Eric Dumazet f54b311142 tcp: auto corking ... With the introduction of TCP Small Queues, TSO auto sizing, and TCP pacing, we can implement Automatic Corking in the kernel, to help applications doing small write()/sendmsg() to TCP sockets. Idea is to change tcp_push() to check if the current skb payload is under skb optimal size (a multiple of MSS bytes) If under 'size_goal', and at least one packet is still in Qdisc or NIC TX queues, set the TCP Small Queue Throttled bit, so that the push will be delayed up to TX completion time. This delay might allow the application to coalesce more bytes in the skb in following write()/sendmsg()/sendfile() system calls. The exact duration of the delay is depending on the dynamics of the system, and might be zero if no packet for this flow is actually held in Qdisc or NIC TX ring. Using FQ/pacing is a way to increase the probability of autocorking being triggered. Add a new sysctl (/proc/sys/net/ipv4/tcp_autocorking) to control this feature and default it to 1 (enabled) Add a new SNMP counter : nstat -a | grep TcpExtTCPAutoCorking This counter is incremented every time we detected skb was under used and its flush was deferred. Tested: Interesting effects when using line buffered commands under ssh. Excellent performance results in term of cpu usage and total throughput. lpq83:~# echo 1 >/proc/sys/net/ipv4/tcp_autocorking lpq83:~# perf stat ./super_netperf 4 -t TCP_STREAM -H lpq84 -- -m 128 9410.39 Performance counter stats for './super_netperf 4 -t TCP_STREAM -H lpq84 -- -m 128': 35209.439626 task-clock # 2.901 CPUs utilized 2,294 context-switches # 0.065 K/sec 101 CPU-migrations # 0.003 K/sec 4,079 page-faults # 0.116 K/sec 97,923,241,298 cycles # 2.781 GHz [83.31%] 51,832,908,236 stalled-cycles-frontend # 52.93% frontend cycles idle [83.30%] 25,697,986,603 stalled-cycles-backend # 26.24% backend cycles idle [66.70%] 102,225,978,536 instructions # 1.04 insns per cycle # 0.51 stalled cycles per insn [83.38%] 18,657,696,819 branches # 529.906 M/sec [83.29%] 91,679,646 branch-misses # 0.49% of all branches [83.40%] 12.136204899 seconds time elapsed lpq83:~# echo 0 >/proc/sys/net/ipv4/tcp_autocorking lpq83:~# perf stat ./super_netperf 4 -t TCP_STREAM -H lpq84 -- -m 128 6624.89 Performance counter stats for './super_netperf 4 -t TCP_STREAM -H lpq84 -- -m 128': 40045.864494 task-clock # 3.301 CPUs utilized 171 context-switches # 0.004 K/sec 53 CPU-migrations # 0.001 K/sec 4,080 page-faults # 0.102 K/sec 111,340,458,645 cycles # 2.780 GHz [83.34%] 61,778,039,277 stalled-cycles-frontend # 55.49% frontend cycles idle [83.31%] 29,295,522,759 stalled-cycles-backend # 26.31% backend cycles idle [66.67%] 108,654,349,355 instructions # 0.98 insns per cycle # 0.57 stalled cycles per insn [83.34%] 19,552,170,748 branches # 488.244 M/sec [83.34%] 157,875,417 branch-misses # 0.81% of all branches [83.34%] 12.130267788 seconds time elapsed Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>		2013-12-06 12:51:41 -05:00
..
