR_SIZE) {
-		memcpy(__skb_put(skb, size), va, ALIGN(size, sizeof(long)));
-
-		/* page is to be freed, increase pagecnt_bias instead of
-		 * decreasing page count.
-		 */
-		rx_buffer->pagecnt_bias++;
-		return;
-	}
-
-	/* 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, I40E_RX_HDR_SIZE);
-
-	/* align pull length to size of long to optimize
-	 * memcpy performance
-	 */
-	memcpy(__skb_put(skb, pull_len), va, ALIGN(pull_len, sizeof(long)));
-
-	/* update all of the pointers */
-	va += pull_len;
-	size -= pull_len;
 
-add_tail_frag:
-	skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, page,
-			(unsigned long)va & ~PAGE_MASK, size, truesize);
+	skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, rx_buffer->page,
+			rx_buffer->page_offset, size, truesize);
 
 	/* page is being used so we must update the page offset */
 #if (PAGE_SIZE < 8192)

@@ -1781,45 +1743,66 @@ static struct i40e_rx_buffer *i40e_get_rx_buffer(struct i40e_ring *rx_ring,
 }
 
 /**
- * i40e_fetch_rx_buffer - Allocate skb and populate it
+ * i40e_construct_skb - Allocate skb and populate it
  * @rx_ring: rx descriptor ring to transact packets on
  * @rx_buffer: rx buffer to pull data from
  * @size: size of buffer to add to skb
  *
- * 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.
+ * This function allocates an skb.  It then populates it with the page
+ * data from the current receive descriptor, taking care to set up the
+ * skb correctly.
  */
-static inline
-struct sk_buff *i40e_fetch_rx_buffer(struct i40e_ring *rx_ring,
-				     struct i40e_rx_buffer *rx_buffer,
-				     struct sk_buff *skb,
-				     unsigned int size)
+static struct sk_buff *i40e_construct_skb(struct i40e_ring *rx_ring,
+					  struct i40e_rx_buffer *rx_buffer,
+					  unsigned int size)
 {
-	if (likely(!skb)) {
-		void *page_addr = page_address(rx_buffer->page) +
-				  rx_buffer->page_offset;
+	void *va = page_address(rx_buffer->page) + rx_buffer->page_offset;
+#if (PAGE_SIZE < 8192)
+	unsigned int truesize = I40E_RXBUFFER_2048;
+#else
+	unsigned int truesize = SKB_DATA_ALIGN(size);
+#endif
+	unsigned int headlen;
+	struct sk_buff *skb;
 
-		/* prefetch first cache line of first page */
-		prefetch(page_addr);
+	/* prefetch first cache line of first page */
+	prefetch(va);
 #if L1_CACHE_BYTES < 128
-		prefetch(page_addr + L1_CACHE_BYTES);
+	prefetch(va + L1_CACHE_BYTES);
 #endif
 
-		/* allocate a skb to store the frags */
-		skb = __napi_alloc_skb(&rx_ring->q_vector->napi,
-				       I40E_RX_HDR_SIZE,
-				       GFP_ATOMIC | __GFP_NOWARN);
-		if (unlikely(!skb)) {
-			rx_ring->rx_stats.alloc_buff_failed++;
-			rx_buffer->pagecnt_bias++;
-			return NULL;
-		}
-	}
+	/* allocate a skb to store the frags */
+	skb = __napi_alloc_skb(&rx_ring->q_vector->napi,
+			       I40E_RX_HDR_SIZE,
+			       GFP_ATOMIC | __GFP_NOWARN);
+	if (unlikely(!skb))
+		return NULL;
+
+	/* Determine available headroom for copy */
+	headlen = size;
+	if (headlen > I40E_RX_HDR_SIZE)
+		headlen = eth_get_headlen(va, I40E_RX_HDR_SIZE);
 
-	/* pull page into skb */
-	i40e_add_rx_frag(rx_ring, rx_buffer, skb, size);
+	/* align pull length to size of long to optimize memcpy performance */
+	memcpy(__skb_put(skb, headlen), va, ALIGN(headlen, sizeof(long)));
+
+	/* update all of the pointers */
+	size -= headlen;
+	if (size) {
+		skb_add_rx_frag(skb, 0, rx_buffer->page,
+				rx_buffer->page_offset + headlen,
+				size, truesize);
+
+		/* buffer is used by skb, update page_offset */
+#if (PAGE_SIZE < 8192)
+		rx_buffer->page_offset ^= truesize;
+#else
+		rx_buffer->page_offset += truesize;
+#endif
+	} else {
+		/* buffer is unused, reset bias back to rx_buffer */
+		rx_buffer->pagecnt_bias++;
+	}
 
 	return skb;
 }

@@ -1944,9 +1927,18 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget)
 
 		rx_buffer = i40e_get_rx_buffer(rx_ring, size);
 
