mm: place pages to the freelist tail when onling and undoing isolation
Subsystems (6)
cover mm: place pages to the freelist tail when onling and undoing isolation
When adding separate memory blocks via add_memory*() and onlining them immediately, the metadata (especially the memmap) of the next block will be placed onto one of the just added+onlined block. This creates a chain of unmovable allocations: If the last memory block cannot get offlined+removed() so will all dependant ones. We directly have unmovable allocations all over the place.
This can be observed quite easily using virtio-mem, however, it can also be observed when using DIMMs. The freshly onlined pages will usually be placed to the head of the freelists, meaning they will be allocated next, turning the just-added memory usually immediately un-removable. The fresh pages are cold, prefering to allocate others (that might be hot) also feels to be the natural thing to do.
It also applies to the hyper-v balloon xen-balloon, and ppc64 dlpar: when adding separate, successive memory blocks, each memory block will have unmovable allocations on them - for example gigantic pages will fail to allocate.
While the ZONE_NORMAL doesn't provide any guarantees that memory can get offlined+removed again (any kind of fragmentation with unmovable allocations is possible), there are many scenarios (hotplugging a lot of memory, running workload, hotunplug some memory/as much as possible) where we can offline+remove quite a lot with this patchset.
a) To visualize the problem, a very simple example:
Start a VM with 4GB and 8GB of virtio-mem memory:
[root@localhost ~]# lsmem
RANGE SIZE STATE REMOVABLE BLOCK
0x0000000000000000-0x00000000bfffffff 3G online yes 0-23
0x0000000100000000-0x000000033fffffff 9G online yes 32-103
Memory block size: 128M
Total online memory: 12G
Total offline memory: 0B
Then try to unplug as much as possible using virtio-mem. Observe which memory blocks are still around. Without this patch set:
[root@localhost ~]# lsmem RANGE SIZE STATE REMOVABLE BLOCK 0x0000000000000000-0x00000000bfffffff 3G online yes 0-23 0x0000000100000000-0x000000013fffffff 1G online yes 32-39 0x0000000148000000-0x000000014fffffff 128M online yes 41 0x0000000158000000-0x000000015fffffff 128M online yes 43 0x0000000168000000-0x000000016fffffff 128M online yes 45 0x0000000178000000-0x000000017fffffff 128M online yes 47 0x0000000188000000-0x0000000197ffffff 256M online yes 49-50 0x00000001a0000000-0x00000001a7ffffff 128M online yes 52 0x00000001b0000000-0x00000001b7ffffff 128M online yes 54 0x00000001c0000000-0x00000001c7ffffff 128M online yes 56 0x00000001d0000000-0x00000001d7ffffff 128M online yes 58 0x00000001e0000000-0x00000001e7ffffff 128M online yes 60 0x00000001f0000000-0x00000001f7ffffff 128M online yes 62 0x0000000200000000-0x0000000207ffffff 128M online yes 64 0x0000000210000000-0x0000000217ffffff 128M online yes 66 0x0000000220000000-0x0000000227ffffff 128M online yes 68 0x0000000230000000-0x0000000237ffffff 128M online yes 70 0x0000000240000000-0x0000000247ffffff 128M online yes 72 0x0000000250000000-0x0000000257ffffff 128M online yes 74 0x0000000260000000-0x0000000267ffffff 128M online yes 76 0x0000000270000000-0x0000000277ffffff 128M online yes 78 0x0000000280000000-0x0000000287ffffff 128M online yes 80 0x0000000290000000-0x0000000297ffffff 128M online yes 82 0x00000002a0000000-0x00000002a7ffffff 128M online yes 84 0x00000002b0000000-0x00000002b7ffffff 128M online yes 86 0x00000002c0000000-0x00000002c7ffffff 128M online yes 88 0x00000002d0000000-0x00000002d7ffffff 128M online yes 90 0x00000002e0000000-0x00000002e7ffffff 128M online yes 92 0x00000002f0000000-0x00000002f7ffffff 128M online yes 94 0x0000000300000000-0x0000000307ffffff 128M online yes 96 0x0000000310000000-0x0000000317ffffff 128M online yes 98 0x0000000320000000-0x0000000327ffffff 128M online yes 100 0x0000000330000000-0x000000033fffffff 256M online yes 102-103 Memory block size: 128M Total online memory: 8.1G Total offline memory: 0B
With this patch set:
[root@localhost ~]# lsmem
RANGE SIZE STATE REMOVABLE BLOCK
0x0000000000000000-0x00000000bfffffff 3G online yes 0-23
0x0000000100000000-0x000000013fffffff 1G online yes 32-39
Memory block size: 128M
Total online memory: 4G
Total offline memory: 0B
All memory can get unplugged, all memory block can get removed. Of course, no workload ran and the system was basically idle, but it highlights the issue - the fairly deterministic chain of unmovable allocations. When a huge page for the 2MB memmap is needed, a just-onlined 4MB page will be split. The remaining 2MB page will be used for the memmap of the next memory block. So one memory block will hold the memmap of the two following memory blocks. Finally the pages of the last-onlined memory block will get used for the next bigger allocations - if any allocation is unmovable, all dependent memory blocks cannot get unplugged and removed until that allocation is gone.
