d346a3fae3
This patch introduces a PACKET_QDISC_BYPASS socket option, that
allows for using a similar xmit() function as in pktgen instead
of taking the dev_queue_xmit() path. This can be very useful when
PF_PACKET applications are required to be used in a similar
scenario as pktgen, but with full, flexible packet payload that
needs to be provided, for example.
On default, nothing changes in behaviour for normal PF_PACKET
TX users, so everything stays as is for applications. New users,
however, can now set PACKET_QDISC_BYPASS if needed to prevent
own packets from i) reentering packet_rcv() and ii) to directly
push the frame to the driver.
In doing so we can increase pps (here 64 byte packets) for
PF_PACKET a bit:
# CPUs -- QDISC_BYPASS -- qdisc path -- qdisc path[**]
1 CPU == 1,509,628 pps -- 1,208,708 -- 1,247,436
2 CPUs == 3,198,659 pps -- 2,536,012 -- 1,605,779
3 CPUs == 4,787,992 pps -- 3,788,740 -- 1,735,610
4 CPUs == 6,173,956 pps -- 4,907,799 -- 1,909,114
5 CPUs == 7,495,676 pps -- 5,956,499 -- 2,014,422
6 CPUs == 9,001,496 pps -- 7,145,064 -- 2,155,261
7 CPUs == 10,229,776 pps -- 8,190,596 -- 2,220,619
8 CPUs == 11,040,732 pps -- 9,188,544 -- 2,241,879
9 CPUs == 12,009,076 pps -- 10,275,936 -- 2,068,447
10 CPUs == 11,380,052 pps -- 11,265,337 -- 1,578,689
11 CPUs == 11,672,676 pps -- 11,845,344 -- 1,297,412
[...]
20 CPUs == 11,363,192 pps -- 11,014,933 -- 1,245,081
[**]: qdisc path with packet_rcv(), how probably most people
seem to use it (hopefully not anymore if not needed)
The test was done using a modified trafgen, sending a simple
static 64 bytes packet, on all CPUs. The trick in the fast
"qdisc path" case, is to avoid reentering packet_rcv() by
setting the RAW socket protocol to zero, like:
socket(PF_PACKET, SOCK_RAW, 0);
Tradeoffs are documented as well in this patch, clearly, if
queues are busy, we will drop more packets, tc disciplines are
ignored, and these packets are not visible to taps anymore. For
a pktgen like scenario, we argue that this is acceptable.
The pointer to the xmit function has been placed in packet
socket structure hole between cached_dev and prot_hook that
is hot anyway as we're working on cached_dev in each send path.
Done in joint work together with Jesper Dangaard Brouer.
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
||
|---|---|---|
| .. | ||
| byteorder | ||
| caif | ||
| can | ||
| cifs | ||
| dvb | ||
| hdlc | ||
| hsi | ||
| isdn | ||
| mmc | ||
| netfilter | ||
| netfilter_arp | ||
| netfilter_bridge | ||
| netfilter_ipv4 | ||
| netfilter_ipv6 | ||
| nfsd | ||
| raid | ||
| spi | ||
| sunrpc | ||
| tc_act | ||
| tc_ematch | ||
| usb | ||
| wimax | ||
| a.out.h | ||
| acct.h | ||
| adb.h | ||
| adfs_fs.h | ||
| affs_hardblocks.h | ||
| agpgart.h | ||
| aio_abi.h | ||
| apm_bios.h | ||
| arcfb.h | ||
| atalk.h | ||
| atm.h | ||
| atm_eni.h | ||
| atm_he.h | ||
| atm_idt77105.h | ||
| atm_nicstar.h | ||
| atm_tcp.h | ||
| atm_zatm.h | ||
| atmapi.h | ||
| atmarp.h | ||
| atmbr2684.h | ||
| atmclip.h | ||
| atmdev.h | ||
| atmioc.h | ||
| atmlec.h | ||
| atmmpc.h | ||
| atmppp.h | ||
| atmsap.h | ||
| atmsvc.h | ||
