kernel/linux
0
Fork 0
mirror of https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ synced 2026-10-24 19:10:26 +00:00
Code Issues Releases Packages Activity
linux/drivers
History
Davidlohr Bueso b980077899 mm: introduce per-node proactive reclaim interface 
This adds support for allowing proactive reclaim in general on a NUMA
system.  A per-node interface extends support for beyond a memcg-specific
interface, respecting the current semantics of memory.reclaim: respecting
aging LRU and not supporting artificially triggering eviction on nodes
belonging to non-bottom tiers.

This patch allows userspace to do:

     echo "512M swappiness=10" > /sys/devices/system/node/nodeX/reclaim

One of the premises for this is to semantically align as best as possible
with memory.reclaim.  During a brief time memcg did support nodemask until
55ab834a86 (Revert "mm: add nodes= arg to memory.reclaim"), for which
semantics around reclaim (eviction) vs demotion were not clear, rendering
charging expectations to be broken.

With this approach:

1. Users who do not use memcg can benefit from proactive reclaim.  The
   memcg interface is not NUMA aware and there are usecases that are
   focusing on NUMA balancing rather than workload memory footprint.

2. Proactive reclaim on top tiers will trigger demotion, for which
   memory is still byte-addressable.  Reclaiming on the bottom nodes will
   trigger evicting to swap (the traditional sense of reclaim).  This
   follows the semantics of what is today part of the aging process on
   tiered memory, mirroring what every other form of reclaim does
   (reactive and memcg proactive reclaim).  Furthermore per-node proactive
   reclaim is not as susceptible to the memcg charging problem mentioned
   above.

3. Unlike the nodes= arg, this interface avoids confusing semantics,
   such as what exactly the user wants when mixing top-tier and low-tier
   nodes in the nodemask.  Further per-node interface is less exposed to
   "free up memory in my container" usecases, where eviction is intended.

4. Users that *really* want to free up memory can use proactive
   reclaim on nodes knowingly to be on the bottom tiers to force eviction
   in a natural way - higher access latencies are still better than swap. 
   If compelled, while no guarantees and perhaps not worth the effort,
   users could also also potentially follow a ladder-like approach to
   eventually free up the memory.  Alternatively, perhaps an 'evict'
   option could be added to the parameters for both memory.reclaim and
   per-node interfaces to force this action unconditionally.