Note that with bigger memory blocks (e.g., 256MB), *all* memory blocks are dependent and none can get unplugged again!
b) Experiment with memory intensive workload
I performed an experiment with an older version of this patch set (before we used undo_isolate_page_range() in online_pages():
Hotplug 56GB to a VM with an initial 4GB, onlining all memory to ZONE_NORMAL right from the kernel when adding it. I then run various memory intensive workloads that consume most system memory for a total of 45 minutes. Once finished, I try to unplug as much memory as possible.
With this change, I am able to remove via virtio-mem (adding individual 128MB memory blocks) 413 out of 448 added memory blocks. Via individual (256MB) DIMMs 380 out of 448 added memory blocks. (I don't have any numbers without this patchset, but looking at the above example, it's at most half of the 448 memory blocks for virtio-mem, and most probably none for DIMMs).
Again, there are workloads that might behave very differently due to the nature of ZONE_NORMAL.
c) Future work:
- I'll be looking into avoiding reporting freshly onlined pages via the free page reporting framework. They are unbacked in the hypervisor, so reporting them isn't necessary (and might actually be bad for performance in some future use cases in the hypervisor).
- I'll be looking into being able to tell the OS that some pages are fresh (e.g., via alloc_contig_range() in virito-mem, freeing balloon inflated memory in a ballooning driver), such that we will skip reporting them via free page reporting (marking them reported), and placing them to the tail of the freelist.
- virtio-mem will soon also support ZONE_MOVABLE, however, especially when hotplugging a lot of memory (as in the experiment), a considerable amount of memory will have to remain in ZONE_NORMAL - so this change is relevant in any case.
I'm sending this as RFC as it also in its current form for simplicity affects not only memory onlining but also
- Other users of undo_isolate_page_range(): Pages are always placed to the tail.
-- When memory offlining fails
-- When memory isolation fails after having isolated some pageblocks -- When alloc_contig_range() either succeeds or fails
- Other users of __putback_isolated_page(): Pages are always placed to the tail.
-- Free page reporting
- Other users of __free_pages_core()
-- AFAIKs, any memory that is getting exposed to the buddy during boot.
IIUC we will now usually allocate memory from lower addresses within a zone first (especially during boot).
- Other users of generic_online_page()
-- Hyper-V balloon
Let's see if there are concerns for these users with this approach.
David Hildenbrand (4):
mm/page_alloc: convert "report" flag of __free_one_page() to a proper
flag
mm/page_alloc: place pages to tail in __putback_isolated_page() mm/page_alloc: always move pages to the tail of the freelist in
unset_migratetype_isolate()
mm/page_alloc: place pages to tail in __free_pages_core()
include/linux/page-isolation.h | 2 + mm/page_alloc.c | 102 +++++++++++++++++++++++++-------- mm/page_isolation.c | 8 ++- 3 files changed, 86 insertions(+), 26 deletions(-)
Signature · 1 line
2.26.2
RFC PATCH 1/4 mm/page_alloc: convert "report" flag of __free_one_page() to a proper flag
Let's prepare for additional flags and avoid long parameter lists of bools. Follow-up patches will also make use of the flags in __free_pages_ok(), however, I wasn't able to come up with a better name for the type - should be good enough for internal purposes.
--- mm/page_alloc.c | 28 ++++++++++++++++++++-------- 1 file changed, 20 insertions(+), 8 deletions(-)
mm/page_alloc.c +20 -8
@@ -77,6 +77,18 @@#include "shuffle.h" #include "page_reporting.h" +/* Free One Page flags: for internal, non-pcp variants of free_pages(). */ +typedef int __bitwise fop_t ; + +/* No special request */ +#define FOP_NONE ((__force fop_t)0) + +/* +* Skip free page reporting notification after buddy merging (will *not* mark +* the page reported, only skip the notification). +*/ +#define FOP_SKIP_REPORT_NOTIFY ((__force fop_t)BIT(0)) +/* prevent >1 _updater_ of zone percpu pageset ->high and ->batch fields */ static DEFINE_MUTEX (pcp_batch_high_lock);#define MIN_PERCPU_PAGELIST_FRACTION (8)
@@ -948,10 +960,9 @@ buddy_merge_likely(unsigned long pfn, unsigned long buddy_pfn,* -- nyc */ -static inline void __free_one_page (struct page *page, -unsigned long pfn, -struct zone *zone,unsigned int order, -int migratetype,bool report) +static inline void __free_one_page (struct page *page,unsigned long pfn, +struct zone *zone,unsigned int order, +int migratetype,fop_t fop_flags) {struct capture_control *capc =task_capc (zone);unsigned long buddy_pfn;
@@ -1038,7 +1049,7 @@ static inline void __free_one_page(struct page *page,add_to_free_list (page, zone, order, migratetype);/* Notify page reporting subsystem of freed page */ -if (report) +if (!(fop_flags &FOP_SKIP_REPORT_NOTIFY ))page_reporting_notify_free (order); }
@@ -1368,7 +1379,7 @@ static void free_pcppages_bulk(struct zone *zone, int count,if (unlikely (isolated_pageblocks)) mt =get_pageblock_migratetype (page); -__free_one_page (page,page_to_pfn (page), zone,0 , mt, true); +__free_one_page (page,page_to_pfn (page), zone,0 , mt,FOP_NONE );trace_mm_page_pcpu_drain (page,0 , mt); }spin_unlock (&zone->lock);
@@ -1384,7 +1395,7 @@ static void free_one_page(struct zone *zone, is_migrate_isolate(migratetype))) { migratetype =get_pfnblock_migratetype (page, pfn); } -__free_one_page (page, pfn, zone, order, migratetype, true); +__free_one_page (page, pfn, zone, order, migratetype,FOP_NONE );spin_unlock (&zone->lock); }
@@ -3277,7 +3288,8 @@ void __putback_isolated_page(struct page *page, unsigned int order, int mt)lockdep_assert_held (&zone->lock);/* Return isolated page to tail of freelist. */ -__free_one_page (page,page_to_pfn (page), zone, order, mt, false); +__free_one_page (page,page_to_pfn (page), zone, order, mt, +FOP_SKIP_REPORT_NOTIFY ); }/*
Signature · 1 line
2.26.2
RFC PATCH 2/4 mm/page_alloc: place pages to tail in __putback_isolated_page()
In reply to David Hildenbrand__putback_isolated_page() already documents that pages will be placed to the tail of the freelist - this is, however, not the case for "order >= MAX_ORDER - 2" (see buddy_merge_likely()) - which should be the case for all existing users.
