Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities

2017-09-18 Thread John Ferlan 


On 09/18/2017 10:57 AM, Ján Tomko wrote:
> On Thu, Sep 14, 2017 at 11:23:02AM -0400, Laine Stump wrote:
>> (Almost all of my comments result in "ok, no action needed". There are
>> just three items I would like to see changed (2 trivial, 1 also small
>> but Edan or John may think it prudent to re-test with the change before
>> pushing) - I marked those comments with [**].
>>
>> On 08/21/2017 05:19 AM, Edan David wrote:
>>> Adding functionality to libvirt that will allow querying the interface
>>> for the availability of switchdev Offloading NIC capabilities.
>>> The switchdev mode was introduced in kernel 4.8, the iproute2-devlink
>>> command to retrieve the swtichdev NIC feature,
>>> Command example:  devlink dev eswitch show pci/:03:00.0
>>> This feature is needed for Openstack so we can do a scheduling decision
>>> if the NIC is in Hardware Offload (switchdev) or regular SR-IOV
>>> (legacy) mode.
>>> And select the appropriate hypervisors with the requested capability
>>> see [1].
>>>
>>> [1] -
>>> https://specs.openstack.org/openstack/nova-specs/specs/pike/approved/enable-sriov-nic-features.html
>>>
>>> ---
>>>  configure.ac  |  13 ++
>>>  docs/formatnode.html.in   |   1 +
>>>  src/util/virnetdev.c  | 187
>>> +-
>>>  src/util/virnetdev.h  |   1 +
>>>  tests/nodedevschemadata/net_00_13_02_b9_f9_d3.xml |   1 +
>>>  tests/nodedevschemadata/net_00_15_58_2f_e9_55.xml |   1 +
>>>  6 files changed, 203 insertions(+), 1 deletion(-)
>>>
> 
> This patch fails to compile on a Gentoo machine with
> sys-kernel/linux-headers-4.8:
> 
> util/virnetdev.c:3248:18: error: use of undeclared identifier
> 'DEVLINK_CMD_ESWITCH_GET'; did you mean 'DEVLINK_CMD_ESWITCH_MODE_GET'?
>    gmsgh->cmd = DEVLINK_CMD_ESWITCH_GET;
>     ^~~
>     DEVLINK_CMD_ESWITCH_MODE_GET
> /usr/include/linux/devlink.h:60:2: note: 'DEVLINK_CMD_ESWITCH_MODE_GET'
> declared here
>    DEVLINK_CMD_ESWITCH_MODE_GET,
>    ^
> 1 error generated.
> 
>>> diff --git a/configure.ac b/configure.ac
>>> index b12b7fa..c089798 100644
>>> --- a/configure.ac
>>> +++ b/configure.ac
>>> @@ -627,6 +627,19 @@ if test "$with_linux" = "yes"; then
>>>  AC_CHECK_HEADERS([linux/btrfs.h])
>>>  fi
>>>
>>> +dnl
>>> +dnl check for kernel headers required by devlink
>>> +dnl
>>> +if test "$with_linux" = "yes"; then
>>> +    AC_CHECK_HEADERS([linux/devlink.h])
>>> +    AC_CHECK_DECLS([DEVLINK_GENL_VERSION, DEVLINK_GENL_NAME,
>>> DEVLINK_ATTR_MAX, DEVLINK_CMD_ESWITCH_GET, DEVLINK_ATTR_BUS_NAME,
>>> DEVLINK_ATTR_DEV_NAME, DEVLINK_ATTR_ESWITCH_MODE,
>>> DEVLINK_ESWITCH_MODE_SWITCHDEV],
>>
>> That's very . thorough, and potentially misleading if someone ever
>> wanted to use devlink to check for something other than switchdev (e.g.
>> [mythical feature X]) on some system that didn't have the proper stuff
>> defined for switchdev, but did have it for [other mythical feature X].
>> But that's all just hypothetical, so this is fine with me.
>>
>>
>>> +   [AC_DEFINE([HAVE_DECL_DEVLINK],
>>> +  [1],
>>> +  [whether devlink declarations is
>>> available])],
>>
>> [**]
>> s/is/are/
>>
> 
> It seems the [action-if-found] is executed for every found symbol:
> 
> #define HAVE_DECL_DEVLINK_GENL_VERSION 1
> #define HAVE_DECL_DEVLINK 1
> #define HAVE_DECL_DEVLINK_GENL_NAME 1
> #define HAVE_DECL_DEVLINK 1
> #define HAVE_DECL_DEVLINK_ATTR_MAX 1
> #define HAVE_DECL_DEVLINK 1
> #define HAVE_DECL_DEVLINK_CMD_ESWITCH_GET 0
> #define HAVE_DECL_DEVLINK_ATTR_BUS_NAME 1
> #define HAVE_DECL_DEVLINK 1
> #define HAVE_DECL_DEVLINK_ATTR_DEV_NAME 1
> #define HAVE_DECL_DEVLINK 1
> #define HAVE_DECL_DEVLINK_ATTR_ESWITCH_MODE 1
> #define HAVE_DECL_DEVLINK 1
> #define HAVE_DECL_DEVLINK_ESWITCH_MODE_SWITCHDEV 1
> #define HAVE_DECL_DEVLINK 1
> 
> so this usage of AC_CHECK_DECLS is bogus.
> 
> Jan
> 

So then the question becomes is it "safe enough" to only look for
HAVE_DECL_DEVLINK_CMD_ESWITCH_GET and assume the others are present or
does there have to be some sort of AC_CHECK_DECLS for every one of the
symbols that are going to be used and then some way to only allow that
code to be compiled if all are true?

I'd say the former - just check for CMD_ESWITCH_GET and assume the
others, but you know what can be said about assumptions... And this is
why I didn't want to push this on Saturday!  bad enough that I've been
in and out of the office today ;-)

John

>>> +   [],
>>> +   [[#include ]])
>>> +fi
>>> +
>>>  dnl Allow perl/python overrides
>>>  AC_PATH_PROGS([PYTHON], [python2 python])
>>>  if test -z "$PYTHON"; then
> 
> 
> --
> libvir-list mailing list
> libvir-list@redhat.com
> https://www.redhat.com/mailman/listinfo/libvir-list
> 

--
libvir-list mailing list
libvir-list@redhat.com

Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities

2017-09-18 Thread Ján Tomko 

On Thu, Sep 14, 2017 at 11:23:02AM -0400, Laine Stump wrote:

(Almost all of my comments result in "ok, no action needed". There are
just three items I would like to see changed (2 trivial, 1 also small
but Edan or John may think it prudent to re-test with the change before
pushing) - I marked those comments with [**].