[akpm@linux-foundation.org: user_proactive_reclaim(): return -EBUSY on PGDAT_RECLAIM_LOCKED contention, per Roman]
[dave@stgolabs.net: memcg && node is also a bogus case, per Shakeel]
  Link: https://lkml.kernel.org/r/20250717235604.2atyx2aobwowpge3@offworld
Link: https://lkml.kernel.org/r/20250623185851.830632-5-dave@stgolabs.net
Signed-off-by: Davidlohr Bueso <dave@stgolabs.net>
Acked-by: Shakeel Butt <shakeel.butt@linux.dev>
Acked-by: Roman Gushchin <roman.gushchin@linux.dev>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Yosry Ahmed <yosryahmed@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
History 2025-07-19 18:59:53 -07:00
..
accel accel/amdxdna: Fix incorrect PSP firmware size 2025-06-09 07:16:32 -07:00
accessibility treewide: Switch/rename to timer_delete[_sync]() 2025-04-05 10:30:12 +02:00
acpi drivers,hmat: use node-notifier instead of memory-notifier 2025-07-13 16:38:16 -07:00
amba iommu: Get DT/ACPI parsing into the proper probe path 2025-03-11 14:05:43 +01:00
android Char/Misc/IIO pull request for 6.16-rc1 2025-06-06 11:50:47 -07:00
ata ata: ahci: Use correct DMI identifier for ASUSPRO-D840SA LPM quirk 2025-06-25 15:17:57 +02:00
atm atm: idt77252: Add missing dma_map_error() 2025-06-25 15:28:57 -07:00
auxdisplay treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
base mm: introduce per-node proactive reclaim interface 2025-07-19 18:59:53 -07:00
bcma bcma: use new GPIO line value setter callbacks 2025-04-23 15:28:24 +02:00
block null_blk: use memzero_page() 2025-07-09 22:42:08 -07:00
bluetooth driver: bluetooth: hci_qca:fix unable to load the BT driver 2025-06-20 11:55:03 -04:00
bus treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
cache cache: sifive_ccache: Add ESWIN EIC7700 support 2025-04-07 16:53:46 +01:00
cdrom cdrom: Remove unnecessary NULL check before unregister_sysctl_table() 2025-05-15 16:25:20 -06:00
cdx irqdomain: cdx: Switch to of_fwnode_handle() 2025-04-25 15:53:38 +02:00
char treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
clk I've recently moved computers (among other things) so I'm sending this from a 2025-05-30 09:15:40 -07:00
clocksource MFD for v6.16 2025-06-03 11:53:55 -07:00
comedi treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
connector connector/cn_proc: revert "connector: Fix proc_event_num_listeners count not cleared" 2024-02-13 11:15:44 +01:00
counter Second set of Counter updates for 6.16 2025-05-24 08:29:32 +02:00
cpufreq rust: Use CpuId in place of raw CPU numbers 2025-06-12 10:31:28 +05:30
cpuidle Merge branch 'pm-cpuidle' 2025-05-30 20:21:36 +02:00
crypto treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
cxl drivers,cxl: use node-notifier instead of memory-notifier 2025-07-13 16:38:15 -07:00
dax mm: remove callers of pfn_t functionality 2025-07-09 22:42:19 -07:00
dca Driver core changes for 6.11-rc1 2024-07-25 10:42:22 -07:00
devfreq PM / devfreq: rockchip-dfi: Switch to use hrtimer_setup() 2025-02-18 10:32:34 +01:00
dio dio: Have dio_bus_match() callback take a const * 2024-07-10 15:38:14 +02:00
dma treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
dma-buf dma-buf: fix timeout handling in dma_resv_wait_timeout v2 2025-06-30 13:15:44 +02:00
dpll Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2025-03-20 21:38:01 +01:00
edac EDAC: Initialize EDAC features sysfs attributes 2025-06-30 10:57:24 +02:00
eisa EISA: Move devlist.h out of obj to always 2025-05-07 07:01:01 +02:00
extcon Update extcon next for v6.14 2025-01-12 13:44:27 +01:00
firewire treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
firmware Samsung SoC fixes for v6.16 2025-07-03 16:23:53 +02:00
fpga FPGA Manager changes for 6.16-rc1 2025-05-21 14:08:44 +02:00
fsi fsi: core: Use const 'struct bin_attribute' callbacks 2025-02-21 09:20:31 +01:00
fwctl fwctl: Fix repeated device word in log message 2025-04-11 20:47:45 -03:00
gnss [tree-wide] finally take no_llseek out 2024-09-27 08:18:43 -07:00
gpio gpio: mlxbf3: only get IRQ for device instance 0 2025-06-18 12:19:39 +02:00
