Re: [PATCH -V14 2/3] NUMA balancing: optimize page placement for memory tiering system

From: Yang Shi
Date: Tue Mar 01 2022 - 16:59:19 EST


On Tue, Mar 1, 2022 at 12:54 AM Huang Ying <ying.huang@xxxxxxxxx> wrote:
>
> With the advent of various new memory types, some machines will have
> multiple types of memory, e.g. DRAM and PMEM (persistent memory). The
> memory subsystem of these machines can be called memory tiering
> system, because the performance of the different types of memory are
> usually different.
>
> In such system, because of the memory accessing pattern changing etc,
> some pages in the slow memory may become hot globally. So in this
> patch, the NUMA balancing mechanism is enhanced to optimize the page
> placement among the different memory types according to hot/cold
> dynamically.
>
> In a typical memory tiering system, there are CPUs, fast memory and
> slow memory in each physical NUMA node. The CPUs and the fast memory
> will be put in one logical node (called fast memory node), while the
> slow memory will be put in another (faked) logical node (called slow
> memory node). That is, the fast memory is regarded as local while the
> slow memory is regarded as remote. So it's possible for the recently
> accessed pages in the slow memory node to be promoted to the fast
> memory node via the existing NUMA balancing mechanism.
>
> The original NUMA balancing mechanism will stop to migrate pages if
> the free memory of the target node becomes below the high watermark.
> This is a reasonable policy if there's only one memory type. But this
> makes the original NUMA balancing mechanism almost do not work to
> optimize page placement among different memory types. Details are as
> follows.
>
> It's the common cases that the working-set size of the workload is
> larger than the size of the fast memory nodes. Otherwise, it's
> unnecessary to use the slow memory at all. So, there are almost
> always no enough free pages in the fast memory nodes, so that the
> globally hot pages in the slow memory node cannot be promoted to the
> fast memory node. To solve the issue, we have 2 choices as follows,
>
> a. Ignore the free pages watermark checking when promoting hot pages
> from the slow memory node to the fast memory node. This will
> create some memory pressure in the fast memory node, thus trigger
> the memory reclaiming. So that, the cold pages in the fast memory
> node will be demoted to the slow memory node.
>
> b. Define a new watermark called wmark_promo which is higher than
> wmark_high, and have kswapd reclaiming pages until free pages reach
> such watermark. The scenario is as follows: when we want to promote
> hot-pages from a slow memory to a fast memory, but fast memory's free
> pages would go lower than high watermark with such promotion, we wake
> up kswapd with wmark_promo watermark in order to demote cold pages and
> free us up some space. So, next time we want to promote hot-pages we
> might have a chance of doing so.
>
> The choice "a" may create high memory pressure in the fast memory
> node. If the memory pressure of the workload is high, the memory
> pressure may become so high that the memory allocation latency of the
> workload is influenced, e.g. the direct reclaiming may be triggered.
>
> The choice "b" works much better at this aspect. If the memory
> pressure of the workload is high, the hot pages promotion will stop
> earlier because its allocation watermark is higher than that of the
> normal memory allocation. So in this patch, choice "b" is
> implemented. A new zone watermark (WMARK_PROMO) is added. Which is
> larger than the high watermark and can be controlled via
> watermark_scale_factor.
>
> In addition to the original page placement optimization among sockets,
> the NUMA balancing mechanism is extended to be used to optimize page
> placement according to hot/cold among different memory types. So the
> sysctl user space interface (numa_balancing) is extended in a backward
> compatible way as follow, so that the users can enable/disable these
> functionality individually.
>
> The sysctl is converted from a Boolean value to a bits field. The
> definition of the flags is,
>
> - 0: NUMA_BALANCING_DISABLED
> - 1: NUMA_BALANCING_NORMAL
> - 2: NUMA_BALANCING_MEMORY_TIERING
>
> We have tested the patch with the pmbench memory accessing benchmark
> with the 80:20 read/write ratio and the Gauss access address
> distribution on a 2 socket Intel server with Optane DC Persistent
> Memory Model. The test results shows that the pmbench score can
> improve up to 95.9%.
