On 02/11/2022 17:33, Ilya Maximets wrote:
> On 11/2/22 16:22, Han Zhou wrote:
>>
>>
>> On Wed, Nov 2, 2022 at 5:59 AM Simon Horman <[email protected] 
>> <mailto:[email protected]>> wrote:
>>>
>>> On Wed, Nov 02, 2022 at 12:07:22PM +0100, Simon Horman wrote:
>>>> On Wed, Nov 02, 2022 at 10:18:03AM +0200, Roi Dayan wrote:
>>>>>
>>>>>
>>>>> On 31/10/2022 15:12, Simon Horman wrote:
>>>>>> On Mon, Oct 31, 2022 at 02:10:58PM +0100, Ilya Maximets wrote:
>>>>>>> On 10/31/22 13:49, Simon Horman wrote:
>>>>>>>> On Mon, Oct 31, 2022 at 11:22:48AM +0100, Simon Horman wrote:
>>>>>>>>> On Sun, Oct 30, 2022 at 09:55:55AM +0200, Roi Dayan wrote:
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> On 26/10/2022 11:38, Simon Horman wrote:
>>>>>>>>>>> On Sun, Oct 23, 2022 at 09:27:10AM +0300, Roi Dayan wrote:
>>>>>>>>>>>> OVN is setting ct drop rule with a ct clear action.
>>>>>>>>>>>> OVS datapath behavior is if there is no forward action
>>>>>>>>>>>> the default is drop.
>>>>>>>>>>>> TC behavior is to continue with next match.
>>>>>>>>>>>> Fix to match tc to ovs behavior by setting last action
>>>>>>>>>>>> attribute as drop instead of pipe.
>>>>>>>>>>>
>>>>>>>>>>> This change looks good to me.
>>>>>>>>>>> And we will try to exercise it.
>>>>>>>>>>>
>>>>>>>>>>>> Also update lastused when parsing ct action.
>>>>>>>>>>>
>>>>>>>>>>> Could you split this into a separate patch?
>>>>>>>>>>> It seems like a distinct change.
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> it's kind of related. now CT is last action so we update the stats.
>>>>>>>>>> we only use the stats on the last action. i.e. drop, fwd.
>>>>>>>>>> do you still want to separate?
>>>>>>>>>
>>>>>>>>> Thanks. I think I am fine with this the way it is.
>>>>>>>>> And will plan to apply this patch.
>>>>>>>>
>>>>>>>> Thanks again, applied:
>>>>>>>>
>>>>>>>> * 7a5ee32518df ("tc: On last action use drop action attribute instead 
>>>>>>>> of pipe")
>>>>>>>>   
>>>>>>>> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fopenvswitch%2Fovs%2Fcommit%2F7a5ee32518df&amp;data=05%7C01%7Croid%40nvidia.com%7C862050724b0e4cd106b908dabce7981e%7C43083d15727340c1b7db39efd9ccc17a%7C0%7C0%7C638030000127719304%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&amp;sdata=9VvJUr4u%2BBNioUyG5QBny%2BjZoXydmK4V2tbrY3dhZBs%3D&amp;reserved=0
>>>>>>>>  
>>>>>>>> <https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fopenvswitch%2Fovs%2Fcommit%2F7a5ee32518df&amp;data=05%7C01%7Croid%40nvidia.com%7C862050724b0e4cd106b908dabce7981e%7C43083d15727340c1b7db39efd9ccc17a%7C0%7C0%7C638030000127719304%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&amp;sdata=9VvJUr4u%2BBNioUyG5QBny%2BjZoXydmK4V2tbrY3dhZBs%3D&amp;reserved=0>
>>>>>>>
>>>>>>> Roi, Simon, do you think we also need to backport this change?
>>>>>>> Sounds like a bug fix, but I'm not sure what is the actual impact
>>>>>>> of the issue.
>>>>>>
>>>>>> Sorry, I forgot to ask: I have the same question.
>>>>>
>>>>> Hi,
>>>>>
>>>>> yes you can take it to branch-3.0. i checked and it applies cleanly.
>>>>> branch-2.9 doesn't have the tc conntrack support so the patch is not 
>>>>> relevant.
>>>>
>>
>> Thanks for checking, but why branch-2.9?
> 
> Yeah, that looked strange to me as well.
> 

right. my mistake checking the branches. when i used git branch
the ordering i noticed branch-2.9 was just before branch-3.0 so
I miss-looked the other branch numbers.

> In any case, 2.9 is EOL.  The oldest supported branch right now is 2.13.
> 2.13 will reach EOL in February.
> 
>> May it be backported to branch-2.17?
> 
> This will require a separate patch specifically for branch-2.17 as the
> code changed significantly.
> 
>>
>>>> Thanks Roi,
>>>>
>>>> I will work on making this so.
>>>
>>> Backport applied to branch-3.0
>>>
>>> * eb8ebf8c43b5 ("doc: Add meter offload topic document")
>>>   
>>> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fopenvswitch%2Fovs%2Fcommit%2Feb8ebf8c43b5&amp;data=05%7C01%7Croid%40nvidia.com%7C862050724b0e4cd106b908dabce7981e%7C43083d15727340c1b7db39efd9ccc17a%7C0%7C0%7C638030000127719304%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&amp;sdata=jtkulPRKO8y4b5tQEmKp7%2FqLMpE7%2FsOwtYo3hExYJdI%3D&amp;reserved=0
>>>  
>>> <https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fopenvswitch%2Fovs%2Fcommit%2Feb8ebf8c43b5&amp;data=05%7C01%7Croid%40nvidia.com%7C862050724b0e4cd106b908dabce7981e%7C43083d15727340c1b7db39efd9ccc17a%7C0%7C0%7C638030000127719304%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&amp;sdata=jtkulPRKO8y4b5tQEmKp7%2FqLMpE7%2FsOwtYo3hExYJdI%3D&amp;reserved=0>
>>
>> This seems to be an unrelated patch, probably a copy/paste problem, and I do 
>> see the patch in branch-3.0.
>>
>> Thanks,
>> Han
>>
>>> _______________________________________________
>>> dev mailing list
>>> [email protected] <mailto:[email protected]>
>>> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.openvswitch.org%2Fmailman%2Flistinfo%2Fovs-dev&amp;data=05%7C01%7Croid%40nvidia.com%7C862050724b0e4cd106b908dabce7981e%7C43083d15727340c1b7db39efd9ccc17a%7C0%7C0%7C638030000127875526%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&amp;sdata=P2wOVqjDg2eWTnQsFB4OChEKbMjDlLeKCWo%2BmLh5dN0%3D&amp;reserved=0
>>>  
>>> <https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.openvswitch.org%2Fmailman%2Flistinfo%2Fovs-dev&amp;data=05%7C01%7Croid%40nvidia.com%7C862050724b0e4cd106b908dabce7981e%7C43083d15727340c1b7db39efd9ccc17a%7C0%7C0%7C638030000127875526%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&amp;sdata=P2wOVqjDg2eWTnQsFB4OChEKbMjDlLeKCWo%2BmLh5dN0%3D&amp;reserved=0>
> 
_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to