linux/drivers
Ming Lei 396eaf21ee blk-mq: improve DM's blk-mq IO merging via blk_insert_cloned_request feedback
blk_insert_cloned_request() is called in the fast path of a dm-rq driver
(e.g. blk-mq request-based DM mpath).  blk_insert_cloned_request() uses
blk_mq_request_bypass_insert() to directly append the request to the
blk-mq hctx->dispatch_list of the underlying queue.

1) This way isn't efficient enough because the hctx spinlock is always
used.

2) With blk_insert_cloned_request(), we completely bypass underlying
queue's elevator and depend on the upper-level dm-rq driver's elevator
to schedule IO.  But dm-rq currently can't get the underlying queue's
dispatch feedback at all.  Without knowing whether a request was issued
or not (e.g. due to underlying queue being busy) the dm-rq elevator will
not be able to provide effective IO merging (as a side-effect of dm-rq
currently blindly destaging a request from its elevator only to requeue
it after a delay, which kills any opportunity for merging).  This
obviously causes very bad sequential IO performance.

Fix this by updating blk_insert_cloned_request() to use
blk_mq_request_direct_issue().  blk_mq_request_direct_issue() allows a
request to be issued directly to the underlying queue and returns the
dispatch feedback (blk_status_t).  If blk_mq_request_direct_issue()
returns BLK_SYS_RESOURCE the dm-rq driver will now use DM_MAPIO_REQUEUE
to _not_ destage the request.  Whereby preserving the opportunity to
merge IO.

With this, request-based DM's blk-mq sequential IO performance is vastly
improved (as much as 3X in mpath/virtio-scsi testing).

