►
From YouTube: SONiC DASH Workgroup Community Meeting Sept 13 2023
Description
Closed out 3 Issues this week
Intel contributed a fix to the P4 DPDK which was merged into the P4C main branch, which we hope to leverage for future testing purposes – thank you @Fingerhut, John Andy and Intel!
A
Where
do
I
want
to
start
so
this
yeah,
so
today
is
our
84th
meeting
and
last
week
was
our
83rd,
and
so
last
week
we
were
talking
a
lot.
I
was
surprised.
We
took
that
much
time.
It
was
really
interesting
to
talk
about
the
service
tunnel
inbound
processing.
So
we
did
that
and
I
actually
created
a
couple
of
PR's
off
of
that
here
and
here
to
make
changes
to
part
of
the
GitHub
documentation.
A
So
yeah,
and
so
we
need
to
follow
up
about
what's
the
trigger
for
the
flow,
all
that
kind
of
thing
and
then
of
course,
Camille's
work
and
the
more
detailed
notes
were
here
and
I
noticed
a
ping
on
Camille
had
a
ping
on
415
today,
so
I
guess
we
could
talk
about
that
and
then
we
we
had
a
wrap
up
on
387.
So
that's
what
we
did
there
and
I
guess
we
can
get
started
for
today.
A
So
Camille
did
you
want
to
talk
about
4,
15
I
noticed
you
sent
a
ping
this
morning.
A
On
I
sort
of
remember,
let
me
I
can
grab
it.
If
you
need,
was
it
an
issue
or
a
yeah,
oh
yeah,
so
let
me
see
like
waiting
to
think
for
Mariam
to.
B
A
B
A
A
A
I'm
working
on
this
one
we
just
had
in
the
so
we
still
there's
something.
A
And
so
this
one
here,
I
hope
I'm
doing
this
correctly,
but
I
went
ahead
and
just
created
these
URLs
here
too,
identify
these
that
either
I
need
to
work
on
or
I'll
work
with
the
Sonic
Team.
To
do
the
clarification
here
so
do
I
have
anyone
from
the
Sonic
team
on
not
quite
yet
okay,
so
I
did
that
and
then
anyone
else
want
to
talk
about.
One
of
these
I
noticed
a
note
from
Usha
about
this.
One
I
believe.
C
Sure
there
were
a
couple
of
things:
issues
with
the
open
source,
P4
compiler,
dvdk,
backend,
that
you
could
compile
the
old
Dash
code
from
March,
but
not
load
it
into
the
software
switch.
It
would
give
errors
which
has
fixed
two
of
those
recently,
the
ones
that
I
called
blockers.
So
as
of
today,
if
you
built
the
latest
P4
cdpk,
you
can
now
compile
and
load
in
the
March
30th
version
of
the
dash
P4
code
successfully
whoa.
A
A
B
A
Progress,
thank
you
yeah,
and
so
what
that
means
is
possibly
the
some
of
the
tests
that
haven't
been
able
to
run
in
bmv2
might
be
able
to
run
now.
A
B
No
and
Andy
I
think
I
think
she
meant
something
different
by
our
question.
Sorry,
there
are
test
cases
right
now
which
are
disabled
for
bmv2
they're
in
the
repository,
but
they've
only
been
run
on,
let's
say
actual
Hardware
dpus,
because
the
bmv2
can't
support
them.
So
the
hope
is
that
we
can
actually
run
those
tests
on
this
model
someday.
C
Yep
and
and
that
day
is
not
today,
but
it
that
day
is
probably
closer
yep.
A
A
B
I'm
I'm
reviewing
or
two
right
now.
A
B
Just
about
just
got
it:
yeah
Camille
just
give
us
another
Omnibus
PR.
A
B
Might
be
someone
else
on
the
call
besides
Myron,
because
the
answer
is
I
couldn't
give
it
a
good
answer,
but
the
implementation
has
like
there's
some
macros
in
the
code
that
generate
multiple
tables.
Well,
it
looks
like
one
logical
table
for
Psy
I.
Believe
I,
think
this
question.
A
Okay,
so
I've
been
talking
to
Paul,
also
at
Nvidia,
and
he
thought
he
might
be
able
to
find
me
like
a
second
P4
resource.
Possibly
so
maybe
we'll
have
some
extra
help
in
the
next
couple
of
weeks
too.
So
I'll
I'll
show
him
these
kind
of
issues
where
we're
looking
for
a
little
bit
more
explanation
and
see
if
he
can
find
us
somebody
Paul
Cummins,
so
I'll
do
the
best.
I
can
Camille
okay,
great.
A
I
guess
it's
open,
then,
for
q
a
for
this
week.
Did
you
guys
have
anything
specific?
You
wanted
to
talk
about.
A
B
Everybody
yeah
I'm,
no
problem
with
that
expressing
there's
a
lot
of
work
in
progress
going
on
yeah.
A
I
think
so,
how
about
the
rest
of
you
guys
I'll
do
the
the
Andy
seven
second
rule.