Re: [vpp-dev] system func in vpp

[yug...@telincn.com]

Thanks,  that is helpful.
Can we use  rte_spinlock_lock directly on vpp? Or any other options?

Regards,
Ewan.


yug...@telincn.com
 
From: Dave Barach (dbarach)
Date: 2017-03-02 20:10
To: yug...@telincn.com; Luke, Chris; Ed Warnicke
CC: vpp-dev
Subject: RE: RE: [vpp-dev] system func in vpp
Do you have a specific reason for calling that function? 
 
If the answer amounts to “I need some memory,” please use clib_mem_alloc(...) 
instead. 
 
Thanks… Dave
 
From: yug...@telincn.com [mailto:yug...@telincn.com] 
Sent: Thursday, March 2, 2017 2:43 AM
To: Luke, Chris ; Dave Barach (dbarach) 
; Ed Warnicke 
Cc: vpp-dev 
Subject: Re: RE: [vpp-dev] system func in vpp
 
Is func rte_malloc safe to be called between   rte_spinlock_lock  adn 
rte_spinlock_unlock   or can it be only called  at main thread?
 
Ewan


yug...@telincn.com
 
From: Luke, Chris
Date: 2017-03-01 20:14
To: yug...@telincn.com; Dave Barach (dbarach); Ed Warnicke
CC: vpp-dev
Subject: RE: RE: [vpp-dev] system func in vpp
A Unix domain socket would work, if you use it only from the main thread.
 
Instead though, VPP already has a performant binary API that uses shared 
memory. Could you not use this? What sort of messages are you trying to send?
 
Chris.
 
From: yug...@telincn.com [mailto:yug...@telincn.com] 
Sent: Wednesday, March 1, 2017 4:38
To: Dave Barach (dbarach) ; Luke, Chris 
; Ed Warnicke 
Cc: vpp-dev 
Subject: Re: RE: [vpp-dev] system func in vpp
 
Thanks.
So it's not security to use unix socket too? If i want to send some data  from 
vpp to linux or to another process, any options? 
 
Regards,
Ewan


yug...@telincn.com
 
From: Dave Barach (dbarach)
Date: 2017-02-28 21:12
To: Luke, Chris; yug...@telincn.com; Ed Warnicke
CC: vpp-dev
Subject: RE: [vpp-dev] system func in vpp
See also src/vnet/unix/tuntap.c, which seems to already do pretty much what 
you’re describing. It’s disabled by default. Use the command-line argument 
“tuntap { enable }” to kick the tires.
 
In addition to the concerns which Chris mentioned, adding ‘system(“foo”)’ calls 
to vpp is not a security best practice.
 
Thanks… Dave
 
From: vpp-dev-boun...@lists.fd.io [mailto:vpp-dev-boun...@lists.fd.io] On 
Behalf Of Luke, Chris
Sent: Tuesday, February 28, 2017 7:18 AM
To: yug...@telincn.com; Ed Warnicke 
Cc: vpp-dev 
Subject: Re: [vpp-dev] system func in vpp
 
Using any syscall needs thinking through from inside VPP. Using system() or 
similar is simply ill-advised; what you’re doing is best done outside of VPP 
where the penalty of executing fork() is much diminished.
 
If you really must create these interfaces from inside VPP then I suggest you 
look up the netlink ABI in Linux, which is what the ‘ip’ program uses, and 
program it directly.
 
Chris.
 
From: vpp-dev-boun...@lists.fd.io [mailto:vpp-dev-boun...@lists.fd.io] On 
Behalf Of yug...@telincn.com
Sent: Monday, February 27, 2017 22:56
To: Ed Warnicke 
Cc: vpp-dev 
Subject: Re: [vpp-dev] system func in vpp
 
I would like to add veth interface in vpp for every phy interface, " system("ip 
link add vpp type veth peer name kernel-vpp") ".
It's convenient  to do this from vpp.
 
Regards,
Ewan.


yug...@telincn.com
 
From: Ed Warnicke
Date: 2017-02-28 11:44
To: yug...@telincn.com
CC: vpp-dev
Subject: Re: Re: [vpp-dev] system func in vpp
Why would you do that from within vpp?
 
