[ovs-dev] [PATCH] ofproto: Return error codes for Rule insertions

[Aravind.Sridharan at dell.com]

Hi All,

Sorry for resending the mail, the link appears wrongly in previous mail (due to mingling of 'dot' with github link). 

Correct GitHub pull request link - https://github.com/openvswitch/ovs/pull/282   

Kindly review.

Thanks,
Aravind Prasad S

-----Original Message-----
From: Sridharan, Aravind 
Sent: Wednesday, April 24, 2019 1:26 AM
To: 'Ben Pfaff'; ovs-dev at openvswitch.org
Subject: RE: [ovs-dev] [PATCH] ofproto: Return error codes for Rule insertions

Hi All,

I have raised the pull request in Github - https://github.com/openvswitch/ovs/pull/282. Kindly review and let me know your views. 

Hi Ben,
Also, I raised the similar patch sometime back and you had requested for the vendor name. Had to get approval from my Organization. Hence, requesting your review for the changes now. 

> Hi Ben,
> 
> By "which" switches, you mean to ask to which vendor switches ?
> In general, this patch will be very useful for all Switches which 
> implement OVS for their HW implementations, specifically, will be 
> useful for Switches which had implemented ofproto and netdev provider 
> APIs for their respective HW/NPU and Kernel. And, in case of HW, the 
> possibility of dynamic rule insertion failures are usually higher. 
> Static checks during rule-construct phase are not sufficient and 
> predicting future failure predictions in construct phase may not be 
> possible.
> Hence, this patch will make OVS more flexible to handle such rule 
> insertion failures.

> So, basically, you're not willing to name any switches.  I'm tired of  
>the vendor secrecy game.  They get to see all of our code but they're 
>never willing even to name themselves.

> Sorry, I'm not going to willingly take this patch.


Thanks,
S. Aravind Prasad

On Tue, Apr 23, 2019 at 02:40:55PM +0000, Aravind.Sridharan at dell.com wrote:
> Currently, rule_insert() API does not have return value. There are 
> some possible scenarios where rule insertions can fail at run-time 
> even though the static checks during rule_construct() had passed 
> previously.
> Some possible scenarios for failure of rule insertions:
> **) Rule insertions can fail dynamically in Hybrid mode (both Openflow 
> and Normal switch functioning coexist) where the CAM space could get 
> suddenly filled up by Normal switch functioning and Openflow gets 
> devoid of available space.
> **) Some deployments could have separate independent layers for HW 
> rule insertions and application layer to interact with OVS.
> HW layer could face any dynamic issue during rule handling which 
> application could not have predicted/captured in rule-construction 
> phase.
> Rule-insert errors for bundles are handled too in this pull-request.

> It looks like you're not having much luck with posting patches.  If you  > haven't already tried "git send-email", please try that; otherwise, you > can always submit a pull request on Github.

Thanks,

Ben.