[dpdk-stable] request backporting some patches to 16.07.2

[dpdk-stable] request backporting some patches to 16.07.2

[Yuanhan Liu]

Hi commit authors (and maintainers),

I tried to pick following commits from upstream to stable branch 16.07,
but I failed: they can't be applied. I'm wondering can the authors check
the list and backport some patches belong to them?

FYI, branch 16.07 at git://dpdk.org/dpdk-stable is the place those patches
should be packported to. If you are willing to do that, it'd be great if
you could do it in one week (or two). Please drop me an email if you are
willing to do that, or if you need more time, or even, if you think it's
worthless to packport it.

Thanks.

	-yliu

---
10313a4 ethdev: fix vendor id in debug message
243e93a hash: fix unlimited cuckoo path
94eaeba hash: fix bucket size usage
6d3cbd5 net/thunderx: remove private data to ethdev link
b16e60a net/enic: document how to configure vNIC parameters
cf32c2e net/bnx2x: serialize access to VF mailbox
316e412 net/bnxt: fix crash when closing
893fbab examples/l2fwd-crypto: fix verify with decrypt in chain
6445198 kni: fix build with kernel 4.9
e8a71c4 net/enic: fix Rx queue index when not using Rx scatter
ceeb00b net/enic: fix crash on MTU update or Rx queue reconfigure
ce16fd7 net/enic: fix multi-queue Rx performance
ff6edd3 net/i40e: fix out of order Rx read
6949188 net/mlx5: fix initialization in secondary process
dbac54c net/qede: fix port (re)configuration
d6cb175 net/qede: fix VLAN filters
f1e4b6c net/qede: fix status block index for VF queues
75ef62a net/mlx5: fix link speed capability information
1884087 net/mlx5: fix support for newer link speeds

Re: [dpdk-stable] request backporting some patches to 16.07.2

[Yuanhan Liu]

Hi,

I still havn't got any response for following commits. Just want to let
you know, that if you can make a backport in this week, we could still
make it in 16.07.2 release.

And please do me a favor by adding a heading line like following before
the commit log body if you are willing to offer help:

    [ backported from upstream commit xxx ]

Thanks.

	-yliu

---
6d3cbd5  Kamil Rytarowski net/thunderx: remove private data to ethdev link
cf32c2e  Chas Williams  net/bnx2x: serialize access to VF mailbox
316e412  Ajit Khaparde  net/bnxt: fix crash when closing
893fbab  Piotr Azarewicz examples/l2fwd-crypto: fix verify with decrypt in chain
6445198  Ferruh Yigit   kni: fix build with kernel 4.9
e8a71c4  Nelson Escobar net/enic: fix Rx queue index when not using Rx scatter
dbac54c  Harish Patil   net/qede: fix port (re)configuration
d6cb175  Harish Patil   net/qede: fix VLAN filters
f1e4b6c  Harish Patil   net/qede: fix status block index for VF queues

On Sun, Nov 06, 2016 at 03:22:13PM +0800, Yuanhan Liu wrote:
> Hi commit authors (and maintainers),
> 
> I tried to pick following commits from upstream to stable branch 16.07,
> but I failed: they can't be applied. I'm wondering can the authors check
> the list and backport some patches belong to them?
> 
> FYI, branch 16.07 at git://dpdk.org/dpdk-stable is the place those patches
> should be packported to. If you are willing to do that, it'd be great if
> you could do it in one week (or two). Please drop me an email if you are
> willing to do that, or if you need more time, or even, if you think it's
> worthless to packport it.
> 
> Thanks.
> 
> 	-yliu
> 
> ---
> 10313a4 ethdev: fix vendor id in debug message
> 243e93a hash: fix unlimited cuckoo path
> 94eaeba hash: fix bucket size usage
> 6d3cbd5 net/thunderx: remove private data to ethdev link
> b16e60a net/enic: document how to configure vNIC parameters
> cf32c2e net/bnx2x: serialize access to VF mailbox
> 316e412 net/bnxt: fix crash when closing
> 893fbab examples/l2fwd-crypto: fix verify with decrypt in chain
> 6445198 kni: fix build with kernel 4.9
> e8a71c4 net/enic: fix Rx queue index when not using Rx scatter
> ceeb00b net/enic: fix crash on MTU update or Rx queue reconfigure
> ce16fd7 net/enic: fix multi-queue Rx performance
> ff6edd3 net/i40e: fix out of order Rx read
> 6949188 net/mlx5: fix initialization in secondary process
> dbac54c net/qede: fix port (re)configuration
> d6cb175 net/qede: fix VLAN filters
> f1e4b6c net/qede: fix status block index for VF queues
> 75ef62a net/mlx5: fix link speed capability information
> 1884087 net/mlx5: fix support for newer link speeds