Signed-off-by: Ming Lei <ming.lei@redhat.com>
[blk-mq.c changes heavily influenced by Ming Lei's initial solution, but
they were refactored to make them less fragile and easier to read/review]
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-01-17 09:46:54 -07:00
..
accessibility License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
acpi PM / sleep: Avoid excess pm_runtime_enable() calls in device_resume() 2017-12-11 14:32:56 +01:00
amba A couple of dma-mapping updates: 2017-11-14 16:54:12 -08:00
android ANDROID: binder: fix transaction leak. 2017-11-28 16:54:45 +01:00
ata libata: sata_down_spd_limit should return if driver has not recorded sstatus speed 2017-12-04 13:57:03 -08:00
atm Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net 2017-11-29 13:10:25 -08:00
auxdisplay auxdisplay: img-ascii-lcd: Only build on archs that have IOMEM 2017-11-27 12:36:45 -08:00
base PM / sleep: Avoid excess pm_runtime_enable() calls in device_resume() 2017-12-11 14:32:56 +01:00
bcma Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next 2017-11-15 11:56:19 -08:00
block aoe: use ktime_t instead of timeval 2018-01-17 08:41:07 -07:00
bluetooth Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net 2017-11-04 09:26:51 +09:00
bus bus: arm-ccn: fix module unloading Error: Removing state 147 which has instances left. 2017-12-04 17:15:20 +00:00
cdrom Merge branch 'for-4.15/block' of git://git.kernel.dk/linux-block 2017-11-14 15:32:19 -08:00
char The big changes for IPMI that just went in had a few problems. These 2017-12-11 17:01:59 -08:00
clk We have two changes to the core framework this time around. The first being a 2017-11-17 20:04:24 -08:00
clocksource - final batch of "non trivial" timer conversions (multi-tree dependencies, 2017-11-23 16:29:05 +01:00
connector
cpufreq cpufreq: mediatek: add missing MODULE_DESCRIPTION/AUTHOR/LICENSE 2017-11-22 00:00:14 +01:00
cpuidle powerpc updates for 4.15 2017-11-16 12:47:46 -08:00
crypto powerpc updates for 4.15 2017-11-16 12:47:46 -08:00
dax device-dax: implement ->split() to catch invalid munmap attempts 2017-11-29 18:40:42 -08:00
dca
devfreq Merge branches 'pm-devfreq' and 'pm-tools' 2017-11-13 01:41:39 +01:00
dio License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
dma dmaengine: fsl-edma: disable clks on all error paths 2017-12-15 09:53:04 +05:30
dma-buf Tracing updates for 4.15: 2017-11-17 14:58:01 -08:00
edac Modules updates for v4.15 2017-11-15 13:46:33 -08:00
eisa License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
extcon USB/PHY patches for 4.15-rc1 2017-11-13 21:14:07 -08:00
firewire Merge branch 'timers-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2017-11-13 17:56:58 -08:00
firmware ARM: SoC fixes for 4.15-rc 2017-12-10 08:26:59 -08:00
fmc License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
fpga Char/Misc patches for 4.15-rc1 2017-11-16 09:10:59 -08:00
fsi
gpio gpio: pca953x: fix vendor prefix for PCA9654 2017-12-02 22:41:43 +01:00
gpu Merge branch 'akpm' (patches from Andrew) 2017-12-14 16:35:20 -08:00
hid treewide: setup_timer() -> timer_setup() 2017-11-21 15:57:07 -08:00
hsi HSI changes for the v4.15 series 2017-11-15 13:35:43 -08:00
hv Drivers: hv: vmbus: Fix a rescind issue 2017-11-28 16:56:26 +01:00
hwmon hwmon: (jc42) optionally try to disable the SMBUS timeout 2017-11-30 13:12:44 -08:00
hwspinlock hwspinlock update for v4.15 2017-11-17 20:16:20 -08:00
hwtracing tracing: Pass export pointer as argument to ->write() 2017-12-04 07:14:30 -05:00
i2c i2c: piix4: Fix port number check on release 2017-12-12 23:27:04 +01:00
ide Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/ide 2017-11-19 08:04:41 -10:00
idle Merge branch 'pm-cpuidle' 2017-11-13 01:34:14 +01:00
iio iio: health: max30102: Temperature should be in milli Celsius 2017-12-02 11:15:14 +00:00
infiniband Second pull request for 4.15-rc 2017-12-16 13:43:08 -08:00
input treewide: Remove TIMER_FUNC_TYPE and TIMER_DATA_TYPE casts 2017-11-21 16:35:54 -08:00
iommu IOMMU fixes for Linux v4.15-rc3 2017-12-06 10:53:02 -08:00
ipack
irqchip Merge branch 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2017-11-26 14:39:20 -08:00
isdn treewide: setup_timer() -> timer_setup() (2 field) 2017-11-21 15:57:09 -08:00
leds LED updates for 4.15rc1 2017-11-14 18:09:31 -08:00
lightnvm lightnvm: pblk: refactor pblk_ppa_comp function 2018-01-05 08:50:12 -07:00
macintosh Merge branch 'timers-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2017-11-13 17:56:58 -08:00
mailbox Change to POLL api and fixes for FlexRM and OMAP driver 2017-11-15 13:39:18 -08:00
mcb License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
md blk-mq: improve DM's blk-mq IO merging via blk_insert_cloned_request feedback 2018-01-17 09:46:54 -07:00
media media fixes for v4.15-rc3 2017-12-08 13:18:47 -08:00