This change affects two users:
- free page reporting
- page isolation, when undoing the isolation.
This behavior is desireable for pages that haven't really been touched lately, so exactly the two users that don't actually read/write page content, but rather move untouched pages.
The new behavior is especially desirable for memory onlining, where we allow allocation of newly onlined pages via undo_isolate_page_range() in online_pages(). Right now, we always place them to the head of the free list, resulting in undesireable behavior: Assume we add individual memory chunks via add_memory() and online them right away to the NORMAL zone. We create a dependency chain of unmovable allocations e.g., via the memmap. The memmap of the next chunk will be placed onto previous chunks - if the last block cannot get offlined+removed, all dependent ones cannot get offlined+removed. While this can already be observed with individual DIMMs, it's more of an issue for virtio-mem (and I suspect also ppc DLPAR).
Note: If we observe a degradation due to the changed page isolation behavior (which I doubt), we can always make this configurable by the instance triggering undo of isolation (e.g., alloc_contig_range(), memory onlining, memory offlining).
--- mm/page_alloc.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-)
mm/page_alloc.c +9 -1
@@ -89,6 +89,12 @@ typedef int __bitwise fop_t;*/ #define FOP_SKIP_REPORT_NOTIFY ((__force fop_t)BIT(0)) +/* +* Place the freed page to the tail of the freelist after buddy merging. Will +* get ignored with page shuffling enabled. +*/ +#define FOP_TO_TAIL ((__force fop_t)BIT(1)) +/* prevent >1 _updater_ of zone percpu pageset ->high and ->batch fields */ static DEFINE_MUTEX (pcp_batch_high_lock);#define MIN_PERCPU_PAGELIST_FRACTION (8)
@@ -1040,6 +1046,8 @@ static inline void __free_one_page(struct page *page, unsigned long pfn,if (is_shuffle_order (order)) to_tail =shuffle_pick_tail (); +else if (fop_flags &FOP_TO_TAIL ) + to_tail = true;else to_tail =buddy_merge_likely (pfn, buddy_pfn, page, order);
@@ -3289,7 +3297,7 @@ void __putback_isolated_page(struct page *page, unsigned int order, int mt)/* Return isolated page to tail of freelist. */ __free_one_page (page,page_to_pfn (page), zone, order, mt, -FOP_SKIP_REPORT_NOTIFY ); +FOP_SKIP_REPORT_NOTIFY |FOP_TO_TAIL ); }/*
Signature · 1 line
2.26.2
RFC PATCH 3/4 mm/page_alloc: always move pages to the tail of the freelist in unset_migratetype_isolate()
In reply to David HildenbrandPage isolation doesn't actually touch the pages, it simply isolates pageblocks and moves all free pages to the MIGRATE_ISOLATE freelist.
We already place pages to the tail of the freelists when undoing isolation via __putback_isolated_page(), let's do it in any case (e.g., if order == pageblock_order) and document the behavior.
This change results in all pages getting onlined via online_pages() to be placed to the tail of the freelist.