-		skb = i40e_fetch_rx_buffer(rx_ring, rx_buffer, skb, size);
-		if (!skb)
+		/* retrieve a buffer from the ring */
+		if (skb)
+			i40e_add_rx_frag(rx_ring, rx_buffer, skb, size);
+		else
+			skb = i40e_construct_skb(rx_ring, rx_buffer, size);
+
+		/* exit if we failed to retrieve a buffer */
+		if (!skb) {
+			rx_ring->rx_stats.alloc_buff_failed++;
+			rx_buffer->pagecnt_bias++;
 			break;
+		}
 
 		i40e_put_rx_buffer(rx_ring, rx_buffer);
 		cleaned_count++;

180 lines not on this page · raw message

index dcba258..561fe1f 100644
@@ -430,6 +430,7 @@ struct i40e_pf {
 #define I40E_FLAG_TEMP_LINK_POLLING		BIT_ULL(55)
 #define I40E_FLAG_CLIENT_L2_CHANGE		BIT_ULL(56)
 #define I40E_FLAG_WOL_MC_MAGIC_PKT_WAKE		BIT_ULL(57)
+#define I40E_FLAG_LEGACY_RX			BIT_ULL(57)
 
        /* Tracks features that are disabled due to hw limitations.
 	* If a bit is set here, it means that the corresponding

index 56dccf8..0db6b36 100644
@@ -226,6 +226,7 @@ static const struct i40e_priv_flags i40e_gstrings_priv_flags[] = {
 	I40E_PRIV_FLAG("flow-director-atr", I40E_FLAG_FD_ATR_ENABLED, 0),
 	I40E_PRIV_FLAG("veb-stats", I40E_FLAG_VEB_STATS_ENABLED, 0),
 	I40E_PRIV_FLAG("hw-atr-eviction", I40E_FLAG_HW_ATR_EVICT_CAPABLE, 0),
+	I40E_PRIV_FLAG("legacy-rx", I40E_FLAG_LEGACY_RX, 0),
 };
 
 #define I40E_PRIV_FLAGS_STR_LEN ARRAY_SIZE(i40e_gstrings_priv_flags)

@@ -4048,6 +4049,7 @@ static int i40e_set_priv_flags(struct net_device *dev, u32 flags)
 	}
 
 flags_complete:
+	/* check for flags that changed */
 	changed_flags ^= pf->flags;
 
 	/* Process any additional changes needed as a result of flag changes.

@@ -4088,7 +4090,8 @@ static int i40e_set_priv_flags(struct net_device *dev, u32 flags)
 	/* Issue reset to cause things to take effect, as additional bits
 	 * are added we will need to create a mask of bits requiring reset
 	 */
-	if (changed_flags & I40E_FLAG_VEB_STATS_ENABLED)
+	if ((changed_flags & I40E_FLAG_VEB_STATS_ENABLED) ||
+	    ((changed_flags & I40E_FLAG_LEGACY_RX) && netif_running(dev)))
 		i40e_do_reset(pf, BIT(__I40E_PF_RESET_REQUESTED));
 
 	return 0;

index b2b4851..e60cbfa 100644
@@ -222,6 +222,7 @@ struct i40evf_adapter {
 #define I40EVF_FLAG_CLIENT_NEEDS_L2_PARAMS	BIT(17)
 #define I40EVF_FLAG_PROMISC_ON			BIT(18)
 #define I40EVF_FLAG_ALLMULTI_ON			BIT(19)
+#define I40EVF_FLAG_LEGACY_RX			BIT(20)
 /* duplicates for common code */
 #define I40E_FLAG_FDIR_ATR_ENABLED		0
 #define I40E_FLAG_DCB_ENABLED			0