ABI	Main batch of InfiniBand/RDMA changes for 3.13:	2013-11-18 15:36:04 -08:00
DocBook	doc: fix generation of device-drivers	2013-11-27 20:40:04 -08:00
EDID	drm: Add 1600x1200 (UXGA) screen resolution to the built-in EDIDs	2013-04-12 14:06:16 +10:00
PCI	PCI: Update pci_find_slot() description in pci.txt	2013-09-25 15:43:33 -06:00
RCU	rcu: Fix occurrence of "the the" in checklist.txt	2013-09-25 10:07:02 -07:00
accounting	Documentation/accounting/getdelays.c: avoid strncpy in accounting tool	2013-07-03 16:08:06 -07:00
acpi	gpiolib / ACPI: document the GPIO descriptor based interface	2013-10-19 23:32:50 +02:00
aoe	aoe: remove do-nothing NAME="%k" term from example udev rules	2013-09-11 15:59:28 -07:00
arm	Allwinner sunXi SoCs machine additions for 3.13	2013-10-28 10:19:38 -07:00
arm64	arm64: Use 42-bit address space with 64K pages	2013-11-05 17:23:52 +00:00
auxdisplay	…
backlight	backlight: lp855x_bl: support new LP8555 device	2013-11-13 12:09:14 +09:00
blackfin	…
block	block: change config option name for cmdline partition parsing	2013-09-30 14:31:02 -07:00
blockdev	floppy: Correct documentation of driver options when used as a module.	2013-11-08 09:10:31 -07:00
bus-devices	…
cdrom	…
cgroups	memcg: support hierarchical memory.numa_stats	2013-11-13 12:09:06 +09:00
connector	connector - documentation: simplify netlink message length assignment	2013-10-02 16:03:51 -04:00
console	TTY:console: update document console.txt	2013-05-21 10:21:57 -07:00
cpu-freq	cpufreq: Implement light weight ->target_index() routine	2013-10-25 22:42:24 +02:00
cpuidle	cpuidle: remove cpuidle_unregister_governor()	2013-10-30 01:21:24 +01:00
cris	…
crypto	drivers/dma: remove unused support for MEMSET operations	2013-07-03 16:07:42 -07:00
development-process	Documentation: development-process: Update -mm and -next URLs	2013-07-25 12:37:24 +02:00
device-mapper	dm cache: resolve small nits and improve Documentation	2013-11-12 13:11:09 -05:00
devicetree	ARM: SoC fixes for 3.13-rc	2013-11-26 11:18:37 -08:00
driver-model	spi: Updates for v3.13	2013-11-12 15:01:39 +09:00
dvb	…
early-userspace	Documentation: remove reference to 2.7 kernel in early-userspace	2013-08-20 12:47:28 +02:00
extcon	extcon: Simplify extcon_dev_register() prototype by removing unnecessary parameter	2013-09-27 09:37:01 +09:00
fault-injection	…
fb	Documentation/fb/viafb.modes fix a typo	2013-08-20 12:41:11 +02:00
filesystems	Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs	2013-11-22 08:38:55 -08:00
firmware_class	…
fmc	doc: Fix typo "is is" in Documentations	2013-08-27 10:50:52 +02:00
frv	…
gpio	Documentation: gpiolib: document new interface	2013-11-25 09:02:30 +01:00
hid	HID: uhid: use generic hidinput_input_event()	2013-07-31 10:33:05 +02:00
hwmon	Merge branch 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging	2013-11-15 16:35:10 -08:00
i2c	i2c: i801: Add Device IDs for Intel Wildcat Point-LP PCH	2013-11-14 18:38:04 +01:00
i2o	…
ia64	Fix example error_injection_tool	2013-04-02 09:39:55 -07:00
ide	…
infiniband	…
input	Input: clarify gamepad API ABS values	2013-10-15 23:42:07 -07:00
ioctl	ALSA: add DICE driver	2013-10-17 21:18:32 +02:00
isdn	…
ja_JP	HOWTO ja_JP sync	2013-07-24 22:06:34 -07:00
kbuild	Documentation/kbuild/kconfig.txt: 'make listnewconfig' replaces: yes "" | make oldconfig	2013-10-08 23:51:50 +02:00