Ed
 
On Mon, Feb 27, 2017 at 8:36 PM, yug...@telincn.com  wrote:
"int system(const char *command)"
 
I would like to use this func to start another dameon , such as "  
system("zebra -d")  ".
 
Regards,
Ewan


yug...@telincn.com
 
From: Ed Warnicke
Date: 2017-02-28 11:15
To: yug...@telincn.com
CC: vpp-dev
Subject: Re: [vpp-dev] system func in vpp
I'm not quite sure what you mean by the 'func system'...
 
Ed
 
On Mon, Feb 27, 2017 at 7:26 PM, yug...@telincn.com  wrote:
Hi, all
 
Does vpp can't use func  system? Any reason to this or what can i do  if i 
really need ?
 


yug...@telincn.com

___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev
 
 
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] TCP stack in master

[Florin Coras]

Hi Alec, 

The code is still very much under development. We pushed it out early to make 
the API freeze window but our goal is to have something stable for the 17.04 
release (due in about a month or so). We’ll try to have some decent 
documentation by then, but in the mean time, running the code will probably 
result in pain. :)

Thanks,
Florin

 
> On Mar 2, 2017, at 6:51 AM, Alec Hothan (ahothan)  wrote:
> 
>  
> Impressive!
> Is there any document describing the features, how is it integrated with the 
> rest of VPP, CLI, compile flags, APIs and short term use cases?
>  
> Thanks
>  
>Alec
>  
>  
> From:  on behalf of "Damjan Marion (damarion)" 
> 
> Date: Thursday, March 2, 2017 at 5:45 AM
> To: vpp-dev 
> Subject: [vpp-dev] TCP stack in master
>  
>  
> In case people didn’t notice, since yesterday we have full TCP stack
> in master. It is “just” 17 KLOCs.
>  
> Thanks to Dave, Florin and all other folks participated in development of 
> this great addition to VPP.
>  
> Damjan
> ___
> vpp-dev mailing list
> vpp-dev@lists.fd.io 
> https://lists.fd.io/mailman/listinfo/vpp-dev 
> ___
> vpp-dev mailing list
> vpp-dev@lists.fd.io 
> https://lists.fd.io/mailman/listinfo/vpp-dev 
> 
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] sw_interface_set_vxlan_bypass API message - 2x incompatible handlers...

[Dave Barach (dbarach)]

Cool... Thanks... Dave

From: John Lo (loj)
Sent: Thursday, March 2, 2017 4:32 PM
To: Dave Barach (dbarach) 
Cc: vpp-dev 
Subject: RE: sw_interface_set_vxlan_bypass API message - 2x incompatible 
handlers...

Thanks Dave, I will clean up the one in api.c which is obsolete.

Thanks for letting me know,
John

From: Dave Barach (dbarach)
Sent: Thursday, March 02, 2017 12:38 PM
To: John Lo (loj) >
Cc: vpp-dev >
Subject: sw_interface_set_vxlan_bypass API message - 2x incompatible handlers...

Dear John,

I imagine that you might be the right person to warn about this...

There are two incompatible implementations of 
vl_api_sw_interface_set_vxlan_bypass_t_handler. One in .../src/vpp/api/api.c, 
and the other one is in .../src/vnet/vxlan.c.

One is feature-arc based, the other is not. Only one of them is being called. 
This can't be good.

D.
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] sw_interface_set_vxlan_bypass API message - 2x incompatible handlers...

[John Lo (loj)]

Thanks Dave, I will clean up the one in api.c which is obsolete.

Thanks for letting me know,
John

From: Dave Barach (dbarach)
Sent: Thursday, March 02, 2017 12:38 PM
To: John Lo (loj) 
Cc: vpp-dev 
Subject: sw_interface_set_vxlan_bypass API message - 2x incompatible handlers...

Dear John,

I imagine that you might be the right person to warn about this...

There are two incompatible implementations of 
vl_api_sw_interface_set_vxlan_bypass_t_handler. One in .../src/vpp/api/api.c, 
and the other one is in .../src/vnet/vxlan.c.

One is feature-arc based, the other is not. Only one of them is being called. 
This can't be good.

D.
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

[vpp-dev] sw_interface_set_vxlan_bypass API message - 2x incompatible handlers...