On 08/21/2017 05:19 AM, Edan David wrote:

Adding functionality to libvirt that will allow querying the interface
for the availability of switchdev Offloading NIC capabilities.
The switchdev mode was introduced in kernel 4.8, the iproute2-devlink
command to retrieve the swtichdev NIC feature,
Command example:  devlink dev eswitch show pci/:03:00.0
This feature is needed for Openstack so we can do a scheduling decision
if the NIC is in Hardware Offload (switchdev) or regular SR-IOV (legacy) mode.
And select the appropriate hypervisors with the requested capability see [1].

[1] - 
https://specs.openstack.org/openstack/nova-specs/specs/pike/approved/enable-sriov-nic-features.html
---
 configure.ac  |  13 ++
 docs/formatnode.html.in   |   1 +
 src/util/virnetdev.c  | 187 +-
 src/util/virnetdev.h  |   1 +
 tests/nodedevschemadata/net_00_13_02_b9_f9_d3.xml |   1 +
 tests/nodedevschemadata/net_00_15_58_2f_e9_55.xml |   1 +
 6 files changed, 203 insertions(+), 1 deletion(-)



This patch fails to compile on a Gentoo machine with
sys-kernel/linux-headers-4.8:

util/virnetdev.c:3248:18: error: use of undeclared identifier 
'DEVLINK_CMD_ESWITCH_GET'; did you mean 'DEVLINK_CMD_ESWITCH_MODE_GET'?
   gmsgh->cmd = DEVLINK_CMD_ESWITCH_GET;
^~~
DEVLINK_CMD_ESWITCH_MODE_GET
/usr/include/linux/devlink.h:60:2: note: 'DEVLINK_CMD_ESWITCH_MODE_GET' 
declared here
   DEVLINK_CMD_ESWITCH_MODE_GET,
   ^
1 error generated.


diff --git a/configure.ac b/configure.ac
index b12b7fa..c089798 100644
--- a/configure.ac
+++ b/configure.ac
@@ -627,6 +627,19 @@ if test "$with_linux" = "yes"; then
 AC_CHECK_HEADERS([linux/btrfs.h])
 fi

+dnl
+dnl check for kernel headers required by devlink
+dnl
+if test "$with_linux" = "yes"; then
+AC_CHECK_HEADERS([linux/devlink.h])
+AC_CHECK_DECLS([DEVLINK_GENL_VERSION, DEVLINK_GENL_NAME, DEVLINK_ATTR_MAX, 
DEVLINK_CMD_ESWITCH_GET, DEVLINK_ATTR_BUS_NAME, DEVLINK_ATTR_DEV_NAME, 
DEVLINK_ATTR_ESWITCH_MODE, DEVLINK_ESWITCH_MODE_SWITCHDEV],


That's very . thorough, and potentially misleading if someone ever
wanted to use devlink to check for something other than switchdev (e.g.
[mythical feature X]) on some system that didn't have the proper stuff
defined for switchdev, but did have it for [other mythical feature X].
But that's all just hypothetical, so this is fine with me.



+   [AC_DEFINE([HAVE_DECL_DEVLINK],
+  [1],
+  [whether devlink declarations is available])],


[**]
s/is/are/



It seems the [action-if-found] is executed for every found symbol:

#define HAVE_DECL_DEVLINK_GENL_VERSION 1
#define HAVE_DECL_DEVLINK 1
#define HAVE_DECL_DEVLINK_GENL_NAME 1
#define HAVE_DECL_DEVLINK 1
#define HAVE_DECL_DEVLINK_ATTR_MAX 1
#define HAVE_DECL_DEVLINK 1
#define HAVE_DECL_DEVLINK_CMD_ESWITCH_GET 0
#define HAVE_DECL_DEVLINK_ATTR_BUS_NAME 1
#define HAVE_DECL_DEVLINK 1
#define HAVE_DECL_DEVLINK_ATTR_DEV_NAME 1
#define HAVE_DECL_DEVLINK 1
#define HAVE_DECL_DEVLINK_ATTR_ESWITCH_MODE 1
#define HAVE_DECL_DEVLINK 1
#define HAVE_DECL_DEVLINK_ESWITCH_MODE_SWITCHDEV 1
#define HAVE_DECL_DEVLINK 1

so this usage of AC_CHECK_DECLS is bogus.

Jan


+   [],
+   [[#include ]])
+fi
+
 dnl Allow perl/python overrides
 AC_PATH_PROGS([PYTHON], [python2 python])
 if test -z "$PYTHON"; then


signature.asc
Description: Digital signature
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities

2017-09-18 Thread John Ferlan 


On 09/17/2017 05:32 AM, Edan David wrote:
> I removed the memset from virNetDevPutExtraHeader and tested on a card 
> supporting switchdev.
> It looks fine to me :)
> Great review guys, thanks!
> 

Thanks for checking... This is now pushed.