>
> Thanks Andrew Morton to help fix the document format error.
>
> Signed-off-by: "Huang, Ying" <ying.huang@xxxxxxxxx>
> Tested-by: Baolin Wang <baolin.wang@xxxxxxxxxxxxxxxxx>
> Reviewed-by: Baolin Wang <baolin.wang@xxxxxxxxxxxxxxxxx>
> Reviewed-by: Oscar Salvador <osalvador@xxxxxxx>
> Acked-by: Johannes Weiner <hannes@xxxxxxxxxxx>
> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> Cc: Michal Hocko <mhocko@xxxxxxxx>
> Cc: Rik van Riel <riel@xxxxxxxxxxx>
> Cc: Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx>
> Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> Cc: Dave Hansen <dave.hansen@xxxxxxxxxxxxxxx>
> Cc: Yang Shi <shy828301@xxxxxxxxx>
> Cc: Zi Yan <ziy@xxxxxxxxxx>
> Cc: Wei Xu <weixugc@xxxxxxxxxx>
> Cc: Shakeel Butt <shakeelb@xxxxxxxxxx>
> Cc: zhongjiang-ali <zhongjiang-ali@xxxxxxxxxxxxxxxxx>
> Cc: Randy Dunlap <rdunlap@xxxxxxxxxxxxx>
> Cc: linux-kernel@xxxxxxxxxxxxxxx
> Cc: linux-mm@xxxxxxxxx
> ---
> Documentation/admin-guide/sysctl/kernel.rst | 29 ++++++++++++++-------
> include/linux/mmzone.h | 1 +
> include/linux/sched/sysctl.h | 10 +++++++
> kernel/sched/core.c | 21 ++++++++++++---
> kernel/sysctl.c | 2 +-
> mm/migrate.c | 16 ++++++++++--
> mm/page_alloc.c | 3 ++-
> mm/vmscan.c | 6 ++++-
> 8 files changed, 70 insertions(+), 18 deletions(-)
>
> diff --git a/Documentation/admin-guide/sysctl/kernel.rst b/Documentation/admin-guide/sysctl/kernel.rst
> index d359bcfadd39..fdfd2b684822 100644
> --- a/Documentation/admin-guide/sysctl/kernel.rst
> +++ b/Documentation/admin-guide/sysctl/kernel.rst
> @@ -595,16 +595,23 @@ Documentation/admin-guide/kernel-parameters.rst).
> numa_balancing
> ==============
>
> -Enables/disables automatic page fault based NUMA memory
> -balancing. Memory is moved automatically to nodes
> -that access it often.
> +Enables/disables and configures automatic page fault based NUMA memory
> +balancing. Memory is moved automatically to nodes that access it often.
> +The value to set can be the result of ORing the following:
>
> -Enables/disables automatic NUMA memory balancing. On NUMA machines, there
> -is a performance penalty if remote memory is accessed by a CPU. When this
> -feature is enabled the kernel samples what task thread is accessing memory
> -by periodically unmapping pages and later trapping a page fault. At the
> -time of the page fault, it is determined if the data being accessed should
> -be migrated to a local memory node.
> += =================================
> +0 NUMA_BALANCING_DISABLED
> +1 NUMA_BALANCING_NORMAL
> +2 NUMA_BALANCING_MEMORY_TIERING
> += =================================
> +
> +Or NUMA_BALANCING_NORMAL to optimize page placement among different
> +NUMA nodes to reduce remote accessing. On NUMA machines, there is a
> +performance penalty if remote memory is accessed by a CPU. When this
> +feature is enabled the kernel samples what task thread is accessing
> +memory by periodically unmapping pages and later trapping a page
> +fault. At the time of the page fault, it is determined if the data
> +being accessed should be migrated to a local memory node.