| audit.h | ||
| auto_fs.h | ||
| auto_fs4.h | ||
| auxvec.h | ||
| ax25.h | ||
| b1lli.h | ||
| baycom.h | ||
| bcache.h | ||
| bcm933xx_hcs.h | ||
| bfs_fs.h | ||
| binfmts.h | ||
| blkpg.h | ||
| blktrace_api.h | ||
| bpqether.h | ||
| bsg.h | ||
| btrfs.h | ||
| can.h | ||
| capability.h | ||
| capi.h | ||
| cciss_defs.h | ||
| cciss_ioctl.h | ||
| cdrom.h | ||
| cgroupstats.h | ||
| chio.h | ||
| cm4000_cs.h | ||
| cn_proc.h | ||
| coda.h | ||
| coda_psdev.h | ||
| coff.h | ||
| connector.h | ||
| const.h | ||
| cramfs_fs.h | ||
| cuda.h | ||
| cyclades.h | ||
| cycx_cfm.h | ||
| dcbnl.h | ||
| dccp.h | ||
| dlm.h | ||
| dlm_device.h | ||
| dlm_netlink.h | ||
| dlm_plock.h | ||
| dlmconstants.h | ||
| dm-ioctl.h | ||
| dm-log-userspace.h | ||
| dn.h | ||
| dqblk_xfs.h | ||
| edd.h | ||
| efs_fs_sb.h | ||
| elf-em.h | ||
| elf-fdpic.h | ||
| elf.h | ||
| elfcore.h | ||
| errno.h | ||
| errqueue.h | ||
| ethtool.h | ||
| eventpoll.h | ||
| fadvise.h | ||
| falloc.h | ||
| fanotify.h | ||
| fb.h | ||
| fcntl.h | ||
| fd.h | ||
| fdreg.h | ||
| fib_rules.h | ||
| fiemap.h | ||
| filter.h | ||
| firewire-cdev.h | ||
| firewire-constants.h | ||
| flat.h | ||
| fs.h | ||
| fsl_hypervisor.h | ||
| fuse.h | ||
| futex.h | ||
| gameport.h | ||
| gen_stats.h | ||
| genetlink.h | ||
| gfs2_ondisk.h | ||
| gigaset_dev.h | ||
| hash_info.h | ||
| hdlc.h | ||
| hdlcdrv.h | ||
| hdreg.h | ||
| hid.h | ||
| hiddev.h | ||
| hidraw.h | ||
| hpet.h | ||
| hsr_netlink.h | ||
| hw_breakpoint.h | ||
| hysdn_if.h | ||
| i2c-dev.h | ||
| i2c.h | ||
| i2o-dev.h | ||
| i8k.h | ||
| icmp.h | ||
| icmpv6.h | ||
| if.h | ||
| if_addr.h | ||
| if_addrlabel.h | ||
| if_alg.h | ||
| if_arcnet.h | ||
| if_arp.h | ||
| if_bonding.h | ||
| if_bridge.h | ||
| if_cablemodem.h | ||
| if_eql.h | ||
| if_ether.h | ||
| if_fc.h | ||
| if_fddi.h | ||
| if_frad.h | ||
| if_hippi.h | ||
| if_infiniband.h | ||
| if_link.h | ||
| if_ltalk.h | ||
| if_packet.h | ||
| if_phonet.h | ||
| if_plip.h | ||
| if_ppp.h | ||
| if_pppol2tp.h | ||
| if_pppox.h | ||
| if_slip.h | ||
| if_team.h | ||
| if_tun.h | ||
| if_tunnel.h | ||
| if_vlan.h | ||
| if_x25.h | ||
| igmp.h | ||
| in.h | ||
| in6.h | ||
| in_route.h | ||
| inet_diag.h | ||
| inotify.h | ||
| input.h | ||
| ioctl.h | ||
| ip.h | ||
| ip6_tunnel.h | ||
| ip_vs.h | ||
| ipc.h | ||
| ipmi.h | ||
| ipmi_msgdefs.h | ||
| ipsec.h | ||
| ipv6.h | ||
| ipv6_route.h | ||
| ipx.h | ||
| irda.h | ||
| irqnr.h | ||
| isdn.h | ||
| isdn_divertif.h | ||
| isdn_ppp.h | ||
| isdnif.h | ||
| iso_fs.h | ||
| ivtv.h | ||
| ivtvfb.h | ||
| ixjuser.h | ||
| jffs2.h | ||
| joystick.h | ||
| Kbuild | ||
| kd.h | ||
| kdev_t.h | ||
| kernel-page-flags.h | ||
| kernel.h | ||
| kernelcapi.h | ||
| kexec.h | ||
| keyboard.h | ||
| keyctl.h | ||
| kvm.h | ||
| kvm_para.h | ||
| l2tp.h | ||
| libc-compat.h | ||
| limits.h | ||
| llc.h | ||
| loop.h | ||
| lp.h | ||
| magic.h | ||
| major.h | ||
| map_to_7segment.h | ||
| matroxfb.h | ||
| mdio.h | ||
| media.h | ||
| mei.h | ||
| mempolicy.h | ||
| meye.h | ||
| mic_common.h | ||
| mic_ioctl.h | ||
| mii.h | ||
| minix_fs.h | ||