@@ -229,6 +230,7 @@ struct i40evf_adapter {
 #define I40E_FLAG_RX_CSUM_ENABLED		I40EVF_FLAG_RX_CSUM_ENABLED
 #define I40E_FLAG_WB_ON_ITR_CAPABLE		I40EVF_FLAG_WB_ON_ITR_CAPABLE
 #define I40E_FLAG_OUTER_UDP_CSUM_CAPABLE	I40EVF_FLAG_OUTER_UDP_CSUM_CAPABLE
+#define I40E_FLAG_LEGACY_RX			I40EVF_FLAG_LEGACY_RX
 	/* flags for admin queue service task */
 	u32 aq_required;
 #define I40EVF_FLAG_AQ_ENABLE_QUEUES		BIT(0)

135 lines not on this page · raw message

index 4cfe9491..724a5a4 100644
@@ -3002,7 +3002,8 @@ static int i40e_configure_rx_ring(struct i40e_ring *ring)
 
 	ring->rx_buf_len = vsi->rx_buf_len;
 
-	rx_ctx.dbuff = ring->rx_buf_len >> I40E_RXQ_CTX_DBUFF_SHIFT;
+	rx_ctx.dbuff = DIV_ROUND_UP(ring->rx_buf_len,
+				    BIT_ULL(I40E_RXQ_CTX_DBUFF_SHIFT));
 
 	rx_ctx.base = (ring->dma / 128);
 	rx_ctx.qlen = ring->count;

@@ -3082,17 +3083,18 @@ static int i40e_vsi_configure_rx(struct i40e_vsi *vsi)
 	int err = 0;
 	u16 i;
 
-	if (vsi->netdev && (vsi->netdev->mtu > ETH_DATA_LEN))
-		vsi->max_frame = vsi->netdev->mtu + ETH_HLEN
-			       + ETH_FCS_LEN + VLAN_HLEN;
-	else
-		vsi->max_frame = I40E_RXBUFFER_2048;
-
-	vsi->rx_buf_len = I40E_RXBUFFER_2048;
-
-	/* round up for the chip's needs */
-	vsi->rx_buf_len = ALIGN(vsi->rx_buf_len,
-				BIT_ULL(I40E_RXQ_CTX_DBUFF_SHIFT));
+	if (!vsi->netdev || (vsi->back->flags & I40E_FLAG_LEGACY_RX)) {
+		vsi->max_frame = I40E_MAX_RXBUFFER;
+		vsi->rx_buf_len = I40E_RXBUFFER_2048;
+#if (PAGE_SIZE < 8192)
+	} else if (vsi->netdev->mtu <= ETH_DATA_LEN) {
+		vsi->max_frame = I40E_RXBUFFER_1536 - NET_IP_ALIGN;
+		vsi->rx_buf_len = I40E_RXBUFFER_1536 - NET_IP_ALIGN;
+#endif
+	} else {
+		vsi->max_frame = I40E_MAX_RXBUFFER;
+		vsi->rx_buf_len = I40E_RXBUFFER_2048;
+	}
 
 	/* set up individual rings */
 	for (i = 0; i < vsi->num_queue_pairs && !err; i++)

index eb73372..d6609de 100644
@@ -117,10 +117,8 @@ enum i40e_dyn_idx_t {
 
 /* Supported Rx Buffer Sizes (a multiple of 128) */
 #define I40E_RXBUFFER_256   256
+#define I40E_RXBUFFER_1536  1536  /* 128B aligned standard Ethernet frame */
 #define I40E_RXBUFFER_2048  2048
-#define I40E_RXBUFFER_3072  3072   /* For FCoE MTU of 2158 */
-#define I40E_RXBUFFER_4096  4096
-#define I40E_RXBUFFER_8192  8192
 #define I40E_MAX_RXBUFFER   9728  /* largest size for single descriptor */
 
 /* NOTE: netdev_alloc_skb reserves up to 64 bytes, NET_IP_ALIGN means we

index aba40ed..3bb4d73 100644
@@ -104,10 +104,8 @@ enum i40e_dyn_idx_t {
 
 /* Supported Rx Buffer Sizes (a multiple of 128) */
 #define I40E_RXBUFFER_256   256
+#define I40E_RXBUFFER_1536  1536  /* 128B aligned standard Ethernet frame */
 #define I40E_RXBUFFER_2048  2048
-#define I40E_RXBUFFER_3072  3072   /* For FCoE MTU of 2158 */
-#define I40E_RXBUFFER_4096  4096
-#define I40E_RXBUFFER_8192  8192
 #define I40E_MAX_RXBUFFER   9728  /* largest size for single descriptor */
 
