[ovs-discuss] patch

Ashok Chippa a.n.chippa at gmail.com
Tue Feb 3 20:14:07 UTC 2015


Yes, you don't work for me... But do you support the code you write?  If
not, I will stop asking (or pestering as you call it)...

And yes, it is obviously not important for you... but it is for me...
Unfortunately we (in our company) were asked to base our product on
openvswitch, otherwise why would I even bother with it?

Alright, don't help... I have coded products that are infinitely more
complex than openvswitch... it is just a matter of time which I do not have
at this time... hence my urgency and dependency on others...

Lori was kind enough to suggest using the master branch...which is what
I'll do...


On Tue, Feb 3, 2015 at 9:11 AM, Ben Pfaff <blp at nicira.com> wrote:

> Ashok, you think that this is important.  Therefore, if I worked for
> you, then I would scramble to get it fixed.  But I don't work for you
> and I don't think it is important.  Therefore, it's your job to get it
> fixed, or to find someone to get it fixed.  All this pestering just
> makes people less likely to help you now and in the future.
>
> On Mon, Feb 02, 2015 at 11:19:05PM -0800, Ashok Chippa wrote:
> > Can someone please help?... I AM BLOCKED ON THIS... One hunk has failed
> and
> > I shared that with you. The code base I have (2.3.0) seems to have
> changed
> > significantly where I cannot find a place to add these 4 lines:
> >
> > +            if (i == 0) {
> > +                s->active_count -= connmgr_count_hidden_rules(
> > +                    ofproto->connmgr);
> > +            }
> >
> > This is your code I assume, and it should not take you any time to
> provide
> > a patch that works... I am underwhelmed by your responses... This is not
> my
> > code base and I do not have the luxury to go deep into ofproto.c and
> track
> > all your changes and how the code morphed and then figure out where (and
> > how) to add the above 4 lines....
> >
> > The only place in 2.3.0 that references s->active_count is in
> > handle_table_stats_request() function...:
> >
> > handle_table_stats_request(...)
> > {
> >     ots[i].active_count = htonl(classifier_count(&p->tables[i].cls));
> > ...
> > }
> >
> > I am sure this is where it goes...
> >
> > p.s.: It does not seem like you guys are interested in fixing your
> bugs...
> > I have volunteered to test it. Yet, I am not getting any help. If you
> think
> > it is that easy to fix somone else's code, would you like to try patching
> > newer version of my code with an older version of my patch? Let's see how
> > easy it is for you guys to find and resolve it...
> >
> >
> >
> >
> > On Mon, Feb 2, 2015 at 7:41 PM, Ashok Chippa <a.n.chippa at gmail.com>
> wrote:
> >
> > > Hi Ben,
> > >
> > > Hoping you will provide me with a patch that will work... Here's the
> > > ofproto.c.rej:
> > >
> > > --- ofproto/ofproto.c
> > > +++ ofproto/ofproto.c
> > > @@ -1,5 +1,5 @@
> > >  /*
> > > - * Copyright (c) 2009, 2010, 2011, 2012, 2013, 2014, 2015 Nicira, Inc.
> > > + * Copyright (c) 2009-2015 Nicira, Inc.
> > >   * Copyright (c) 2010 Jean Tourrilhes - HP-Labs.
> > >   *
> > >   * Licensed under the Apache License, Version 2.0 (the "License");
> > > @@ -2963,6 +2963,10 @@
> > >
> > >              s->table_id = i;
> > >              s->active_count = classifier_count(cls);
> > > <<<<==== cannot find this in my branch (2.3.0)
> > > +            if (i == 0) {
> > > +                s->active_count -= connmgr_count_hidden_rules(
> > > +                    ofproto->connmgr);
> > > +            }
> > >          }
> > >      } else {
> > >          stats = NULL;
> > >
> > > s->active_count = classifier_count(cls); <<<<<===== cannot find this
> > > anywhere in my 2.3.0 branch. Please either provide a patch with more
> > > context (so I know where to patch this change) or provide a patch
> based on
> > > 2.3.0 that will work... I am blocked on this....
> > >
> > > Thanks
> > >
> > > On Mon, Feb 2, 2015 at 12:21 PM, Ashok Chippa <a.n.chippa at gmail.com>
> > > wrote:
> > >
> > >> Still waiting :) I cannot find the context (around the change) I saw
> in
> > >> ofproto.c.rej in 2.3.0 ofproto.c. That's the reason for asking you
> for the
> > >> patch (based on 2.3.0). I do not need the entire patch, just the
> small hunk
> > >> that failed (in ofproto.c.rej I sent earlier). The faster I get it,
> the
> > >> faster I can test it :)
> > >>
> > >> Thanks
> > >>
> > >> On Sun, Feb 1, 2015 at 11:05 PM, Ashok Chippa <a.n.chippa at gmail.com>
> > >> wrote:
> > >>
> > >>> Can one of you please apply these diffs to 2.3.0 and provide me with
> a
> > >>> patch. I am not able to readily figure out where (and how) in
> ofproto.c
> > >>> this needs to be applied: It will take me more time to figure this
> out, as
> > >>> I am not the original author of this code...
> > >>>
> > >>> Appreciate your help. I will then patch it and test it. Thanks.
> > >>>
> > >>> Here's ofproto.c.rej:
> > >>>
> > >>>
> > >>>
> > >>> On Sun, Feb 1, 2015 at 3:07 PM, Ashok Chippa <a.n.chippa at gmail.com>
> > >>> wrote:
> > >>>
> > >>>> Thanks, I figure so... Will do and let you know if it worked...
> > >>>>
> > >>>> On Sun, Feb 1, 2015 at 1:55 PM, Ben Pfaff <blp at nicira.com> wrote:
> > >>>>
> > >>>>> On Fri, Jan 30, 2015 at 05:13:42PM -0800, Ashok Chippa wrote:
> > >>>>> > root at ashok-vb:/usr/local/bin# ovs-ofctl del-flows br-int
> > >>>>> > root at ashok-vb:/usr/local/bin# ovs-ofctl dump-tables br-int
> > >>>>> > OFPST_TABLE reply (xid=0x2): 254 tables
> > >>>>> >   0: classifier: wild=0x3fffff, max=1000000, active=10
> > >>>>> >                lookup=0, matched=0
> > >>>>> >   1: table1  : wild=0x3fffff, max=1000000, active=0
> > >>>>>
> > >>>>> If you apply the patch, then it will show active=0.
> > >>>>>
> > >>>>
> > >>>>
> > >>>
> > >>
> > >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://openvswitch.org/pipermail/ovs-discuss/attachments/20150203/ac6aef11/attachment-0002.html>


More information about the discuss mailing list