--- include/linux/page-isolation.h | 2 ++ mm/page_alloc.c | 36 +++++++++++++++++++++++++++++----- mm/page_isolation.c | 8 ++++++-- 3 files changed, 39 insertions(+), 7 deletions(-)
include/linux/page-isolation.h +2 -0
@@ -38,6 +38,8 @@ struct page *has_unmovable_pages(struct zone *zone, struct page *page,void set_pageblock_migratetype (struct page *page,int migratetype);int move_freepages_block (struct zone *zone,struct page *page,int migratetype,int *num_movable); +int move_freepages_block_tail (struct zone *zone,struct page *page, +int migratetype);/* * Changes migrate type in [start_pfn, end_pfn) to be MIGRATE_ISOLATE.
mm/page_alloc.c +31 -5
@@ -899,6 +899,15 @@ static inline void move_to_free_list(struct page *page, struct zone *zone,list_move (&page->lru, &area->free_list[migratetype]); } +/* Used for pages which are on another list */ +static inline void move_to_free_list_tail (struct page *page,struct zone *zone, +unsigned int order,int migratetype) +{ +struct free_area *area = &zone->free_area[order]; + +list_move_tail (&page->lru, &area->free_list[migratetype]); +} +static inline void del_page_from_free_list (struct page *page,struct zone *zone,unsigned int order) {
@@ -2323,7 +2332,7 @@ static inline struct page *__rmqueue_cma_fallback(struct zone *zone,*/ static int move_freepages (struct zone *zone,struct page *start_page,struct page *end_page, -int migratetype,int *num_movable) +int migratetype,int *num_movable,bool to_tail) {struct page *page;unsigned int order;
@@ -2354,7 +2363,10 @@ static int move_freepages(struct zone *zone,VM_BUG_ON_PAGE (page_zone (page) != zone, page); order =page_order (page); -move_to_free_list (page, zone, order, migratetype); +if (to_tail) +move_to_free_list_tail (page, zone, order, migratetype); +else +move_to_free_list (page, zone, order, migratetype); page +=1 << order; pages_moved +=1 << order; }
@@ -2362,8 +2374,9 @@ static int move_freepages(struct zone *zone,return pages_moved; } -int move_freepages_block (struct zone *zone,struct page *page, -int migratetype,int *num_movable) +static int __move_freepages_block (struct zone *zone,struct page *page, +int migratetype,int *num_movable, +bool to_tail) {unsigned long start_pfn, end_pfn;struct page *start_page, *end_page;
@@ -2384,7 +2397,20 @@ int move_freepages_block(struct zone *zone, struct page *page,return 0 ;return move_freepages (zone, start_page, end_page, migratetype, - num_movable); + num_movable, to_tail); +} + +int move_freepages_block (struct zone *zone,struct page *page, +int migratetype,int *num_movable) +{ +return __move_freepages_block (zone, page, migratetype, num_movable, + false); +} + +int move_freepages_block_tail (struct zone *zone,struct page *page, +int migratetype) +{ +return __move_freepages_block (zone, page, migratetype,NULL , true); } static void change_pageblock_range(struct page *pageblock_page,
mm/page_isolation.c +6 -2
@@ -83,7 +83,7 @@ static void unset_migratetype_isolate(struct page *page, unsigned migratetype)* Because freepage with more than pageblock_order on isolated * pageblock is restricted to merge due to freepage counting problem, * it is possible that there is free buddy page. -* move_freepages_block() doesn't care of merge so we need other +* move_freepages_block*() don't care about merging, so we need another * approach in order to merge them. Isolation and free will make * these pages to be merged. */
@@ -106,9 +106,13 @@ static void unset_migratetype_isolate(struct page *page, unsigned migratetype)* If we isolate freepage with more than pageblock_order, there * should be no freepage in the range, so we could avoid costly * pageblock scanning for freepage moving. +* +* We didn't actually touch any of the isolated pages, so place them +* to the tail of the freelists. This is especially relevant during +* memory onlining. */ if (!isolated_page) { - nr_pages =move_freepages_block (zone, page, migratetype,NULL ); + nr_pages =move_freepages_block_tail (zone, page, migratetype);__mod_zone_freepage_state (zone, nr_pages, migratetype); }set_pageblock_migratetype (page, migratetype);
Signature · 1 line
2.26.2
RFC PATCH 4/4 mm/page_alloc: place pages to tail in __free_pages_core()
In reply to David Hildenbrand__free_pages_core() is used when exposing fresh memory to the buddy during system boot and when onlining memory in generic_online_page().
generic_online_page() is used in two cases:
1. Direct memory onlining in online_pages().
2. Deferred memory onlining in memory-ballooning-like mechanisms (HyperV balloon and virtio-mem), when parts of a section are kept fake-offline to be fake-onlined later on.
In 1, we already place pages to the tail of the freelist. Pages will be freed to MIGRATE_ISOLATE lists first and moved to the tail of the freelists via undo_isolate_page_range().
In 2, we currently don't implement a proper rule. In case of virtio-mem, where we currently always online MAX_ORDER - 1 pages, the pages will be placed to the HEAD of the freelist - undesireable. While the hyper-v balloon calls generic_online_page() with single pages, usually it will call it on successive single pages in a larger block.
The pages are fresh, so place them to the tail of the freelists and avoid the PCP.
Note: If we detect that the new behavior is undesireable for __free_pages_core() during boot, we can let the caller specify the behavior.