| mman.h | ||
| mmtimer.h | ||
| module.h | ||
| mqueue.h | ||
| mroute.h | ||
| mroute6.h | ||
| msdos_fs.h | ||
| msg.h | ||
| mtio.h | ||
| n_r3964.h | ||
| nbd.h | ||
| ncp.h | ||
| ncp_fs.h | ||
| ncp_mount.h | ||
| ncp_no.h | ||
| neighbour.h | ||
| net.h | ||
| net_dropmon.h | ||
| net_tstamp.h | ||
| netconf.h | ||
| netdevice.h | ||
| netfilter.h | ||
| netfilter_arp.h | ||
| netfilter_bridge.h | ||
| netfilter_decnet.h | ||
| netfilter_ipv4.h | ||
| netfilter_ipv6.h | ||
| netlink.h | ||
| netlink_diag.h | ||
| netrom.h | ||
| nfc.h | ||
| nfs.h | ||
| nfs2.h | ||
| nfs3.h | ||
| nfs4.h | ||
| nfs4_mount.h | ||
| nfs_fs.h | ||
| nfs_idmap.h | ||
| nfs_mount.h | ||
| nfsacl.h | ||
| nl80211.h | ||
| nubus.h | ||
| nvme.h | ||
| nvram.h | ||
| omap3isp.h | ||
| omapfb.h | ||
| oom.h | ||
| openvswitch.h | ||
| packet_diag.h | ||
| param.h | ||
| parport.h | ||
| patchkey.h | ||
| pci.h | ||
| pci_regs.h | ||
| perf_event.h | ||
| personality.h | ||
| pfkeyv2.h | ||
| pg.h | ||
| phantom.h | ||
| phonet.h | ||
| pkt_cls.h | ||
| pkt_sched.h | ||
| pktcdvd.h | ||
| pmu.h | ||
| poll.h | ||
| posix_types.h | ||
| ppdev.h | ||
| ppp-comp.h | ||
| ppp-ioctl.h | ||
| ppp_defs.h | ||
| pps.h | ||
| prctl.h | ||
| ptp_clock.h | ||
| ptrace.h | ||
| qnx4_fs.h | ||
| qnxtypes.h | ||
| quota.h | ||
| radeonfb.h | ||
| random.h | ||
| raw.h | ||
| rds.h | ||
| reboot.h | ||
| reiserfs_fs.h | ||
| reiserfs_xattr.h | ||
| resource.h | ||
| rfkill.h | ||
| romfs_fs.h | ||
| rose.h | ||
| route.h | ||
| rtc.h | ||
| rtnetlink.h | ||
| scc.h | ||
| sched.h | ||
| screen_info.h | ||
| sctp.h | ||
| sdla.h | ||
| seccomp.h | ||
| securebits.h | ||
| selinux_netlink.h | ||
| sem.h | ||
| serial.h | ||
| serial_core.h | ||
| serial_reg.h | ||
| serio.h | ||
| shm.h | ||
| signal.h | ||
| signalfd.h | ||
| snmp.h | ||
| sock_diag.h | ||
| socket.h | ||
| sockios.h | ||
| som.h | ||
| sonet.h | ||
| sonypi.h | ||
| sound.h | ||
| soundcard.h | ||
| stat.h | ||
| stddef.h | ||
| string.h | ||
| suspend_ioctls.h | ||
| swab.h | ||
| synclink.h | ||
| sysctl.h | ||
| sysinfo.h | ||
| taskstats.h | ||
| tcp.h | ||
| tcp_metrics.h | ||
| telephony.h | ||
| termios.h | ||
| time.h | ||
| times.h | ||
| timex.h | ||
| tiocl.h | ||
| tipc.h | ||
| tipc_config.h | ||
| toshiba.h | ||
| tty.h | ||
| tty_flags.h | ||
| types.h | ||
| udf_fs_i.h | ||
| udp.h | ||
| uhid.h | ||
| uinput.h | ||
| uio.h | ||
| ultrasound.h | ||
| un.h | ||
| unistd.h | ||
| unix_diag.h | ||
| usbdevice_fs.h | ||
| utime.h | ||
| utsname.h | ||
| uuid.h | ||
| uvcvideo.h | ||
| v4l2-common.h | ||
| v4l2-controls.h | ||
| v4l2-dv-timings.h | ||
| v4l2-mediabus.h | ||
| v4l2-subdev.h | ||
| veth.h | ||
| vfio.h | ||
| vhost.h | ||
| videodev2.h | ||
| virtio_9p.h | ||
| virtio_balloon.h | ||
| virtio_blk.h | ||
| virtio_config.h | ||
| virtio_console.h | ||
| virtio_ids.h | ||
| virtio_net.h | ||
| virtio_pci.h | ||
| virtio_ring.h | ||
| virtio_rng.h | ||
| vm_sockets.h | ||
| vt.h | ||
| wait.h | ||
| wanrouter.h | ||
| watchdog.h | ||
| wimax.h | ||
| wireless.h | ||
| x25.h | ||
| xattr.h | ||
| xfrm.h | ||