[net-next 00/15][pull request] 10GbE Intel Wired LAN Driver Updates 2018-01-26

[net-next 00/15][pull request] 10GbE Intel Wired LAN Driver Updates 2018-01-26

[Jeff Kirsher]

This series contains updates to ixgbe and ixgbevf.

Emil updates ixgbevf to match ixgbe functionality, starting with the
consolidating of functions that represent logical steps in the receive
process so we can later update them more easily.  Updated ixgbevf to
only synchronize the length of the frame, which will typically be the
MTU or smaller.  Updated the VF driver to use the length of the packet
instead of the DD status bit to determine if a new descriptor is ready
to be processed, which saves on reads and we can save time on
initialization.  Added support for DMA_ATTR_SKIP_CPU_SYNC/WEAK_ORDERING
to help improve performance on some platforms.  Updated the VF driver to
do bulk updates of the page reference count instead of just incrementing
it by one reference at a time.  Updated the VF driver to only go through
the region of the receive ring that was designated to be cleaned up,
rather than process the entire ring.

Colin Ian King adds the use of ARRAY_SIZE() on various arrays.

Miroslav Lichvar fixes an issue where ethtool was reporting timestamping
filters unsupported for X550, which is incorrect.

Paul adds support for reporting 5G link speed for some devices.

Dan Carpenter fixes a typo where && was used when it should have been
||.

The following are changes since commit 9515a2e082f91457db0ecff4b65371d0fb5d9aad:
  net/ipv4: Allow send to local broadcast from a socket bound to a VRF
and are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queue 10GbE

Colin Ian King (2):
  ixgbevf: use ARRAY_SIZE for various array sizing calculations
  ixgbe: use ARRAY_SIZE for array sizing calculation on array buf

Dan Carpenter (1):
  ixgbe: Fix && vs || typo

Emil Tantilov (10):
  ixgbevf: add function for checking if we can reuse page
  ixgbevf: only DMA sync frame length
  ixgbevf: use length to determine if descriptor is done
  ixgbevf: add support for DMA_ATTR_SKIP_CPU_SYNC/WEAK_ORDERING
  ixgbevf: update code to better handle incrementing page count
  ixgbevf: add counters for Rx page allocations
  ixgbevf: clear rx_buffer_info in configure instead of clean
  ixgbevf: improve performance and reduce size of ixgbevf_tx_map()
  ixgbevf: don't bother clearing tx_buffer_info in
    ixgbevf_clean_tx_ring()
  ixgbe: don't set RXDCTL.RLPML for 82599

Miroslav Lichvar (1):
  ixgbe: Don't report unsupported timestamping filters for X550

Paul Greenwalt (1):
  ixgbe: add support for reporting 5G link speed

 drivers/net/ethernet/intel/ixgbe/ixgbe_common.c   |   2 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c  |  37 +--
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c     |  11 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c     |   2 +-
 drivers/net/ethernet/intel/ixgbevf/ethtool.c      |   3 +
 drivers/net/ethernet/intel/ixgbevf/ixgbevf.h      |  16 +-
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 362 +++++++++++++---------
 drivers/net/ethernet/intel/ixgbevf/vf.c           |  17 +-
 8 files changed, 271 insertions(+), 179 deletions(-)

-- 
2.14.3

[net-next 01/15] ixgbevf: add function for checking if we can reuse page

[Jeff Kirsher]

From: Emil Tantilov <emil.s.tantilov@intel.com>

Introduce ixgbevf_can_reuse_page() similar to the change in ixgbe from
commit af43da0dba0b
("ixgbe: Add function for checking to see if we can reuse page")

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 59 +++++++++++++----------
 1 file changed, 33 insertions(+), 26 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index ed5c3aea7939..7ffd429d8e40 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -754,6 +754,38 @@ static inline bool ixgbevf_page_is_reserved(struct page *page)
 	return (page_to_nid(page) != numa_mem_id()) || page_is_pfmemalloc(page);
 }
 
+static bool ixgbevf_can_reuse_rx_page(struct ixgbevf_rx_buffer *rx_buffer,
+				      struct page *page,
+				      const unsigned int truesize)
+{
+	/* avoid re-using remote pages */
+	if (unlikely(ixgbevf_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_buffer->page_offset ^= IXGBEVF_RX_BUFSZ;
+
+#else
+	/* move offset up to the next cache line */
+	rx_buffer->page_offset += truesize;
+
+	if (rx_buffer->page_offset > (PAGE_SIZE - IXGBEVF_RX_BUFSZ))
+		return false;
+
+#endif
+	/* Even if we own the page, we are not allowed to use atomic_set()
+	 * This would break get_page_unless_zero() users.
+	 */
+	page_ref_inc(page);
+
+	return true;
+}
+
 /**
  * ixgbevf_add_rx_frag - Add contents of Rx buffer to sk_buff
  * @rx_ring: rx descriptor ring to transact packets on
@@ -815,32 +847,7 @@ static bool ixgbevf_add_rx_frag(struct ixgbevf_ring *rx_ring,
 	skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, page,
 			(unsigned long)va & ~PAGE_MASK, size, truesize);
 
-	/* avoid re-using remote pages */
-	if (unlikely(ixgbevf_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_buffer->page_offset ^= IXGBEVF_RX_BUFSZ;
-
-#else
-	/* move offset up to the next cache line */
-	rx_buffer->page_offset += truesize;
-
-	if (rx_buffer->page_offset > (PAGE_SIZE - IXGBEVF_RX_BUFSZ))
-		return false;
-
-#endif
-	/* Even if we own the page, we are not allowed to use atomic_set()
-	 * This would break get_page_unless_zero() users.
-	 */
-	page_ref_inc(page);
-
-	return true;
+	return ixgbevf_can_reuse_rx_page(rx_buffer, page, truesize);
 }
 
 static struct sk_buff *ixgbevf_fetch_rx_buffer(struct ixgbevf_ring *rx_ring,
-- 
2.14.3

[net-next 02/15] ixgbevf: only DMA sync frame length

[Jeff Kirsher]

From: Emil Tantilov <emil.s.tantilov@intel.com>

Based on commit 64f2525ca4e7 ("igb: Only DMA sync frame length")

On some architectures synching a buffer for DMA may be expensive.
Instead of the entire 2K receive buffer only synchronize the length of
the frame, which will typically be the MTU or smaller.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 7ffd429d8e40..0cc2688671ec 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -803,12 +803,12 @@ static bool ixgbevf_can_reuse_rx_page(struct ixgbevf_rx_buffer *rx_buffer,
  **/
 static bool ixgbevf_add_rx_frag(struct ixgbevf_ring *rx_ring,
 				struct ixgbevf_rx_buffer *rx_buffer,
+				u16 size,
 				union ixgbe_adv_rx_desc *rx_desc,
 				struct sk_buff *skb)
 {
 	struct page *page = rx_buffer->page;
 	unsigned char *va = page_address(page) + rx_buffer->page_offset;
-	unsigned int size = le16_to_cpu(rx_desc->wb.upper.length);
 #if (PAGE_SIZE < 8192)
 	unsigned int truesize = IXGBEVF_RX_BUFSZ;
 #else
@@ -856,6 +856,7 @@ static struct sk_buff *ixgbevf_fetch_rx_buffer(struct ixgbevf_ring *rx_ring,
 {
 	struct ixgbevf_rx_buffer *rx_buffer;
 	struct page *page;
+	u16 size = le16_to_cpu(rx_desc->wb.upper.length);
 
 	rx_buffer = &rx_ring->rx_buffer_info[rx_ring->next_to_clean];
 	page = rx_buffer->page;
@@ -890,11 +891,11 @@ static struct sk_buff *ixgbevf_fetch_rx_buffer(struct ixgbevf_ring *rx_ring,
 	dma_sync_single_range_for_cpu(rx_ring->dev,
 				      rx_buffer->dma,
 				      rx_buffer->page_offset,
-				      IXGBEVF_RX_BUFSZ,
+				      size,
 				      DMA_FROM_DEVICE);
 