kdump	Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial	2013-07-04 11:40:58 -07:00
ko_KR	Correct unfaithful translation on HOWTO in ko_KR	2013-08-12 17:43:13 -07:00
laptops	thinkpad-acpi: Add mute and mic-mute LED functionality	2013-10-17 14:38:44 +02:00
leds	Documentation: leds-lp5521,lp5523: update device attribute information	2013-08-26 17:22:13 -07:00
m68k	block: remove refs to XD disks from documentation	2013-05-17 15:17:12 +02:00
make	…
memory-devices	…
metag	doc: fix misspellings with 'codespell' tool	2013-05-28 12:02:12 +02:00
mic	misc: mic: Enable OSPM suspend and resume support.	2013-10-05 18:01:42 -07:00
mips	…
misc-devices	doc: fix misspellings with 'codespell' tool	2013-05-28 12:02:12 +02:00
mmc	mmc: core: Add in support to expose PRV for v4 MMCs	2013-03-22 12:10:42 -04:00
mn10300	…
mtd	doc: Fix typo "is is" in Documentations	2013-08-27 10:50:52 +02:00
namespaces	…
netlabel	…
networking	tcp: auto corking	2013-12-06 12:51:41 -05:00
nfc	…
parisc	parisc: document the shadow registers	2013-07-09 22:09:19 +02:00
pcmcia	…
power	More ACPI and power management updates for 3.13-rc1	2013-11-20 13:25:04 -08:00
powerpc	powerpc: Update the 00-Index in Documentation/powerpc	2013-08-27 14:44:27 +10:00
pps	USB: serial: invoke dcd_change ldisc's handler.	2013-09-26 09:45:40 -07:00
prctl	…
pti	…
ptp	ptp: add the PTP_SYS_OFFSET ioctl to the testptp program	2013-09-23 16:46:17 -04:00
rapidio	doc: Fix typo in doucmentations	2013-07-25 12:34:15 +02:00
s390	s390/s390dbf: add debug_level_enabled() function	2013-10-24 17:16:53 +02:00
scheduler	H8/300 has been dead for several years, the kernel for it has	2013-11-12 14:13:14 +09:00
scsi	SCSI misc on 20130915	2013-09-15 17:41:30 -04:00
security	ima: new templates management mechanism	2013-10-25 17:17:04 -04:00
serial	serial: core: delete .set_wake() callback	2013-10-16 13:16:19 -07:00
sh	…
sound	ALSA: Fix typo in documentation/alsa	2013-10-29 11:38:04 +01:00
spi	spi/documentation: Fix usage of __initdata	2013-08-20 12:52:28 +02:00
sysctl	vsprintf: check real user/group id for %pK	2013-11-13 12:09:14 +09:00
target	target: Remove TF_CIT_TMPL macro	2013-10-16 13:35:02 -07:00
thermal	thermal: thermal_core: allow binding with limits on bind_params	2013-09-03 09:10:24 -04:00
timers	doc: add missing files to timers/00-INDEX	2013-10-27 21:55:50 +00:00
tpm	drivers/tpm: add xen tpmfront interface	2013-08-09 10:57:06 -04:00
trace	Documentation/trace/tracepoints.txt: add links to TRACE_EVENT documentation	2013-11-13 12:09:32 +09:00
usb	doc: usb: Fix typo in Documentation/usb/gadget_configs.txt	2013-10-31 13:31:39 +01:00
vDSO	…
video4linux	[media] V4L: Add support for integer menu controls with standard menu items	2013-08-18 07:12:59 -03:00
virtual	Merge branch 'kvm-ppc-queue' of git://github.com/agraf/linux-2.6 into queue	2013-11-04 10:20:57 +02:00
vm	x86, mm: do not leak page->ptl for pmd page tables	2013-11-21 16:42:28 -08:00
w1	Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial	2013-07-04 11:40:58 -07:00
watchdog	watchdog: delete mpcore_wdt driver	2013-07-11 21:47:58 +02:00
wimax	…
x86	Merge branch 'x86-efi-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip	2013-11-12 10:48:30 +09:00
