linux_dsm_epyc7002/drivers
Omer Peleg 9257b4a206 iommu/iova: introduce per-cpu caching to iova allocation
IOVA allocation has two problems that impede high-throughput I/O.
First, it can do a linear search over the allocated IOVA ranges.
Second, the rbtree spinlock that serializes IOVA allocations becomes
contended.

Address these problems by creating an API for caching allocated IOVA
ranges, so that the IOVA allocator isn't accessed frequently.  This
patch adds a per-CPU cache, from which CPUs can alloc/free IOVAs
without taking the rbtree spinlock.  The per-CPU caches are backed by
a global cache, to avoid invoking the (linear-time) IOVA allocator
without needing to make the per-CPU cache size excessive.  This design
is based on magazines, as described in "Magazines and Vmem: Extending
the Slab Allocator to Many CPUs and Arbitrary Resources" (currently
available at https://www.usenix.org/legacy/event/usenix01/bonwick.html)

Adding caching on top of the existing rbtree allocator maintains the
property that IOVAs are densely packed in the IO virtual address space,
which is important for keeping IOMMU page table usage low.

To keep the cache size reasonable, we bound the IOVA space a CPU can
cache by 32 MiB (we cache a bounded number of IOVA ranges, and only
ranges of size <= 128 KiB).  The shared global cache is bounded at
4 MiB of IOVA space.