 	/* pull page into skb */
-	if (ixgbevf_add_rx_frag(rx_ring, rx_buffer, rx_desc, skb)) {
+	if (ixgbevf_add_rx_frag(rx_ring, rx_buffer, size, rx_desc, skb)) {
 		/* hand second half of page back to the ring */
 		ixgbevf_reuse_rx_page(rx_ring, rx_buffer);
 	} else {
-- 
2.14.3

[net-next 03/15] ixgbevf: use length to determine if descriptor is done

[Jeff Kirsher]

From: Emil Tantilov <emil.s.tantilov@intel.com>

Based on:
commit 7ec0116c9131 ("igb: Use length to determine if descriptor is done")

This change makes it so that we use the length of the packet instead of the
DD status bit to determine if a new descriptor is ready to be processed.
The obvious advantage is that it cuts down on reads as we don't really even
need the DD bit if going from a 0 to a non-zero value on size is enough to
inform us that the packet has been completed.

In addition we only reset the Rx descriptor length for descriptor zero when
resetting a ring instead of having to do a memset with 0 over the entire
ring. By doing this we can save some time on initialization.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 0cc2688671ec..725fe2dca868 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -653,8 +653,8 @@ static void ixgbevf_alloc_rx_buffers(struct ixgbevf_ring *rx_ring,
 			i -= rx_ring->count;
 		}
 
-		/* clear the hdr_addr for the next_to_use descriptor */
-		rx_desc->read.hdr_addr = 0;
+		/* clear the length for the next_to_use descriptor */
+		rx_desc->wb.upper.length = 0;
 
 		cleaned_count--;
 	} while (cleaned_count);
@@ -938,7 +938,7 @@ static int ixgbevf_clean_rx_irq(struct ixgbevf_q_vector *q_vector,
 
 		rx_desc = IXGBEVF_RX_DESC(rx_ring, rx_ring->next_to_clean);
 
-		if (!ixgbevf_test_staterr(rx_desc, IXGBE_RXD_STAT_DD))
+		if (!rx_desc->wb.upper.length)
 			break;
 
 		/* This memory barrier is needed to keep us from reading
@@ -1729,6 +1729,7 @@ static void ixgbevf_configure_rx_ring(struct ixgbevf_adapter *adapter,
 				      struct ixgbevf_ring *ring)
 {
 	struct ixgbe_hw *hw = &adapter->hw;
+	union ixgbe_adv_rx_desc *rx_desc;
 	u64 rdba = ring->dma;
 	u32 rxdctl;
 	u8 reg_idx = ring->reg_idx;
@@ -1757,6 +1758,10 @@ static void ixgbevf_configure_rx_ring(struct ixgbevf_adapter *adapter,
 	IXGBE_WRITE_REG(hw, IXGBE_VFRDT(reg_idx), 0);
 	ring->tail = adapter->io_addr + IXGBE_VFRDT(reg_idx);
 
+	/* initialize Rx descriptor 0 */
+	rx_desc = IXGBEVF_RX_DESC(ring, 0);
+	rx_desc->wb.upper.length = 0;
+
 	/* reset ntu and ntc to place SW in sync with hardwdare */
 	ring->next_to_clean = 0;
 	ring->next_to_use = 0;
@@ -2141,9 +2146,6 @@ static void ixgbevf_clean_rx_ring(struct ixgbevf_ring *rx_ring)
 
 	size = sizeof(struct ixgbevf_rx_buffer) * rx_ring->count;
 	memset(rx_ring->rx_buffer_info, 0, size);
-
-	/* Zero out the descriptor ring */
-	memset(rx_ring->desc, 0, rx_ring->size);
 }
 
 /**
-- 
2.14.3

[net-next 04/15] ixgbevf: add support for DMA_ATTR_SKIP_CPU_SYNC/WEAK_ORDERING

[Jeff Kirsher]

From: Emil Tantilov <emil.s.tantilov@intel.com>

Based on commit 5be5955425c2
("igb: update driver to make use of DMA_ATTR_SKIP_CPU_SYNC")
and
commit 7bd175928280 ("igb: Add support for DMA_ATTR_WEAK_ORDERING")

Convert the calls to dma_map/unmap_page() to the attributes version
and add DMA_ATTR_SKIP_CPU_SYNC/WEAK_ORDERING which should help
improve performance on some platforms.

Move sync_for_cpu call before we perform a prefetch to avoid
invalidating the first 128 bytes of the packet on architectures where
that call may invalidate the cache.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbevf/ixgbevf.h      |  3 ++
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 57 ++++++++++++++---------
 2 files changed, 38 insertions(+), 22 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
index 581f44bbd7b3..b1da9f41c1dc 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
@@ -260,6 +260,9 @@ static inline void ixgbevf_write_tail(struct ixgbevf_ring *ring, u32 value)
 #define MIN_MSIX_Q_VECTORS	1
 #define MIN_MSIX_COUNT		(MIN_MSIX_Q_VECTORS + NON_Q_VECTORS)
 
+#define IXGBEVF_RX_DMA_ATTR \
+	(DMA_ATTR_SKIP_CPU_SYNC | DMA_ATTR_WEAK_ORDERING)
+
 /* board specific private data structure */
 struct ixgbevf_adapter {
 	/* this field must be first, see ixgbevf_process_skb_fields */
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 725fe2dca868..fbd493efd14e 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -595,8 +595,8 @@ static bool ixgbevf_alloc_mapped_page(struct ixgbevf_ring *rx_ring,
 	}
 
 	/* map page for use */
-	dma = dma_map_page(rx_ring->dev, page, 0,
-			   PAGE_SIZE, DMA_FROM_DEVICE);
+	dma = dma_map_page_attrs(rx_ring->dev, page, 0, PAGE_SIZE,
+				 DMA_FROM_DEVICE, IXGBEVF_RX_DMA_ATTR);
 
 	/* if mapping failed free memory back to system since
 	 * there isn't much point in holding memory we can't use
@@ -639,6 +639,12 @@ static void ixgbevf_alloc_rx_buffers(struct ixgbevf_ring *rx_ring,
 		if (!ixgbevf_alloc_mapped_page(rx_ring, bi))
 			break;
 