xtensa	xtensa: document MMUv3 setup sequence	2013-05-09 01:07:09 -07:00
zh_CN	Documentation/zh_CN/SubmittingPatches fix a typo	2013-08-20 12:41:25 +02:00
.gitignore	…
00-INDEX	doc: fix a typo in Documentation/00-INDEX	2013-08-27 10:53:07 +02:00
BUG-HUNTING	…
Changes	remove obsolete references to powertweak	2013-11-27 20:34:32 -08:00
CodingStyle	Documentation/CodingStyle: allow multiple return statements per function	2013-07-03 16:08:01 -07:00
DMA-API-HOWTO.txt	DMA-API: provide a helper to set both DMA and coherent DMA masks	2013-09-17 15:32:37 +01:00
DMA-API.txt	DMA-API: provide a helper to set both DMA and coherent DMA masks	2013-09-17 15:32:37 +01:00
DMA-ISA-LPC.txt	…
DMA-attributes.txt	doc: Documentation/DMA-attributes.txt fix typo	2013-10-14 15:50:53 +02:00
HOWTO	Documentation: Updated broken link in HOWTO	2013-06-03 14:22:57 -07:00
IPMI.txt	ipmi: add options to disable openfirmware and PCI scanning	2013-02-27 19:10:21 -08:00
IRQ-affinity.txt	doc: fix a typo about irq affinity	2013-08-20 12:59:18 +02:00
IRQ-domain.txt	…
IRQ.txt	…
Intel-IOMMU.txt	…
Makefile	…
ManagementStyle	…
SAK.txt	…
SM501.txt	…
SecurityBugs	…
SubmitChecklist	Finally eradicate CONFIG_HOTPLUG	2013-06-03 14:20:18 -07:00
SubmittingDrivers	…
SubmittingPatches	Documentation/SubmittingPatches: Request summaries for commit references	2013-08-20 12:58:15 +02:00
VGA-softcursor.txt	…
applying-patches.txt	…
assoc_array.txt	Add a generic associative array implementation.	2013-09-24 10:35:17 +01:00
atomic_ops.txt	…
bad_memory.txt	…
basic_profiling.txt	…
bcache.txt	Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial	2013-07-04 11:40:58 -07:00
binfmt_misc.txt	…
braille-console.txt	…
bt8xxgpio.txt	…
btmrvl.txt	…
bus-virt-phys-mapping.txt	…
cachetlb.txt	Documentation: fix typo and update version in cachetlb.txt	2013-08-20 12:46:52 +02:00
circular-buffers.txt	…
clk.txt	clk: add support for clock reparent on set_rate	2013-08-19 12:27:17 -07:00
coccinelle.txt	Coccinelle: Update information about the minimal version required	2013-07-03 22:58:20 +02:00
cpu-hotplug.txt	MAINTAINERS: update Zwane Mwaikambo's e-mail address	2013-11-13 12:09:14 +09:00
cpu-load.txt	…
cputopology.txt	doc: Documentation/cputopology.txt fix typo	2013-09-04 12:59:47 +02:00
crc32.txt	…
dcdbas.txt	…
debugging-modules.txt	…
debugging-via-ohci1394.txt	…
dell_rbu.txt	…
devices.txt	cuse: add fix minor number to /dev/cuse	2013-10-01 16:44:54 +02:00
digsig.txt	…
dma-buf-sharing.txt	dma-buf: Expose buffer size to userspace (v2)	2013-09-10 11:36:45 +05:30
dmaengine.txt	…
dmatest.txt	dmatest: add a 'wait' parameter	2013-11-14 11:04:40 -08:00
dontdiff	…
dynamic-debug-howto.txt	doc: fix misspellings with 'codespell' tool	2013-05-28 12:02:12 +02:00
edac.txt	…
efi-stub.txt	EFI stub documentation updates	2013-09-25 12:34:32 +01:00
eisa.txt	…
email-clients.txt	…
flexible-arrays.txt	…
futex-requeue-pi.txt	…
gcov.txt	gcov: compile specific gcov implementation based on gcc version	2013-11-13 12:09:34 +09:00
highuid.txt	…
hw_random.txt	hwrng: Fix a wrong comment in Documentation/hw_random.txt	2013-03-10 18:16:36 +08:00
hwspinlock.txt	doc: documentation/hwspinlock.txt fix typo	2013-08-27 10:46:02 +02:00
init.txt	…
initrd.txt	…
intel_txt.txt	…
io-mapping.txt	…
io_ordering.txt	…