--- mm/page_alloc.c | 32 ++++++++++++++++++++------------ 1 file changed, 20 insertions(+), 12 deletions(-)
mm/page_alloc.c +20 -12
@@ -264,7 +264,8 @@ bool pm_suspended_storage(void)unsigned int pageblock_order__read_mostly ;#endif -static void __free_pages_ok (struct page *page,unsigned int order); +static void __free_pages_ok (struct page *page,unsigned int order, +fop_t fop_flags);/* * results with 256, 32 in the lowmem_reserve sysctl:
@@ -676,7 +677,7 @@ static void bad_page(struct page *page, const char *reason)void free_compound_page (struct page *page) {mem_cgroup_uncharge (page); -__free_pages_ok (page,compound_order (page)); +__free_pages_ok (page,compound_order (page),FOP_NONE ); }void prep_compound_page (struct page *page,unsigned int order)
@@ -1402,17 +1403,15 @@ static void free_pcppages_bulk(struct zone *zone, int count,spin_unlock (&zone->lock); } -static void free_one_page (struct zone *zone, -struct page *page,unsigned long pfn, -unsigned int order, -int migratetype) +static void free_one_page (struct zone *zone,struct page *page,unsigned long pfn, +unsigned int order,int migratetype,fop_t fop_flags) {spin_lock (&zone->lock);if (unlikely (has_isolate_pageblock (zone) ||is_migrate_isolate (migratetype))) { migratetype =get_pfnblock_migratetype (page, pfn); } -__free_one_page (page, pfn, zone, order, migratetype,FOP_NONE ); +__free_one_page (page, pfn, zone, order, migratetype, fop_flags);spin_unlock (&zone->lock); }
@@ -1490,7 +1489,8 @@ void __meminit reserve_bootmem_region(phys_addr_t start, phys_addr_t end) } } -static void __free_pages_ok (struct page *page,unsigned int order) +static void __free_pages_ok (struct page *page,unsigned int order, +fop_t fop_flags) {unsigned long flags;int migratetype;
@@ -1502,7 +1502,8 @@ static void __free_pages_ok(struct page *page, unsigned int order) migratetype =get_pfnblock_migratetype (page, pfn);local_irq_save (flags);__count_vm_events (PGFREE ,1 << order); -free_one_page (page_zone (page), page, pfn, order, migratetype); +free_one_page (page_zone (page), page, pfn, order, migratetype, + fop_flags);local_irq_restore (flags); }
@@ -1523,7 +1524,13 @@ void __free_pages_core(struct page *page, unsigned int order)atomic_long_add (nr_pages, &page_zone (page)->managed_pages);set_page_refcounted (page); -__free_pages (page, order); + +/* +* Bypass PCP and place fresh pages right to the tail, primarily +* relevant for memory onlining. +*/ +page_ref_dec (page); +__free_pages_ok (page, order,FOP_TO_TAIL ); }#ifdef CONFIG_NEED_MULTIPLE_NODES
@@ -3167,7 +3174,8 @@ static void free_unref_page_commit(struct page *page, unsigned long pfn)*/ if (migratetype >=MIGRATE_PCPTYPES ) {if (unlikely (is_migrate_isolate (migratetype))) { -free_one_page (zone, page, pfn,0 , migratetype); +free_one_page (zone, page, pfn,0 , migratetype, +FOP_NONE );return ; } migratetype =MIGRATE_MOVABLE ;
@@ -4984,7 +4992,7 @@ static inline void free_the_page(struct page *page, unsigned int order)if (order ==0 )/* Via pcp? */ free_unref_page (page);else -__free_pages_ok (page, order); +__free_pages_ok (page, order,FOP_NONE ); }void __free_pages (struct page *page,unsigned int order)
Signature · 1 line
2.26.2
Quoted 19 lines · David Hildenbrand
When adding separate memory blocks via add_memory*() and onlining them immediately, the metadata (especially the memmap) of the next block will be placed onto one of the just added+onlined block. This creates a chain of unmovable allocations: If the last memory block cannot get offlined+removed() so will all dependant ones. We directly have unmovable allocations all over the place.
This can be observed quite easily using virtio-mem, however, it can also be observed when using DIMMs. The freshly onlined pages will usually be placed to the head of the freelists, meaning they will be allocated next, turning the just-added memory usually immediately un-removable. The fresh pages are cold, prefering to allocate others (that might be hot) also feels to be the natural thing to do.
It also applies to the hyper-v balloon xen-balloon, and ppc64 dlpar: when adding separate, successive memory blocks, each memory block will have unmovable allocations on them - for example gigantic pages will fail to allocate.
Quoted 5 lines
While the
ZONE_NORMALdoesn't provide any guarantees that memory can get offlined+removed again (any kind of fragmentation with unmovable allocations is possible), there are many scenarios (hotplugging a lot of memory, running workload, hotunplug some memory/as much as possible) where we can offline+remove quite a lot with this patchset.
Hi David,
I did not read through the patchset yet, so sorry if the question is nonsense, but is this not trying to fix the same issue the vmemmap patches did? [1]
I was about to give it a new respin now that thw hwpoison stuff has been settled.
[1] https://patchwork.kernel.org/cover/11059175/
Quoted 1 line
Quoted 27 lines
Am 16.09.2020 um 20:50 schrieb [email protected]:
Quoted 13 lines · David Hildenbrand
When adding separate memory blocks via add_memory*() and onlining them immediately, the metadata (especially the memmap) of the next block will be placed onto one of the just added+onlined block. This creates a chain of unmovable allocations: If the last memory block cannot get offlined+removed() so will all dependant ones. We directly have unmovable allocations all over the place.