+		/* sync the buffer for use by the device */
+		dma_sync_single_range_for_device(rx_ring->dev, bi->dma,
+						 bi->page_offset,
+						 IXGBEVF_RX_BUFSZ,
+						 DMA_FROM_DEVICE);
+
 		/* Refresh the desc even if pkt_addr didn't change
 		 * because each write-back erases this info.
 		 */
@@ -741,12 +747,6 @@ static void ixgbevf_reuse_rx_page(struct ixgbevf_ring *rx_ring,
 	new_buff->page = old_buff->page;
 	new_buff->dma = old_buff->dma;
 	new_buff->page_offset = old_buff->page_offset;
-
-	/* sync the buffer for use by the device */
-	dma_sync_single_range_for_device(rx_ring->dev, new_buff->dma,
-					 new_buff->page_offset,
-					 IXGBEVF_RX_BUFSZ,
-					 DMA_FROM_DEVICE);
 }
 
 static inline bool ixgbevf_page_is_reserved(struct page *page)
@@ -862,6 +862,13 @@ static struct sk_buff *ixgbevf_fetch_rx_buffer(struct ixgbevf_ring *rx_ring,
 	page = rx_buffer->page;
 	prefetchw(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);
+
 	if (likely(!skb)) {
 		void *page_addr = page_address(page) +
 				  rx_buffer->page_offset;
@@ -887,21 +894,15 @@ static struct sk_buff *ixgbevf_fetch_rx_buffer(struct ixgbevf_ring *rx_ring,
 		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 (ixgbevf_add_rx_frag(rx_ring, rx_buffer, size, rx_desc, skb)) {
 		/* hand second half of page back to the ring */
 		ixgbevf_reuse_rx_page(rx_ring, rx_buffer);
 	} else {
 		/* we are not reusing the buffer so unmap it */
-		dma_unmap_page(rx_ring->dev, rx_buffer->dma,
-			       PAGE_SIZE, DMA_FROM_DEVICE);
+		dma_unmap_page_attrs(rx_ring->dev, rx_buffer->dma,
+				     PAGE_SIZE, DMA_FROM_DEVICE,
+				     IXGBEVF_RX_DMA_ATTR);
 	}
 
 	/* clear contents of buffer_info */
@@ -2116,7 +2117,6 @@ void ixgbevf_up(struct ixgbevf_adapter *adapter)
  **/
 static void ixgbevf_clean_rx_ring(struct ixgbevf_ring *rx_ring)
 {
-	struct device *dev = rx_ring->dev;
 	unsigned long size;
 	unsigned int i;
 
@@ -2135,10 +2135,23 @@ static void ixgbevf_clean_rx_ring(struct ixgbevf_ring *rx_ring)
 		struct ixgbevf_rx_buffer *rx_buffer;
 
 		rx_buffer = &rx_ring->rx_buffer_info[i];
-		if (rx_buffer->dma)
-			dma_unmap_page(dev, rx_buffer->dma,
-				       PAGE_SIZE, DMA_FROM_DEVICE);
-		rx_buffer->dma = 0;
+
+		/* Invalidate cache lines that may have been written to by
+		 * device so that we avoid corrupting memory.
+		 */
+		dma_sync_single_range_for_cpu(rx_ring->dev,
+					      rx_buffer->dma,
+					      rx_buffer->page_offset,
+					      IXGBEVF_RX_BUFSZ,
+					      DMA_FROM_DEVICE);
+
+		/* free resources associated with mapping */
+		dma_unmap_page_attrs(rx_ring->dev,
+				     rx_buffer->dma,
+				     PAGE_SIZE,
+				     DMA_FROM_DEVICE,
+				     IXGBEVF_RX_DMA_ATTR);
+
 		if (rx_buffer->page)
 			__free_page(rx_buffer->page);
 		rx_buffer->page = NULL;
-- 
2.14.3

[net-next 05/15] ixgbevf: update code to better handle incrementing page count

[Jeff Kirsher]

From: Emil Tantilov <emil.s.tantilov@intel.com>

Based on commit bd4171a5d4c2
("igb: update code to better handle incrementing page count")

Update the driver code so that we do bulk updates of the page reference
count instead of just incrementing it by one reference at a time.  The
advantage to doing this is that we cut down on atomic operations and
this in turn should give us a slight improvement in cycles per packet.
In addition if we eventually move this over to using build_skb the gains
will be more noticeable.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbevf/ixgbevf.h      |  7 +++++-
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 30 +++++++++++++++++------
 2 files changed, 28 insertions(+), 9 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
index b1da9f41c1dc..c70a789035ae 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
@@ -62,7 +62,12 @@ struct ixgbevf_tx_buffer {
 struct ixgbevf_rx_buffer {
 	dma_addr_t dma;
 	struct page *page;
-	unsigned int page_offset;
+#if (BITS_PER_LONG > 32) || (PAGE_SIZE >= 65536)
+	__u32 page_offset;
+#else
+	__u16 page_offset;
+#endif
+	__u16 pagecnt_bias;
 };
 
 struct ixgbevf_stats {
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index fbd493efd14e..ae2402ddd9fb 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -611,6 +611,7 @@ static bool ixgbevf_alloc_mapped_page(struct ixgbevf_ring *rx_ring,
 	bi->dma = dma;
 	bi->page = page;
 	bi->page_offset = 0;
+	bi->pagecnt_bias = 1;
 
 	return true;
 }
@@ -747,6 +748,7 @@ static void ixgbevf_reuse_rx_page(struct ixgbevf_ring *rx_ring,
 	new_buff->page = old_buff->page;
 	new_buff->dma = old_buff->dma;
 	new_buff->page_offset = old_buff->page_offset;
+	new_buff->pagecnt_bias = old_buff->pagecnt_bias;
 }
 
 static inline bool ixgbevf_page_is_reserved(struct page *page)
@@ -758,13 +760,15 @@ static bool ixgbevf_can_reuse_rx_page(struct ixgbevf_rx_buffer *rx_buffer,
 				      struct page *page,
 				      const unsigned int truesize)
 {
+	unsigned int pagecnt_bias = rx_buffer->pagecnt_bias--;
+
 	/* avoid re-using remote pages */
 	if (unlikely(ixgbevf_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))
+	if (unlikely(page_ref_count(page) != pagecnt_bias))
 		return false;
 
 	/* flip page offset to other buffer */
@@ -778,10 +782,15 @@ static bool ixgbevf_can_reuse_rx_page(struct ixgbevf_rx_buffer *rx_buffer,
 		return false;
 
 #endif
-	/* Even if we own the page, we are not allowed to use atomic_set()
-	 * This would break get_page_unless_zero() users.
+
+	/* If we have drained the page fragment pool we need to update
+	 * the pagecnt_bias and page count so that we fully restock the
+	 * number of references the driver holds.
 	 */
-	page_ref_inc(page);
+	if (unlikely(pagecnt_bias == 1)) {
+		page_ref_add(page, USHRT_MAX);
+		rx_buffer->pagecnt_bias = USHRT_MAX;
+	}
 
 	return true;
 }
@@ -827,7 +836,6 @@ static bool ixgbevf_add_rx_frag(struct ixgbevf_ring *rx_ring,
 			return true;
 
 		/* this page cannot be reused so discard it */
-		put_page(page);
 		return false;
 	}
 
