Download raw body.
SoftLRO diff for cnmac/octeon
On Thu, 29 Jan 2026 16:00:48 +0100, Janne Johansson <icepic.dz@gmail.com> wrote: > > This is an update for the diff I sent when Jan Klemkow first sent out > the proposal for SoftLRO last year for certain interfaces. > It still gives me large speedups (*) for incoming TCP and I have been > running it for a long while. > > In case gmail as usual mangles the whitespaces, the diff is also available here: > > https://s3.sto3.safedc.net/jj-public-files/cnmac.softlro.diff > > Index: dev/if_cnmac.c > =================================================================== > RCS file: /cvs/src/sys/arch/octeon/dev/if_cnmac.c,v > diff -u -p -u -r1.86 if_cnmac.c > --- dev/if_cnmac.c 20 May 2024 23:13:33 -0000 1.86 > +++ dev/if_cnmac.c 29 Jan 2026 14:35:19 -0000 > @@ -55,6 +55,9 @@ > #include <net/if_media.h> > #include <netinet/in.h> > #include <netinet/if_ether.h> > +#include <netinet/tcp.h> > +#include <netinet/tcp_timer.h> > +#include <netinet/tcp_var.h> > > #if NBPFILTER > 0 > #include <net/bpf.h> > @@ -106,6 +109,7 @@ > > CTASSERT(MCLBYTES >= OCTEON_POOL_SIZE_PKT + CACHELINESIZE); > > + > void cnmac_buf_init(struct cnmac_softc *); > > int cnmac_match(struct device *, void *, void *); > @@ -306,7 +310,7 @@ cnmac_attach(struct device *parent, stru > strncpy(ifp->if_xname, sc->sc_dev.dv_xname, sizeof(ifp->if_xname)); > ifp->if_softc = sc; > ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST; > - ifp->if_xflags = IFXF_MPSAFE; > + ifp->if_xflags = IFXF_MPSAFE|IFXF_LRO; > ifp->if_ioctl = cnmac_ioctl; > ifp->if_qstart = cnmac_start; > ifp->if_watchdog = cnmac_watchdog; > @@ -314,7 +318,8 @@ cnmac_attach(struct device *parent, stru > ifq_init_maxlen(&ifp->if_snd, max(GATHER_QUEUE_SIZE, IFQ_MAXLEN)); > > ifp->if_capabilities = IFCAP_VLAN_MTU | IFCAP_CSUM_TCPv4 | > - IFCAP_CSUM_UDPv4 | IFCAP_CSUM_TCPv6 | IFCAP_CSUM_UDPv6; > + IFCAP_CSUM_UDPv4 | IFCAP_CSUM_TCPv6 | IFCAP_CSUM_UDPv6 | > + IFCAP_LRO; > > cn30xxgmx_set_filter(sc->sc_gmx_port); > > @@ -1246,7 +1251,10 @@ cnmac_recv(struct cnmac_softc *sc, uint6 > M_TCP_CSUM_IN_OK | M_UDP_CSUM_IN_OK; > } > > - ml_enqueue(ml, m); > + if (ISSET(ifp->if_xflags, IFXF_LRO)) > + tcp_softlro_glue(ml, m, ifp); > + else > + ml_enqueue(ml, m); > > return nmbuf; > > > > > Benching with iperf3 -s running on an Edgerouter 8-Pro with "ifconfig > cnmac0 -tcplro" > > jj@jjs-MacBook-Air:~$ iperf3 -c 172.31.31.102 > Connecting to host 172.31.31.102, port 5201 > [ 5] local 172.31.31.1 port 53101 connected to 172.31.31.102 port 5201 > [ ID] Interval Transfer Bitrate > [ 5] 