Re: [dpdk-stable] request backporting some patches to 16.07.2

[Maciej Czekaj]

Hi,

Excuse me for late response.
I tried to apply the patch

6d3cbd5  Kamil Rytarowski net/thunderx: remove private data to ethdev link


but it strongly depends on previous patches from the whole patchset 
submitted to 16.11 so I think it does not make sense to apply that fix 
to 16.07.

Thanks
Maciej


On Tue, 15 Nov 2016 21:58:14 +0800 Yuanhan Liu wrote:
> Hi,
>
> I still havn't got any response for following commits. Just want to let
> you know, that if you can make a backport in this week, we could still
> make it in 16.07.2 release.
>
> And please do me a favor by adding a heading line like following before
> the commit log body if you are willing to offer help:
>
>      [ backported from upstream commit xxx ]
>
> Thanks.
>
> 	-yliu
>
> ---
> 6d3cbd5  Kamil Rytarowski net/thunderx: remove private data to ethdev link
> cf32c2e  Chas Williams  net/bnx2x: serialize access to VF mailbox
> 316e412  Ajit Khaparde  net/bnxt: fix crash when closing
> 893fbab  Piotr Azarewicz examples/l2fwd-crypto: fix verify with decrypt in chain
> 6445198  Ferruh Yigit   kni: fix build with kernel 4.9
> e8a71c4  Nelson Escobar net/enic: fix Rx queue index when not using Rx scatter
> dbac54c  Harish Patil   net/qede: fix port (re)configuration
> d6cb175  Harish Patil   net/qede: fix VLAN filters
> f1e4b6c  Harish Patil   net/qede: fix status block index for VF queues
>
> On Sun, Nov 06, 2016 at 03:22:13PM +0800, Yuanhan Liu wrote:
>> Hi commit authors (and maintainers),
>>
>> I tried to pick following commits from upstream to stable branch 16.07,
>> but I failed: they can't be applied. I'm wondering can the authors check
>> the list and backport some patches belong to them?
>>
>> FYI, branch 16.07 at git://dpdk.org/dpdk-stable is the place those patches
>> should be packported to. If you are willing to do that, it'd be great if
>> you could do it in one week (or two). Please drop me an email if you are
>> willing to do that, or if you need more time, or even, if you think it's
>> worthless to packport it.
>>
>> Thanks.
>>
>> 	-yliu
>>
>> ---
>> 10313a4 ethdev: fix vendor id in debug message
>> 243e93a hash: fix unlimited cuckoo path
>> 94eaeba hash: fix bucket size usage
>> 6d3cbd5 net/thunderx: remove private data to ethdev link
>> b16e60a net/enic: document how to configure vNIC parameters
>> cf32c2e net/bnx2x: serialize access to VF mailbox
>> 316e412 net/bnxt: fix crash when closing
>> 893fbab examples/l2fwd-crypto: fix verify with decrypt in chain
>> 6445198 kni: fix build with kernel 4.9
>> e8a71c4 net/enic: fix Rx queue index when not using Rx scatter
>> ceeb00b net/enic: fix crash on MTU update or Rx queue reconfigure
>> ce16fd7 net/enic: fix multi-queue Rx performance
>> ff6edd3 net/i40e: fix out of order Rx read
>> 6949188 net/mlx5: fix initialization in secondary process
>> dbac54c net/qede: fix port (re)configuration
>> d6cb175 net/qede: fix VLAN filters
>> f1e4b6c net/qede: fix status block index for VF queues
>> 75ef62a net/mlx5: fix link speed capability information
>> 1884087 net/mlx5: fix support for newer link speeds

Re: [dpdk-stable] request backporting some patches to 16.07.2

[Yuanhan Liu]

On Wed, Nov 16, 2016 at 11:48:58AM +0100, Maciej Czekaj wrote:
> 
> Hi,
> 
> Excuse me for late response.
> I tried to apply the patch
> 
> 
> 6d3cbd5  Kamil Rytarowski net/thunderx: remove private data to ethdev link
> 
> 
> but it strongly depends on previous patches from the whole patchset submitted
> to 16.11 so I think it does not make sense to apply that fix to 16.07.