@@ -899,10 +907,13 @@ static struct sk_buff *ixgbevf_fetch_rx_buffer(struct ixgbevf_ring *rx_ring,
 		/* hand second half of page back to the ring */
 		ixgbevf_reuse_rx_page(rx_ring, rx_buffer);
 	} else {
-		/* we are not reusing the buffer so unmap it */
+		/* We are not reusing the buffer so unmap it and free
+		 * any references we are holding to it
+		 */
 		dma_unmap_page_attrs(rx_ring->dev, rx_buffer->dma,
 				     PAGE_SIZE, DMA_FROM_DEVICE,
 				     IXGBEVF_RX_DMA_ATTR);
+		__page_frag_cache_drain(page, rx_buffer->pagecnt_bias);
 	}
 
 	/* clear contents of buffer_info */
@@ -2135,6 +2146,8 @@ static void ixgbevf_clean_rx_ring(struct ixgbevf_ring *rx_ring)
 		struct ixgbevf_rx_buffer *rx_buffer;
 
 		rx_buffer = &rx_ring->rx_buffer_info[i];
+		if (!rx_buffer->page)
+			continue;
 
 		/* Invalidate cache lines that may have been written to by
 		 * device so that we avoid corrupting memory.
@@ -2152,8 +2165,9 @@ static void ixgbevf_clean_rx_ring(struct ixgbevf_ring *rx_ring)
 				     DMA_FROM_DEVICE,
 				     IXGBEVF_RX_DMA_ATTR);
 
-		if (rx_buffer->page)
-			__free_page(rx_buffer->page);
+		__page_frag_cache_drain(rx_buffer->page,
+					rx_buffer->pagecnt_bias);
+
 		rx_buffer->page = NULL;
 	}
 
-- 
2.14.3

[net-next 06/15] ixgbevf: add counters for Rx page allocations

[Jeff Kirsher]

From: Emil Tantilov <emil.s.tantilov@intel.com>

We already had placehloders for failed page and buffer allocations.
Added alloc_rx_page and made sure the stats are properly updated and
exposed in ethtool.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbevf/ethtool.c      |  3 +++
 drivers/net/ethernet/intel/ixgbevf/ixgbevf.h      |  6 ++++--
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 23 ++++++++++++++++++-----
 3 files changed, 25 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbevf/ethtool.c b/drivers/net/ethernet/intel/ixgbevf/ethtool.c
index ff9d05f308ee..4400e49090b4 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ethtool.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ethtool.c
@@ -75,6 +75,9 @@ static struct ixgbe_stats ixgbevf_gstrings_stats[] = {
 	IXGBEVF_STAT("tx_timeout_count", tx_timeout_count),
 	IXGBEVF_NETDEV_STAT(multicast),
 	IXGBEVF_STAT("rx_csum_offload_errors", hw_csum_rx_error),
+	IXGBEVF_STAT("alloc_rx_page", alloc_rx_page),
+	IXGBEVF_STAT("alloc_rx_page_failed", alloc_rx_page_failed),
+	IXGBEVF_STAT("alloc_rx_buff_failed", alloc_rx_buff_failed),
 };
 
 #define IXGBEVF_QUEUE_STATS_LEN ( \
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
index c70a789035ae..f6952425c87d 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
@@ -84,6 +84,7 @@ struct ixgbevf_tx_queue_stats {
 struct ixgbevf_rx_queue_stats {
 	u64 alloc_rx_page_failed;
 	u64 alloc_rx_buff_failed;
+	u64 alloc_rx_page;
 	u64 csum_err;
 };
 
@@ -295,8 +296,9 @@ struct ixgbevf_adapter {
 	u64 hw_csum_rx_error;
 	u64 hw_rx_no_dma_resources;
 	int num_msix_vectors;
-	u32 alloc_rx_page_failed;
-	u32 alloc_rx_buff_failed;
+	u64 alloc_rx_page_failed;
+	u64 alloc_rx_buff_failed;
+	u64 alloc_rx_page;
 
 	struct msix_entry *msix_entries;
 
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index ae2402ddd9fb..350afec3dde8 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -604,7 +604,7 @@ static bool ixgbevf_alloc_mapped_page(struct ixgbevf_ring *rx_ring,
 	if (dma_mapping_error(rx_ring->dev, dma)) {
 		__free_page(page);
 
-		rx_ring->rx_stats.alloc_rx_buff_failed++;
+		rx_ring->rx_stats.alloc_rx_page_failed++;
 		return false;
 	}
 
@@ -612,6 +612,7 @@ static bool ixgbevf_alloc_mapped_page(struct ixgbevf_ring *rx_ring,
 	bi->page = page;
 	bi->page_offset = 0;
 	bi->pagecnt_bias = 1;
+	rx_ring->rx_stats.alloc_rx_page++;
 
 	return true;
 }
@@ -963,8 +964,10 @@ static int ixgbevf_clean_rx_irq(struct ixgbevf_q_vector *q_vector,
 		skb = ixgbevf_fetch_rx_buffer(rx_ring, rx_desc, skb);
 
 		/* exit if we failed to retrieve a buffer */
-		if (!skb)
+		if (!skb) {
+			rx_ring->rx_stats.alloc_rx_buff_failed++;
 			break;
+		}
 
 		cleaned_count++;
 
@@ -2749,6 +2752,8 @@ static int ixgbevf_sw_init(struct ixgbevf_adapter *adapter)
 void ixgbevf_update_stats(struct ixgbevf_adapter *adapter)
 {
 	struct ixgbe_hw *hw = &adapter->hw;
+	u64 alloc_rx_page_failed = 0, alloc_rx_buff_failed = 0;
+	u64 alloc_rx_page = 0, hw_csum_rx_error = 0;
 	int i;
 
 	if (test_bit(__IXGBEVF_DOWN, &adapter->state) ||
@@ -2769,10 +2774,18 @@ void ixgbevf_update_stats(struct ixgbevf_adapter *adapter)
 				adapter->stats.vfmprc);
 
 	for (i = 0;  i  < adapter->num_rx_queues;  i++) {
-		adapter->hw_csum_rx_error +=
-			adapter->rx_ring[i]->hw_csum_rx_error;
-		adapter->rx_ring[i]->hw_csum_rx_error = 0;
+		struct ixgbevf_ring *rx_ring = adapter->rx_ring[i];
+
+		hw_csum_rx_error += rx_ring->rx_stats.csum_err;
+		alloc_rx_page_failed += rx_ring->rx_stats.alloc_rx_page_failed;
+		alloc_rx_buff_failed += rx_ring->rx_stats.alloc_rx_buff_failed;
+		alloc_rx_page += rx_ring->rx_stats.alloc_rx_page;
 	}
+
+	adapter->hw_csum_rx_error = hw_csum_rx_error;
+	adapter->alloc_rx_page_failed = alloc_rx_page_failed;
+	adapter->alloc_rx_buff_failed = alloc_rx_buff_failed;
+	adapter->alloc_rx_page = alloc_rx_page;
 }
 
 /**
-- 
2.14.3

[net-next 07/15] ixgbevf: clear rx_buffer_info in configure instead of clean

[Jeff Kirsher]

From: Emil Tantilov <emil.s.tantilov@intel.com>

Based on commit d2bead576e67
("igb: Clear Rx buffer_info in configure instead of clean")

This change makes it so that instead of going through the entire ring on Rx
cleanup we only go through the region that was designated to be cleaned up
and stop when we reach the region where new allocations should start.