0.00-1.00 sec 50.9 MBytes 426 Mbits/sec > [ 5] 1.00-2.01 sec 49.0 MBytes 410 Mbits/sec > [ 5] 2.01-3.00 sec 50.5 MBytes 424 Mbits/sec > [ 5] 3.00-4.00 sec 51.1 MBytes 429 Mbits/sec > [ 5] 4.00-5.00 sec 48.4 MBytes 405 Mbits/sec > [ 5] 5.00-6.00 sec 49.5 MBytes 415 Mbits/sec > [ 5] 6.00-7.00 sec 49.5 MBytes 416 Mbits/sec > [ 5] 7.00-8.01 sec 51.1 MBytes 428 Mbits/sec > [ 5] 8.01-9.00 sec 49.2 MBytes 414 Mbits/sec > [ 5] 9.00-10.00 sec 48.8 MBytes 409 Mbits/sec > - - - - - - - - - - - - - - - - - - - - - - - - - > [ ID] Interval Transfer Bitrate > [ 5] 0.00-10.00 sec 498 MBytes 418 Mbits/sec sender > [ 5] 0.00-10.02 sec 498 MBytes 417 Mbits/sec receiver > > iperf Done. > > After enabling LRO again: > > jj@jjs-MacBook-Air:~$ iperf3 -c 172.31.31.102 > Connecting to host 172.31.31.102, port 5201 > [ 5] local 172.31.31.1 port 53103 connected to 172.31.31.102 port 5201 > [ ID] Interval Transfer Bitrate > [ 5] 0.00-1.00 sec 109 MBytes 910 Mbits/sec > [ 5] 1.00-2.01 sec 113 MBytes 942 Mbits/sec > [ 5] 2.01-3.00 sec 112 MBytes 943 Mbits/sec > [ 5] 3.00-4.00 sec 111 MBytes 928 Mbits/sec > [ 5] 4.00-5.00 sec 111 MBytes 933 Mbits/sec > [ 5] 5.00-6.00 sec 112 MBytes 939 Mbits/sec > [ 5] 6.00-7.00 sec 112 MBytes 937 Mbits/sec > [ 5] 7.00-8.00 sec 112 MBytes 938 Mbits/sec > [ 5] 8.00-9.00 sec 111 MBytes 933 Mbits/sec > [ 5] 9.00-10.00 sec 112 MBytes 935 Mbits/sec > - - - - - - - - - - - - - - - - - - - - - - - - - > [ ID] Interval Transfer Bitrate > [ 5] 0.00-10.00 sec 1.09 GBytes 934 Mbits/sec sender > [ 5] 0.00-10.01 sec 1.09 GBytes 932 Mbits/sec receiver > > iperf Done. > > Almost same diff tested on a friends ER-4 with 7.8 gives 200 MBit/s > extra on those tests. > Tested on my ER-4 which I used as home's gw with multiple vlans. mac mini -> ER-4 [ ID] Interval Transfer Bitrate Retr [ 5] 0.00-30.00 sec 1.35 GBytes 387 Mbits/sec 4344 sender [ 5] 0.00-30.01 sec 1.35 GBytes 387 Mbits/sec receiver mac mini -> ER-4 -> mac mini 2 [ ID] Interval Transfer Bitrate Retr [ 5] 0.00-30.00 sec 1.63 GBytes 467 Mbits/sec 130320 sender [ 5] 0.00-30.01 sec 1.63 GBytes 467 Mbits/sec receiver With this diff: mac mini -> ER-4 [ ID] Interval Transfer Bitrate Retr [ 5] 0.00-30.00 sec 3.00 GBytes 859 Mbits/sec 70989 sender [ 5] 0.00-30.01 sec 3.00 GBytes 858 Mbits/sec receiver mac mini -> ER-4 -> mac mini 2 [ ID] Interval Transfer Bitrate Retr [ 5] 0.00-30.00 sec 1.65 GBytes 473 Mbits/sec 545896 sender [ 5] 0.00-30.01 sec 1.65 GBytes 472 Mbits/sec receiver but, mac minis are in different vlans but on the same cnmac interface. -- wbr, Kirill
SoftLRO diff for cnmac/octeon