Shaohua Li
9dedf60313
md/raid1: read balance chooses idlest disk for SSD ...
SSD hasn't spindle, distance between requests means nothing. And the original
distance based algorithm sometimes can cause severe performance issue for SSD
raid.
Considering two thread groups, one accesses file A, the other access file B.
The first group will access one disk and the second will access the other disk,
because requests are near from one group and far between groups. In this case,
read balance might keep one disk very busy but the other relative idle. For
SSD, we should try best to distribute requests to as many disks as possible.
There isn't spindle move penality anyway.
With below patch, I can see more than 50% throughput improvement sometimes
depending on workloads.
The only exception is small requests can be merged to a big request which
typically can drive higher throughput for SSD too. Such small requests are
sequential reads. Unlike hard disk, sequential read which can't be merged (for
example direct IO, or read without readahead) can be ignored for SSD. Again
there is no spindle move penality. readahead dispatches small requests and such
requests can be merged.
Last patch can help detect sequential read well, at least if concurrent read
number isn't greater than raid disk number. In that case, distance based
algorithm doesn't work well too.
V2: For hard disk and SSD mixed raid, doesn't use distance based algorithm for
random IO too. This makes the algorithm generic for raid with SSD.
Signed-off-by: Shaohua Li <shli@fusionio.com>
Signed-off-by: NeilBrown <neilb@suse.de>
2012-07-31 10:03:53 +10:00
Merge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux
2012-07-14 12:44:26 -07:00
Viresh has moved
2012-06-20 14:39:36 -07:00
PM / Sleep: Prevent waiting forever on asynchronous suspend after abort
2012-06-24 23:31:09 +02:00
bcma: fix null pointer in bcma_core_pci_irq_ctl
2012-06-08 13:47:07 -04:00
blk: fix wrong idr_pre_get() error check in loop.c
2012-07-14 15:39:58 -07:00
Bluetooth: btmrvl: Do not send vendor events to bluetooth stack
2012-06-19 00:19:11 -03:00
Merge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
2012-06-18 12:20:36 -07:00
clk: fix parent validation in __clk_set_parent()
2012-07-03 12:05:14 -07:00
clocksource: sh_tmu: Use clockevents_config_and_register().
2012-06-11 17:10:16 +09:00
Merge branch 'fixes' of git://git.infradead.org/users/vkoul/slave-dma
2012-06-20 22:12:52 -07:00
edac: Do alignment logic properly in edac_align_ptr()
2012-06-11 12:43:16 -03:00
extcon: max8997: Add missing kfree for info->edev in max8997_muic_remove()
2012-06-18 16:30:42 -07:00
gpio/gpio-tps65910: gpio_chip.of_node referenced without CONFIG_OF_GPIO defined
2012-07-10 22:53:31 +02:00
drm/i915: kick any firmware framebuffers before claiming the gtt
2012-07-03 11:18:48 +01:00
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
2012-07-13 10:33:18 -07:00
hwmon: (it87) Preserve configuration register bits on init
2012-07-12 22:47:37 +02:00
hwspinlock/core: use global ID to register hwspinlocks on multiple devices
2012-07-07 22:35:30 +03:00
i2c: Add generic I2C multiplexer using pinctrl API
2012-06-04 16:49:43 +02:00
drivers/ide/ide-cs.c: adjust suspicious bit operation
2012-06-12 15:51:41 -07:00
iio: drop wrong reference from Kconfig
2012-06-14 17:28:46 -07:00
Merge branches 'cma' and 'ocrdma' into for-linus
2012-06-24 04:59:59 -07:00
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
2012-07-13 10:33:18 -07:00
iommu/amd: fix type bug in flush code
2012-07-02 12:11:40 +02:00
leds: heartbeat: fix bug on panic
2012-07-04 15:55:19 +08:00
md/raid1: read balance chooses idlest disk for SSD
2012-07-31 10:03:53 +10:00
Merge branch 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media
2012-07-13 10:29:41 -07:00
Merge branch 'akpm' (Andrew's patch-bomb)
2012-05-31 18:10:18 -07:00
mfd: Add missing hunk to change palmas irq to clear on read
2012-07-09 00:16:26 +02:00
Merge branch 'akpm' (Andrew's patch-bomb)
2012-07-11 16:06:54 -07:00
mmc: cd-gpio: pass IRQF_ONESHOT to request_threaded_irq()
2012-07-10 23:03:57 -04:00
Late MTD fixes for 3.5:
2012-07-13 09:56:26 -07:00
e1000e: remove use of IP payload checksum
2012-07-01 00:25:32 -07:00
ARM: SoC fixes for 3.5-rc
2012-07-11 12:44:25 -07:00
oprofile: perf: use NR_CPUS instead or nr_cpumask_bits for static array
2012-06-21 16:15:11 +02:00
PCI: EHCI: fix crash during suspend on ASUS computers
2012-07-10 09:52:05 -07:00
Merge branch 'akpm' (Andrew's patch-bomb)
2012-06-20 14:41:57 -07:00
drivers/platform/x86/acerhdf.c: correct Boris' mail address
2012-06-07 14:43:55 -07:00
A bunch of fixes for v3.5, nothing extraordinary.
2012-05-31 12:10:15 -07:00
rapidio/tsi721: add DMA engine support
2012-05-31 17:49:31 -07:00
regulator: Fix recursive mutex lockdep warning
2012-07-03 20:25:58 +01:00
remoteproc: fix missing CONFIG_FW_LOADER configurations
2012-07-04 11:01:12 +03:00
rpmsg: make sure inflight messages don't invoke just-removed callbacks
2012-07-04 11:51:59 +03:00
drivers/rtc/rtc-twl.c: fix threaded IRQ to use IRQF_ONESHOT
2012-07-11 16:04:48 -07:00
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux
2012-05-31 10:51:10 -07:00
[SCSI] bnx2i: Removed the reference to the netdev->base_addr
2012-07-08 09:52:03 +01:00
SPI: fix over-eager devm_xxx() conversion
2012-06-18 11:27:04 +01:00
Merge branch 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media
2012-06-25 14:53:09 -07:00
tcm_fc: Resolve suspicious RCU usage warnings
2012-07-06 12:52:09 -07:00
tty/hvc_opal: Fix debug function name
2012-07-10 19:16:25 +10:00
MFD Fixes for 3.5
2012-07-13 09:54:26 -07:00
vhost: use USER_DS in vhost_worker thread
2012-06-26 21:10:56 -07:00
fbdev fixes for 3.5
2012-07-11 16:17:14 -07:00
virtio-balloon: fix add/get API use
2012-07-09 09:07:22 +09:30
watchdog: core: fix WDIOC_GETSTATUS return value
2012-06-28 20:40:56 +02:00
Five bug-fixes:
2012-06-15 17:17:15 -07:00
9dedf60313