Congrats on your first libvirt patch,

John

> 
> -Original Message-
> From: John Ferlan [mailto:jfer...@redhat.com] 
> Sent: Saturday, September 16, 2017 2:04 PM
> To: Laine Stump <la...@laine.org>; libvir-list@redhat.com
> Cc: Edan David <ed...@mellanox.com>
> Subject: Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities
> 
> 
> 
> On 09/14/2017 11:23 AM, Laine Stump wrote:
>> (Almost all of my comments result in "ok, no action needed". There are 
>> just three items I would like to see changed (2 trivial, 1 also small 
>> but Edan or John may think it prudent to re-test with the change 
>> before
>> pushing) - I marked those comments with [**].
>>
>> On 08/21/2017 05:19 AM, Edan David wrote:
>>> Adding functionality to libvirt that will allow querying the 
>>> interface for the availability of switchdev Offloading NIC capabilities.
>>> The switchdev mode was introduced in kernel 4.8, the iproute2-devlink 
>>> command to retrieve the swtichdev NIC feature, Command example:  
>>> devlink dev eswitch show pci/:03:00.0 This feature is needed for 
>>> Openstack so we can do a scheduling decision if the NIC is in 
>>> Hardware Offload (switchdev) or regular SR-IOV (legacy) mode.
>>> And select the appropriate hypervisors with the requested capability see 
>>> [1].
>>>
>>> [1] - 
>>> https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsp
>>> ecs.openstack.org%2Fopenstack%2Fnova-specs%2Fspecs%2Fpike%2Fapproved%
>>> 2Fenable-sriov-nic-features.html=02%7C01%7Cedand%40mellanox.com%
>>> 7C1e30890a5c7a4f9f045f08d4fcf2b479%7Ca652971c7d2e4d9ba6a4d149256f461b
>>> %7C0%7C0%7C636411566714987559=Ri340H6MLap3HpOMDrb%2FFk6D9agQjEh
>>> C9cvR7%2FbV1Ls%3D=0
>>> ---
>>>  configure.ac  |  13 ++
>>>  docs/formatnode.html.in   |   1 +
>>>  src/util/virnetdev.c  | 187 
>>> +-
>>>  src/util/virnetdev.h  |   1 +
>>>  tests/nodedevschemadata/net_00_13_02_b9_f9_d3.xml |   1 +
>>>  tests/nodedevschemadata/net_00_15_58_2f_e9_55.xml |   1 +
>>>  6 files changed, 203 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/configure.ac b/configure.ac index b12b7fa..c089798 
>>> 100644
>>> --- a/configure.ac
>>> +++ b/configure.ac
>>> @@ -627,6 +627,19 @@ if test "$with_linux" = "yes"; then
>>>  AC_CHECK_HEADERS([linux/btrfs.h])  fi
>>>  
>>> +dnl
>>> +dnl check for kernel headers required by devlink dnl if test 
>>> +"$with_linux" = "yes"; then
>>> +AC_CHECK_HEADERS([linux/devlink.h])
>>> +AC_CHECK_DECLS([DEVLINK_GENL_VERSION, DEVLINK_GENL_NAME, 
>>> +DEVLINK_ATTR_MAX, DEVLINK_CMD_ESWITCH_GET, DEVLINK_ATTR_BUS_NAME, 
>>> +DEVLINK_ATTR_DEV_NAME, DEVLINK_ATTR_ESWITCH_MODE, 
>>> +DEVLINK_ESWITCH_MODE_SWITCHDEV],
>>
>> That's very . thorough, and potentially misleading if someone ever 
>> wanted to use devlink to check for something other than switchdev (e.g.
>> [mythical feature X]) on some system that didn't have the proper stuff 
>> defined for switchdev, but did have it for [other mythical feature X].
>> But that's all just hypothetical, so this is fine with me.
>>
>>
>>> +   [AC_DEFINE([HAVE_DECL_DEVLINK],
>>> +  [1],
>>> +  [whether devlink declarations is 
>>> + available])],
>>
>> [**]
>> s/is/are/
>>
> 
> Yep - altered...
> 
>>> +   [],
>>> +   [[#include ]]) fi
>>> +
>>>  dnl Allow perl/python overrides
>>>  AC_PATH_PROGS([PYTHON], [python2 python])  if test -z "$PYTHON"; 
>>> then diff --git a/docs/formatnode.html.in b/docs/formatnode.html.in 
>>> index 4d935b5..29244a8 100644
>>> --- a/docs/formatnode.html.in
>>> +++ b/docs/formatnode.html.in
>>> @@ -227,6 +227,7 @@
>>>  rxhashreceive-hashing
>>>  
>>> rdmaremote-direct-memory-access
>>>  
>>> txudptnltx-udp-tunn

Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities

2017-09-17 Thread Edan David 
I removed the memset from virNetDevPutExtraHeader and tested on a card 
supporting switchdev.
It looks fine to me :)
Great review guys, thanks!


-Original Message-
From: John Ferlan [mailto:jfer...@redhat.com] 
Sent: Saturday, September 16, 2017 2:04 PM
To: Laine Stump <la...@laine.org>; libvir-list@redhat.com
Cc: Edan David <ed...@mellanox.com>
Subject: Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities



On 09/14/2017 11:23 AM, Laine Stump wrote:
> (Almost all of my comments result in "ok, no action needed". There are 
> just three items I would like to see changed (2 trivial, 1 also small 
> but Edan or John may think it prudent to re-test with the change 
> before
> pushing) - I marked those comments with [**].
> 
> On 08/21/2017 05:19 AM, Edan David wrote:
>> Adding functionality to libvirt that will allow querying the 
>> interface for the availability of switchdev Offloading NIC capabilities.
>> The switchdev mode was introduced in kernel 4.8, the iproute2-devlink 
>> command to retrieve the swtichdev NIC feature, Command example:  
>> devlink dev eswitch show pci/:03:00.0 This feature is needed for 
>> Openstack so we can do a scheduling decision if the NIC is in 
>> Hardware Offload (switchdev) or regular SR-IOV (legacy) mode.
>> And select the appropriate hypervisors with the requested capability see [1].
>>
>> [1] - 
>> https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsp
>> ecs.openstack.org%2Fopenstack%2Fnova-specs%2Fspecs%2Fpike%2Fapproved%
>> 2Fenable-sriov-nic-features.html=02%7C01%7Cedand%40mellanox.com%
>> 7C1e30890a5c7a4f9f045f08d4fcf2b479%7Ca652971c7d2e4d9ba6a4d149256f461b
>> %7C0%7C0%7C636411566714987559=Ri340H6MLap3HpOMDrb%2FFk6D9agQjEh
>> C9cvR7%2FbV1Ls%3D=0
>> ---
>>  configure.ac  |  13 ++
>>  docs/formatnode.html.in   |   1 +
>>  src/util/virnetdev.c  | 187 
>> +-
>>  src/util/virnetdev.h  |   1 +
>>  tests/nodedevschemadata/net_00_13_02_b9_f9_d3.xml |   1 +
>>  tests/nodedevschemadata/net_00_15_58_2f_e9_55.xml |   1 +
>>  6 files changed, 203 insertions(+), 1 deletion(-)
>>
>> diff --git a/configure.ac b/configure.ac index b12b7fa..c089798 
>> 100644
>> --- a/configure.ac
>> +++ b/configure.ac
>> @@ -627,6 +627,19 @@ if test "$with_linux" = "yes"; then
>>  AC_CHECK_HEADERS([linux/btrfs.h])  fi
>>  
>> +dnl
>> +dnl check for kernel headers required by devlink dnl if test 
>> +"$with_linux" = "yes"; then
>> +AC_CHECK_HEADERS([linux/devlink.h])
>> +AC_CHECK_DECLS([DEVLINK_GENL_VERSION, DEVLINK_GENL_NAME, 
>> +DEVLINK_ATTR_MAX, DEVLINK_CMD_ESWITCH_GET, DEVLINK_ATTR_BUS_NAME, 
>> +DEVLINK_ATTR_DEV_NAME, DEVLINK_ATTR_ESWITCH_MODE, 
>> +DEVLINK_ESWITCH_MODE_SWITCHDEV],
> 
> That's very . thorough, and potentially misleading if someone ever 
> wanted to use devlink to check for something other than switchdev (e.g.
> [mythical feature X]) on some system that didn't have the proper stuff 
> defined for switchdev, but did have it for [other mythical feature X].
> But that's all just hypothetical, so this is fine with me.
> 
> 
>> +   [AC_DEFINE([HAVE_DECL_DEVLINK],
>> +  [1],
>> +  [whether devlink declarations is 
>> + available])],
> 
> [**]
> s/is/are/
> 

Yep - altered...