[Dave Barach (dbarach)]

Dear John,

I imagine that you might be the right person to warn about this...

There are two incompatible implementations of 
vl_api_sw_interface_set_vxlan_bypass_t_handler. One in .../src/vpp/api/api.c, 
and the other one is in .../src/vnet/vxlan.c.

One is feature-arc based, the other is not. Only one of them is being called. 
This can't be good.

D.
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] vpp w/ dpdk-as-plugin

[Kinsella, Ray]

Good job - is it being benchmarked in CSIT yet.

RAy K

On 02/03/2017 00:07, Luke, Chris wrote:

Ooh, nice





*From:* vpp-dev-boun...@lists.fd.io [mailto:vpp-dev-boun...@lists.fd.io]
*On Behalf Of *Dave Barach (dbarach)
*Sent:* Wednesday, March 1, 2017 18:09
*To:* Damjan Marion (damarion) 
*Cc:* vpp-dev@lists.fd.io
*Subject:* [vpp-dev] vpp w/ dpdk-as-plugin



“If you’re reading this message, vpp w/ dpdk-as-a-plugin is now running
my home gateway.”



No issues noted... Nice job, Damjan!



dbarach@vppgate:~$ telnet 0 5002

Trying 0.0.0.0...

Connected to 0.

Escape character is '^]'.

_____   _  ___

 __/ __/ _ \  (_)__| | / / _ \/ _ \

_/ _// // / / / _ \   | |/ / ___/ ___/

/_/ /(_)_/\___/   |___/_/  /_/



vpp# sh plug

Plugin path is: /usr/lib/vpp_plugins



 Plugin   Version

  1. ioam_plugin.so   17.04-rc0~329-gc3a814b

  2. ila_plugin.so17.04-rc0~329-gc3a814b

  3. dpdk_plugin.so   17.04-rc0~329-gc3a814b

  4. acl_plugin.so17.04-rc0~329-gc3a814b

  5. flowperpkt_plugin.so 17.04-rc0~329-gc3a814b

  6. snat_plugin.so   17.04-rc0~329-gc3a814b

  7. lb_plugin.so 17.04-rc0~329-gc3a814b

  8. libsixrd_plugin.so   17.04-rc0~329-gc3a814b

vpp#



Thanks… Dave





___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev


___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] TCP stack in master

[Alec Hothan (ahothan)]

Impressive!
Is there any document describing the features, how is it integrated with the 
rest of VPP, CLI, compile flags, APIs and short term use cases?

Thanks

   Alec


From:  on behalf of "Damjan Marion (damarion)" 

Date: Thursday, March 2, 2017 at 5:45 AM
To: vpp-dev 
Subject: [vpp-dev] TCP stack in master


In case people didn’t notice, since yesterday we have full TCP stack
in master. It is “just” 17 KLOCs.

Thanks to Dave, Florin and all other folks participated in development of this 
great addition to VPP.

Damjan
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] TCP stack in master

[Luke, Chris]

W00t!

> -Original Message-
> From: vpp-dev-boun...@lists.fd.io [mailto:vpp-dev-boun...@lists.fd.io] On
> Behalf Of Damjan Marion (damarion)
> Sent: Thursday, March 02, 2017 8:45 AM
> To: vpp-dev 
> Subject: [vpp-dev] TCP stack in master
> 
> 
> In case people didn’t notice, since yesterday we have full TCP stack in
> master. It is “just” 17 KLOCs.
> 
> Thanks to Dave, Florin and all other folks participated in development of this
> great addition to VPP.
> 
> Damjan
> ___
> vpp-dev mailing list
> vpp-dev@lists.fd.io
> https://lists.fd.io/mailman/listinfo/vpp-dev
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

[vpp-dev] TCP stack in master

[Damjan Marion (damarion)]

In case people didn’t notice, since yesterday we have full TCP stack 
in master. It is “just” 17 KLOCs.

Thanks to Dave, Florin and all other folks participated in development of this 
great addition to VPP.

Damjan
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] vpp w/ dpdk-as-plugin

[Damjan Marion (damarion)]

this is startup.conf config which disables dpdk plugin and effectively makes 
normal vpp binary acting as vpp_lite.

plugins {
  plugin dpdk_plugin.so { disable }
}

I still need to do few smaller changes before fully depreciating vpp_lite….


