[PATCH iproute2 v2] route: label externally offloaded routes

[PATCH iproute2 v2] route: label externally offloaded routes

[sfeldma]

From: Scott Feldman <sfeldma@gmail.com>

v2:

No clear outcome on polling for choices of "external", "offload", "hw", or
"hardware".  Andy's concern about "external" being confusing for users is
valid.  So, I'm making executive decision to use "external_offload" for label,
as suggested by Jiri.

v1:

On ip route print dump, label externally offloaded routes with "external".
Offloaded routes are flagged with RTNH_F_EXTERNAL, a recent additon to
net-next.  For example:

$ ip route
default via 192.168.0.2 dev eth0
11.0.0.0/30 dev swp1  proto kernel  scope link  src 11.0.0.2 external
11.0.0.4/30 via 11.0.0.1 dev swp1  proto zebra  metric 20 external
11.0.0.8/30 dev swp2  proto kernel  scope link  src 11.0.0.10 external
11.0.0.12/30 via 11.0.0.9 dev swp2  proto zebra  metric 20 external
12.0.0.2  proto zebra  metric 30 external
        nexthop via 11.0.0.1  dev swp1 weight 1
        nexthop via 11.0.0.9  dev swp2 weight 1
12.0.0.3 via 11.0.0.1 dev swp1  proto zebra  metric 20 external
12.0.0.4 via 11.0.0.9 dev swp2  proto zebra  metric 20 external
192.168.0.0/24 dev eth0  proto kernel  scope link  src 192.168.0.15

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Acked-by: Jiri Pirko <jiri@resnulli.us>
---
 include/linux/rtnetlink.h |    1 +
 ip/iproute.c              |    2 ++
 2 files changed, 3 insertions(+)

diff --git a/include/linux/rtnetlink.h b/include/linux/rtnetlink.h
index 3eb7810..c74773c 100644
--- a/include/linux/rtnetlink.h
+++ b/include/linux/rtnetlink.h
@@ -332,6 +332,7 @@ struct rtnexthop {
 #define RTNH_F_DEAD		1	/* Nexthop is dead (used by multipath)	*/
 #define RTNH_F_PERVASIVE	2	/* Do recursive gateway lookup	*/
 #define RTNH_F_ONLINK		4	/* Gateway is forced on link	*/
+#define RTNH_F_EXTERNAL		8	/* Route installed externally   */
 
 /* Macros to handle hexthops */
 
diff --git a/ip/iproute.c b/ip/iproute.c
index b32025f..ffdf4f2 100644
--- a/ip/iproute.c
+++ b/ip/iproute.c
@@ -412,6 +412,8 @@ int print_route(const struct sockaddr_nl *who, struct nlmsghdr *n, void *arg)
 		fprintf(fp, "onlink ");
 	if (r->rtm_flags & RTNH_F_PERVASIVE)
 		fprintf(fp, "pervasive ");
+	if (r->rtm_flags & RTNH_F_EXTERNAL)
+		fprintf(fp, "external_offload ");
 	if (r->rtm_flags & RTM_F_NOTIFY)
 		fprintf(fp, "notify ");
 	if (tb[RTA_MARK]) {
-- 
1.7.10.4

Re: [PATCH iproute2 v2] route: label externally offloaded routes

[Andy Gospodarek]

On Sun, Mar 15, 2015 at 12:31:34PM -0700, sfeldma@gmail.com wrote:
> From: Scott Feldman <sfeldma@gmail.com>
> 
> v2:
> 
> No clear outcome on polling for choices of "external", "offload", "hw", or
> "hardware".  Andy's concern about "external" being confusing for users is
> valid.  So, I'm making executive decision to use "external_offload" for label,
> as suggested by Jiri.

This seems like quite a mouthful.  Just make it 'offload' or 'hardware'
so the output of 'ip route show' isn't so long that it will frequently
be longer than 80 chars.  Being longer than 80 chars is going to be
irritating for _many_ users.