gpu mm: remove callers of pfn_t functionality 2025-07-09 22:42:19 -07:00
greybus treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
hid hid-for-linus-2025070502 2025-07-05 16:14:03 -07:00
hsi treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
hte treewide: Switch/rename to timer_delete[_sync]() 2025-04-05 10:30:12 +02:00
hv hyperv-next for v6.16 2025-06-03 08:39:20 -07:00
hwmon hwmon: (ltc4282) avoid repeated register write 2025-06-16 06:30:58 -07:00
hwspinlock hwspinlock: Remove unused hwspin_lock_get_id() 2025-03-21 17:12:04 -05:00
hwtracing mm: remove callers of pfn_t functionality 2025-07-09 22:42:19 -07:00
i2c i2c-for-6.16-rc5 2025-07-05 12:54:24 -07:00
i3c i3c: controllers do not need to depend on I3C 2025-05-24 22:49:07 +02:00
idle intel_idle: Update arguments of mwait_idle_with_hints() 2025-06-10 21:09:28 +02:00
iio treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
infiniband SCSI fixes on 20250703 2025-07-03 11:52:39 -07:00
input Input updates for v6.16-rc4 2025-07-04 09:54:15 -07:00
interconnect Merge branch 'icc-sa8775p' into icc-next 2025-05-19 17:09:50 +03:00
iommu iommu/vt-d: Assign devtlb cache tag on ATS enablement 2025-07-04 10:33:56 +02:00
ipack driver core: have match() callback in struct bus_type take a const * 2024-07-03 15:16:54 +02:00
irqchip irqchip/irq-msi-lib: Select CONFIG_GENERIC_MSI_IRQ 2025-06-30 16:59:12 +02:00
isdn treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
leds treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
macintosh treewide: Switch/rename to timer_delete[_sync]() 2025-04-05 10:30:12 +02:00
mailbox treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
mcb mcb: fix a double free bug in chameleon_parse_gdd() 2025-04-15 18:21:39 +02:00
md mm: remove callers of pfn_t functionality 2025-07-09 22:42:19 -07:00
media treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
memory treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
memstick treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
message SCSI misc on 20250326 2025-03-26 19:57:34 -07:00
mfd mfd: Fix building without CONFIG_OF 2025-06-19 11:05:30 +01:00
misc mm/balloon_compaction: convert balloon_page_delete() to balloon_page_finalize() 2025-07-13 16:38:25 -07:00
mmc mtk-sd: reset host->mrq on prepare_data() error 2025-06-25 14:42:51 +02:00
most treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
mtd mtd: nand: qpic_common: prevent out of bounds access of BAM arrays 2025-06-29 22:10:47 +01:00
mux mux: adgs1408: fix Wvoid-pointer-to-enum-cast warning 2025-05-08 17:12:08 +02:00
net net: ngbe: specify IRQ vector when the number of VFs is 7 2025-07-03 11:51:40 +02:00
nfc treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
ntb NTB/msi: Switch MSI descriptor locking to lock guard() 2025-04-09 20:47:29 +02:00
nubus nubus: Make nubus_bus_type static and constant 2024-01-03 13:33:59 +01:00
nvdimm mm: remove callers of pfn_t functionality 2025-07-09 22:42:19 -07:00
nvme block-6.16-20250704 2025-07-04 09:33:59 -07:00
nvmem Char/Misc/IIO pull request for 6.16-rc1 2025-06-06 11:50:47 -07:00
of - The 11 patch series "Add folio_mk_pte()" from Matthew Wilcox 2025-05-31 15:44:16 -07:00
opp OPP: switch to use kmemdup_array() 2025-05-19 15:37:53 +05:30
parisc parisc: led: Use scnprintf() to avoid string truncation warning 2025-03-09 22:27:54 +01:00
parport treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
pci pci-v6.16-fixes-2 2025-06-27 20:17:48 -07:00
pcmcia treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
peci module: Convert symbol namespace to string literal 2024-12-02 11:34:44 -08:00
perf arm64 updates for 6.16 2025-05-28 14:55:35 -07:00
phy phy-for-6.16 2025-06-05 08:20:21 -07:00
pinctrl pinctrl: sunxi: dt: Consider pin base when calculating bank number from pin 2025-06-10 14:35:40 +02:00
platform platform-drivers-x86 for v6.16-3 2025-07-04 10:05:31 -07:00
pmdomain pmdomain: ti: Fix STANDBY handling of PER power domain 2025-05-19 16:11:05 +02:00
