diff mbox

[dpdk-dev] net/virtio-user: specify the MAC of the tap

Message ID 1514518722-27302-1-git-send-email-muziding001@163.com (mailing list archive)
State Accepted, archived
Delegated to: Maxime Coquelin
Headers

Checks

Context Check Description
ci/checkpatch success coding style OK
ci/Intel-compilation success Compilation OK

Commit Message

Ning Li Dec. 29, 2017, 3:38 a.m. UTC 
  When using virtio-user with vhost-kernel to exchange
packet with kernel networking stack, application can
set the MAC of the tap interface via parameter.

Signed-off-by: Ning Li <muziding001@163.com>
---
 drivers/net/virtio/virtio_user/vhost_kernel.c     |  3 ++-
 drivers/net/virtio/virtio_user/vhost_kernel_tap.c | 14 +++++++++++++-
 drivers/net/virtio/virtio_user/vhost_kernel_tap.h |  3 ++-
 3 files changed, 17 insertions(+), 3 deletions(-)

Comments

Tiwei Bie Dec. 29, 2017, 9:44 a.m. UTC | #1 
Hi Ning,

On Fri, Dec 29, 2017 at 11:38:42AM +0800, Ning Li wrote:
> When using virtio-user with vhost-kernel to exchange
> packet with kernel networking stack, application can
> set the MAC of the tap interface via parameter.
> 
> Signed-off-by: Ning Li <muziding001@163.com>
> ---

Thanks for the new version.

Just FYI, when sending a new version, you also need to
add the version number, e.g. using -v2 when generating
the patch. You can find more details in the "Contribute
by sending patches" section in below link:

http://dpdk.org/dev

>  drivers/net/virtio/virtio_user/vhost_kernel.c     |  3 ++-
>  drivers/net/virtio/virtio_user/vhost_kernel_tap.c | 14 +++++++++++++-
>  drivers/net/virtio/virtio_user/vhost_kernel_tap.h |  3 ++-
>  3 files changed, 17 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/net/virtio/virtio_user/vhost_kernel.c b/drivers/net/virtio/virtio_user/vhost_kernel.c
> index 68d28b1..dd24b6b 100644
> --- a/drivers/net/virtio/virtio_user/vhost_kernel.c
> +++ b/drivers/net/virtio/virtio_user/vhost_kernel.c
> @@ -380,7 +380,8 @@ struct vhost_memory_kernel {
>  	else
>  		hdr_size = sizeof(struct virtio_net_hdr);
>  
> -	tapfd = vhost_kernel_open_tap(&dev->ifname, hdr_size, req_mq);
> +	tapfd = vhost_kernel_open_tap(&dev->ifname, hdr_size, req_mq,
> +			 (char *)dev->mac_addr);

I think it's better to add a new device argument for
virtio-user to specify the MAC for the corresponding
tap. But I don't have a very strong opinion on this
for now. So I'd like to hear others' opinions.

>  	if (tapfd < 0) {
>  		PMD_DRV_LOG(ERR, "fail to open tap for vhost kernel");
>  		return -1;
> diff --git a/drivers/net/virtio/virtio_user/vhost_kernel_tap.c b/drivers/net/virtio/virtio_user/vhost_kernel_tap.c
> index 689a5cf..d036428 100644
> --- a/drivers/net/virtio/virtio_user/vhost_kernel_tap.c
> +++ b/drivers/net/virtio/virtio_user/vhost_kernel_tap.c
[...]
> +	memset(&ifr, 0, sizeof(ifr));
> +	ifr.ifr_hwaddr.sa_family = ARPHRD_ETHER;

Just to address others' potential doubts on this, as
I was confused about the macro used here after I saw
the macro used by tap PMD is AF_LOCAL. I checked the
tap related code in Linux, sa_family needs to be the
same with dev->type, otherwise it will return -EINVAL.
For tap, dev->type is initialized to ARPHRD_ETHER in
ether_setup().

