Solarflare
Server
Adapter
User
Guide
Solarflare
Adapters
on
VMware
Issue
20
©
Solarflare
Communications
2017
285
NOTE:
The
interrupt
moderation
time
dictates
the
minimum
gap
between
two
consecutive
interrupts.
It
does
not
mandate
a
delay
on
the
triggering
of
an
interrupt
on
the
reception
of
every
packet.
For
example,
an
interrupt
moderation
setting
of
30µs
will
not
delay
the
reception
of
the
first
packet
received,
but
the
interrupt
for
any
following
packets
will
be
delayed
until
30µs
after
the
reception
of
that
first
packet.
TCP/IP
Checksum
Offload
Checksum
offload
moves
calculation
and
verification
of
IP
Header,
TCP
and
UDP
packet
checksums
to
the
adapter.
The
driver
by
default
has
all
checksum
offload
features
enabled.
Therefore,
there
is
no
opportunity
to
improve
performance
from
the
default.
Checksum
offload
is
controlled
using
ethtool:
Receive
Checksum:
#
/sbin/ethtool
–K
<vmnicX>
rx
<on|off>
Transmit
Checksum:
#
/sbin/ethtool
–K
<vmnicX>
tx
<on|off>
Verification
of
the
checksum
settings
may
be
performed
by
running
ethtool
with
the
–k
option.
Solarflare
recommend
you
do
not
disable
checksum
offload.
For
advice
on
configuring
checksum
offload
in
the
guest,
consult
the
relevant
Solarflare
section
for
that
guest,
or
the
documentation
for
the
guest
operating
system.
TCP
Segmentation
Offload
(TSO)
TCP
Segmentation
offload
(TSO)
offloads
the
splitting
of
outgoing
TCP
data
into
packets
to
the
adapter.
TCP
segmentation
offload
benefits
applications
using
TCP.
Non
TCP
protocol
applications
will
not
benefit
(but
will
not
suffer)
from
TSO.
Enabling
TCP
segmentation
offload
will
reduce
CPU
utilization
on
the
transmit
side
of
a
TCP
connection,
and
so
improve
peak
throughput,
if
the
CPU
is
fully
utilized.
Since
TSO
has
no
effect
on
latency,
it
can
be
enabled
at
all
times.
The
driver
has
TSO
enabled
by
default.
Therefore,
there
is
no
opportunity
to
improve
performance
from
the
default.
NOTE:
TSO
cannot
be
controlled
via
the
host
on
VMware
ESX.
It
can
only
be
controlled
via
the
guest
Operating
System.
TCP
Large
Receive
Offload
(LRO)
TCP
Large
Receive
Offload
(LRO)
is
a
feature
whereby
the
adapter
coalesces
multiple
packets
received
on
a
TCP
connection
into
a
single
call
to
the
operating
system
TCP
Stack.
This
reduces
CPU
utilization,
and
so
improves
peak
throughput
when
the
CPU
is
fully
utilized.