>> +   [],
>> +   [[#include ]]) fi
>> +
>>  dnl Allow perl/python overrides
>>  AC_PATH_PROGS([PYTHON], [python2 python])  if test -z "$PYTHON"; 
>> then diff --git a/docs/formatnode.html.in b/docs/formatnode.html.in 
>> index 4d935b5..29244a8 100644
>> --- a/docs/formatnode.html.in
>> +++ b/docs/formatnode.html.in
>> @@ -227,6 +227,7 @@
>>  rxhashreceive-hashing
>>  
>> rdmaremote-direct-memory-access
>>  
>> txudptnltx-udp-tunnel-segmentation
>> +
>> + switchdevkernel-forward-plane-offload> + >
> 
> pretty odd abbreviation. But it is what it is :-)
> >>  
>>
>>capability diff --git 
>> a/src/util/virnetdev.c b/src/util/virnetdev.c index 51a6e42..fc7c961 
>> 100644
>> --- a/src/util/virnetdev.c
>> +++ b/src/util/virnetdev.c
>> @@ -59,6 +59,10 @@
>>  # include 
>>  #endif
>>  
>> +#if HAVE_DECL_DEVLINK
>> +

Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities

2017-09-16 Thread John Ferlan 


On 09/14/2017 11:23 AM, Laine Stump wrote:
> (Almost all of my comments result in "ok, no action needed". There are
> just three items I would like to see changed (2 trivial, 1 also small
> but Edan or John may think it prudent to re-test with the change before
> pushing) - I marked those comments with [**].
> 
> On 08/21/2017 05:19 AM, Edan David wrote:
>> Adding functionality to libvirt that will allow querying the interface
>> for the availability of switchdev Offloading NIC capabilities.
>> The switchdev mode was introduced in kernel 4.8, the iproute2-devlink
>> command to retrieve the swtichdev NIC feature,
>> Command example:  devlink dev eswitch show pci/:03:00.0
>> This feature is needed for Openstack so we can do a scheduling decision
>> if the NIC is in Hardware Offload (switchdev) or regular SR-IOV (legacy) 
>> mode.
>> And select the appropriate hypervisors with the requested capability see [1].
>>
>> [1] - 
>> https://specs.openstack.org/openstack/nova-specs/specs/pike/approved/enable-sriov-nic-features.html
>> ---
>>  configure.ac  |  13 ++
>>  docs/formatnode.html.in   |   1 +
>>  src/util/virnetdev.c  | 187 
>> +-
>>  src/util/virnetdev.h  |   1 +
>>  tests/nodedevschemadata/net_00_13_02_b9_f9_d3.xml |   1 +
>>  tests/nodedevschemadata/net_00_15_58_2f_e9_55.xml |   1 +
>>  6 files changed, 203 insertions(+), 1 deletion(-)
>>
>> diff --git a/configure.ac b/configure.ac
>> index b12b7fa..c089798 100644
>> --- a/configure.ac
>> +++ b/configure.ac
>> @@ -627,6 +627,19 @@ if test "$with_linux" = "yes"; then
>>  AC_CHECK_HEADERS([linux/btrfs.h])
>>  fi
>>  
>> +dnl
>> +dnl check for kernel headers required by devlink
>> +dnl
>> +if test "$with_linux" = "yes"; then
>> +AC_CHECK_HEADERS([linux/devlink.h])
>> +AC_CHECK_DECLS([DEVLINK_GENL_VERSION, DEVLINK_GENL_NAME, 
>> DEVLINK_ATTR_MAX, DEVLINK_CMD_ESWITCH_GET, DEVLINK_ATTR_BUS_NAME, 
>> DEVLINK_ATTR_DEV_NAME, DEVLINK_ATTR_ESWITCH_MODE, 
>> DEVLINK_ESWITCH_MODE_SWITCHDEV],
> 
> That's very . thorough, and potentially misleading if someone ever
> wanted to use devlink to check for something other than switchdev (e.g.
> [mythical feature X]) on some system that didn't have the proper stuff
> defined for switchdev, but did have it for [other mythical feature X].
> But that's all just hypothetical, so this is fine with me.
> 
> 
>> +   [AC_DEFINE([HAVE_DECL_DEVLINK],
>> +  [1],
>> +  [whether devlink declarations is available])],
> 
> [**]
> s/is/are/
> 

Yep - altered...

>> +   [],
>> +   [[#include ]])
>> +fi
>> +
>>  dnl Allow perl/python overrides
>>  AC_PATH_PROGS([PYTHON], [python2 python])
>>  if test -z "$PYTHON"; then
>> diff --git a/docs/formatnode.html.in b/docs/formatnode.html.in
>> index 4d935b5..29244a8 100644
>> --- a/docs/formatnode.html.in
>> +++ b/docs/formatnode.html.in
>> @@ -227,6 +227,7 @@
>>  rxhashreceive-hashing
>>  
>> rdmaremote-direct-memory-access
>>  
>> txudptnltx-udp-tunnel-segmentation
>> +
>> switchdevkernel-forward-plane-offload
> 
> pretty odd abbreviation. But it is what it is :-)
> >>  
>>
>>capability
>> diff --git a/src/util/virnetdev.c b/src/util/virnetdev.c
>> index 51a6e42..fc7c961 100644
>> --- a/src/util/virnetdev.c
>> +++ b/src/util/virnetdev.c
>> @@ -59,6 +59,10 @@
>>  # include 
>>  #endif
>>  
>> +#if HAVE_DECL_DEVLINK
>> +# include 
>> +#endif
>> +
>>  #ifndef IFNAMSIZ
>>  # define IFNAMSIZ 16
>>  #endif
>> @@ -2481,7 +2485,8 @@ VIR_ENUM_IMPL(virNetDevFeature,
>>"ntuple",
>>"rxhash",
>>"rdma",
>> -  "txudptnl")
>> +  "txudptnl",
>> +  "switchdev")
>>  
>>  #ifdef __linux__
>>  int
>> @@ -2936,6 +2941,7 @@ int virNetDevGetRxFilter(const char *ifname,
>>  return ret;
>>  }
>>  
>> +
> 
> [**]
> Added a spurious extra line.
> 

Removed.