Thanks,
Tiwei

> +	memcpy(ifr.ifr_hwaddr.sa_data, mac, ETHER_ADDR_LEN);
> +	if (ioctl(tapfd, SIOCSIFHWADDR, (void *)&ifr) == -1) {
> +		PMD_DRV_LOG(ERR, "SIOCSIFHWADDR failed: %s", strerror(errno));
> +		goto error;
> +	}
> +
>  	if (!(*p_ifname))
>  		*p_ifname = strdup(ifr.ifr_name);
>  
> diff --git a/drivers/net/virtio/virtio_user/vhost_kernel_tap.h b/drivers/net/virtio/virtio_user/vhost_kernel_tap.h
> index eae340c..402f964 100644
> --- a/drivers/net/virtio/virtio_user/vhost_kernel_tap.h
> +++ b/drivers/net/virtio/virtio_user/vhost_kernel_tap.h
> @@ -64,4 +64,5 @@
>  /* Constants */
>  #define PATH_NET_TUN	"/dev/net/tun"
>  
> -int vhost_kernel_open_tap(char **p_ifname, int hdr_size, int req_mq);
> +int vhost_kernel_open_tap(char **p_ifname, int hdr_size, int req_mq,
> +			 const char *mac);
> -- 
> 1.8.3.1
> 
>
Sean Harte March 28, 2018, 4:54 p.m. UTC | #2 
On 29/12/2017 09:44, tiwei.bie at intel.com (Tiwei Bie) wrote:
> Hi Ning,
> 
> On Fri, Dec 29, 2017 at 11:38:42AM +0800, Ning Li wrote:
>> When using virtio-user with vhost-kernel to exchange
>> packet with kernel networking stack, application can
>> set the MAC of the tap interface via parameter.
>>
>> Signed-off-by: Ning Li <muziding001 at 163.com>
>> ---
> 
> Thanks for the new version.
> 
> Just FYI, when sending a new version, you also need to
> add the version number, e.g. using -v2 when generating
> the patch. You can find more details in the "Contribute
> by sending patches" section in below link:
> 
> http://dpdk.org/dev
> 
>>   drivers/net/virtio/virtio_user/vhost_kernel.c     |  3 ++-
>>   drivers/net/virtio/virtio_user/vhost_kernel_tap.c | 14 +++++++++++++-
>>   drivers/net/virtio/virtio_user/vhost_kernel_tap.h |  3 ++-
>>   3 files changed, 17 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/net/virtio/virtio_user/vhost_kernel.c b/drivers/net/virtio/virtio_user/vhost_kernel.c
>> index 68d28b1..dd24b6b 100644
>> --- a/drivers/net/virtio/virtio_user/vhost_kernel.c
>> +++ b/drivers/net/virtio/virtio_user/vhost_kernel.c
>> @@ -380,7 +380,8 @@ struct vhost_memory_kernel {
>>   	else
>>   		hdr_size = sizeof(struct virtio_net_hdr);
>>   
>> -	tapfd = vhost_kernel_open_tap(&dev->ifname, hdr_size, req_mq);
>> +	tapfd = vhost_kernel_open_tap(&dev->ifname, hdr_size, req_mq,
>> +			 (char *)dev->mac_addr);
> 
> I think it's better to add a new device argument for
> virtio-user to specify the MAC for the corresponding
> tap. But I don't have a very strong opinion on this
> for now. So I'd like to hear others' opinions.

No harm if it was a seperate argument, although I can't think of a 
scenario where you care about the MAC address and would want the tap and 
virtio devices to have differnet MAC addresses.

> 

snip...

> 

I'm not sure if this patch is still under consideration, but it looks 
good to me, and works.

Reviewed-by: Seán Harte <seanbh@gmail.com>
Tested-by: Seán Harte <seanbh@gmail.com>
Jianfeng Tan March 29, 2018, 12:50 a.m. UTC | #3 
> -----Original Message-----

> From: Seán Harte [mailto:seanbh@gmail.com]

> Sent: Thursday, March 29, 2018 12:55 AM

> To: Ning Li; Yuanhan Liu; Maxime Coquelin; Bie, Tiwei

> Cc: dev@dpdk.org; Tan, Jianfeng

> Subject: Re: [dpdk-dev] [PATCH] net/virtio-user: specify the MAC of the tap

> 

> On 29/12/2017 09:44, tiwei.bie at intel.com (Tiwei Bie) wrote:

> > Hi Ning,

> >

> > On Fri, Dec 29, 2017 at 11:38:42AM +0800, Ning Li wrote:

> >> When using virtio-user with vhost-kernel to exchange

> >> packet with kernel networking stack, application can

> >> set the MAC of the tap interface via parameter.

> >>

> >> Signed-off-by: Ning Li <muziding001 at 163.com>

> 

> Reviewed-by: Seán Harte <seanbh@gmail.com>

> Tested-by: Seán Harte <seanbh@gmail.com>


Reviewed-by: Jianfeng Tan <jianfeng.tan@intel.com>


> >> ---

> >

> > Thanks for the new version.