pnp x86/platform/amd: Move the <asm/amd_nb.h> header to <asm/amd/nb.h> 2025-04-14 09:34:14 +02:00
power - The 3 patch series "hung_task: extend blocking task stacktrace dump to 2025-05-31 19:12:53 -07:00
powercap powercap: intel_rapl: Do not change CLAMPING bit if ENABLE bit cannot be changed 2025-06-30 20:32:29 +02:00
pps treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
ps3 ps3: Correct some typos in comments 2024-11-15 00:32:29 +11:00
ptp ptp: allow reading of currently dialed frequency to succeed on free-running clocks 2025-06-17 16:13:09 -07:00
pwm pwm: axi-pwmgen: Fix handling of external clock 2025-06-06 13:16:50 -07:00
rapidio drivers/rapidio/rio_cm.c: prevent possible heap overwrite 2025-06-11 22:42:36 -07:00
ras Linux 6.15-rc5 2025-05-06 10:00:58 +02:00
regulator regulator: gpio: Fix the out-of-bounds access to drvdata::gpiods 2025-07-03 12:22:35 +01:00
remoteproc remoteproc updates for v6.16 2025-06-02 11:04:29 -07:00
reset reset: Add USB2PHY port reset driver for Renesas RZ/V2H(P) 2025-05-05 15:18:35 +02:00
rpmsg rpmsg: qcom_smd: Fix uninitialized return variable in __qcom_smd_send() 2025-05-20 21:46:10 -05:00
rtc rtc: pcf2127: add missing semicolon after statement 2025-06-24 16:06:14 +02:00
s390 mm: remove callers of pfn_t functionality 2025-07-09 22:42:19 -07:00
sbus Get rid of 'remove_new' relic from platform driver struct 2024-12-01 15:12:43 -08:00
scsi scsi: core: Enforce unlimited max_segment_size when virt_boundary_mask is set 2025-06-24 21:20:58 -04:00
sh sh: Switch to irq_domain_create_*() 2025-05-16 21:06:11 +02:00
siox Driver core changes for 6.11-rc1 2024-07-25 10:42:22 -07:00
slimbus slimbus: messaging: Free transaction ID in delayed interrupt scenario 2025-02-20 15:19:51 +01:00
soc soc: drivers for 6.16 2025-05-31 07:53:30 -07:00
soundwire soundwire updates for 6.16 2025-06-05 08:07:24 -07:00
spi spi: cadence-quadspi: fix cleanup of rx_chan on failure paths 2025-07-01 14:02:26 +01:00
spmi irqdomain: spmi: Switch to irq_domain_create_tree() 2025-05-21 14:53:17 +02:00
ssb irqdomain: ssb: Switch to irq_domain_create_linear() 2025-04-23 15:08:44 +02:00
staging staging: rtl8723bs: Avoid memset() in aes_cipher() and aes_decipher() 2025-06-19 17:33:43 +02:00
target scsi: target: Fix NULL pointer dereference in core_scsi3_decode_spec_i_port() 2025-06-16 14:35:57 -04:00
tc TC: Fix the wrong format specifier 2024-11-12 15:48:08 +01:00
tee A fix in the OP-TEE driver for v6.16 2025-07-03 16:26:08 +02:00
thermal Thermal control updates for 6.16-rc1 2025-05-27 16:28:02 -07:00
thunderbolt thunderbolt: Changes for v6.16 merge window 2025-05-21 12:26:51 +02:00
tty serial: imx: Restore original RXTL for console to fix data loss 2025-06-24 15:34:21 +01:00
ufs scsi: ufs: core: Fix spelling of a sysfs attribute name 2025-06-24 21:22:20 -04:00
uio Char/Misc/IIO pull request for 6.16-rc1 2025-06-06 11:50:47 -07:00
usb usb: hub: Fix flushing of delayed work used for post resume purposes 2025-06-30 15:36:00 +02:00
vdpa vdpa/octeon_ep: Control PCI dev enabling manually 2025-05-27 10:27:53 -04:00
vfio mm: remove callers of pfn_t functionality 2025-07-09 22:42:19 -07:00
vhost virtio, vhost: features, fixes 2025-05-29 08:15:35 -07:00
video treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
virt treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
virtio mm/balloon_compaction: convert balloon_page_delete() to balloon_page_finalize() 2025-07-13 16:38:25 -07:00
w1 Char/Misc/IIO pull request for 6.16-rc1 2025-06-06 11:50:47 -07:00
watchdog treewide, timers: Rename from_timer() to timer_container_of() 2025-06-08 09:07:37 +02:00
xen xen/x86: fix initial memory balloon target 2025-05-23 07:09:00 +02:00
zorro zorro: Constify 'struct bin_attribute' 2025-01-08 18:04:36 +01:00
Kconfig fwctl: Add basic structure for a class subsystem with a cdev 2025-03-06 15:11:30 -04:00
Makefile fwctl: Add basic structure for a class subsystem with a cdev 2025-03-06 15:11:30 -04:00