Signed-off-by: Omer Peleg <omer@cs.technion.ac.il>
[mad@cs.technion.ac.il: rebased, cleaned up and reworded the commit message]
Signed-off-by: Adam Morrison <mad@cs.technion.ac.il>
Reviewed-by: Shaohua Li <shli@fb.com>
Reviewed-by: Ben Serebrin <serebrin@google.com>
[dwmw2: split out VT-d part into a separate patch]
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2016-04-20 15:42:24 -04:00
..
accessibility
acpi Merge branch 'acpi-processor' 2016-04-02 01:17:36 +02:00
amba
android
ata Merge branch 'for-4.6' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/libata 2016-03-18 20:06:46 -07:00
atm Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial 2016-03-17 21:38:27 -07:00
auxdisplay
base PM / wakeirq: fix wakeirq setting after wakup re-configuration from sysfs 2016-04-07 22:23:47 +02:00
bcma
block Merge branch 'for-linus' of git://git.kernel.dk/linux-block 2016-04-15 15:44:10 -07:00
bluetooth
bus mvebu fixes for 4.6 (part 1) 2016-04-12 12:35:07 -07:00
cdrom
char hwrng: bcm63xx - fix device tree compilation 2016-04-05 20:23:11 +08:00
clk clk: qcom: ipq4019: add some fixed clocks for ddrppl and fepll 2016-03-29 16:31:16 -07:00
clocksource Merge branch 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2016-03-24 10:32:42 -07:00
connector
cpufreq Merge branches 'pm-cpufreq', 'pm-cpuidle' and 'acpi-cppc' 2016-04-08 21:46:05 +02:00
cpuidle cpuidle: menu: Fall back to polling if next timer event is near 2016-03-21 15:50:28 +01:00
crypto Merge branch 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6 2016-03-23 06:12:39 -07:00
dca
devfreq PM / devfreq: Spelling s/frequnecy/frequency/ 2016-03-17 02:30:16 +01:00
dio
dma Merge branch 'fix/edma' into fixes 2016-04-16 22:52:03 +05:30
dma-buf dma-buf: Update docs for SYNC ioctl 2016-03-21 09:26:45 +01:00
edac EDAC queue for 4.6 2016-03-16 08:36:55 -07:00
eisa
extcon extcon: palmas: Drop stray IRQF_EARLY_RESUME flag 2016-04-04 08:32:45 +09:00
firewire IEEE 1394 subsystem patch: 2016-03-25 08:52:25 -07:00
firmware Merge branch 'efi-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2016-04-16 15:37:05 -07:00
fmc
fpga
gpio gpio: pca953x: Use correct u16 value for register word write 2016-04-08 11:49:47 +02:00
gpu Merge branch 'exynos-drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/daeinki/drm-exynos into drm-fixes 2016-04-14 13:06:19 +10:00
hid Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid 2016-04-11 12:31:09 -07:00
hsi
hv Char/Misc patches for 4.6-rc1 2016-03-17 13:47:50 -07:00
hwmon hwmon: (max1111) Return -ENODEV from max1111_read_channel if not instantiated 2016-03-27 10:37:48 -07:00
hwspinlock
hwtracing
i2c i2c: jz4780: really prevent potential division by zero 2016-04-09 08:36:44 +02:00
ide ide: palm_bk3710: test clock rate to avoid division by 0 2016-03-20 16:59:27 -04:00
idle intel_idle: Add KBL support 2016-04-07 22:11:08 +02:00
iio Second set of IIO fixes for the 4.6 cycle. 2016-04-04 13:45:10 -07:00
infiniband Revert "ib_srpt: Convert to percpu_ida tag allocation" 2016-04-07 18:16:20 -07:00
input Merge branch 'akpm' (patches from Andrew) 2016-03-25 16:59:11 -07:00
iommu iommu/iova: introduce per-cpu caching to iova allocation 2016-04-20 15:42:24 -04:00
ipack
irqchip irqchip/mbigen: Make CONFIG_HISILICON_IRQ_MBIGEN a hidden option 2016-03-23 12:02:29 +01:00
isdn Drivers: isdn: hisax: isac.c: Fix assignment and check into one expression. 2016-03-27 22:38:12 -04:00
leds platform-drivers-x86 for 4.6-1 2016-03-23 17:20:59 -07:00
lguest lguest, x86/entry/32: Fix handling of guest syscalls using interrupt gates 2016-04-01 08:58:13 +02:00
lightnvm lightnvm: do not load L2P table if not supported 2016-03-18 18:10:38 -07:00
macintosh
mailbox Merge branch 'mailbox-devel' of git://git.linaro.org/landing-teams/working/fujitsu/integration 2016-04-14 18:40:47 -07:00
mcb
md dm cache metadata: fix cmd_read_lock() acquiring write lock 2016-04-17 11:24:46 -04:00
media media fixes for v4.6-rc2 2016-04-05 06:47:50 -07:00
memory MTD updates for v4.6 2016-03-24 19:57:15 -07:00
memstick drivers/memstick/host/r592.c: avoid gcc-6 warning 2016-03-25 16:37:42 -07:00
message
mfd - New Drivers 2016-03-18 10:15:11 -07:00
misc lkdtm: do not leak free page on kmalloc failure 2016-04-06 16:22:25 -07:00
mmc mmc: tegra: Disable UHS-I modes for Tegra124 2016-04-14 11:40:01 +02:00
mtd One MTD fix for v4.6-rc4: 2016-04-15 15:25:09 -07:00
net tuntap: restore default qdisc 2016-04-08 15:52:45 -04:00
nfc Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next 2016-03-19 10:05:34 -07:00
ntb NTB: Remove _addr functions from ntb_hw_amd 2016-03-26 11:44:33 -04:00
nubus
nvdimm libnvdimm, pmem: clarify the write+clear_poison+write flow 2016-04-15 14:59:41 -06:00
nvme NVMe: Always use MSI/MSI-x interrupts 2016-04-14 14:04:50 -06:00
nvmem
of DeviceTree updates for 4.6: 2016-03-19 15:15:07 -07:00
oprofile mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros 2016-04-04 10:41:08 -07:00
parisc PCI changes for the v4.6 merge window: 2016-03-16 14:45:55 -07:00
parport
pci pci-sysfs: use proper file capability helper function 2016-04-14 12:56:09 -07:00
pcmcia pcmcia: db1xxx_ss: fix last irq_to_gpio user 2016-03-29 22:48:53 +02:00
perf drivers/perf: arm_pmu: avoid NULL dereference when not using devicetree 2016-03-21 11:36:17 +00:00
phy
pinctrl Revert "Revert "pinctrl: lantiq: Implement gpio_chip.to_irq"" 2016-04-01 15:21:27 +02:00
platform Convert straggling drivers to new six-argument get_user_pages() 2016-04-02 18:35:05 -05:00
pnp
power Power management and ACPI material for v4.6-rc1, part 2 2016-03-25 16:55:37 -07:00
powercap powercap: intel_rapl: Add missing Haswell model 2016-04-05 03:44:48 +02:00
pps
ps3
ptp Merge branch 'timers-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2016-03-15 12:13:56 -07:00
pwm pwm: fsl-ftm: Use flat regmap cache 2016-04-14 16:54:00 +02:00
rapidio Convert straggling drivers to new six-argument get_user_pages() 2016-04-02 18:35:05 -05:00
ras
regulator - New Drivers 2016-03-18 10:15:11 -07:00
remoteproc remoteproc: st: fix check of syscon_regmap_lookup_by_phandle() return value 2016-03-28 16:19:00 -07:00
reset
rpmsg
rtc RTC for 4.6 #2 2016-03-24 22:49:08 -07:00
s390 Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux 2016-04-01 07:15:54 -05:00
sbus
scsi SCSI fixes on 20160408 2016-04-09 12:00:42 -07:00
sfi
sh
sn
soc ARM: SoC driver updates for v4.6 2016-03-20 15:40:32 -07:00
spi Merge remote-tracking branches 'spi/fix/omap2' and 'spi/fix/rockchip' into spi-linus 2016-04-04 10:05:49 -07:00
spmi
ssb
staging Staging / IIO driver fixes for 4.6-rc3 2016-04-09 12:09:37 -07:00
target target: add a new add_wwn_groups fabrics method 2016-03-30 20:06:44 -07:00
tc
thermal Thermal: Ignore invalid trip points 2016-03-18 14:10:57 +08:00
thunderbolt
tty tty: Fix merge of "tty: Refactor tty_open()" 2016-03-31 20:49:39 -07:00
uio
usb usb: hcd: out of bounds access in for_each_companion 2016-04-13 12:06:18 -07:00
uwb
vfio VFIO updates for v4.6-rc1 2016-03-17 13:05:09 -07:00
vhost Merge branch 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending 2016-03-22 12:41:14 -07:00
video mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros 2016-04-04 10:41:08 -07:00
virt
virtio virtio: virtio 1.0 cs04 spec compliance for reset 2016-04-07 15:16:39 +03:00
vlynq
vme
w1
watchdog hpwdt: use nmi_panic() when kernel panics in NMI handler 2016-03-22 15:36:02 -07:00
xen xen/events: Mask a moving irq 2016-04-04 11:18:00 +01:00
zorro
Kconfig
Makefile