In addition we can avoid having to perform a memset on the Rx buffer_info
structures until we are about to start using the ring again.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 26 +++++++++++------------
 1 file changed, 13 insertions(+), 13 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 350afec3dde8..a793f9ea05e7 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -1773,6 +1773,10 @@ static void ixgbevf_configure_rx_ring(struct ixgbevf_adapter *adapter,
 	IXGBE_WRITE_REG(hw, IXGBE_VFRDT(reg_idx), 0);
 	ring->tail = adapter->io_addr + IXGBE_VFRDT(reg_idx);
 
+	/* initialize rx_buffer_info */
+	memset(ring->rx_buffer_info, 0,
+	       sizeof(struct ixgbevf_rx_buffer) * ring->count);
+
 	/* initialize Rx descriptor 0 */
 	rx_desc = IXGBEVF_RX_DESC(ring, 0);
 	rx_desc->wb.upper.length = 0;
@@ -2131,8 +2135,7 @@ void ixgbevf_up(struct ixgbevf_adapter *adapter)
  **/
 static void ixgbevf_clean_rx_ring(struct ixgbevf_ring *rx_ring)
 {
-	unsigned long size;
-	unsigned int i;
+	u16 i = rx_ring->next_to_clean;
 
 	/* Free Rx ring sk_buff */
 	if (rx_ring->skb) {
@@ -2140,17 +2143,11 @@ static void ixgbevf_clean_rx_ring(struct ixgbevf_ring *rx_ring)
 		rx_ring->skb = NULL;
 	}
 
-	/* ring already cleared, nothing to do */
-	if (!rx_ring->rx_buffer_info)
-		return;
-
 	/* Free all the Rx ring pages */
-	for (i = 0; i < rx_ring->count; i++) {
+	while (i != rx_ring->next_to_alloc) {
 		struct ixgbevf_rx_buffer *rx_buffer;
 
 		rx_buffer = &rx_ring->rx_buffer_info[i];
-		if (!rx_buffer->page)
-			continue;
 
 		/* Invalidate cache lines that may have been written to by
 		 * device so that we avoid corrupting memory.
@@ -2171,11 +2168,14 @@ static void ixgbevf_clean_rx_ring(struct ixgbevf_ring *rx_ring)
 		__page_frag_cache_drain(rx_buffer->page,
 					rx_buffer->pagecnt_bias);
 
-		rx_buffer->page = NULL;
+		i++;
+		if (i == rx_ring->count)
+			i = 0;
 	}
 
-	size = sizeof(struct ixgbevf_rx_buffer) * rx_ring->count;
-	memset(rx_ring->rx_buffer_info, 0, size);
+	rx_ring->next_to_alloc = 0;
+	rx_ring->next_to_clean = 0;
+	rx_ring->next_to_use = 0;
 }
 
 /**
@@ -3090,7 +3090,7 @@ int ixgbevf_setup_rx_resources(struct ixgbevf_ring *rx_ring)
 	int size;
 
 	size = sizeof(struct ixgbevf_rx_buffer) * rx_ring->count;
-	rx_ring->rx_buffer_info = vzalloc(size);
+	rx_ring->rx_buffer_info = vmalloc(size);
 	if (!rx_ring->rx_buffer_info)
 		goto err;
 
-- 
2.14.3

[net-next 08/15] ixgbevf: improve performance and reduce size of ixgbevf_tx_map()

[Jeff Kirsher]

From: Emil Tantilov <emil.s.tantilov@intel.com>

Based on commit ec718254cbfe
("ixgbe: Improve performance and reduce size of ixgbe_tx_map")

This change is meant to both improve the performance and reduce the size of
ixgbevf_tx_map().

Expand the work done in the main loop by pushing first into tx_buffer.
This allows us to pull in the dma_mapping_error check, the tx_buffer value
assignment, and the initial DMA value assignment to the Tx descriptor.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 45 ++++++++++-------------
 1 file changed, 20 insertions(+), 25 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index a793f9ea05e7..d3415ee38597 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -3532,34 +3532,37 @@ static void ixgbevf_tx_map(struct ixgbevf_ring *tx_ring,
 			   struct ixgbevf_tx_buffer *first,
 			   const u8 hdr_len)
 {
-	dma_addr_t dma;
 	struct sk_buff *skb = first->skb;
 	struct ixgbevf_tx_buffer *tx_buffer;
 	union ixgbe_adv_tx_desc *tx_desc;
-	struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[0];
-	unsigned int data_len = skb->data_len;
-	unsigned int size = skb_headlen(skb);
-	unsigned int paylen = skb->len - hdr_len;
+	struct skb_frag_struct *frag;
+	dma_addr_t dma;
+	unsigned int data_len, size;
 	u32 tx_flags = first->tx_flags;
-	__le32 cmd_type;
+	__le32 cmd_type = ixgbevf_tx_cmd_type(tx_flags);
 	u16 i = tx_ring->next_to_use;
 
 	tx_desc = IXGBEVF_TX_DESC(tx_ring, i);
 
-	ixgbevf_tx_olinfo_status(tx_desc, tx_flags, paylen);
-	cmd_type = ixgbevf_tx_cmd_type(tx_flags);
+	ixgbevf_tx_olinfo_status(tx_desc, tx_flags, skb->len - hdr_len);
+
+	size = skb_headlen(skb);
+	data_len = skb->data_len;
 
 	dma = dma_map_single(tx_ring->dev, skb->data, size, DMA_TO_DEVICE);
-	if (dma_mapping_error(tx_ring->dev, dma))
-		goto dma_error;
 
-	/* record length, and DMA address */
-	dma_unmap_len_set(first, len, size);
-	dma_unmap_addr_set(first, dma, dma);
+	tx_buffer = first;
 
-	tx_desc->read.buffer_addr = cpu_to_le64(dma);
+	for (frag = &skb_shinfo(skb)->frags[0];; frag++) {
+		if (dma_mapping_error(tx_ring->dev, dma))
+			goto dma_error;
+
+		/* record length, and DMA address */
+		dma_unmap_len_set(tx_buffer, len, size);
+		dma_unmap_addr_set(tx_buffer, dma, dma);
+
+		tx_desc->read.buffer_addr = cpu_to_le64(dma);
 