> 
> v1:
> 
> On ip route print dump, label externally offloaded routes with "external".
> Offloaded routes are flagged with RTNH_F_EXTERNAL, a recent additon to
> net-next.  For example:
> 
> $ ip route
> default via 192.168.0.2 dev eth0
> 11.0.0.0/30 dev swp1  proto kernel  scope link  src 11.0.0.2 external
> 11.0.0.4/30 via 11.0.0.1 dev swp1  proto zebra  metric 20 external
> 11.0.0.8/30 dev swp2  proto kernel  scope link  src 11.0.0.10 external
> 11.0.0.12/30 via 11.0.0.9 dev swp2  proto zebra  metric 20 external
> 12.0.0.2  proto zebra  metric 30 external
>         nexthop via 11.0.0.1  dev swp1 weight 1
>         nexthop via 11.0.0.9  dev swp2 weight 1
> 12.0.0.3 via 11.0.0.1 dev swp1  proto zebra  metric 20 external
> 12.0.0.4 via 11.0.0.9 dev swp2  proto zebra  metric 20 external
> 192.168.0.0/24 dev eth0  proto kernel  scope link  src 192.168.0.15

I understand why you did this the way you did (keeping the changelog the
same with the only change being the parts listed in 'v2'), but could you
change the new output to reflect the code change?  Readers of the git
log would probably be confused by this.

(Sorry to debate such a simple patch, but this is the interface most will
use so it seems like a good idea to make sure we are thinking about it
in the proper context.  Healthy debate on the mailing-list seems like a
great way to track the reason behind some of these decisions.)


> Signed-off-by: Scott Feldman <sfeldma@gmail.com>
> Acked-by: Jiri Pirko <jiri@resnulli.us>
> ---
>  include/linux/rtnetlink.h |    1 +
>  ip/iproute.c              |    2 ++
>  2 files changed, 3 insertions(+)
> 
> diff --git a/include/linux/rtnetlink.h b/include/linux/rtnetlink.h
> index 3eb7810..c74773c 100644
> --- a/include/linux/rtnetlink.h
> +++ b/include/linux/rtnetlink.h
> @@ -332,6 +332,7 @@ struct rtnexthop {
>  #define RTNH_F_DEAD		1	/* Nexthop is dead (used by multipath)	*/
>  #define RTNH_F_PERVASIVE	2	/* Do recursive gateway lookup	*/
>  #define RTNH_F_ONLINK		4	/* Gateway is forced on link	*/
> +#define RTNH_F_EXTERNAL		8	/* Route installed externally   */
>  
>  /* Macros to handle hexthops */
>  
> diff --git a/ip/iproute.c b/ip/iproute.c
> index b32025f..ffdf4f2 100644
> --- a/ip/iproute.c
> +++ b/ip/iproute.c
> @@ -412,6 +412,8 @@ int print_route(const struct sockaddr_nl *who, struct nlmsghdr *n, void *arg)
>  		fprintf(fp, "onlink ");
>  	if (r->rtm_flags & RTNH_F_PERVASIVE)
>  		fprintf(fp, "pervasive ");
> +	if (r->rtm_flags & RTNH_F_EXTERNAL)
> +		fprintf(fp, "external_offload ");
>  	if (r->rtm_flags & RTM_F_NOTIFY)
>  		fprintf(fp, "notify ");
>  	if (tb[RTA_MARK]) {
> -- 
> 1.7.10.4
> 

Re: [PATCH iproute2 v2] route: label externally offloaded routes

[roopa]

On 3/16/15, 6:18 AM, Andy Gospodarek wrote:
> On Sun, Mar 15, 2015 at 12:31:34PM -0700, sfeldma@gmail.com wrote:
>> From: Scott Feldman <sfeldma@gmail.com>
>>
>> v2:
>>
>> No clear outcome on polling for choices of "external", "offload", "hw", or
>> "hardware".  Andy's concern about "external" being confusing for users is
>> valid.  So, I'm making executive decision to use "external_offload" for label,
>> as suggested by Jiri.
> This seems like quite a mouthful.  Just make it 'offload' or 'hardware'
> so the output of 'ip route show' isn't so long that it will frequently
> be longer than 80 chars.  Being longer than 80 chars is going to be
> irritating for _many_ users.
>
>
+1

Re: [PATCH iproute2 v2] route: label externally offloaded routes

[Scott Feldman]

On Mon, Mar 16, 2015 at 6:18 AM, Andy Gospodarek
<gospo@cumulusnetworks.com> wrote:
> On Sun, Mar 15, 2015 at 12:31:34PM -0700, sfeldma@gmail.com wrote:
>> From: Scott Feldman <sfeldma@gmail.com>
>>
>> v2:
>>
>> No clear outcome on polling for choices of "external", "offload", "hw", or
>> "hardware".  Andy's concern about "external" being confusing for users is
>> valid.  So, I'm making executive decision to use "external_offload" for label,
>> as suggested by Jiri.
>
> This seems like quite a mouthful.  Just make it 'offload' or 'hardware'
> so the output of 'ip route show' isn't so long that it will frequently
> be longer than 80 chars.  Being longer than 80 chars is going to be
> irritating for _many_ users.

