action stats broken?

Sorry, dont have time to chase this today and it may be fixed in the
latest  tree. Possibly recent patches or likely earlier patches
even.

It seems stats are no longer displaying for actions.
This is a net-next kernel. I have tried kernel 3.16.0-rc2
and 3.17.0-rc4 with same results.

Simple test case:

----
root@moja-1:~# tc actions add  action drop index 4
root@moja-1:~# tc -s actions ls action gact

	action order 0: gact action drop
	 random type none pass val 0
	 index 4 ref 1 bind 0 installed 9 sec used 4 sec
-----

What I would have expected to see is:
---
	action order 0: gact action drop
	 random type none pass val 0
	 index 4 ref 1 bind 0 installed 9 sec used 4 sec
          Sent 108948 bytes 1297 pkts (dropped 1297, overlimits 0)
---


cheers,
jamal

On 09/20/2014 06:17 AM, Jamal Hadi Salim wrote:

Sorry, dont have time to chase this today and it may be fixed in the
latest  tree. Possibly recent patches or likely earlier patches
even.

It seems stats are no longer displaying for actions.
This is a net-next kernel. I have tried kernel 3.16.0-rc2
and 3.17.0-rc4 with same results.

Simple test case:

----
root@moja-1:~# tc actions add  action drop index 4
root@moja-1:~# tc -s actions ls action gact

     action order 0: gact action drop
      random type none pass val 0
      index 4 ref 1 bind 0 installed 9 sec used 4 sec
-----

What I would have expected to see is:
---
     action order 0: gact action drop
      random type none pass val 0
      index 4 ref 1 bind 0 installed 9 sec used 4 sec
          Sent 108948 bytes 1297 pkts (dropped 1297, overlimits 0)
---

I'm looking into it which classifier? A basic u32 classifier attached
to the ingress qdisc seems to work. If your using u32 any hash tables,
or 'filter change' commands before this?

# tc actions add  action drop index 4

# tc filter add dev p3p1 parent ffff:0 protocol ip prio 100 \
	u32 match ip src 16.0.0.2 action gact index 4

#tc -s actions ls action gact

         action order 0: gact action drop
          random type none pass val 0
          index 4 ref 2 bind 1 installed 142 sec used 0 sec
         Action statistics:
         Sent 840 bytes 10 pkt (dropped 10, overlimits 0 requeues 0)
         backlog 0b 0p requeues 0

cheers,
jamal
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

-- 
John Fastabend         Intel Corporation

On 09/20/14 11:21, John Fastabend wrote:

On 09/20/2014 06:17 AM, Jamal Hadi Salim wrote:

I'm looking into it which classifier? A basic u32 classifier attached
to the ingress qdisc seems to work. If your using u32 any hash tables,
or 'filter change' commands before this?

Nothing that exciting.

# tc actions add  action drop index 4

# tc filter add dev p3p1 parent ffff:0 protocol ip prio 100 \
     u32 match ip src 16.0.0.2 action gact index 4

#tc -s actions ls action gact

         action order 0: gact action drop
          random type none pass val 0
          index 4 ref 2 bind 1 installed 142 sec used 0 sec
         Action statistics:
         Sent 840 bytes 10 pkt (dropped 10, overlimits 0 requeues 0)
         backlog 0b 0p requeues 0

Indeed that looks sane. What kernel version + iproute2?

I have time this morning - I will chase it to double check in case
my environment is messed up.

cheers,
jamal

quoted

cheers,
jamal
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

On 09/21/14 08:33, Jamal Hadi Salim wrote:

On 09/20/14 11:21, John Fastabend wrote:

I have time this morning - I will chase it to double check in case
my environment is messed up.

There is something wrong for sure.
Ruling out kernel regression at this point.
It is definetely passing the stats.
Zoning on iproute2.

cheers,
jamal

On 09/21/2014 07:21 AM, Jamal Hadi Salim wrote:

On 09/21/14 08:33, Jamal Hadi Salim wrote:

quoted

On 09/20/14 11:21, John Fastabend wrote:

I have time this morning - I will chase it to double check in case
my environment is messed up.

There is something wrong for sure.
Ruling out kernel regression at this point.
It is definetely passing the stats.
Zoning on iproute2.

cheers,
jamal

I am using the latest iproute2 tip,

commit eb5d01ff38080935a1a528e4c99cab42bf93d762
Author: Stephen Hemminger [off-list ref]
Date:   Sun Sep 14 20:40:37 2014 -0700

     update dsfield file values

     Update the rt_dsfield file to contain values defined in current RFC.
     The days of TOS precedence are gone, even Cisco doesn't refer
     to these in the documents.

I think you are right might need to bisect iproute2.


-- 
John Fastabend         Intel Corporation

On 09/21/14 10:50, John Fastabend wrote:

I think you are right might need to bisect iproute2.

I think i may have found it. Broken by a patch i sent
in May.
Working on a fix ..

cheers,
jamal