iostats.txt	iostats.txt: add easy-to-find description for field 6	2013-04-29 15:18:50 +02:00
irqflags-tracing.txt	…
isapnp.txt	…
java.txt	…
kernel-doc-nano-HOWTO.txt	kernel-doc: Update references to SGML to refs to XML instead.	2013-05-28 12:02:11 +02:00
kernel-docs.txt	…
kernel-parameters.txt	Merge branch 'for-linus2' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security	2013-11-21 19:46:00 -08:00
kernel-per-CPU-kthreads.txt	kthread: Add pointer to vmstat-avoidance patch	2013-09-25 06:49:46 -07:00
kmemcheck.txt	Documentation/kmemcheck: update kmemcheck documentation	2013-08-27 10:47:05 +02:00
kmemleak.txt	…
kobject.txt	…
kprobes.txt	…
kref.txt	…
ldm.txt	…
local_ops.txt	…
lockdep-design.txt	…
lockstat.txt	lockstat: Report avg wait and hold times	2013-10-09 08:19:08 +02:00
lockup-watchdogs.txt	…
logo.gif	…
logo.txt	…
magic-number.txt	wanrouter: completely decouple obsolete code from kernel.	2013-01-31 19:20:33 -05:00
md.txt	md: remove doubled description for sync_max, merging it within sync_min/sync_max	2013-07-03 09:43:28 +10:00
media-framework.txt	Merge branch 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media	2013-07-13 12:09:57 -07:00
memory-barriers.txt	doc: Fix memory-barrier control-dependency example	2013-08-19 21:39:42 -07:00
memory-hotplug.txt	Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial	2013-09-06 09:36:28 -07:00
mono.txt	…
mutex-design.txt	locking/doc: Update references to kernel/mutex.c	2013-11-11 12:41:33 +01:00
nommu-mmap.txt	…
numastat.txt	…
oops-tracing.txt	…
padata.txt	…
parport-lowlevel.txt	…
parport.txt	…
percpu-rw-semaphore.txt	…
phy.txt	drivers: phy: add generic PHY framework	2013-09-27 17:35:41 -07:00
pi-futex.txt	…
pinctrl.txt	pinctrl: add documentation for pinctrl_get_group_pins()	2013-10-16 15:35:21 +02:00
pnp.txt	…
preempt-locking.txt	…
printk-formats.txt	Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial	2013-09-06 09:36:28 -07:00
pwm.txt	Documentation/pwm: Fix trivial typos	2013-10-24 10:51:33 +02:00
ramoops.txt	…
rbtree.txt	…
remoteproc.txt	…
rfkill.txt	…
robust-futex-ABI.txt	…
robust-futexes.txt	…
rpmsg.txt	…
rt-mutex-design.txt	sched: Rename sched.c as sched/core.c in comments and Documentation	2013-06-19 12:58:42 +02:00
rt-mutex.txt	…
rtc.txt	rtc: add ability to push out an existing wakealarm using sysfs	2013-07-03 16:07:54 -07:00
serial-console.txt	…
sgi-ioc4.txt	…
sgi-visws.txt	…
smsc_ece1099.txt	…
sparse.txt	…
spinlocks.txt	sched: Rename sched.c as sched/core.c in comments and Documentation	2013-06-19 12:58:42 +02:00
stable_api_nonsense.txt	…
stable_kernel_rules.txt	…
static-keys.txt	…
svga.txt	…
sysfs-rules.txt	doc: Fix typo in doucmentations	2013-07-25 12:34:15 +02:00
sysrq.txt	sysrq: Allow magic SysRq key functions to be disabled through Kconfig	2013-10-16 13:01:44 -07:00
this_cpu_ops.txt	percpu: add documentation on this_cpu operations	2013-04-04 10:24:53 -07:00
unaligned-memory-access.txt	…
unicode.txt	…
unshare.txt	…
vfio.txt	vfio: fix documentation	2013-09-05 16:36:21 -06:00
vgaarbiter.txt	…
video-output.txt	…
vme_api.txt	…
volatile-considered-harmful.txt	…
workqueue.txt	workqueue: Correct/Drop references to gcwq in Documentation	2013-08-21 10:32:09 -04:00
ww-mutex-design.txt	mutex: Add support for wound/wait style locks	2013-06-26 12:10:56 +02:00
xz.txt	…
zorro.txt	…