I see. Thank you for the notice!

	--yliu
> 
> Thanks
> Maciej
> 
> 
> On Tue, 15 Nov 2016 21:58:14 +0800 Yuanhan Liu wrote:
> 
>     Hi,
> 
>     I still havn't got any response for following commits. Just want to let
>     you know, that if you can make a backport in this week, we could still
>     make it in 16.07.2 release.
> 
>     And please do me a favor by adding a heading line like following before
>     the commit log body if you are willing to offer help:
> 
>         [ backported from upstream commit xxx ]
> 
>     Thanks.
> 
>             -yliu
> 
>     ---
>     6d3cbd5  Kamil Rytarowski net/thunderx: remove private data to ethdev link
>     cf32c2e  Chas Williams  net/bnx2x: serialize access to VF mailbox
>     316e412  Ajit Khaparde  net/bnxt: fix crash when closing
>     893fbab  Piotr Azarewicz examples/l2fwd-crypto: fix verify with decrypt in chain
>     6445198  Ferruh Yigit   kni: fix build with kernel 4.9
>     e8a71c4  Nelson Escobar net/enic: fix Rx queue index when not using Rx scatter
>     dbac54c  Harish Patil   net/qede: fix port (re)configuration
>     d6cb175  Harish Patil   net/qede: fix VLAN filters
>     f1e4b6c  Harish Patil   net/qede: fix status block index for VF queues
> 
>     On Sun, Nov 06, 2016 at 03:22:13PM +0800, Yuanhan Liu wrote:
> 
>         Hi commit authors (and maintainers),
> 
>         I tried to pick following commits from upstream to stable branch 16.07,
>         but I failed: they can't be applied. I'm wondering can the authors check
>         the list and backport some patches belong to them?
> 
>         FYI, branch 16.07 at git://dpdk.org/dpdk-stable is the place those patches
>         should be packported to. If you are willing to do that, it'd be great if
>         you could do it in one week (or two). Please drop me an email if you are
>         willing to do that, or if you need more time, or even, if you think it's
>         worthless to packport it.
> 
>         Thanks.
> 
>                 -yliu
> 
>         ---
>         10313a4 ethdev: fix vendor id in debug message
>         243e93a hash: fix unlimited cuckoo path
>         94eaeba hash: fix bucket size usage
>         6d3cbd5 net/thunderx: remove private data to ethdev link
>         b16e60a net/enic: document how to configure vNIC parameters
>         cf32c2e net/bnx2x: serialize access to VF mailbox
>         316e412 net/bnxt: fix crash when closing
>         893fbab examples/l2fwd-crypto: fix verify with decrypt in chain
>         6445198 kni: fix build with kernel 4.9
>         e8a71c4 net/enic: fix Rx queue index when not using Rx scatter
>         ceeb00b net/enic: fix crash on MTU update or Rx queue reconfigure
>         ce16fd7 net/enic: fix multi-queue Rx performance
>         ff6edd3 net/i40e: fix out of order Rx read
>         6949188 net/mlx5: fix initialization in secondary process
>         dbac54c net/qede: fix port (re)configuration
>         d6cb175 net/qede: fix VLAN filters
>         f1e4b6c net/qede: fix status block index for VF queues
>         75ef62a net/mlx5: fix link speed capability information
>         1884087 net/mlx5: fix support for newer link speeds
> 
> 
> 
>  
> 

Re: [dpdk-stable] request backporting some patches to 16.07.2

[Harish Patil]