-	for (;;) {
 		while (unlikely(size > IXGBE_MAX_DATA_PER_TXD)) {
 			tx_desc->read.cmd_type_len =
 				cmd_type | cpu_to_le32(IXGBE_MAX_DATA_PER_TXD);
@@ -3570,12 +3573,12 @@ static void ixgbevf_tx_map(struct ixgbevf_ring *tx_ring,
 				tx_desc = IXGBEVF_TX_DESC(tx_ring, 0);
 				i = 0;
 			}
+			tx_desc->read.olinfo_status = 0;
 
 			dma += IXGBE_MAX_DATA_PER_TXD;
 			size -= IXGBE_MAX_DATA_PER_TXD;
 
 			tx_desc->read.buffer_addr = cpu_to_le64(dma);
-			tx_desc->read.olinfo_status = 0;
 		}
 
 		if (likely(!data_len))
@@ -3589,23 +3592,15 @@ static void ixgbevf_tx_map(struct ixgbevf_ring *tx_ring,
 			tx_desc = IXGBEVF_TX_DESC(tx_ring, 0);
 			i = 0;
 		}
+		tx_desc->read.olinfo_status = 0;
 
 		size = skb_frag_size(frag);
 		data_len -= size;
 
 		dma = skb_frag_dma_map(tx_ring->dev, frag, 0, size,
 				       DMA_TO_DEVICE);
-		if (dma_mapping_error(tx_ring->dev, dma))
-			goto dma_error;
 
 		tx_buffer = &tx_ring->tx_buffer_info[i];
-		dma_unmap_len_set(tx_buffer, len, size);
-		dma_unmap_addr_set(tx_buffer, dma, dma);
-
-		tx_desc->read.buffer_addr = cpu_to_le64(dma);
-		tx_desc->read.olinfo_status = 0;
-
-		frag++;
 	}
 
 	/* write last descriptor with RS and EOP bits */
-- 
2.14.3

[net-next 09/15] ixgbevf: don't bother clearing tx_buffer_info in ixgbevf_clean_tx_ring()

[Jeff Kirsher]

From: Emil Tantilov <emil.s.tantilov@intel.com>

In the case of the Tx rings we need to only clear the Tx buffer_info when
we are resetting the rings.  Ideally we do this when we configure the ring
to bring it back up instead of when we are taking it down in order to avoid
dirtying pages we don't need to.

In addition we don't need to clear the Tx descriptor ring since we will
fully repopulate it when we begin transmitting frames and next_to_watch can
be cleared to prevent the ring from being cleaned beyond that point instead
of needing to touch anything in the Tx descriptor ring.

Finally with these changes we can avoid having to reset the skb member of
the Tx buffer_info structure in the cleanup path since the skb will always
be associated with the first buffer which has next_to_watch set.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 115 ++++++++++++++--------
 1 file changed, 72 insertions(+), 43 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index d3415ee38597..9b3d43d28106 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -206,28 +206,6 @@ static void ixgbevf_set_ivar(struct ixgbevf_adapter *adapter, s8 direction,
 	}
 }
 
-static void ixgbevf_unmap_and_free_tx_resource(struct ixgbevf_ring *tx_ring,
-					struct ixgbevf_tx_buffer *tx_buffer)
-{
-	if (tx_buffer->skb) {
-		dev_kfree_skb_any(tx_buffer->skb);
-		if (dma_unmap_len(tx_buffer, len))
-			dma_unmap_single(tx_ring->dev,
-					 dma_unmap_addr(tx_buffer, dma),
-					 dma_unmap_len(tx_buffer, len),
-					 DMA_TO_DEVICE);
-	} else if (dma_unmap_len(tx_buffer, len)) {
-		dma_unmap_page(tx_ring->dev,
-			       dma_unmap_addr(tx_buffer, dma),
-			       dma_unmap_len(tx_buffer, len),
-			       DMA_TO_DEVICE);
-	}
-	tx_buffer->next_to_watch = NULL;
-	tx_buffer->skb = NULL;
-	dma_unmap_len_set(tx_buffer, len, 0);
-	/* tx_buffer must be completely set up in the transmit path */
-}
-
 static u64 ixgbevf_get_tx_completed(struct ixgbevf_ring *ring)
 {
 	return ring->stats.packets;
@@ -349,7 +327,6 @@ static bool ixgbevf_clean_tx_irq(struct ixgbevf_q_vector *q_vector,
 				 DMA_TO_DEVICE);
 
 		/* clear tx_buffer data */
-		tx_buffer->skb = NULL;
 		dma_unmap_len_set(tx_buffer, len, 0);
 
 		/* unmap remaining buffers */
@@ -1576,6 +1553,10 @@ static void ixgbevf_configure_tx_ring(struct ixgbevf_adapter *adapter,
 	txdctl |= (1u << 8) |    /* HTHRESH = 1 */
 		   32;           /* PTHRESH = 32 */
 
+	/* reinitialize tx_buffer_info */
+	memset(ring->tx_buffer_info, 0,
+	       sizeof(struct ixgbevf_tx_buffer) * ring->count);
+
 	clear_bit(__IXGBEVF_HANG_CHECK_ARMED, &ring->state);
 
 	IXGBE_WRITE_REG(hw, IXGBE_VFTXDCTL(reg_idx), txdctl);
@@ -2184,23 +2165,57 @@ static void ixgbevf_clean_rx_ring(struct ixgbevf_ring *rx_ring)
  **/
 static void ixgbevf_clean_tx_ring(struct ixgbevf_ring *tx_ring)
 {
-	struct ixgbevf_tx_buffer *tx_buffer_info;
-	unsigned long size;
-	unsigned int i;
+	u16 i = tx_ring->next_to_clean;
+	struct ixgbevf_tx_buffer *tx_buffer = &tx_ring->tx_buffer_info[i];
 
-	if (!tx_ring->tx_buffer_info)
-		return;
+	while (i != tx_ring->next_to_use) {
+		union ixgbe_adv_tx_desc *eop_desc, *tx_desc;
+
+		/* Free all the Tx ring sk_buffs */
+		dev_kfree_skb_any(tx_buffer->skb);
+
+		/* unmap skb header data */
+		dma_unmap_single(tx_ring->dev,
+				 dma_unmap_addr(tx_buffer, dma),
+				 dma_unmap_len(tx_buffer, len),
+				 DMA_TO_DEVICE);
+
+		/* check for eop_desc to determine the end of the packet */
+		eop_desc = tx_buffer->next_to_watch;
+		tx_desc = IXGBEVF_TX_DESC(tx_ring, i);
+
+		/* unmap remaining buffers */
+		while (tx_desc != eop_desc) {
+			tx_buffer++;
+			tx_desc++;
+			i++;
+			if (unlikely(i == tx_ring->count)) {
+				i = 0;
+				tx_buffer = tx_ring->tx_buffer_info;
+				tx_desc = IXGBEVF_TX_DESC(tx_ring, 0);
+			}
+
+			/* unmap any remaining paged data */
+			if (dma_unmap_len(tx_buffer, len))
+				dma_unmap_page(tx_ring->dev,
+					       dma_unmap_addr(tx_buffer, dma),
+					       dma_unmap_len(tx_buffer, len),
+					       DMA_TO_DEVICE);
+		}
 
-	/* Free all the Tx ring sk_buffs */
-	for (i = 0; i < tx_ring->count; i++) {
-		tx_buffer_info = &tx_ring->tx_buffer_info[i];
-		ixgbevf_unmap_and_free_tx_resource(tx_ring, tx_buffer_info);
+		/* move us one more past the eop_desc for start of next pkt */
+		tx_buffer++;
+		i++;
+		if (unlikely(i == tx_ring->count)) {
+			i = 0;
+			tx_buffer = tx_ring->tx_buffer_info;
+		}
 	}
 