> >

> > Just FYI, when sending a new version, you also need to

> > add the version number, e.g. using -v2 when generating

> > the patch. You can find more details in the "Contribute

> > by sending patches" section in below link:

> >

> > http://dpdk.org/dev

> >

> >>   drivers/net/virtio/virtio_user/vhost_kernel.c     |  3 ++-

> >>   drivers/net/virtio/virtio_user/vhost_kernel_tap.c | 14 +++++++++++++-

> >>   drivers/net/virtio/virtio_user/vhost_kernel_tap.h |  3 ++-

> >>   3 files changed, 17 insertions(+), 3 deletions(-)

> >>

> >> diff --git a/drivers/net/virtio/virtio_user/vhost_kernel.c

> b/drivers/net/virtio/virtio_user/vhost_kernel.c

> >> index 68d28b1..dd24b6b 100644

> >> --- a/drivers/net/virtio/virtio_user/vhost_kernel.c

> >> +++ b/drivers/net/virtio/virtio_user/vhost_kernel.c

> >> @@ -380,7 +380,8 @@ struct vhost_memory_kernel {

> >>   	else

> >>   		hdr_size = sizeof(struct virtio_net_hdr);

> >>

> >> -	tapfd = vhost_kernel_open_tap(&dev->ifname, hdr_size, req_mq);

> >> +	tapfd = vhost_kernel_open_tap(&dev->ifname, hdr_size, req_mq,

> >> +			 (char *)dev->mac_addr);

> >

> > I think it's better to add a new device argument for

> > virtio-user to specify the MAC for the corresponding

> > tap. But I don't have a very strong opinion on this

> > for now. So I'd like to hear others' opinions.

> 

> No harm if it was a seperate argument, although I can't think of a

> scenario where you care about the MAC address and would want the tap and

> virtio devices to have differnet MAC addresses.


I would also prefer to keep it as simple as possible before we find a real use case for different mac addr.


> 

> >

> 

> snip...

> 

> >

> 

> I'm not sure if this patch is still under consideration, but it looks

> good to me, and works.


I think yes. Maxime, would you like to consider merging this patch?

Thanks,
Jianfeng
Tiwei Bie March 29, 2018, 1:15 a.m. UTC | #4 
On Thu, Mar 29, 2018 at 08:50:37AM +0800, Tan, Jianfeng wrote:
> > -----Original Message-----
> > From: Seán Harte [mailto:seanbh@gmail.com]
> > Sent: Thursday, March 29, 2018 12:55 AM
> > To: Ning Li; Yuanhan Liu; Maxime Coquelin; Bie, Tiwei
> > Cc: dev@dpdk.org; Tan, Jianfeng
> > Subject: Re: [dpdk-dev] [PATCH] net/virtio-user: specify the MAC of the tap
> > 
> > On 29/12/2017 09:44, tiwei.bie at intel.com (Tiwei Bie) wrote:
> > > Hi Ning,
> > >
> > > On Fri, Dec 29, 2017 at 11:38:42AM +0800, Ning Li wrote:
> > >> When using virtio-user with vhost-kernel to exchange
> > >> packet with kernel networking stack, application can
> > >> set the MAC of the tap interface via parameter.
> > >>
> > >> Signed-off-by: Ning Li <muziding001 at 163.com>
> > 
> > Reviewed-by: Seán Harte <seanbh@gmail.com>
> > Tested-by: Seán Harte <seanbh@gmail.com>
> 
> Reviewed-by: Jianfeng Tan <jianfeng.tan@intel.com>
> 
> > >> ---
[...]
> > >> --- a/drivers/net/virtio/virtio_user/vhost_kernel.c
> > >> +++ b/drivers/net/virtio/virtio_user/vhost_kernel.c
> > >> @@ -380,7 +380,8 @@ struct vhost_memory_kernel {
> > >>   	else
> > >>   		hdr_size = sizeof(struct virtio_net_hdr);
> > >>
> > >> -	tapfd = vhost_kernel_open_tap(&dev->ifname, hdr_size, req_mq);
> > >> +	tapfd = vhost_kernel_open_tap(&dev->ifname, hdr_size, req_mq,
> > >> +			 (char *)dev->mac_addr);
> > >
> > > I think it's better to add a new device argument for
> > > virtio-user to specify the MAC for the corresponding
> > > tap. But I don't have a very strong opinion on this
> > > for now. So I'd like to hear others' opinions.
> > 
> > No harm if it was a seperate argument, although I can't think of a
> > scenario where you care about the MAC address and would want the tap and
> > virtio devices to have differnet MAC addresses.
> 
> I would also prefer to keep it as simple as possible before we find a real use case for different mac addr.
> 

Logically they're different virtual ports, so my first thought
is that they deserve different devargs and MACs. But after a
second thought, I also can't find a scenario where we would want
the tap and virtio device to have different MACs. So I'd like to
hear others' opinions. If that's the case, maybe we can take it.