Andy, since you have passion on this topic, would you mind taking it
from here to get consensus, and post an updated patch, if needed?

I was about to suggest a new made-up work like "keroff" and someday
the new word is added to Websters .  Daddy, what's "keroff"?  Oh son,
that means "offloaded from kernel to external device".  Ya, actually,
that's my vote: keroff.

("koff" was my first choice, but apparently that already has a
not-so-nice meaning).

-scott

Re: [PATCH iproute2 v2] route: label externally offloaded routes

[Jiri Pirko]

Tue, Mar 17, 2015 at 05:14:50AM CET, sfeldma@gmail.com wrote:
>On Mon, Mar 16, 2015 at 6:18 AM, Andy Gospodarek
><gospo@cumulusnetworks.com> wrote:
>> On Sun, Mar 15, 2015 at 12:31:34PM -0700, sfeldma@gmail.com wrote:
>>> From: Scott Feldman <sfeldma@gmail.com>
>>>
>>> v2:
>>>
>>> No clear outcome on polling for choices of "external", "offload", "hw", or
>>> "hardware".  Andy's concern about "external" being confusing for users is
>>> valid.  So, I'm making executive decision to use "external_offload" for label,
>>> as suggested by Jiri.
>>
>> This seems like quite a mouthful.  Just make it 'offload' or 'hardware'
>> so the output of 'ip route show' isn't so long that it will frequently
>> be longer than 80 chars.  Being longer than 80 chars is going to be
>> irritating for _many_ users.
>
>Andy, since you have passion on this topic, would you mind taking it
>from here to get consensus, and post an updated patch, if needed?
>
>I was about to suggest a new made-up work like "keroff" and someday
>the new word is added to Websters .  Daddy, what's "keroff"?  Oh son,
>that means "offloaded from kernel to external device".  Ya, actually,
>that's my vote: keroff.

I like it :)

>
>("koff" was my first choice, but apparently that already has a
>not-so-nice meaning).
>
>-scott

Re: [PATCH iproute2 v2] route: label externally offloaded routes

[Andy Gospodarek]

On Mon, Mar 16, 2015 at 09:14:50PM -0700, Scott Feldman wrote:
> On Mon, Mar 16, 2015 at 6:18 AM, Andy Gospodarek
> <gospo@cumulusnetworks.com> wrote:
> > On Sun, Mar 15, 2015 at 12:31:34PM -0700, sfeldma@gmail.com wrote:
> >> From: Scott Feldman <sfeldma@gmail.com>
> >>
> >> v2:
> >>
> >> No clear outcome on polling for choices of "external", "offload", "hw", or
> >> "hardware".  Andy's concern about "external" being confusing for users is
> >> valid.  So, I'm making executive decision to use "external_offload" for label,
> >> as suggested by Jiri.
> >
> > This seems like quite a mouthful.  Just make it 'offload' or 'hardware'
> > so the output of 'ip route show' isn't so long that it will frequently
> > be longer than 80 chars.  Being longer than 80 chars is going to be
> > irritating for _many_ users.
> 
> Andy, since you have passion on this topic, would you mind taking it
> from here to get consensus, and post an updated patch, if needed?

Absolutely.  I felt bad responding with more complaints yesterday
without proposing an alternative patch.

I'll be happy to handle this -- it will elimininate at least one
complainer!


> I was about to suggest a new made-up work like "keroff" and someday
> the new word is added to Websters .  Daddy, what's "keroff"?  Oh son,
> that means "offloaded from kernel to external device".  Ya, actually,
> that's my vote: keroff.

Bless you!