This can be observed quite easily using virtio-mem, however, it can also be observed when using DIMMs. The freshly onlined pages will usually be placed to the head of the freelists, meaning they will be allocated next, turning the just-added memory usually immediately un-removable. The fresh pages are cold, prefering to allocate others (that might be hot) also feels to be the natural thing to do.
It also applies to the hyper-v balloon xen-balloon, and ppc64 dlpar: whenQuoted 8 lines
adding separate, successive memory blocks, each memory block will have unmovable allocations on them - for example gigantic pages will fail to allocate.
While theZONE_NORMALdoesn't provide any guarantees that memory can get offlined+removed again (any kind of fragmentation with unmovable allocations is possible), there are many scenarios (hotplugging a lot of memory, running workload, hotunplug some memory/as much as possible) where we can offline+remove quite a lot with this patchset.
Quoted 2 lines
Hi David,
Hi Oscar.
Quoted 1 line
I did not read through the patchset yet, so sorry if the question is nonsense, but is this not trying to fix the same issue the vmemmap patches did? [1]
Not nonesense at all. It only helps to some degree, though. It solves the dependencies due to the memmap. However, it‘s not completely ideal, especially for single memory blocks.
With single memory blocks (virtio-mem, xen-balloon, hv balloon, ppc dlpar) you still have unmovable (vmemmap chunks) all over the physical address space. Consider the gigantic page example after hotplug. You directly fragmented all hotplugged memory.
Of course, there might be (less extreme) dependencies due page tables for the identity mapping, extended struct pages and similar.
Having that said, there are other benefits when preferring other memory over just hotplugged memory. Think about adding+onlining memory during boot (dimms under QEMU, virtio-mem), once the system is up you will have most (all) of that memory completely untouched.
So while vmemmap on hotplugged memory would tackle some part of the issue, there are cases where this approach is better, and there are even benefits when combining both.
Thanks!
David
Quoted 5 lines
I was about to give it a new respin now that thw hwpoison stuff has been settled.
Re: [PATCH RFC 1/4] mm/page_alloc: convert "report" flag of __free_one_page() to a proper flag
In reply to David HildenbrandQuoted 85 lines · David Hildenbrand
Let's prepare for additional flags and avoid long parameter lists of bools. Follow-up patches will also make use of the flags in
__free_pages_ok(), however, I wasn't able to come up with a better name for the type - should be good enough for internal purposes.Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Signed-off-by David Hildenbrand <[email protected]>--- mm/page_alloc.c | 28 ++++++++++++++++++++-------- 1 file changed, 20 insertions(+), 8 deletions(-)mm/page_alloc.c +18
-7@@ -77,6 +77,18 @@#include "shuffle.h" #include "page_reporting.h" +/* Free One Page flags: for internal, non-pcp variants of free_pages(). */ +typedef int __bitwise fop_t ; + +/* No special request */ +#define FOP_NONE ((__force fop_t)0) + +/* +* Skip free page reporting notification after buddy merging (will *not* mark +* the page reported, only skip the notification). +*/ +#define FOP_SKIP_REPORT_NOTIFY ((__force fop_t)BIT(0)) +/* prevent >1 _updater_ of zone percpu pageset ->high and ->batch fields */ static DEFINE_MUTEX (pcp_batch_high_lock);#define MIN_PERCPU_PAGELIST_FRACTION (8)@@ -948,10 +960,9 @@ buddy_merge_likely(unsigned long pfn, unsigned long buddy_pfn,* -- nyc */ -static inline void __free_one_page (struct page *page, -unsigned long pfn, -struct zone *zone,unsigned int order, -int migratetype,bool report) +static inline void __free_one_page (struct page *page,unsigned long pfn, +struct zone *zone,unsigned int order, +int migratetype,fop_t fop_flags) {struct capture_control *capc =task_capc (zone);unsigned long buddy_pfn;@@ -1038,7 +1049,7 @@ static inline void __free_one_page(struct page *page,add_to_free_list (page, zone, order, migratetype);/* Notify page reporting subsystem of freed page */ -if (report) +if (!(fop_flags &FOP_SKIP_REPORT_NOTIFY ))page_reporting_notify_free (order); }@@ -1368,7 +1379,7 @@ static void free_pcppages_bulk(struct zone *zone, int count,if (unlikely (isolated_pageblocks)) mt =get_pageblock_migratetype (page); -__free_one_page (page,page_to_pfn (page), zone,0 , mt, true); +__free_one_page (page,page_to_pfn (page), zone,0 , mt,FOP_NONE );trace_mm_page_pcpu_drain (page,0 , mt); }spin_unlock (&zone->lock);@@ -1384,7 +1395,7 @@ static void free_one_page(struct zone *zone, is_migrate_isolate(migratetype))) { migratetype =get_pfnblock_migratetype (page, pfn); } -__free_one_page (page, pfn, zone, order, migratetype, true); +__free_one_page (page, pfn, zone, order, migratetype,FOP_NONE );spin_unlock (&zone->lock); }
mm/page_alloc.c void __putback_isolated_page(struct page *page, unsigned int order, int mt) · line 3295
Quoted 9 lines
mm/page_alloc.c +2
-1@@ -3277,7 +3288,8 @@ void __putback_isolated_page(struct page *page, unsigned int order, int mt)lockdep_assert_held (&zone->lock);/* Return isolated page to tail of freelist. */ -__free_one_page (page,page_to_pfn (page), zone, order, mt, false); +__free_one_page (page,page_to_pfn (page), zone, order, mt, +FOP_SKIP_REPORT_NOTIFY ); }/*
Seems pretty straight forward. So we are basically flipping the logic and replacing !report with FOP_SKIP_REPORT_NOTIFY.