>>  #if defined(SIOCETHTOOL) && defined(HAVE_STRUCT_IFREQ)
>>  
>>  /**
>> @@ -3115,6 +3121,182 @@ virNetDevGetEthtoolFeatures(virBitmapPtr bitmap,
>>  }
>>  
>>  
>> +#if HAVE_DECL_DEVLINK
>> +/**
>> + * virNetDevPutExtraHeader
>> + * reserve and prepare room for an extra header
>> + * This function sets to zero the room that is required to put the extra
>> + * header after the initial Netlink header. This function also increases
>> + * the nlmsg_len field.
>> + *
>> + * @nlh: pointer to Netlink header
>> + * @size: size of the extra header that we want to put
>> + *
>> + * Returns pointer to the start of the extended header
>> + */
>> +static void *
>> +virNetDevPutExtraHeader(struct nlmsghdr *nlh,
>> +size_t size)
>> +{
>> +char *ptr = (char *)nlh + nlh->nlmsg_len;
>> +

Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities

2017-09-14 Thread Laine Stump 
(Almost all of my comments result in "ok, no action needed". There are
just three items I would like to see changed (2 trivial, 1 also small
but Edan or John may think it prudent to re-test with the change before
pushing) - I marked those comments with [**].

On 08/21/2017 05:19 AM, Edan David wrote:
> Adding functionality to libvirt that will allow querying the interface
> for the availability of switchdev Offloading NIC capabilities.
> The switchdev mode was introduced in kernel 4.8, the iproute2-devlink
> command to retrieve the swtichdev NIC feature,
> Command example:  devlink dev eswitch show pci/:03:00.0
> This feature is needed for Openstack so we can do a scheduling decision
> if the NIC is in Hardware Offload (switchdev) or regular SR-IOV (legacy) mode.
> And select the appropriate hypervisors with the requested capability see [1].
>
> [1] - 
> https://specs.openstack.org/openstack/nova-specs/specs/pike/approved/enable-sriov-nic-features.html
> ---
>  configure.ac  |  13 ++
>  docs/formatnode.html.in   |   1 +
>  src/util/virnetdev.c  | 187 
> +-
>  src/util/virnetdev.h  |   1 +
>  tests/nodedevschemadata/net_00_13_02_b9_f9_d3.xml |   1 +
>  tests/nodedevschemadata/net_00_15_58_2f_e9_55.xml |   1 +
>  6 files changed, 203 insertions(+), 1 deletion(-)
>
> diff --git a/configure.ac b/configure.ac
> index b12b7fa..c089798 100644
> --- a/configure.ac
> +++ b/configure.ac
> @@ -627,6 +627,19 @@ if test "$with_linux" = "yes"; then
>  AC_CHECK_HEADERS([linux/btrfs.h])
>  fi
>  
> +dnl
> +dnl check for kernel headers required by devlink
> +dnl
> +if test "$with_linux" = "yes"; then
> +AC_CHECK_HEADERS([linux/devlink.h])
> +AC_CHECK_DECLS([DEVLINK_GENL_VERSION, DEVLINK_GENL_NAME, 
> DEVLINK_ATTR_MAX, DEVLINK_CMD_ESWITCH_GET, DEVLINK_ATTR_BUS_NAME, 
> DEVLINK_ATTR_DEV_NAME, DEVLINK_ATTR_ESWITCH_MODE, 
> DEVLINK_ESWITCH_MODE_SWITCHDEV],

That's very . thorough, and potentially misleading if someone ever
wanted to use devlink to check for something other than switchdev (e.g.
[mythical feature X]) on some system that didn't have the proper stuff
defined for switchdev, but did have it for [other mythical feature X].
But that's all just hypothetical, so this is fine with me.


> +   [AC_DEFINE([HAVE_DECL_DEVLINK],
> +  [1],
> +  [whether devlink declarations is available])],

[**]
s/is/are/

> +   [],
> +   [[#include ]])
> +fi
> +
>  dnl Allow perl/python overrides
>  AC_PATH_PROGS([PYTHON], [python2 python])
>  if test -z "$PYTHON"; then
> diff --git a/docs/formatnode.html.in b/docs/formatnode.html.in
> index 4d935b5..29244a8 100644
> --- a/docs/formatnode.html.in
> +++ b/docs/formatnode.html.in
> @@ -227,6 +227,7 @@
>  rxhashreceive-hashing
>  
> rdmaremote-direct-memory-access
>  
> txudptnltx-udp-tunnel-segmentation
> +
> switchdevkernel-forward-plane-offload

pretty odd abbreviation. But it is what it is :-)

>  
>
>capability
> diff --git a/src/util/virnetdev.c b/src/util/virnetdev.c
> index 51a6e42..fc7c961 100644
> --- a/src/util/virnetdev.c
> +++ b/src/util/virnetdev.c
> @@ -59,6 +59,10 @@
>  # include 
>  #endif
>  
> +#if HAVE_DECL_DEVLINK
> +# include 
> +#endif
> +
>  #ifndef IFNAMSIZ
>  # define IFNAMSIZ 16
>  #endif
> @@ -2481,7 +2485,8 @@ VIR_ENUM_IMPL(virNetDevFeature,
>"ntuple",
>"rxhash",
>"rdma",
> -  "txudptnl")
> +  "txudptnl",
> +  "switchdev")
>  
>  #ifdef __linux__
>  int
> @@ -2936,6 +2941,7 @@ int virNetDevGetRxFilter(const char *ifname,
>  return ret;
>  }
>  
> +

[**]
Added a spurious extra line.

>  #if defined(SIOCETHTOOL) && defined(HAVE_STRUCT_IFREQ)
>  
>  /**
> @@ -3115,6 +3121,182 @@ virNetDevGetEthtoolFeatures(virBitmapPtr bitmap,
>  }
>  
>  
> +#if HAVE_DECL_DEVLINK
> +/**
> + * virNetDevPutExtraHeader
> + * reserve and prepare room for an extra header
> + * This function sets to zero the room that is required to put the extra
> + * header after the initial Netlink header. This function also increases
> + * the nlmsg_len field.
> + *
> + * @nlh: pointer to Netlink header
> + * @size: size of the extra header that we want to put
> + *
> + * Returns pointer to the start of the extended header
> + */
> +static void *
> +virNetDevPutExtraHeader(struct nlmsghdr *nlh,
> +size_t size)
> +{
> +char *ptr = (char *)nlh + nlh->nlmsg_len;
> +size_t len = NLMSG_ALIGN(size);
> +nlh->nlmsg_len += len;
> +memset(ptr, 0, len);

[**]
I'm fairly confident this memset() is unnecessary. The buffer the "extra
header" is being added to is allocated with nlmsg_alloc_simple(); I
looked

Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities

2017-09-05 Thread Laine Stump 
On 09/05/2017 08:18 AM, John Ferlan wrote:
>
> On 09/03/2017 03:06 AM, Edan David wrote:
>> Hi John,
>>
>> Can we merge this commit? Does anyone else need to review?
>>
>> Thanks.
>>
> Like I pointed out - I would prefer to have Laine give them at least a
> lookover since he's a bit more familiar with the interface/netlink code
> than I am and may "see" something I missed. He's CC'd here to remind him
> again (I had pinged him internally last week).