 /* NOTE: netdev_alloc_skb reserves up to 64 bytes, NET_IP_ALIGN means we

index e60cbfa..d61ecf6 100644
@@ -72,10 +72,6 @@ struct i40e_vsi {
 #define I40EVF_MAX_RXD		4096
 #define I40EVF_MIN_RXD		64
 #define I40EVF_REQ_DESCRIPTOR_MULTIPLE	32
-
-/* Supported Rx Buffer Sizes */
-#define I40EVF_RXBUFFER_2048	2048
-#define I40EVF_MAX_RXBUFFER	16384  /* largest size for single descriptor */
 #define I40EVF_MAX_AQ_BUF_SIZE	4096
 #define I40EVF_AQ_LEN		32
 #define I40EVF_AQ_MAX_ERR	20 /* times to try before resetting AQ */

index 6d666bd..fb2811c 100644
@@ -686,12 +686,26 @@ static void i40evf_configure_tx(struct i40evf_adapter *adapter)
  **/
 static void i40evf_configure_rx(struct i40evf_adapter *adapter)
 {
+	unsigned int rx_buf_len = I40E_RXBUFFER_2048;
+	struct net_device *netdev = adapter->netdev;
 	struct i40e_hw *hw = &adapter->hw;
 	int i;
 
+	/* Legacy Rx will always default to a 2048 buffer size. */
+#if (PAGE_SIZE < 8192)
+	if (!(adapter->flags & I40EVF_FLAG_LEGACY_RX)) {
+		/* We use a 1536 buffer size for configurations with
+		 * standard Ethernet mtu.  On x86 this gives us enough room
+		 * for shared info and 192 bytes of padding.
+		 */
+		if (netdev->mtu <= ETH_DATA_LEN)
+			rx_buf_len = I40E_RXBUFFER_1536 - NET_IP_ALIGN;
+	}
+#endif
+
 	for (i = 0; i < adapter->num_active_queues; i++) {
 		adapter->rx_rings[i].tail = hw->hw_addr + I40E_QRX_TAIL1(i);
-		adapter->rx_rings[i].rx_buf_len = I40EVF_RXBUFFER_2048;
+		adapter->rx_rings[i].rx_buf_len = rx_buf_len;
 	}
 }
 

index 4bc2488..032be8d 100644
@@ -234,7 +234,7 @@ void i40evf_configure_queues(struct i40evf_adapter *adapter)
 	struct i40e_virtchnl_vsi_queue_config_info *vqci;
 	struct i40e_virtchnl_queue_pair_info *vqpi;
 	int pairs = adapter->num_active_queues;
-	int i, len;
+	int i, len, max_frame = I40E_MAX_RXBUFFER;
 
 	if (adapter->current_op != I40E_VIRTCHNL_OP_UNKNOWN) {
 		/* bail because we already have a command pending */

@@ -249,6 +249,11 @@ void i40evf_configure_queues(struct i40evf_adapter *adapter)
 	if (!vqci)
 		return;
 
+	/* Limit maximum frame size when jumbo frames is not enabled */
+	if (!(adapter->flags & I40EVF_FLAG_LEGACY_RX) &&
+	    (adapter->netdev->mtu <= ETH_DATA_LEN))
+		max_frame = I40E_RXBUFFER_1536 - NET_IP_ALIGN;
+
 	vqci->vsi_id = adapter->vsi_res->vsi_id;
 	vqci->num_queue_pairs = pairs;
 	vqpi = vqci->qpair;

@@ -264,9 +269,10 @@ void i40evf_configure_queues(struct i40evf_adapter *adapter)
 		vqpi->rxq.queue_id = i;
 		vqpi->rxq.ring_len = adapter->rx_rings[i].count;
 		vqpi->rxq.dma_ring_addr = adapter->rx_rings[i].dma;
-		vqpi->rxq.max_pkt_size = adapter->netdev->mtu
-					+ ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN;
-		vqpi->rxq.databuffer_size = adapter->rx_rings[i].rx_buf_len;
+		vqpi->rxq.max_pkt_size = max_frame;
+		vqpi->rxq.databuffer_size =
+			ALIGN(adapter->rx_rings[i].rx_buf_len,
+			      BIT_ULL(I40E_RXQ_CTX_DBUFF_SHIFT));
 		vqpi++;
 	}