memory ARM: SoC driver updates for v4.15 2017-11-16 16:05:01 -08:00
memstick treewide: setup_timer() -> timer_setup() 2017-11-21 15:57:07 -08:00
message Modules updates for v4.15 2017-11-15 13:46:33 -08:00
mfd treewide: setup_timer() -> timer_setup() 2017-11-21 15:57:07 -08:00
misc Merge branch 'WIP.x86-pti.base.prep-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2017-12-17 13:54:31 -08:00
mmc mmc: core: apply NO_CMD23 quirk to some specific cards 2017-12-11 13:43:27 +01:00
mtd Rename superblock flags (MS_xyz -> SB_xyz) 2017-11-27 13:05:09 -08:00
mux
net net: qcom/emac: Reduce timeout for mdio read/write 2017-12-15 15:46:19 -05:00
nfc treewide: setup_timer() -> timer_setup() (2 field) 2017-11-21 15:57:09 -08:00
ntb treewide: setup_timer() -> timer_setup() 2017-11-21 15:57:07 -08:00
nubus m68k updates for 4.15 2017-11-13 12:10:24 -08:00
nvdimm libnvdimm for 4.15 2017-11-17 09:51:57 -08:00
nvme nvme/multipath: Use blk_path_error 2018-01-10 10:52:18 -07:00
nvmem Char/Misc patches for 4.15-rc1 2017-11-16 09:10:59 -08:00
of of_mdio / mdiobus: ensure mdio devices have fwnode correctly populated 2017-12-13 15:01:47 -05:00
opp
oprofile License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
parisc License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
parport Char/Misc patches for 4.15-rc1 2017-11-16 09:10:59 -08:00
pci Power management fix for v4.15-rc4 2017-12-14 18:25:03 -08:00
pcmcia drivers/pcmcia/sa1111_badge4.c: avoid unused function warning 2017-11-17 16:10:04 -08:00
perf arm64 updates for 4.15 2017-11-15 10:56:56 -08:00
phy USB/PHY patches for 4.15-rc1 2017-11-13 21:14:07 -08:00
pinctrl pinctrl: sunxi: Disable strict mode for H5 driver 2017-11-30 16:50:43 +01:00
platform platform/x86: dell-wmi: check for kmalloc() errors 2017-12-11 17:26:03 -08:00
pnp License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
power power supply and reset changes for the v4.15 series 2017-11-15 13:37:15 -08:00
powercap
pps treewide: setup_timer() -> timer_setup() 2017-11-21 15:57:07 -08:00
ps3
ptp xen: features and fixes for v4.15-rc1 2017-11-16 13:06:27 -08:00
pwm pwm: Changes for v4.15-rc1 2017-11-22 21:09:18 -10:00
rapidio Merge branch 'akpm' (patches from Andrew) 2017-11-17 16:56:17 -08:00
ras Merge branch 'timers-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2017-11-13 17:56:58 -08:00
regulator - New Drivers 2017-11-16 09:15:57 -08:00
remoteproc remoteproc updates for v4.15 2017-11-17 20:14:10 -08:00
reset ARM: SoC driver updates for v4.15 2017-11-16 16:05:01 -08:00
rpmsg rpmsg updates for v4.15 2017-11-17 20:12:08 -08:00
rtc Merge branch 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2017-11-25 08:37:16 -10:00
s390 s390/qeth: update takeover IPs after configuration change 2017-12-15 11:29:43 -05:00
sbus Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc 2017-11-17 20:21:44 -08:00
scsi SCSI fixes on 20171215 2017-12-15 12:51:42 -08:00
sfi
sh A couple of dma-mapping updates: 2017-11-14 16:54:12 -08:00
sn
soc meson-gx-socinfo: Fix package id parsing 2017-11-30 15:29:44 -08:00
spi Merge remote-tracking branches 'spi/topic/sh-msiof', 'spi/topic/slave', 'spi/topic/spreadtrum' and 'spi/topic/tegra114' into spi-next 2017-11-10 21:33:51 +00:00
spmi
ssb License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
staging Merge tag 'staging-4.15-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging 2017-12-15 12:59:48 -08:00
target target: Use sgl_alloc_order() and sgl_free() 2018-01-06 09:18:00 -07:00
tc
tee optee: fix invalid of_node_put() in optee_driver_init() 2017-11-29 10:24:57 +01:00
thermal Merge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/rzhang/linux 2017-11-17 14:31:27 -08:00
thunderbolt Char/Misc patches for 4.15-rc1 2017-11-16 09:10:59 -08:00
tty serdev: ttyport: fix tty locking in close 2017-11-28 16:00:50 +01:00
uio License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
usb USB: core: prevent malicious bNumInterfaces overflow 2017-12-13 12:28:43 +01:00
uwb treewide: setup_timer() -> timer_setup() 2017-11-21 15:57:07 -08:00
vfio VFIO Updates for Linux v4.15 2017-11-14 16:47:47 -08:00
vhost vhost: fix skb leak in handle_rx() 2017-12-02 21:31:03 -05:00
video fbdev changes for v4.15: 2017-11-20 21:50:24 -10:00
virt
virtio virtio_mmio: fix devm cleanup 2017-12-14 21:01:40 +02:00
vlynq
vme Char/Misc patches for 4.15-rc1 2017-11-16 09:10:59 -08:00
w1 Char/Misc patches for 4.15-rc1 2017-11-16 09:10:59 -08:00
watchdog treewide: setup_timer() -> timer_setup() 2017-11-21 15:57:07 -08:00
xen xen: fixes for 4.15-rc4 2017-12-15 11:32:09 -08:00
zorro License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
Kconfig Merge branches 'pm-cpufreq-sched' and 'pm-opp' 2017-11-13 01:40:52 +01:00
Makefile usb: build drivers/usb/common/ when USB_SUPPORT is set 2017-11-28 15:17:49 +01:00