Re: [PATCH RFC 2/4] mm/page_alloc: place pages to tail in __putback_isolated_page()
In reply to David HildenbrandQuoted 11 lines · David Hildenbrand
__putback_isolated_page()already documents that pages will be placed to the tail of the freelist - this is, however, not the case for "order >=MAX_ORDER- 2" (seebuddy_merge_likely()) - which should be the case for all existing users.This change affects two users:
- free page reporting
- page isolation, when undoing the isolation.This behavior is desireable for pages that haven't really been touched
I think "desirable" is misspelled here.
Quoted 2 lines
lately, so exactly the two users that don't actually read/write page content, but rather move untouched pages.
So in reality we were already dealing with this for page reporting, but not in the most direct way. If I recall we were adding the pages to the head of the list and then when we would go back to pull more pages we were doing list rotation in the report function so they were technically being added to the head, but usually would end up back on the tail anyway. If anything the benefit for page reporting is that it should be more direct this way as we will only have to rescan the pages now when we have consumed all of the reported ones on the list.
Quoted 60 lines
The new behavior is especially desirable for memory onlining, where we allow allocation of newly onlined pages via
undo_isolate_page_range()inonline_pages(). Right now, we always place them to the head of the free list, resulting in undesireable behavior: Assume we add individual memory chunks viaadd_memory()and online them right away to the NORMAL zone. We create a dependency chain of unmovable allocations e.g., via the memmap. The memmap of the next chunk will be placed onto previous chunks - if the last block cannot get offlined+removed, all dependent ones cannot get offlined+removed. While this can already be observed with individual DIMMs, it's more of an issue for virtio-mem (and I suspect also ppc DLPAR).Note: If we observe a degradation due to the changed page isolation behavior (which I doubt), we can always make this configurable by the instance triggering undo of isolation (e.g.,
alloc_contig_range(), memory onlining, memory offlining).Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Cc Scott Cheloha <[email protected]>Cc Michael Ellerman <[email protected]>Signed-off-by David Hildenbrand <[email protected]>--- mm/page_alloc.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-)mm/page_alloc.c +8
-0@@ -89,6 +89,12 @@ typedef int __bitwise fop_t;*/ #define FOP_SKIP_REPORT_NOTIFY ((__force fop_t)BIT(0)) +/* +* Place the freed page to the tail of the freelist after buddy merging. Will +* get ignored with page shuffling enabled. +*/ +#define FOP_TO_TAIL ((__force fop_t)BIT(1)) +/* prevent >1 _updater_ of zone percpu pageset ->high and ->batch fields */ static DEFINE_MUTEX (pcp_batch_high_lock);#define MIN_PERCPU_PAGELIST_FRACTION (8)@@ -1040,6 +1046,8 @@ static inline void __free_one_page(struct page *page, unsigned long pfn,if (is_shuffle_order (order)) to_tail =shuffle_pick_tail (); +else if (fop_flags &FOP_TO_TAIL ) + to_tail = true;else to_tail =buddy_merge_likely (pfn, buddy_pfn, page, order);
mm/page_alloc.c void __putback_isolated_page(struct page *page, unsigned int order, int mt) · line 3303
Quoted 8 lines
mm/page_alloc.c +1
-1@@ -3289,7 +3297,7 @@ void __putback_isolated_page(struct page *page, unsigned int order, int mt)/* Return isolated page to tail of freelist. */ __free_one_page (page,page_to_pfn (page), zone, order, mt, -FOP_SKIP_REPORT_NOTIFY ); +FOP_SKIP_REPORT_NOTIFY |FOP_TO_TAIL ); }/*
The code looks good to me.
Re: [PATCH RFC 1/4] mm/page_alloc: convert "report" flag of __free_one_page() to a proper flag
In reply to David HildenbrandQuoted 25 lines · David Hildenbrand
Let's prepare for additional flags and avoid long parameter lists of bools. Follow-up patches will also make use of the flags in
__free_pages_ok(), however, I wasn't able to come up with a better name for the type - should be good enough for internal purposes.Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Signed-off-by David Hildenbrand <[email protected]>---
mm/page_alloc.c| 28 ++++++++++++++++++++--------1 file changed, 20 insertions(+), 8 deletions(-)mm/page_alloc.c +0
-0@@ -77,6 +77,18 @@#include "shuffle.h"
mm/page_alloc.c line 87
Quoted 10 lines
mm/page_alloc.c +8
-0#include "page_reporting.h" +/* Free One Page flags: for internal, non-pcp variants of free_pages(). */ +typedef int __bitwise fop_t ; + +/* No special request */ +#define FOP_NONE ((__force fop_t)0) + +/* +* Skip free page reporting notification after buddy merging (will *not* mark
__free_one_page() may not merge buddy when its buddy is not available.
Would this comment be a little confusing?