-	size = sizeof(struct ixgbevf_tx_buffer) * tx_ring->count;
-	memset(tx_ring->tx_buffer_info, 0, size);
+	/* reset next_to_use and next_to_clean */
+	tx_ring->next_to_use = 0;
+	tx_ring->next_to_clean = 0;
 
-	memset(tx_ring->desc, 0, tx_ring->size);
 }
 
 /**
@@ -3030,7 +3045,7 @@ int ixgbevf_setup_tx_resources(struct ixgbevf_ring *tx_ring)
 	int size;
 
 	size = sizeof(struct ixgbevf_tx_buffer) * tx_ring->count;
-	tx_ring->tx_buffer_info = vzalloc(size);
+	tx_ring->tx_buffer_info = vmalloc(size);
 	if (!tx_ring->tx_buffer_info)
 		goto err;
 
@@ -3634,18 +3649,32 @@ static void ixgbevf_tx_map(struct ixgbevf_ring *tx_ring,
 	return;
 dma_error:
 	dev_err(tx_ring->dev, "TX DMA map failed\n");
+	tx_buffer = &tx_ring->tx_buffer_info[i];
 
 	/* clear dma mappings for failed tx_buffer_info map */
-	for (;;) {
+	while (tx_buffer != first) {
+		if (dma_unmap_len(tx_buffer, len))
+			dma_unmap_page(tx_ring->dev,
+				       dma_unmap_addr(tx_buffer, dma),
+				       dma_unmap_len(tx_buffer, len),
+				       DMA_TO_DEVICE);
+		dma_unmap_len_set(tx_buffer, len, 0);
+
+		if (i-- == 0)
+			i += tx_ring->count;
 		tx_buffer = &tx_ring->tx_buffer_info[i];
-		ixgbevf_unmap_and_free_tx_resource(tx_ring, tx_buffer);
-		if (tx_buffer == first)
-			break;
-		if (i == 0)
-			i = tx_ring->count;
-		i--;
 	}
 
+	if (dma_unmap_len(tx_buffer, len))
+		dma_unmap_single(tx_ring->dev,
+				 dma_unmap_addr(tx_buffer, dma),
+				 dma_unmap_len(tx_buffer, len),
+				 DMA_TO_DEVICE);
+	dma_unmap_len_set(tx_buffer, len, 0);
+
+	dev_kfree_skb_any(tx_buffer->skb);
+	tx_buffer->skb = NULL;
+
 	tx_ring->next_to_use = i;
 }
 
-- 
2.14.3

[net-next 10/15] ixgbevf: use ARRAY_SIZE for various array sizing calculations

[Jeff Kirsher]

From: Colin Ian King <colin.king@canonical.com>

Use the ARRAY_SIZE macro on various arrays to determine
size of the arrays. Improvement suggested by coccinelle.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbevf/vf.c | 17 +++++++----------
 1 file changed, 7 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbevf/vf.c b/drivers/net/ethernet/intel/ixgbevf/vf.c
index 64c93e8becc6..38d3a327c1bc 100644
--- a/drivers/net/ethernet/intel/ixgbevf/vf.c
+++ b/drivers/net/ethernet/intel/ixgbevf/vf.c
@@ -286,7 +286,7 @@ static s32 ixgbevf_set_uc_addr_vf(struct ixgbe_hw *hw, u32 index, u8 *addr)
 		ether_addr_copy(msg_addr, addr);
 
 	ret_val = ixgbevf_write_msg_read_ack(hw, msgbuf, msgbuf,
-					     sizeof(msgbuf) / sizeof(u32));
+					     ARRAY_SIZE(msgbuf));
 	if (!ret_val) {
 		msgbuf[0] &= ~IXGBE_VT_MSGTYPE_CTS;
 
@@ -456,8 +456,7 @@ static s32 ixgbevf_set_rar_vf(struct ixgbe_hw *hw, u32 index, u8 *addr,
 	ether_addr_copy(msg_addr, addr);
 
 	ret_val = ixgbevf_write_msg_read_ack(hw, msgbuf, msgbuf,
-					     sizeof(msgbuf) / sizeof(u32));
-
+					     ARRAY_SIZE(msgbuf));
 	msgbuf[0] &= ~IXGBE_VT_MSGTYPE_CTS;
 
 	/* if nacked the address was rejected, use "perm_addr" */
@@ -574,7 +573,7 @@ static s32 ixgbevf_update_xcast_mode(struct ixgbe_hw *hw, int xcast_mode)
 	msgbuf[1] = xcast_mode;
 
 	err = ixgbevf_write_msg_read_ack(hw, msgbuf, msgbuf,
-					 sizeof(msgbuf) / sizeof(u32));
+					 ARRAY_SIZE(msgbuf));
 	if (err)
 		return err;
 
@@ -614,7 +613,7 @@ static s32 ixgbevf_set_vfta_vf(struct ixgbe_hw *hw, u32 vlan, u32 vind,
 	msgbuf[0] |= vlan_on << IXGBE_VT_MSGINFO_SHIFT;
 
 	err = ixgbevf_write_msg_read_ack(hw, msgbuf, msgbuf,
-					 sizeof(msgbuf) / sizeof(u32));
+					 ARRAY_SIZE(msgbuf));
 	if (err)
 		goto mbx_err;
 
@@ -826,7 +825,7 @@ static s32 ixgbevf_set_rlpml_vf(struct ixgbe_hw *hw, u16 max_size)
 	msgbuf[1] = max_size;
 
 	ret_val = ixgbevf_write_msg_read_ack(hw, msgbuf, msgbuf,
-					     sizeof(msgbuf) / sizeof(u32));
+					     ARRAY_SIZE(msgbuf));
 	if (ret_val)
 		return ret_val;
 	if ((msgbuf[0] & IXGBE_VF_SET_LPE) &&
@@ -872,8 +871,7 @@ static int ixgbevf_negotiate_api_version_vf(struct ixgbe_hw *hw, int api)
 	msg[1] = api;
 	msg[2] = 0;
 
-	err = ixgbevf_write_msg_read_ack(hw, msg, msg,
-					 sizeof(msg) / sizeof(u32));
+	err = ixgbevf_write_msg_read_ack(hw, msg, msg, ARRAY_SIZE(msg));
 	if (!err) {
 		msg[0] &= ~IXGBE_VT_MSGTYPE_CTS;
 
@@ -924,8 +922,7 @@ int ixgbevf_get_queues(struct ixgbe_hw *hw, unsigned int *num_tcs,
 	msg[0] = IXGBE_VF_GET_QUEUE;
 	msg[1] = msg[2] = msg[3] = msg[4] = 0;
 
-	err = ixgbevf_write_msg_read_ack(hw, msg, msg,
-					 sizeof(msg) / sizeof(u32));
+	err = ixgbevf_write_msg_read_ack(hw, msg, msg, ARRAY_SIZE(msg));
 	if (!err) {
 		msg[0] &= ~IXGBE_VT_MSGTYPE_CTS;
 
-- 
2.14.3

[net-next 11/15] ixgbe: use ARRAY_SIZE for array sizing calculation on array buf

[Jeff Kirsher]

From: Colin Ian King <colin.king@canonical.com>

Use the ARRAY_SIZE macro on array buf to determine size of the array.
Improvement suggested by coccinelle.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c
index 3bce26e77090..f470d0204771 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c
@@ -949,7 +949,7 @@ static s32 ixgbe_checksum_ptr_x550(struct ixgbe_hw *hw, u16 ptr,
 	u16 length, bufsz, i, start;
 	u16 *local_buffer;
 
-	bufsz = sizeof(buf) / sizeof(buf[0]);
+	bufsz = ARRAY_SIZE(buf);
 
 	/* Read a chunk at the pointer location */
 	if (!buffer) {
-- 
2.14.3

[net-next 12/15] ixgbe: Don't report unsupported timestamping filters for X550

[Jeff Kirsher]

From: Miroslav Lichvar <mlichvar@redhat.com>

The current code enables on X550 timestamping of all packets for any
filter, which means ethtool should not report any PTP-specific filters
as unsupported.