> On 2 Mar 2017, at 01:07, Luke, Chris  wrote:
> 
> Ooh, nice
>  
>  
> From: vpp-dev-boun...@lists.fd.io [mailto:vpp-dev-boun...@lists.fd.io] On 
> Behalf Of Dave Barach (dbarach)
> Sent: Wednesday, March 1, 2017 18:09
> To: Damjan Marion (damarion) 
> Cc: vpp-dev@lists.fd.io
> Subject: [vpp-dev] vpp w/ dpdk-as-plugin
>  
> “If you’re reading this message, vpp w/ dpdk-as-a-plugin is now running my 
> home gateway.”
>  
> No issues noted... Nice job, Damjan!
>  
> dbarach@vppgate:~$ telnet 0 5002
> Trying 0.0.0.0...
> Connected to 0.
> Escape character is '^]'.
> _____   _  ___
>  __/ __/ _ \  (_)__| | / / _ \/ _ \
> _/ _// // / / / _ \   | |/ / ___/ ___/
> /_/ /(_)_/\___/   |___/_/  /_/   
>  
> vpp# sh plug
> Plugin path is: /usr/lib/vpp_plugins
>  
>  Plugin   Version
>   1. ioam_plugin.so   17.04-rc0~329-gc3a814b
>   2. ila_plugin.so17.04-rc0~329-gc3a814b
>   3. dpdk_plugin.so   17.04-rc0~329-gc3a814b
>   4. acl_plugin.so17.04-rc0~329-gc3a814b
>   5. flowperpkt_plugin.so 17.04-rc0~329-gc3a814b
>   6. snat_plugin.so   17.04-rc0~329-gc3a814b
>   7. lb_plugin.so 17.04-rc0~329-gc3a814b
>   8. libsixrd_plugin.so   17.04-rc0~329-gc3a814b
> vpp#
>  
> Thanks… Dave

___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] VPP API Freeze date coming up -- March 8

[Dave Barach (dbarach)]

+1, exactly right, thanks Ole... Dave

-Original Message-
From: vpp-dev-boun...@lists.fd.io [mailto:vpp-dev-boun...@lists.fd.io] On 
Behalf Of otr...@employees.org
Sent: Thursday, March 2, 2017 7:17 AM
To: vpp-dev 
Subject: [vpp-dev] VPP API Freeze date coming up -- March 8

All,

Please refer to the 17.04 release plan: 
https://wiki.fd.io/view/Projects/vpp/Release_Plans/Release_Plan_17.04

Any/all binary API changes must be merged no later than the F0 date, March 8th. 
Otherwise they will not make the 17.04 release. Other projects depend on that 
for their release planning. We do not have the option of slipping the F0 date 
to accommodate “last minute rush features.”

Between F0 and RC1 (March 22), the goal is to settle down the master branch. 
Housekeeping activities such as eliminating residual Coverity warnings occur 
between F0 and RC1.

Low-risk patches with contained blast radii - and no effect on binary APIs - 
may be accepted at committer's discretion between F0 and RC1.

The master branch reopens for normal work shortly after RC1.

Best regards,
Ole
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

[vpp-dev] VPP API Freeze date coming up -- March 8

[otroan]

All,

Please refer to the 17.04 release plan: 
https://wiki.fd.io/view/Projects/vpp/Release_Plans/Release_Plan_17.04

Any/all binary API changes must be merged no later than the F0 date, March 8th. 
Otherwise they will not make the 17.04 release. Other projects depend on that 
for their release planning. We do not have the option of slipping the F0 date 
to accommodate “last minute rush features.”

Between F0 and RC1 (March 22), the goal is to settle down the master branch. 
Housekeeping activities such as eliminating residual Coverity warnings occur 
between F0 and RC1.

Low-risk patches with contained blast radii - and no effect on binary APIs - 
may be accepted at committer's discretion between F0 and RC1.

The master branch reopens for normal work shortly after RC1.

Best regards,
Ole


signature.asc
Description: Message signed with OpenPGP
___
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev

Re: [vpp-dev] system func in vpp

[Dave Barach (dbarach)]

Do you have a specific reason for calling that function?

