On Wed, Dec 07, 2016 at 12:51:12PM -0500, Lance Richardson wrote:
> > From: "Ben Pfaff" <b...@ovn.org>
> > To: "Lance Richardson" <lrich...@redhat.com>
> > Cc: d...@openvswitch.org
> > Sent: Wednesday, December 7, 2016 12:38:42 PM
> > Subject: Re: [ovs-dev] [PATCH 4/4] [RFC] ofproto-dpif: Make ofproto/trace
> > output easier to read.
> >
> > On Wed, Dec 07, 2016 at 12:31:24PM -0500, Lance Richardson wrote:
> > > > From: "Ben Pfaff" <b...@ovn.org>
> > > > To: d...@openvswitch.org
> > > > Sent: Wednesday, December 7, 2016 11:36:13 AM
> > > > Subject: Re: [ovs-dev] [PATCH 4/4] [RFC] ofproto-dpif: Make
> > > > ofproto/trace
> > > > output easier to read.
> > > >
> > > > On Tue, Dec 06, 2016 at 11:28:43PM -0800, Ben Pfaff wrote:
> > > > > "ovs-appctl ofproto/trace" is invaluable for debugging, but as the
> > > > > users of
> > > > > Open vSwitch have evolved it has failed to keep up with the times.
> > > > > It's
> > > > > pretty easy to design OpenFlow tables and pipelines that resubmit
> > > > > dozens of
> > > > > times. Each resubmit causes an additional tab of indentation, so the
> > > > > output wraps around, sometimes again and again, and makes the output
> > > > > close
> > > > > to unreadable.
> > > > >
> > > > > ovn-trace pioneered better formatting for tracing in OVN logical
> > > > > datapaths,
> > > > > mostly by not increasing indentation for tail recursion, which in
> > > > > practice
> > > > > gets rid of almost all indentation.
> > > > >
> > > > > This commit experiments with redoing ofproto/trace the same way. Try
> > > > > looking at, for example, the testsuite output for test 2280 "ovn -- 3
> > > > > HVs,
> > > > > 3 LRs connected via LS, source IP based routes". Without this commit,
> > > > > it
> > > > > indents 61 levels (488 spaces!). With this commit, it indents 1 level
> > > > > (4 spaces) and it's possible to actually understand what's going on
> > > > > almost
> > > > > at a glance.
> > > >
> > > > Here's the output with this patch, followed by the output without this
> > > > patch:
> > > >
> > >
> > > Having tried to make sense of the current format before (successfully,
> > > eventually,
> > > but with great effort): thank you for doing this, it should make debugging
> > > and
> > > understanding OVN much easier.
> > >
> > > However, with this 4-patch series applied I'm getting test failures that I
> > > don't
> > > see without these patches:
> > >
> > > 1169: ofproto-dpif - trace (unchanged) FAILED
> > > (ofproto-dpif.at:8321)
> > > 1052: ofproto-dpif - fragment handling - trace FAILED
> > > (ofproto-dpif.at:3799)
> > > 2161: RSTP - dummy interface FAILED (rstp.at:215)
> > > 2154: STP - dummy interface FAILED (stp.at:430)
> > > 1082: ofproto-dpif - ofproto/trace command 1 FAILED
> > > (ofproto-dpif.at:4842)
> > > 1084: ofproto-dpif - ofproto/trace from dpctl output FAILED
> > > (ofproto-dpif.at:5000)
> >
> > Yes, fixing these are on the to-do list that will need to be finished
> > before it becomes non-RFC. (There may be genuine bugs, but these are
> > mostly for tests that are very picky about the output format, so they
> > break because the output format changed.)
> >
>
> OK, I had somehow missed the [RFC] tag.
>
> FWIW, the 1052 failure was caused by the removal of this output from
> trace_resubmit():
>
> - } else if (rule == xin->ofproto->drop_frags_rule) {
> - trace_report(xin, indentation,
> - "Packets dropped because they are IP fragments and "
> - "the fragment handling mode is \"drop\".");
I've got that one fixed in my local tree now.
_______________________________________________
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