Signed-off-by: Miroslav Lichvar <mlichvar@redhat.com>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Acked-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c | 37 ++++++++++++------------
 1 file changed, 19 insertions(+), 18 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
index 317351025fd7..221f15803480 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
@@ -3085,26 +3085,9 @@ static int ixgbe_get_ts_info(struct net_device *dev,
 	case ixgbe_mac_X550EM_x:
 	case ixgbe_mac_x550em_a:
 		info->rx_filters |= BIT(HWTSTAMP_FILTER_ALL);
-		/* fallthrough */
+		break;
 	case ixgbe_mac_X540:
 	case ixgbe_mac_82599EB:
-		info->so_timestamping =
-			SOF_TIMESTAMPING_TX_SOFTWARE |
-			SOF_TIMESTAMPING_RX_SOFTWARE |
-			SOF_TIMESTAMPING_SOFTWARE |
-			SOF_TIMESTAMPING_TX_HARDWARE |
-			SOF_TIMESTAMPING_RX_HARDWARE |
-			SOF_TIMESTAMPING_RAW_HARDWARE;
-
-		if (adapter->ptp_clock)
-			info->phc_index = ptp_clock_index(adapter->ptp_clock);
-		else
-			info->phc_index = -1;
-
-		info->tx_types =
-			BIT(HWTSTAMP_TX_OFF) |
-			BIT(HWTSTAMP_TX_ON);
-
 		info->rx_filters |=
 			BIT(HWTSTAMP_FILTER_PTP_V1_L4_SYNC) |
 			BIT(HWTSTAMP_FILTER_PTP_V1_L4_DELAY_REQ) |
@@ -3113,6 +3096,24 @@ static int ixgbe_get_ts_info(struct net_device *dev,
 	default:
 		return ethtool_op_get_ts_info(dev, info);
 	}
+
+	info->so_timestamping =
+		SOF_TIMESTAMPING_TX_SOFTWARE |
+		SOF_TIMESTAMPING_RX_SOFTWARE |
+		SOF_TIMESTAMPING_SOFTWARE |
+		SOF_TIMESTAMPING_TX_HARDWARE |
+		SOF_TIMESTAMPING_RX_HARDWARE |
+		SOF_TIMESTAMPING_RAW_HARDWARE;
+
+	if (adapter->ptp_clock)
+		info->phc_index = ptp_clock_index(adapter->ptp_clock);
+	else
+		info->phc_index = -1;
+
+	info->tx_types =
+		BIT(HWTSTAMP_TX_OFF) |
+		BIT(HWTSTAMP_TX_ON);
+
 	return 0;
 }
 
-- 
2.14.3

[net-next 13/15] ixgbe: add support for reporting 5G link speed

[Jeff Kirsher]

From: Paul Greenwalt <paul.greenwalt@intel.com>

Since 5G link speed is supported by some devices, add reporting of 5G link
speed.

Signed-off-by: Paul Greenwalt <paul.greenwalt@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index bbb622f15a77..643c7288ea0f 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -7259,6 +7259,9 @@ static void ixgbe_watchdog_link_is_up(struct ixgbe_adapter *adapter)
 	case IXGBE_LINK_SPEED_10GB_FULL:
 		speed_str = "10 Gbps";
 		break;
+	case IXGBE_LINK_SPEED_5GB_FULL:
+		speed_str = "5 Gbps";
+		break;
 	case IXGBE_LINK_SPEED_2_5GB_FULL:
 		speed_str = "2.5 Gbps";
 		break;
-- 
2.14.3

[net-next 14/15] ixgbe: Fix && vs || typo

[Jeff Kirsher]

From: Dan Carpenter <dan.carpenter@oracle.com>

"offset" can't be both 0x0 and 0xFFFF so presumably || was intended
instead of &&.  That matches with how this check is done in other
functions.

Fixes: 73834aec7199 ("ixgbe: extend firmware version support")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_common.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c
index 7ac7ef9b37ff..61188f343955 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c
@@ -4087,7 +4087,7 @@ void ixgbe_get_oem_prod_version(struct ixgbe_hw *hw,
 	hw->eeprom.ops.read(hw, NVM_OEM_PROD_VER_PTR, &offset);
 
 	/* Return is offset to OEM Product Version block is invalid */
-	if (offset == 0x0 && offset == NVM_INVALID_PTR)
+	if (offset == 0x0 || offset == NVM_INVALID_PTR)
 		return;
 
 	/* Read product version block */
-- 
2.14.3

[net-next 15/15] ixgbe: don't set RXDCTL.RLPML for 82599

[Jeff Kirsher]

From: Emil Tantilov <emil.s.tantilov@intel.com>

commit 2de6aa3a666e ("ixgbe: Add support for padding packet")

Uses RXDCTL.RLPML to limit the maximum frame size on Rx when using
build_skb. Unfortunately that register does not work on 82599.

Added an explicit check to avoid setting this register on 82599 MAC.

Extended the comment related to the setting of RXDCTL.RLPML to better
explain its purpose.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index 643c7288ea0f..0da5aa2c8aba 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -4133,11 +4133,15 @@ void ixgbe_configure_rx_ring(struct ixgbe_adapter *adapter,
 		rxdctl &= ~0x3FFFFF;
 		rxdctl |=  0x080420;
 #if (PAGE_SIZE < 8192)
-	} else {
+	/* RXDCTL.RLPML does not work on 82599 */
+	} else if (hw->mac.type != ixgbe_mac_82599EB) {
 		rxdctl &= ~(IXGBE_RXDCTL_RLPMLMASK |
 			    IXGBE_RXDCTL_RLPML_EN);
 
-		/* Limit the maximum frame size so we don't overrun the skb */
+		/* Limit the maximum frame size so we don't overrun the skb.
+		 * This can happen in SRIOV mode when the MTU of the VF is
+		 * higher than the MTU of the PF.
+		 */
 		if (ring_uses_build_skb(ring) &&
 		    !test_bit(__IXGBE_RX_3K_BUFFER, &ring->state))
 			rxdctl |= IXGBE_MAX_2K_FRAME_BUILD_SKB |
-- 
2.14.3

Re: [net-next 00/15][pull request] 10GbE Intel Wired LAN Driver Updates 2018-01-26

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Fri, 26 Jan 2018 10:26:11 -0800

> This series contains updates to ixgbe and ixgbevf.

Pulled, thanks Jeff.