>
>Hi commit authors (and maintainers),
>
>I tried to pick following commits from upstream to stable branch 16.07,
>but I failed: they can't be applied. I'm wondering can the authors check
>the list and backport some patches belong to them?
>
>FYI, branch 16.07 at git://dpdk.org/dpdk-stable is the place those patches
>should be packported to. If you are willing to do that, it'd be great if
>you could do it in one week (or two). Please drop me an email if you are
>willing to do that, or if you need more time, or even, if you think it's
>worthless to packport it.
>
>Thanks.
>
>	-yliu
>
>---
>10313a4 ethdev: fix vendor id in debug message
>243e93a hash: fix unlimited cuckoo path
>94eaeba hash: fix bucket size usage
>6d3cbd5 net/thunderx: remove private data to ethdev link
>b16e60a net/enic: document how to configure vNIC parameters
>cf32c2e net/bnx2x: serialize access to VF mailbox
>316e412 net/bnxt: fix crash when closing
>893fbab examples/l2fwd-crypto: fix verify with decrypt in chain
>6445198 kni: fix build with kernel 4.9
>e8a71c4 net/enic: fix Rx queue index when not using Rx scatter
>ceeb00b net/enic: fix crash on MTU update or Rx queue reconfigure
>ce16fd7 net/enic: fix multi-queue Rx performance
>ff6edd3 net/i40e: fix out of order Rx read
>6949188 net/mlx5: fix initialization in secondary process
>dbac54c net/qede: fix port (re)configuration
>d6cb175 net/qede: fix VLAN filters
>f1e4b6c net/qede: fix status block index for VF queues
>75ef62a net/mlx5: fix link speed capability information
>1884087 net/mlx5: fix support for newer link speeds
>

Adding Rasesh, another maintainer for qede PMD.

Yuanhan,
We will get back to you ASAP.

Thanks,
Harish

Re: [dpdk-stable] request backporting some patches to 16.07.2

[John Daley (johndale)]

Hi,
For Enic PMD, these 3 seem to already be in the stable branch, so no issues:
>> net/enic: document how to configure vNIC parameters 
> >net/enic: fix crash on MTU update or Rx queue reconfigure
> >net/enic: fix multi-queue Rx performance

For this one, I submitted another patch yesterday to stable@ which was rebased on top of the 16.07 stable branch and should apply clean:
> >net/enic: fix Rx queue index when not using Rx scatter

Other than that, enic is good (at least for now ;)

Thanks for managing the branch!

JohnD

[dpdk-stable] [PATCH 1/2] kni: fix build with kernel 4.8

[Ferruh Yigit]

[ backported from upstream commit 7f5565592c5ab2f3541951d2b4e65188d01d5d00 ]

Linux kernel v4.8 removes macro DEFINE_PCI_DEVICE_TABLE

Linux: 7e9321599011 ("treewide: remove references to the now unnecessary
DEFINE_PCI_DEVICE_TABLE")

Replaced macro with its value in kni ethtool drivers.

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Acked-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
---
 lib/librte_eal/linuxapp/kni/ethtool/igb/igb_main.c     | 2 +-
 lib/librte_eal/linuxapp/kni/ethtool/ixgbe/ixgbe_main.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/lib/librte_eal/linuxapp/kni/ethtool/igb/igb_main.c b/lib/librte_eal/linuxapp/kni/ethtool/igb/igb_main.c
index 96acec5..efd39c6 100644
--- a/lib/librte_eal/linuxapp/kni/ethtool/igb/igb_main.c
+++ b/lib/librte_eal/linuxapp/kni/ethtool/igb/igb_main.c
@@ -76,7 +76,7 @@ static const char igb_driver_string[] =
 static const char igb_copyright[] =
 				"Copyright (c) 2007-2013 Intel Corporation.";
 
-static DEFINE_PCI_DEVICE_TABLE(igb_pci_tbl) = {
+const struct pci_device_id igb_pci_tbl[] = {
 	{ PCI_VDEVICE(INTEL, E1000_DEV_ID_I354_BACKPLANE_1GBPS) },
 	{ PCI_VDEVICE(INTEL, E1000_DEV_ID_I354_SGMII) },
 	{ PCI_VDEVICE(INTEL, E1000_DEV_ID_I354_BACKPLANE_2_5GBPS) },
diff --git a/lib/librte_eal/linuxapp/kni/ethtool/ixgbe/ixgbe_main.c b/lib/librte_eal/linuxapp/kni/ethtool/ixgbe/ixgbe_main.c
index 92fc9fc..238028d 100644
--- a/lib/librte_eal/linuxapp/kni/ethtool/ixgbe/ixgbe_main.c
+++ b/lib/librte_eal/linuxapp/kni/ethtool/ixgbe/ixgbe_main.c
@@ -86,7 +86,7 @@ const char ixgbe_driver_version[] = DRV_VERSION;
  * { Vendor ID, Device ID, SubVendor ID, SubDevice ID,
  *   Class, Class Mask, private data (not used) }
  */
-DEFINE_PCI_DEVICE_TABLE(ixgbe_pci_tbl) = {
+const struct pci_device_id ixgbe_pci_tbl[] = {
 	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598)},
 	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598AF_DUAL_PORT)},
 	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598AF_SINGLE_PORT)},
-- 
2.9.3