If the answer amounts to “I need some memory,” please use clib_mem_alloc(...) 
instead.

Thanks… Dave

From: yug...@telincn.com [mailto:yug...@telincn.com]
Sent: Thursday, March 2, 2017 2:43 AM
To: Luke, Chris ; Dave Barach (dbarach) 
; Ed Warnicke 
Cc: vpp-dev 
Subject: Re: RE: [vpp-dev] system func in vpp

Is func rte_malloc safe to be called between   rte_spinlock_lock  adn 
rte_spinlock_unlock   or can it be only called  at main thread?

Ewan

yug...@telincn.com

From: Luke, Chris
Date: 2017-03-01 20:14
To: yug...@telincn.com; Dave Barach 
(dbarach); Ed Warnicke
CC: vpp-dev
Subject: RE: RE: [vpp-dev] system func in vpp
A Unix domain socket would work, if you use it only from the main thread.

Instead though, VPP already has a performant binary API that uses shared 
memory. Could you not use this? What sort of messages are you trying to send?

Chris.

From: yug...@telincn.com [mailto:yug...@telincn.com]
Sent: Wednesday, March 1, 2017 4:38
To: Dave Barach (dbarach) >; Luke, 
Chris >; Ed 
Warnicke >
Cc: vpp-dev >
Subject: Re: RE: [vpp-dev] system func in vpp

Thanks.
So it's not security to use unix socket too? If i want to send some data  from 
vpp to linux or to another process, any options?

Regards,
Ewan

yug...@telincn.com

From: Dave Barach (dbarach)
Date: 2017-02-28 21:12
To: Luke, Chris; 
yug...@telincn.com; Ed 
Warnicke
CC: vpp-dev
Subject: RE: [vpp-dev] system func in vpp
See also src/vnet/unix/tuntap.c, which seems to already do pretty much what 
you’re describing. It’s disabled by default. Use the command-line argument 
“tuntap { enable }” to kick the tires.

In addition to the concerns which Chris mentioned, adding ‘system(“foo”)’ calls 
to vpp is not a security best practice.

Thanks… Dave

From: vpp-dev-boun...@lists.fd.io 
[mailto:vpp-dev-boun...@lists.fd.io] On Behalf Of Luke, Chris
Sent: Tuesday, February 28, 2017 7:18 AM
To: yug...@telincn.com; Ed Warnicke 
>
Cc: vpp-dev >
Subject: Re: [vpp-dev] system func in vpp

Using any syscall needs thinking through from inside VPP. Using system() or 
similar is simply ill-advised; what you’re doing is best done outside of VPP 
where the penalty of executing fork() is much diminished.

If you really must create these interfaces from inside VPP then I suggest you 
look up the netlink ABI in Linux, which is what the ‘ip’ program uses, and 
program it directly.

Chris.

From: vpp-dev-boun...@lists.fd.io 
[mailto:vpp-dev-boun...@lists.fd.io] On Behalf Of 
yug...@telincn.com
Sent: Monday, February 27, 2017 22:56
To: Ed Warnicke >
Cc: vpp-dev >
Subject: Re: [vpp-dev] system func in vpp

I would like to add veth interface in vpp for every phy interface, " system("ip 
link add vpp type veth peer name kernel-vpp") ".
It's convenient  to do this from vpp.

Regards,
Ewan.

yug...@telincn.com

From: Ed Warnicke
Date: 2017-02-28 11:44
To: yug...@telincn.com
CC: vpp-dev
Subject: Re: Re: [vpp-dev] system func in vpp
Why would you do that from within vpp?

Ed

On Mon, Feb 27, 2017 at 8:36 PM, yug...@telincn.com 
> wrote:
"int system(const char *command)"

I would like to use this func to start another dameon , such as "  
system("zebra -d")  ".

Regards,
Ewan

yug...@telincn.com

From: Ed Warnicke
Date: 2017-02-28 11:15
To: yug...@telincn.com
CC: vpp-dev
Subject: Re: [vpp-dev] system func in vpp
I'm not quite sure what you mean by the 'func system'...

Ed

On Mon, Feb 27, 2017 at 7:26 PM, yug...@telincn.com 
> wrote:
Hi, all

Does vpp can't use func  system? Any reason to this or what can i