Yeah, sorry - I forgot again. I'll look at it in a couple hours.


>
> John
>
>> -Original Message-
>> From: Edan David 
>> Sent: Thursday, August 24, 2017 10:18 AM
>> To: 'John Ferlan' <jfer...@redhat.com>; libvir-list@redhat.com
>> Subject: RE: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities
>>
>> Hi John,
>>
>> For some reason I don't see these errors in my environment, I would 
>> appreciate it if you could fix them before merging.
>> Thank you!!
>>
>> -Original Message-
>> From: John Ferlan [mailto:jfer...@redhat.com]
>> Sent: Wednesday, August 23, 2017 8:44 PM
>> To: Edan David <ed...@mellanox.com>; libvir-list@redhat.com
>> Subject: Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities
>>
>>
>>
>> On 08/21/2017 05:19 AM, Edan David wrote:
>>> Adding functionality to libvirt that will allow querying the interface 
>>> for the availability of switchdev Offloading NIC capabilities.
>>> The switchdev mode was introduced in kernel 4.8, the iproute2-devlink 
>>> command to retrieve the swtichdev NIC feature, Command example:  
>>> devlink dev eswitch show pci/:03:00.0 This feature is needed for 
>>> Openstack so we can do a scheduling decision if the NIC is in Hardware 
>>> Offload (switchdev) or regular SR-IOV (legacy) mode.
>>> And select the appropriate hypervisors with the requested capability see 
>>> [1].
>>>
>>> [1] - 
>>> https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fspe
>>> cs.openstack.org%2Fopenstack%2Fnova-specs%2Fspecs%2Fpike%2Fapproved%2F
>>> enable-sriov-nic-features.html=02%7C01%7Cedand%40mellanox.com%7C7
>>> 58a68d1796c435a8af108d4ea4e889b%7Ca652971c7d2e4d9ba6a4d149256f461b%7C0
>>> %7C0%7C636391070397000461=dbe8Z6P71ZEqC3u46YB1hkmmm8pQgYpWjjVxvB
>>> fu5Ko%3D=0
>>> ---
>>>  configure.ac  |  13 ++
>>>  docs/formatnode.html.in   |   1 +
>>>  src/util/virnetdev.c  | 187 
>>> +-
>>>  src/util/virnetdev.h  |   1 +
>>>  tests/nodedevschemadata/net_00_13_02_b9_f9_d3.xml |   1 +
>>>  tests/nodedevschemadata/net_00_15_58_2f_e9_55.xml |   1 +
>>>  6 files changed, 203 insertions(+), 1 deletion(-)
>>>
>> Again - make check syntax-check had a failure because properly indent "#if" 
>> to be "# if" (also "# else" and "# endif")
>>
>> I can fix that easily before pushing though.
>>
>> Reviewed-by: John Ferlan <jfer...@redhat.com>
>>
>> I'll wait a day before pushing - just in case laine comes back from chasing 
>> the eclipse and has a comment to make...
>>
>> Tks -
>>
>> John
>>

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities

2017-09-05 Thread John Ferlan 


On 09/03/2017 03:06 AM, Edan David wrote:
> Hi John,
> 
> Can we merge this commit? Does anyone else need to review?
> 
> Thanks.
> 

Like I pointed out - I would prefer to have Laine give them at least a
lookover since he's a bit more familiar with the interface/netlink code
than I am and may "see" something I missed. He's CC'd here to remind him
again (I had pinged him internally last week).

John

> -Original Message-
> From: Edan David 
> Sent: Thursday, August 24, 2017 10:18 AM
> To: 'John Ferlan' <jfer...@redhat.com>; libvir-list@redhat.com
> Subject: RE: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities
> 
> Hi John,
> 
> For some reason I don't see these errors in my environment, I would 
> appreciate it if you could fix them before merging.
> Thank you!!
> 
> -Original Message-
> From: John Ferlan [mailto:jfer...@redhat.com]
> Sent: Wednesday, August 23, 2017 8:44 PM
> To: Edan David <ed...@mellanox.com>; libvir-list@redhat.com
> Subject: Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities
> 
> 
> 
> On 08/21/2017 05:19 AM, Edan David wrote:
>> Adding functionality to libvirt that will allow querying the interface 
>> for the availability of switchdev Offloading NIC capabilities.
>> The switchdev mode was introduced in kernel 4.8, the iproute2-devlink 
>> command to retrieve the swtichdev NIC feature, Command example:  
>> devlink dev eswitch show pci/:03:00.0 This feature is needed for 
>> Openstack so we can do a scheduling decision if the NIC is in Hardware 
>> Offload (switchdev) or regular SR-IOV (legacy) mode.
>> And select the appropriate hypervisors with the requested capability see [1].
>>
>> [1] - 
>> https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fspe
>> cs.openstack.org%2Fopenstack%2Fnova-specs%2Fspecs%2Fpike%2Fapproved%2F
>> enable-sriov-nic-features.html=02%7C01%7Cedand%40mellanox.com%7C7
>> 58a68d1796c435a8af108d4ea4e889b%7Ca652971c7d2e4d9ba6a4d149256f461b%7C0
>> %7C0%7C636391070397000461=dbe8Z6P71ZEqC3u46YB1hkmmm8pQgYpWjjVxvB
>> fu5Ko%3D=0
>> ---
>>  configure.ac  |  13 ++
>>  docs/formatnode.html.in   |   1 +
>>  src/util/virnetdev.c  | 187 
>> +-
>>  src/util/virnetdev.h  |   1 +
>>  tests/nodedevschemadata/net_00_13_02_b9_f9_d3.xml |   1 +
>>  tests/nodedevschemadata/net_00_15_58_2f_e9_55.xml |   1 +
>>  6 files changed, 203 insertions(+), 1 deletion(-)
>>
> 
> Again - make check syntax-check had a failure because properly indent "#if" 
> to be "# if" (also "# else" and "# endif")
> 
> I can fix that easily before pushing though.
> 
> Reviewed-by: John Ferlan <jfer...@redhat.com>
> 
> I'll wait a day before pushing - just in case laine comes back from chasing 
> the eclipse and has a comment to make...
> 
> Tks -
> 
> John
> 

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities

2017-09-03 Thread Edan David 
Hi John,

Can we merge this commit? Does anyone else need to review?