[dpdk-stable] [PATCH 2/2] kni: fix build with kernel 4.9

[Ferruh Yigit]

[ backported from upstream commit 6445198f802d993c73f4b246353b2ceb2dfafc32 ]

compile error:
  CC [M]  .../lib/librte_eal/linuxapp/kni/igb_main.o
.../lib/librte_eal/linuxapp/kni/igb_main.c:2317:21:
error: initialization from incompatible pointer type
	[-Werror=incompatible-pointer-types]
  .ndo_set_vf_vlan = igb_ndo_set_vf_vlan,
                     ^~~~~~~~~~~~~~~~~~~

Linux kernel 4.9 updates API for ndo_set_vf_vlan:
Linux: 79aab093a0b5 ("net: Update API for VF vlan protocol 802.1ad support")

Use new API for Linux kernels >= 4.9

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Tested-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
---
 lib/librte_eal/linuxapp/kni/ethtool/igb/igb_main.c | 19 +++++++++++++++++++
 lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h  |  5 +++++
 2 files changed, 24 insertions(+)

diff --git a/lib/librte_eal/linuxapp/kni/ethtool/igb/igb_main.c b/lib/librte_eal/linuxapp/kni/ethtool/igb/igb_main.c
index efd39c6..f1dcc95 100644
--- a/lib/librte_eal/linuxapp/kni/ethtool/igb/igb_main.c
+++ b/lib/librte_eal/linuxapp/kni/ethtool/igb/igb_main.c
@@ -195,7 +195,11 @@ static void igb_process_mdd_event(struct igb_adapter *);
 #ifdef IFLA_VF_MAX
 static int igb_ndo_set_vf_mac( struct net_device *netdev, int vf, u8 *mac);
 static int igb_ndo_set_vf_vlan(struct net_device *netdev,
+#ifdef HAVE_VF_VLAN_PROTO
+				int vf, u16 vlan, u8 qos, __be16 vlan_proto);
+#else
 				int vf, u16 vlan, u8 qos);
+#endif
 #ifdef HAVE_VF_SPOOFCHK_CONFIGURE
 static int igb_ndo_set_vf_spoofchk(struct net_device *netdev, int vf,
 				bool setting);
@@ -6411,7 +6415,11 @@ static void igb_set_vmvir(struct igb_adapter *adapter, u32 vid, u32 vf)
 }
 
 static int igb_ndo_set_vf_vlan(struct net_device *netdev,
+#ifdef HAVE_VF_VLAN_PROTO
+			       int vf, u16 vlan, u8 qos, __be16 vlan_proto)
+#else
 			       int vf, u16 vlan, u8 qos)
+#endif
 {
 	int err = 0;
 	struct igb_adapter *adapter = netdev_priv(netdev);
@@ -6419,6 +6427,12 @@ static int igb_ndo_set_vf_vlan(struct net_device *netdev,
 	/* VLAN IDs accepted range 0-4094 */
 	if ((vf >= adapter->vfs_allocated_count) || (vlan > VLAN_VID_MASK-1) || (qos > 7))
 		return -EINVAL;
+
+#ifdef HAVE_VF_VLAN_PROTO
+	if (vlan_proto != htons(ETH_P_8021Q))
+		return -EPROTONOSUPPORT;
+#endif
+
 	if (vlan || qos) {
 		err = igb_vlvf_set(adapter, vlan, !!vlan, vf);
 		if (err)
@@ -6579,7 +6593,12 @@ static inline void igb_vf_reset(struct igb_adapter *adapter, u32 vf)
 	if (adapter->vf_data[vf].pf_vlan)
 		igb_ndo_set_vf_vlan(adapter->netdev, vf,
 				    adapter->vf_data[vf].pf_vlan,
+#ifdef HAVE_VF_VLAN_PROTO
+				    adapter->vf_data[vf].pf_qos,
+				    htons(ETH_P_8021Q));
+#else
 				    adapter->vf_data[vf].pf_qos);
+#endif
 	else
 		igb_clear_vf_vfta(adapter, vf);
 #endif
diff --git a/lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h b/lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h
index e2cf71e..de3b8dc 100644
--- a/lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h
+++ b/lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h
@@ -3915,4 +3915,9 @@ skb_set_hash(struct sk_buff *skb, __u32 hash, __always_unused int type)
 /* ndo_bridge_getlink adds new filter_mask and vlan_fill parameters */
 #define HAVE_NDO_BRIDGE_GETLINK_FILTER_MASK_VLAN_FILL
 #endif /* >= 4.2.0 */
+
+#if ( LINUX_VERSION_CODE >= KERNEL_VERSION(4,9,0) )
+#define HAVE_VF_VLAN_PROTO
+#endif /* >= 4.9.0 */
+
 #endif /* _KCOMPAT_H_ */
-- 
2.9.3

