[v3,1/1] ixgbe: sync the first fragment unconditionally
diff mbox series

Message ID 20190808040312.21719-1-firo.yang@suse.com
State Under Review
Delegated to: Jeff Kirsher
Headers show
Series
  • [v3,1/1] ixgbe: sync the first fragment unconditionally
Related show

Commit Message

Firo Yang Aug. 8, 2019, 4:03 a.m. UTC
In Xen environment, if Xen-swiotlb is enabled, ixgbe driver
could possibly allocate a page, DMA memory buffer, for the first
fragment which is not suitable for Xen-swiotlb to do DMA operations.
Xen-swiotlb have to internally allocate another page for doing DMA
operations. This mechanism requires syncing the data from the internal
page to the page which ixgbe sends to upper network stack. However,
since commit f3213d932173 ("ixgbe: Update driver to make use of DMA
attributes in Rx path"), the unmap operation is performed with
DMA_ATTR_SKIP_CPU_SYNC. As a result, the sync is not performed.
Since the sync isn't performed, the upper network stack could receive
a incomplete network packet. By incomplete, it means the linear data
on the first fragment(between skb->head and skb->end) is invalid. So 
we have to copy the data from the internal xen-swiotlb page to the page 
which ixgbe sends to upper network stack through the sync operation.

More details from Alexander Duyck:
Specifically since we are mapping the frame with
DMA_ATTR_SKIP_CPU_SYNC we have to unmap with that as well. As a result
a sync is not performed on an unmap and must be done manually as we
skipped it for the first frag. As such we need to always sync before
possibly performing a page unmap operation.

Fixes: f3213d932173 ("ixgbe: Update driver to make use of DMA
attributes in Rx path")
Reviewed-by: Alexander Duyck <alexander.h.duyck@linux.intel.com>
Signed-off-by: Firo Yang <firo.yang@suse.com>
---
Changes from v2:
 * Added details on the problem caused by skipping the sync.
 * Added more explanation from Alexander Duyck.

Changes from v1:
 * Imporved the patch description.
 * Added Reviewed-by: and Fixes: as suggested by Alexander Duyck.

 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

Comments

Bowers, AndrewX Aug. 8, 2019, 6:42 p.m. UTC | #1
> -----Original Message-----
> From: Intel-wired-lan [mailto:intel-wired-lan-bounces@osuosl.org] On
> Behalf Of Firo Yang
> Sent: Wednesday, August 7, 2019 9:04 PM
> To: netdev@vger.kernel.org
> Cc: maciejromanfijalkowski@gmail.com; Firo Yang <firo.yang@suse.com>;
> linux-kernel@vger.kernel.org; intel-wired-lan@lists.osuosl.org;
> jian.w.wen@oracle.com; alexander.h.duyck@linux.intel.com;
> davem@davemloft.net
> Subject: [Intel-wired-lan] [PATCH v3 1/1] ixgbe: sync the first fragment
> unconditionally
> 
> In Xen environment, if Xen-swiotlb is enabled, ixgbe driver could possibly
> allocate a page, DMA memory buffer, for the first fragment which is not
> suitable for Xen-swiotlb to do DMA operations.
> Xen-swiotlb have to internally allocate another page for doing DMA
> operations. This mechanism requires syncing the data from the internal page
> to the page which ixgbe sends to upper network stack. However, since
> commit f3213d932173 ("ixgbe: Update driver to make use of DMA attributes
> in Rx path"), the unmap operation is performed with
> DMA_ATTR_SKIP_CPU_SYNC. As a result, the sync is not performed.
> Since the sync isn't performed, the upper network stack could receive a
> incomplete network packet. By incomplete, it means the linear data on the
> first fragment(between skb->head and skb->end) is invalid. So we have to
> copy the data from the internal xen-swiotlb page to the page which ixgbe
> sends to upper network stack through the sync operation.
> 
> More details from Alexander Duyck:
> Specifically since we are mapping the frame with
> DMA_ATTR_SKIP_CPU_SYNC we have to unmap with that as well. As a result
> a sync is not performed on an unmap and must be done manually as we
> skipped it for the first frag. As such we need to always sync before possibly
> performing a page unmap operation.
> 
> Fixes: f3213d932173 ("ixgbe: Update driver to make use of DMA attributes in
> Rx path")
> Reviewed-by: Alexander Duyck <alexander.h.duyck@linux.intel.com>
> Signed-off-by: Firo Yang <firo.yang@suse.com>
> ---
> Changes from v2:
>  * Added details on the problem caused by skipping the sync.
>  * Added more explanation from Alexander Duyck.
> 
> Changes from v1:
>  * Imporved the patch description.
>  * Added Reviewed-by: and Fixes: as suggested by Alexander Duyck.
> 
>  drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 16 +++++++++-------
>  1 file changed, 9 insertions(+), 7 deletions(-)

Tested-by: Andrew Bowers <andrewx.bowers@intel.com>

Patch
diff mbox series

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index cbaf712d6529..200de9838096 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -1825,13 +1825,7 @@  static void ixgbe_pull_tail(struct ixgbe_ring *rx_ring,
 static void ixgbe_dma_sync_frag(struct ixgbe_ring *rx_ring,
 				struct sk_buff *skb)
 {
-	/* if the page was released unmap it, else just sync our portion */
-	if (unlikely(IXGBE_CB(skb)->page_released)) {
-		dma_unmap_page_attrs(rx_ring->dev, IXGBE_CB(skb)->dma,
-				     ixgbe_rx_pg_size(rx_ring),
-				     DMA_FROM_DEVICE,
-				     IXGBE_RX_DMA_ATTR);
-	} else if (ring_uses_build_skb(rx_ring)) {
+	if (ring_uses_build_skb(rx_ring)) {
 		unsigned long offset = (unsigned long)(skb->data) & ~PAGE_MASK;
 
 		dma_sync_single_range_for_cpu(rx_ring->dev,
@@ -1848,6 +1842,14 @@  static void ixgbe_dma_sync_frag(struct ixgbe_ring *rx_ring,
 					      skb_frag_size(frag),
 					      DMA_FROM_DEVICE);
 	}
+
+	/* If the page was released, just unmap it. */
+	if (unlikely(IXGBE_CB(skb)->page_released)) {
+		dma_unmap_page_attrs(rx_ring->dev, IXGBE_CB(skb)->dma,
+				     ixgbe_rx_pg_size(rx_ring),
+				     DMA_FROM_DEVICE,
+				     IXGBE_RX_DMA_ATTR);
+	}
 }
 
 /**