Thanks.

-Original Message-
From: Edan David 
Sent: Thursday, August 24, 2017 10:18 AM
To: 'John Ferlan' <jfer...@redhat.com>; libvir-list@redhat.com
Subject: RE: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities

Hi John,

For some reason I don't see these errors in my environment, I would appreciate 
it if you could fix them before merging.
Thank you!!

-Original Message-
From: John Ferlan [mailto:jfer...@redhat.com]
Sent: Wednesday, August 23, 2017 8:44 PM
To: Edan David <ed...@mellanox.com>; libvir-list@redhat.com
Subject: Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities



On 08/21/2017 05:19 AM, Edan David wrote:
> Adding functionality to libvirt that will allow querying the interface 
> for the availability of switchdev Offloading NIC capabilities.
> The switchdev mode was introduced in kernel 4.8, the iproute2-devlink 
> command to retrieve the swtichdev NIC feature, Command example:  
> devlink dev eswitch show pci/:03:00.0 This feature is needed for 
> Openstack so we can do a scheduling decision if the NIC is in Hardware 
> Offload (switchdev) or regular SR-IOV (legacy) mode.
> And select the appropriate hypervisors with the requested capability see [1].
> 
> [1] - 
> https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fspe
> cs.openstack.org%2Fopenstack%2Fnova-specs%2Fspecs%2Fpike%2Fapproved%2F
> enable-sriov-nic-features.html=02%7C01%7Cedand%40mellanox.com%7C7
> 58a68d1796c435a8af108d4ea4e889b%7Ca652971c7d2e4d9ba6a4d149256f461b%7C0
> %7C0%7C636391070397000461=dbe8Z6P71ZEqC3u46YB1hkmmm8pQgYpWjjVxvB
> fu5Ko%3D=0
> ---
>  configure.ac  |  13 ++
>  docs/formatnode.html.in   |   1 +
>  src/util/virnetdev.c  | 187 
> +-
>  src/util/virnetdev.h  |   1 +
>  tests/nodedevschemadata/net_00_13_02_b9_f9_d3.xml |   1 +
>  tests/nodedevschemadata/net_00_15_58_2f_e9_55.xml |   1 +
>  6 files changed, 203 insertions(+), 1 deletion(-)
> 

Again - make check syntax-check had a failure because properly indent "#if" to 
be "# if" (also "# else" and "# endif")

I can fix that easily before pushing though.

Reviewed-by: John Ferlan <jfer...@redhat.com>

I'll wait a day before pushing - just in case laine comes back from chasing the 
eclipse and has a comment to make...

Tks -

John


--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities

2017-08-24 Thread Edan David 
Hi John,

For some reason I don't see these errors in my environment,
I would appreciate it if you could fix them before merging.
Thank you!!

-Original Message-
From: John Ferlan [mailto:jfer...@redhat.com] 
Sent: Wednesday, August 23, 2017 8:44 PM
To: Edan David <ed...@mellanox.com>; libvir-list@redhat.com
Subject: Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities



On 08/21/2017 05:19 AM, Edan David wrote:
> Adding functionality to libvirt that will allow querying the interface 
> for the availability of switchdev Offloading NIC capabilities.
> The switchdev mode was introduced in kernel 4.8, the iproute2-devlink 
> command to retrieve the swtichdev NIC feature, Command example:  
> devlink dev eswitch show pci/:03:00.0 This feature is needed for 
> Openstack so we can do a scheduling decision if the NIC is in Hardware 
> Offload (switchdev) or regular SR-IOV (legacy) mode.
> And select the appropriate hypervisors with the requested capability see [1].
> 
> [1] - 
> https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fspe
> cs.openstack.org%2Fopenstack%2Fnova-specs%2Fspecs%2Fpike%2Fapproved%2F
> enable-sriov-nic-features.html=02%7C01%7Cedand%40mellanox.com%7C7
> 58a68d1796c435a8af108d4ea4e889b%7Ca652971c7d2e4d9ba6a4d149256f461b%7C0
> %7C0%7C636391070397000461=dbe8Z6P71ZEqC3u46YB1hkmmm8pQgYpWjjVxvB
> fu5Ko%3D=0
> ---
>  configure.ac  |  13 ++
>  docs/formatnode.html.in   |   1 +
>  src/util/virnetdev.c  | 187 
> +-
>  src/util/virnetdev.h  |   1 +
>  tests/nodedevschemadata/net_00_13_02_b9_f9_d3.xml |   1 +
>  tests/nodedevschemadata/net_00_15_58_2f_e9_55.xml |   1 +
>  6 files changed, 203 insertions(+), 1 deletion(-)
> 

Again - make check syntax-check had a failure because properly indent "#if" to 
be "# if" (also "# else" and "# endif")

I can fix that easily before pushing though.

Reviewed-by: John Ferlan <jfer...@redhat.com>

I'll wait a day before pushing - just in case laine comes back from chasing the 
eclipse and has a comment to make...

Tks -

John


--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

Re: [libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities

2017-08-23 Thread John Ferlan 


On 08/21/2017 05:19 AM, Edan David wrote:
> Adding functionality to libvirt that will allow querying the interface
> for the availability of switchdev Offloading NIC capabilities.
> The switchdev mode was introduced in kernel 4.8, the iproute2-devlink
> command to retrieve the swtichdev NIC feature,
> Command example:  devlink dev eswitch show pci/:03:00.0
> This feature is needed for Openstack so we can do a scheduling decision
> if the NIC is in Hardware Offload (switchdev) or regular SR-IOV (legacy) mode.
> And select the appropriate hypervisors with the requested capability see [1].
> 
> [1] - 
> https://specs.openstack.org/openstack/nova-specs/specs/pike/approved/enable-sriov-nic-features.html
> ---
>  configure.ac  |  13 ++
>  docs/formatnode.html.in   |   1 +
>  src/util/virnetdev.c  | 187 
> +-
>  src/util/virnetdev.h  |   1 +
>  tests/nodedevschemadata/net_00_13_02_b9_f9_d3.xml |   1 +
>  tests/nodedevschemadata/net_00_15_58_2f_e9_55.xml |   1 +
>  6 files changed, 203 insertions(+), 1 deletion(-)
> 

Again - make check syntax-check had a failure because properly indent
"#if" to be "# if" (also "# else" and "# endif")

I can fix that easily before pushing though.