Re: [dpdk-stable] [PATCH 1/2] kni: fix build with kernel 4.8

[Yuanhan Liu]

On Fri, Nov 18, 2016 at 12:23:25PM +0000, Ferruh Yigit wrote:
> [ backported from upstream commit 7f5565592c5ab2f3541951d2b4e65188d01d5d00 ]
> 
> Linux kernel v4.8 removes macro DEFINE_PCI_DEVICE_TABLE
> 
> Linux: 7e9321599011 ("treewide: remove references to the now unnecessary
> DEFINE_PCI_DEVICE_TABLE")
> 
> Replaced macro with its value in kni ethtool drivers.

Series applied to dpdk stable branch 16.07.

Thanks.

	--yliu


> Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
> Acked-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
> Acked-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
> Acked-by: Stephen Hemminger <stephen@networkplumber.org>
> ---
>  lib/librte_eal/linuxapp/kni/ethtool/igb/igb_main.c     | 2 +-
>  lib/librte_eal/linuxapp/kni/ethtool/ixgbe/ixgbe_main.c | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/lib/librte_eal/linuxapp/kni/ethtool/igb/igb_main.c b/lib/librte_eal/linuxapp/kni/ethtool/igb/igb_main.c
> index 96acec5..efd39c6 100644
> --- a/lib/librte_eal/linuxapp/kni/ethtool/igb/igb_main.c
> +++ b/lib/librte_eal/linuxapp/kni/ethtool/igb/igb_main.c
> @@ -76,7 +76,7 @@ static const char igb_driver_string[] =
>  static const char igb_copyright[] =
>  				"Copyright (c) 2007-2013 Intel Corporation.";
>  
> -static DEFINE_PCI_DEVICE_TABLE(igb_pci_tbl) = {
> +const struct pci_device_id igb_pci_tbl[] = {
>  	{ PCI_VDEVICE(INTEL, E1000_DEV_ID_I354_BACKPLANE_1GBPS) },
>  	{ PCI_VDEVICE(INTEL, E1000_DEV_ID_I354_SGMII) },
>  	{ PCI_VDEVICE(INTEL, E1000_DEV_ID_I354_BACKPLANE_2_5GBPS) },
> diff --git a/lib/librte_eal/linuxapp/kni/ethtool/ixgbe/ixgbe_main.c b/lib/librte_eal/linuxapp/kni/ethtool/ixgbe/ixgbe_main.c
> index 92fc9fc..238028d 100644
> --- a/lib/librte_eal/linuxapp/kni/ethtool/ixgbe/ixgbe_main.c
> +++ b/lib/librte_eal/linuxapp/kni/ethtool/ixgbe/ixgbe_main.c
> @@ -86,7 +86,7 @@ const char ixgbe_driver_version[] = DRV_VERSION;
>   * { Vendor ID, Device ID, SubVendor ID, SubDevice ID,
>   *   Class, Class Mask, private data (not used) }
>   */
> -DEFINE_PCI_DEVICE_TABLE(ixgbe_pci_tbl) = {
> +const struct pci_device_id ixgbe_pci_tbl[] = {
>  	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598)},
>  	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598AF_DUAL_PORT)},
>  	{PCI_VDEVICE(INTEL, IXGBE_DEV_ID_82598AF_SINGLE_PORT)},
> -- 
> 2.9.3