>
> The unmapping of pages and trapping faults incur additional overhead that
> ideally is offset by improved memory locality but there is no universal
> @@ -615,6 +622,10 @@ faults may be controlled by the `numa_balancing_scan_period_min_ms,
> numa_balancing_scan_delay_ms, numa_balancing_scan_period_max_ms,
> numa_balancing_scan_size_mb`_, and numa_balancing_settle_count sysctls.
>
> +Or NUMA_BALANCING_MEMORY_TIERING to optimize page placement among
> +different types of memory (represented as different NUMA nodes) to
> +place the hot pages in the fast memory. This is implemented based on
> +unmapping and page fault too.
>
> numa_balancing_scan_period_min_ms, numa_balancing_scan_delay_ms, numa_balancing_scan_period_max_ms, numa_balancing_scan_size_mb
> ===============================================================================================================================
> diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
> index 44bd054ca12b..06bc55db19bf 100644
> --- a/include/linux/mmzone.h
> +++ b/include/linux/mmzone.h
> @@ -342,6 +342,7 @@ enum zone_watermarks {
> WMARK_MIN,
> WMARK_LOW,
> WMARK_HIGH,
> + WMARK_PROMO,

TBH I'm not a fan of another water mark since we already have quite a
few water marks (regular water mark, water mark boost, water mark
promo). But it is not a big deal and gated problem for now since it is
not user visible. We definitely could try to consolidate some of them
later.

The patch looks fine to me. Reviewed-by: Yang Shi <shy828301@xxxxxxxxx>

> NR_WMARK
> };
>
> diff --git a/include/linux/sched/sysctl.h b/include/linux/sched/sysctl.h
> index c19dd5a2c05c..b5eec8854c5a 100644
> --- a/include/linux/sched/sysctl.h
> +++ b/include/linux/sched/sysctl.h
> @@ -23,6 +23,16 @@ enum sched_tunable_scaling {
> SCHED_TUNABLESCALING_END,
> };
>
> +#define NUMA_BALANCING_DISABLED 0x0
> +#define NUMA_BALANCING_NORMAL 0x1
> +#define NUMA_BALANCING_MEMORY_TIERING 0x2
> +
> +#ifdef CONFIG_NUMA_BALANCING
> +extern int sysctl_numa_balancing_mode;
> +#else
> +#define sysctl_numa_balancing_mode 0
> +#endif
> +
> /*
> * control realtime throttling:
> *
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index fcf0c180617c..c25348e9ae3a 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -4280,7 +4280,9 @@ DEFINE_STATIC_KEY_FALSE(sched_numa_balancing);
>
> #ifdef CONFIG_NUMA_BALANCING
>
> -void set_numabalancing_state(bool enabled)
> +int sysctl_numa_balancing_mode;
> +
> +static void __set_numabalancing_state(bool enabled)
> {
> if (enabled)
> static_branch_enable(&sched_numa_balancing);
> @@ -4288,13 +4290,22 @@ void set_numabalancing_state(bool enabled)
> static_branch_disable(&sched_numa_balancing);
> }
>
> +void set_numabalancing_state(bool enabled)
> +{
> + if (enabled)
> + sysctl_numa_balancing_mode = NUMA_BALANCING_NORMAL;
> + else
> + sysctl_numa_balancing_mode = NUMA_BALANCING_DISABLED;
> + __set_numabalancing_state(enabled);
> +}
> +
> #ifdef CONFIG_PROC_SYSCTL
> int sysctl_numa_balancing(struct ctl_table *table, int write,
> void *buffer, size_t *lenp, loff_t *ppos)
> {
> struct ctl_table t;
> int err;
> - int state = static_branch_likely(&sched_numa_balancing);
> + int state = sysctl_numa_balancing_mode;
>
> if (write && !capable(CAP_SYS_ADMIN))
> return -EPERM;