Re: [PATCH iproute2 v2] route: label externally offloaded routes

[Jiri Pirko]

Tue, Mar 17, 2015 at 02:24:56PM CET, gospo@cumulusnetworks.com wrote:
>On Mon, Mar 16, 2015 at 09:14:50PM -0700, Scott Feldman wrote:
>> On Mon, Mar 16, 2015 at 6:18 AM, Andy Gospodarek
>> <gospo@cumulusnetworks.com> wrote:
>> > On Sun, Mar 15, 2015 at 12:31:34PM -0700, sfeldma@gmail.com wrote:
>> >> From: Scott Feldman <sfeldma@gmail.com>
>> >>
>> >> v2:
>> >>
>> >> No clear outcome on polling for choices of "external", "offload", "hw", or
>> >> "hardware".  Andy's concern about "external" being confusing for users is
>> >> valid.  So, I'm making executive decision to use "external_offload" for label,
>> >> as suggested by Jiri.
>> >
>> > This seems like quite a mouthful.  Just make it 'offload' or 'hardware'
>> > so the output of 'ip route show' isn't so long that it will frequently
>> > be longer than 80 chars.  Being longer than 80 chars is going to be
>> > irritating for _many_ users.
>> 
>> Andy, since you have passion on this topic, would you mind taking it
>> from here to get consensus, and post an updated patch, if needed?
>
>Absolutely.  I felt bad responding with more complaints yesterday
>without proposing an alternative patch.
>
>I'll be happy to handle this -- it will elimininate at least one
>complainer!
>
>
>> I was about to suggest a new made-up work like "keroff" and someday
>> the new word is added to Websters .  Daddy, what's "keroff"?  Oh son,
>> that means "offloaded from kernel to external device".  Ya, actually,
>> that's my vote: keroff.

Hmm, on the second thought, shouldn't it be rather "extoff" ?

>
>Bless you!
>

Re: [PATCH iproute2 v2] route: label externally offloaded routes

[Andy Gospodarek]

On Tue, Mar 17, 2015 at 02:49:21PM +0100, Jiri Pirko wrote:
> Tue, Mar 17, 2015 at 02:24:56PM CET, gospo@cumulusnetworks.com wrote:
> >On Mon, Mar 16, 2015 at 09:14:50PM -0700, Scott Feldman wrote:
> >> On Mon, Mar 16, 2015 at 6:18 AM, Andy Gospodarek
> >> <gospo@cumulusnetworks.com> wrote:
> >> > On Sun, Mar 15, 2015 at 12:31:34PM -0700, sfeldma@gmail.com wrote:
> >> >> From: Scott Feldman <sfeldma@gmail.com>
> >> >>
> >> >> v2:
> >> >>
> >> >> No clear outcome on polling for choices of "external", "offload", "hw", or
> >> >> "hardware".  Andy's concern about "external" being confusing for users is
> >> >> valid.  So, I'm making executive decision to use "external_offload" for label,
> >> >> as suggested by Jiri.
> >> >
> >> > This seems like quite a mouthful.  Just make it 'offload' or 'hardware'
> >> > so the output of 'ip route show' isn't so long that it will frequently
> >> > be longer than 80 chars.  Being longer than 80 chars is going to be
> >> > irritating for _many_ users.
> >> 
> >> Andy, since you have passion on this topic, would you mind taking it
> >> from here to get consensus, and post an updated patch, if needed?
> >
> >Absolutely.  I felt bad responding with more complaints yesterday
> >without proposing an alternative patch.
> >
> >I'll be happy to handle this -- it will elimininate at least one
> >complainer!
> >
> >
> >> I was about to suggest a new made-up work like "keroff" and someday
> >> the new word is added to Websters .  Daddy, what's "keroff"?  Oh son,
> >> that means "offloaded from kernel to external device".  Ya, actually,
> >> that's my vote: keroff.
> 
> Hmm, on the second thought, shouldn't it be rather "extoff" ?
> 

I'm really liking the sound of 'extoffk'  ;)

Re: [PATCH iproute2 v2] route: label externally offloaded routes

[Stephen Hemminger]

On Tue, 17 Mar 2015 15:15:57 -0400
Andy Gospodarek <gospo@cumulusnetworks.com> wrote:

> On Tue, Mar 17, 2015 at 02:49:21PM +0100, Jiri Pirko wrote:
> > Tue, Mar 17, 2015 at 02:24:56PM CET, gospo@cumulusnetworks.com wrote:
> > >On Mon, Mar 16, 2015 at 09:14:50PM -0700, Scott Feldman wrote:
> > >> On Mon, Mar 16, 2015 at 6:18 AM, Andy Gospodarek
> > >> <gospo@cumulusnetworks.com> wrote:
> > >> > On Sun, Mar 15, 2015 at 12:31:34PM -0700, sfeldma@gmail.com wrote:
> > >> >> From: Scott Feldman <sfeldma@gmail.com>
> > >> >>
> > >> >> v2:
> > >> >>
> > >> >> No clear outcome on polling for choices of "external", "offload", "hw", or
> > >> >> "hardware".  Andy's concern about "external" being confusing for users is
> > >> >> valid.  So, I'm making executive decision to use "external_offload" for label,
> > >> >> as suggested by Jiri.
> > >> >
> > >> > This seems like quite a mouthful.  Just make it 'offload' or 'hardware'
> > >> > so the output of 'ip route show' isn't so long that it will frequently
> > >> > be longer than 80 chars.  Being longer than 80 chars is going to be
> > >> > irritating for _many_ users.
> > >> 
> > >> Andy, since you have passion on this topic, would you mind taking it
> > >> from here to get consensus, and post an updated patch, if needed?
> > >
> > >Absolutely.  I felt bad responding with more complaints yesterday
> > >without proposing an alternative patch.
> > >
> > >I'll be happy to handle this -- it will elimininate at least one
> > >complainer!
> > >
> > >
> > >> I was about to suggest a new made-up work like "keroff" and someday
> > >> the new word is added to Websters .  Daddy, what's "keroff"?  Oh son,
> > >> that means "offloaded from kernel to external device".  Ya, actually,
> > >> that's my vote: keroff.
> > 
> > Hmm, on the second thought, shouldn't it be rather "extoff" ?
> > 
> 
> I'm really liking the sound of 'extoffk'  ;)
> 

I am bouncing this patch, come back when you decide on a name.

Re: [PATCH iproute2 v2] route: label externally offloaded routes

[Andy Gospodarek]

On Tue, Mar 24, 2015 at 03:51:37PM -0700, Stephen Hemminger wrote:
jG> On Tue, 17 Mar 2015 15:15:57 -0400
> Andy Gospodarek <gospo@cumulusnetworks.com> wrote:
> 
> > On Tue, Mar 17, 2015 at 02:49:21PM +0100, Jiri Pirko wrote:
> > > Tue, Mar 17, 2015 at 02:24:56PM CET, gospo@cumulusnetworks.com wrote:
> > > >On Mon, Mar 16, 2015 at 09:14:50PM -0700, Scott Feldman wrote:
> > > >> On Mon, Mar 16, 2015 at 6:18 AM, Andy Gospodarek
> > > >> <gospo@cumulusnetworks.com> wrote:
> > > >> > On Sun, Mar 15, 2015 at 12:31:34PM -0700, sfeldma@gmail.com wrote:
> > > >> >> From: Scott Feldman <sfeldma@gmail.com>
> > > >> >>
> > > >> >> v2:
> > > >> >>
> > > >> >> No clear outcome on polling for choices of "external", "offload", "hw", or
> > > >> >> "hardware".  Andy's concern about "external" being confusing for users is
> > > >> >> valid.  So, I'm making executive decision to use "external_offload" for label,
> > > >> >> as suggested by Jiri.
> > > >> >
> > > >> > This seems like quite a mouthful.  Just make it 'offload' or 'hardware'
> > > >> > so the output of 'ip route show' isn't so long that it will frequently
> > > >> > be longer than 80 chars.  Being longer than 80 chars is going to be
> > > >> > irritating for _many_ users.
> > > >> 
> > > >> Andy, since you have passion on this topic, would you mind taking it
> > > >> from here to get consensus, and post an updated patch, if needed?
> > > >
> > > >Absolutely.  I felt bad responding with more complaints yesterday
> > > >without proposing an alternative patch.
> > > >
> > > >I'll be happy to handle this -- it will elimininate at least one
> > > >complainer!
> > > >
> > > >
> > > >> I was about to suggest a new made-up work like "keroff" and someday
> > > >> the new word is added to Websters .  Daddy, what's "keroff"?  Oh son,
> > > >> that means "offloaded from kernel to external device".  Ya, actually,
> > > >> that's my vote: keroff.
> > > 
> > > Hmm, on the second thought, shouldn't it be rather "extoff" ?
> > > 
> > 
> > I'm really liking the sound of 'extoffk'  ;)
> > 
> 
> I am bouncing this patch, come back when you decide on a name.

Based on the polling I did, the concensus was that people preferred to
use the string 'offload' to describe these entries.  I'll submit the
patch I had prepared now.