Reviewed-by: John Ferlan 

I'll wait a day before pushing - just in case laine comes back from
chasing the eclipse and has a comment to make...

Tks -

John

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

[libvirt] [PATCHv2] nodedev: add switchdev to NIC capabilities

2017-08-21 Thread Edan David 
Adding functionality to libvirt that will allow querying the interface
for the availability of switchdev Offloading NIC capabilities.
The switchdev mode was introduced in kernel 4.8, the iproute2-devlink
command to retrieve the swtichdev NIC feature,
Command example:  devlink dev eswitch show pci/:03:00.0
This feature is needed for Openstack so we can do a scheduling decision
if the NIC is in Hardware Offload (switchdev) or regular SR-IOV (legacy) mode.
And select the appropriate hypervisors with the requested capability see [1].

[1] - 
https://specs.openstack.org/openstack/nova-specs/specs/pike/approved/enable-sriov-nic-features.html
---
 configure.ac  |  13 ++
 docs/formatnode.html.in   |   1 +
 src/util/virnetdev.c  | 187 +-
 src/util/virnetdev.h  |   1 +
 tests/nodedevschemadata/net_00_13_02_b9_f9_d3.xml |   1 +
 tests/nodedevschemadata/net_00_15_58_2f_e9_55.xml |   1 +
 6 files changed, 203 insertions(+), 1 deletion(-)

diff --git a/configure.ac b/configure.ac
index b12b7fa..c089798 100644
--- a/configure.ac
+++ b/configure.ac
@@ -627,6 +627,19 @@ if test "$with_linux" = "yes"; then
 AC_CHECK_HEADERS([linux/btrfs.h])
 fi
 
+dnl
+dnl check for kernel headers required by devlink
+dnl
+if test "$with_linux" = "yes"; then
+AC_CHECK_HEADERS([linux/devlink.h])
+AC_CHECK_DECLS([DEVLINK_GENL_VERSION, DEVLINK_GENL_NAME, DEVLINK_ATTR_MAX, 
DEVLINK_CMD_ESWITCH_GET, DEVLINK_ATTR_BUS_NAME, DEVLINK_ATTR_DEV_NAME, 
DEVLINK_ATTR_ESWITCH_MODE, DEVLINK_ESWITCH_MODE_SWITCHDEV],
+   [AC_DEFINE([HAVE_DECL_DEVLINK],
+  [1],
+  [whether devlink declarations is available])],
+   [],
+   [[#include ]])
+fi
+
 dnl Allow perl/python overrides
 AC_PATH_PROGS([PYTHON], [python2 python])
 if test -z "$PYTHON"; then
diff --git a/docs/formatnode.html.in b/docs/formatnode.html.in
index 4d935b5..29244a8 100644
--- a/docs/formatnode.html.in
+++ b/docs/formatnode.html.in
@@ -227,6 +227,7 @@
 rxhashreceive-hashing
 
rdmaremote-direct-memory-access
 
txudptnltx-udp-tunnel-segmentation
+
switchdevkernel-forward-plane-offload
 
   
   capability
diff --git a/src/util/virnetdev.c b/src/util/virnetdev.c
index 51a6e42..fc7c961 100644
--- a/src/util/virnetdev.c
+++ b/src/util/virnetdev.c
@@ -59,6 +59,10 @@
 # include 
 #endif
 
+#if HAVE_DECL_DEVLINK
+# include 
+#endif
+
 #ifndef IFNAMSIZ
 # define IFNAMSIZ 16
 #endif
@@ -2481,7 +2485,8 @@ VIR_ENUM_IMPL(virNetDevFeature,
   "ntuple",
   "rxhash",
   "rdma",
-  "txudptnl")
+  "txudptnl",
+  "switchdev")
 
 #ifdef __linux__
 int
@@ -2936,6 +2941,7 @@ int virNetDevGetRxFilter(const char *ifname,
 return ret;
 }
 
+
 #if defined(SIOCETHTOOL) && defined(HAVE_STRUCT_IFREQ)
 
 /**
@@ -3115,6 +3121,182 @@ virNetDevGetEthtoolFeatures(virBitmapPtr bitmap,
 }
 
 
+#if HAVE_DECL_DEVLINK
+/**
+ * virNetDevPutExtraHeader
+ * reserve and prepare room for an extra header
+ * This function sets to zero the room that is required to put the extra
+ * header after the initial Netlink header. This function also increases
+ * the nlmsg_len field.
+ *
+ * @nlh: pointer to Netlink header
+ * @size: size of the extra header that we want to put
+ *
+ * Returns pointer to the start of the extended header
+ */
+static void *
+virNetDevPutExtraHeader(struct nlmsghdr *nlh,
+size_t size)
+{
+char *ptr = (char *)nlh + nlh->nlmsg_len;
+size_t len = NLMSG_ALIGN(size);
+nlh->nlmsg_len += len;
+memset(ptr, 0, len);
+return ptr;
+}
+
+
+/**
+ * virNetDevGetFamilyId:
+ * This function supplies the devlink family id
+ *
+ * @family_name: the name of the family to query
+ *
+ * Returns family id or 0 on failure.
+ */
+static uint32_t
+virNetDevGetFamilyId(const char *family_name)
+{
+struct nl_msg *nl_msg = NULL;
+struct nlmsghdr *resp = NULL;
+struct genlmsghdr* gmsgh = NULL;
+struct nlattr *tb[CTRL_ATTR_MAX + 1] = {NULL, };
+unsigned int recvbuflen;
+uint32_t family_id = 0;
+
+if (!(nl_msg = nlmsg_alloc_simple(GENL_ID_CTRL,
+  NLM_F_REQUEST | NLM_F_ACK))) {
+virReportOOMError();
+goto cleanup;
+}
+
+if (!(gmsgh = virNetDevPutExtraHeader(nlmsg_hdr(nl_msg), sizeof(struct 
genlmsghdr
+goto cleanup;
+
+gmsgh->cmd = CTRL_CMD_GETFAMILY;
+gmsgh->version = DEVLINK_GENL_VERSION;
+
+if (nla_put_string(nl_msg, CTRL_ATTR_FAMILY_NAME, family_name) < 0) {
+virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+   _("allocated netlink buffer is too small"));
+goto cleanup;
+}
+
+if (virNetlinkCommand(nl_msg, , ,