> @@ -4304,8 +4315,10 @@ int sysctl_numa_balancing(struct ctl_table *table, int write,
> err = proc_dointvec_minmax(&t, write, buffer, lenp, ppos);
> if (err < 0)
> return err;
> - if (write)
> - set_numabalancing_state(state);
> + if (write) {
> + sysctl_numa_balancing_mode = state;
> + __set_numabalancing_state(state);
> + }
> return err;
> }
> #endif
> diff --git a/kernel/sysctl.c b/kernel/sysctl.c
> index 5ae443b2882e..c90a564af720 100644
> --- a/kernel/sysctl.c
> +++ b/kernel/sysctl.c
> @@ -1689,7 +1689,7 @@ static struct ctl_table kern_table[] = {
> .mode = 0644,
> .proc_handler = sysctl_numa_balancing,
> .extra1 = SYSCTL_ZERO,
> - .extra2 = SYSCTL_ONE,
> + .extra2 = SYSCTL_FOUR,
> },
> #endif /* CONFIG_NUMA_BALANCING */
> {
> diff --git a/mm/migrate.c b/mm/migrate.c
> index cdeaf01e601a..08ca9b9b142e 100644
> --- a/mm/migrate.c
> +++ b/mm/migrate.c
> @@ -51,6 +51,7 @@
> #include <linux/oom.h>
> #include <linux/memory.h>
> #include <linux/random.h>
> +#include <linux/sched/sysctl.h>
>
> #include <asm/tlbflush.h>
>
> @@ -2034,16 +2035,27 @@ static int numamigrate_isolate_page(pg_data_t *pgdat, struct page *page)
> {
> int page_lru;
> int nr_pages = thp_nr_pages(page);
> + int order = compound_order(page);
>
> - VM_BUG_ON_PAGE(compound_order(page) && !PageTransHuge(page), page);
> + VM_BUG_ON_PAGE(order && !PageTransHuge(page), page);
>
> /* Do not migrate THP mapped by multiple processes */
> if (PageTransHuge(page) && total_mapcount(page) > 1)
> return 0;
>
> /* Avoid migrating to a node that is nearly full */
> - if (!migrate_balanced_pgdat(pgdat, nr_pages))
> + if (!migrate_balanced_pgdat(pgdat, nr_pages)) {
> + int z;
> +
> + if (!(sysctl_numa_balancing_mode & NUMA_BALANCING_MEMORY_TIERING))
> + return 0;
> + for (z = pgdat->nr_zones - 1; z >= 0; z--) {
> + if (populated_zone(pgdat->node_zones + z))
> + break;
> + }
> + wakeup_kswapd(pgdat->node_zones + z, 0, order, ZONE_MOVABLE);
> return 0;
> + }
>
> if (isolate_lru_page(page))
> return 0;
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 3589febc6d31..295b8f1fc31d 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -8474,7 +8474,8 @@ static void __setup_per_zone_wmarks(void)
>
> zone->watermark_boost = 0;
> zone->_watermark[WMARK_LOW] = min_wmark_pages(zone) + tmp;
> - zone->_watermark[WMARK_HIGH] = min_wmark_pages(zone) + tmp * 2;
> + zone->_watermark[WMARK_HIGH] = low_wmark_pages(zone) + tmp;
> + zone->_watermark[WMARK_PROMO] = high_wmark_pages(zone) + tmp;
>
> spin_unlock_irqrestore(&zone->lock, flags);
> }
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index 6dd8f455bb82..199b8aadbdd6 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -56,6 +56,7 @@
>
> #include <linux/swapops.h>
> #include <linux/balloon_compaction.h>
> +#include <linux/sched/sysctl.h>
>
> #include "internal.h"
>
> @@ -3988,7 +3989,10 @@ static bool pgdat_balanced(pg_data_t *pgdat, int order, int highest_zoneidx)
> if (!managed_zone(zone))
> continue;
>
> - mark = high_wmark_pages(zone);
> + if (sysctl_numa_balancing_mode & NUMA_BALANCING_MEMORY_TIERING)
> + mark = wmark_pages(zone, WMARK_PROMO);
> + else
> + mark = high_wmark_pages(zone);
> if (zone_watermark_ok_safe(zone, order, mark, highest_zoneidx))
> return true;
> }
> --
> 2.30.2
>