Thanks

> 
> > 
> > >
> > 
> > snip...
> > 
> > >
> > 
> > I'm not sure if this patch is still under consideration, but it looks
> > good to me, and works.
> 
> I think yes. Maxime, would you like to consider merging this patch?
> 
> Thanks,
> Jianfeng
Maxime Coquelin March 29, 2018, 11:45 a.m. UTC | #5 
On 12/29/2017 04:38 AM, Ning Li wrote:
> When using virtio-user with vhost-kernel to exchange
> packet with kernel networking stack, application can
> set the MAC of the tap interface via parameter.
> 
> Signed-off-by: Ning Li <muziding001@163.com>
> ---
>   drivers/net/virtio/virtio_user/vhost_kernel.c     |  3 ++-
>   drivers/net/virtio/virtio_user/vhost_kernel_tap.c | 14 +++++++++++++-
>   drivers/net/virtio/virtio_user/vhost_kernel_tap.h |  3 ++-
>   3 files changed, 17 insertions(+), 3 deletions(-)
> 

Applied to dpdk-next-virtio/master.

Thanks!
Maxime
diff mbox

Patch

diff --git a/drivers/net/virtio/virtio_user/vhost_kernel.c b/drivers/net/virtio/virtio_user/vhost_kernel.c
index 68d28b1..dd24b6b 100644
--- a/drivers/net/virtio/virtio_user/vhost_kernel.c
+++ b/drivers/net/virtio/virtio_user/vhost_kernel.c
@@ -380,7 +380,8 @@  struct vhost_memory_kernel {
 	else
 		hdr_size = sizeof(struct virtio_net_hdr);
 
-	tapfd = vhost_kernel_open_tap(&dev->ifname, hdr_size, req_mq);
+	tapfd = vhost_kernel_open_tap(&dev->ifname, hdr_size, req_mq,
+			 (char *)dev->mac_addr);
 	if (tapfd < 0) {
 		PMD_DRV_LOG(ERR, "fail to open tap for vhost kernel");
 		return -1;
diff --git a/drivers/net/virtio/virtio_user/vhost_kernel_tap.c b/drivers/net/virtio/virtio_user/vhost_kernel_tap.c
index 689a5cf..d036428 100644
--- a/drivers/net/virtio/virtio_user/vhost_kernel_tap.c
+++ b/drivers/net/virtio/virtio_user/vhost_kernel_tap.c
@@ -36,15 +36,19 @@ 
 #include <sys/stat.h>
 #include <fcntl.h>
 #include <net/if.h>
+#include <net/if_arp.h>
 #include <errno.h>
 #include <string.h>
 #include <limits.h>
 
+#include <rte_ether.h>
+
 #include "vhost_kernel_tap.h"
 #include "../virtio_logs.h"
 
 int
-vhost_kernel_open_tap(char **p_ifname, int hdr_size, int req_mq)
+vhost_kernel_open_tap(char **p_ifname, int hdr_size, int req_mq,
+			 const char *mac)
 {
 	unsigned int tap_features;
 	int sndbuf = INT_MAX;
@@ -123,6 +127,14 @@ 
 		PMD_DRV_LOG(ERR, "TUNSETOFFLOAD ioctl() failed: %s",
 			   strerror(errno));
 
+	memset(&ifr, 0, sizeof(ifr));
+	ifr.ifr_hwaddr.sa_family = ARPHRD_ETHER;
+	memcpy(ifr.ifr_hwaddr.sa_data, mac, ETHER_ADDR_LEN);
+	if (ioctl(tapfd, SIOCSIFHWADDR, (void *)&ifr) == -1) {
+		PMD_DRV_LOG(ERR, "SIOCSIFHWADDR failed: %s", strerror(errno));
+		goto error;
+	}
+
 	if (!(*p_ifname))
 		*p_ifname = strdup(ifr.ifr_name);
 
diff --git a/drivers/net/virtio/virtio_user/vhost_kernel_tap.h b/drivers/net/virtio/virtio_user/vhost_kernel_tap.h
index eae340c..402f964 100644
--- a/drivers/net/virtio/virtio_user/vhost_kernel_tap.h
+++ b/drivers/net/virtio/virtio_user/vhost_kernel_tap.h
@@ -64,4 +64,5 @@ 
 /* Constants */
 #define PATH_NET_TUN	"/dev/net/tun"
 
-int vhost_kernel_open_tap(char **p_ifname, int hdr_size, int req_mq);
+int vhost_kernel_open_tap(char **p_ifname, int hdr_size, int req_mq,
+			 const char *mac);