Quoted 60 lines
mm/page_alloc.c +12
-8+* the page reported, only skip the notification). +*/ +#define FOP_SKIP_REPORT_NOTIFY ((__force fop_t)BIT(0)) +/* prevent >1 _updater_ of zone percpu pageset ->high and ->batch fields */ static DEFINE_MUTEX (pcp_batch_high_lock);#define MIN_PERCPU_PAGELIST_FRACTION (8)@@ -948,10 +960,9 @@ buddy_merge_likely(unsigned long pfn, unsigned long buddy_pfn,* -- nyc */ -static inline void __free_one_page (struct page *page, -unsigned long pfn, -struct zone *zone,unsigned int order, -int migratetype,bool report) +static inline void __free_one_page (struct page *page,unsigned long pfn, +struct zone *zone,unsigned int order, +int migratetype,fop_t fop_flags) {struct capture_control *capc =task_capc (zone);unsigned long buddy_pfn;@@ -1038,7 +1049,7 @@ static inline void __free_one_page(struct page *page,add_to_free_list (page, zone, order, migratetype);/* Notify page reporting subsystem of freed page */ -if (report) +if (!(fop_flags &FOP_SKIP_REPORT_NOTIFY ))page_reporting_notify_free (order); }@@ -1368,7 +1379,7 @@ static void free_pcppages_bulk(struct zone *zone, int count,if (unlikely (isolated_pageblocks)) mt =get_pageblock_migratetype (page); -__free_one_page (page,page_to_pfn (page), zone,0 , mt, true); +__free_one_page (page,page_to_pfn (page), zone,0 , mt,FOP_NONE );trace_mm_page_pcpu_drain (page,0 , mt); }spin_unlock (&zone->lock);@@ -1384,7 +1395,7 @@ static void free_one_page(struct zone *zone, is_migrate_isolate(migratetype))) { migratetype =get_pfnblock_migratetype (page, pfn); } -__free_one_page (page, pfn, zone, order, migratetype, true); +__free_one_page (page, pfn, zone, order, migratetype,FOP_NONE );spin_unlock (&zone->lock); }@@ -3277,7 +3288,8 @@ void __putback_isolated_page(struct page *page, unsigned int order, int mt)lockdep_assert_held (&zone->lock);/* Return isolated page to tail of freelist. */ -__free_one_page (page,page_to_pfn (page), zone, order, mt, false); +__free_one_page (page,page_to_pfn (page), zone, order, mt, +FOP_SKIP_REPORT_NOTIFY ); }/* Signature · 1 line
2.26.2
Signature · 2 lines
Wei Yang Help you, Help me
Re: [PATCH RFC 2/4] mm/page_alloc: place pages to tail in __putback_isolated_page()
In reply to David HildenbrandQuoted 9 lines · David Hildenbrand
__putback_isolated_page()already documents that pages will be placed to the tail of the freelist - this is, however, not the case for "order >=MAX_ORDER- 2" (seebuddy_merge_likely()) - which should be the case for all existing users.This change affects two users:
- free page reporting
- page isolation, when undoing the isolation.
Quoted 7 lines
This behavior is desireable for pages that haven't really been touched lately, so exactly the two users that don't actually read/write page content, but rather move untouched pages.
The new behavior is especially desirable for memory onlining, where we allow allocation of newly onlined pages via
undo_isolate_page_range()inonline_pages(). Right now, we always place them to the head of the
The code looks good, while I don't fully understand the log here.
undo_isolate_page_range() is used in __offline_pages and alloc_contig_range. I don't connect them with online_pages(). Do I miss something?
Quoted 67 lines
free list, resulting in undesireable behavior: Assume we add individual memory chunks via
add_memory()and online them right away to the NORMAL zone. We create a dependency chain of unmovable allocations e.g., via the memmap. The memmap of the next chunk will be placed onto previous chunks - if the last block cannot get offlined+removed, all dependent ones cannot get offlined+removed. While this can already be observed with individual DIMMs, it's more of an issue for virtio-mem (and I suspect also ppc DLPAR).Note: If we observe a degradation due to the changed page isolation behavior (which I doubt), we can always make this configurable by the instance triggering undo of isolation (e.g.,
alloc_contig_range(), memory onlining, memory offlining).Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Cc Scott Cheloha <[email protected]>Cc Michael Ellerman <[email protected]>Signed-off-by David Hildenbrand <[email protected]>---
mm/page_alloc.c| 10 +++++++++-1 file changed, 9 insertions(+), 1 deletion(-)mm/page_alloc.c +9
-1@@ -89,6 +89,12 @@ typedef int __bitwise fop_t;*/ #define FOP_SKIP_REPORT_NOTIFY ((__force fop_t)BIT(0)) +/* +* Place the freed page to the tail of the freelist after buddy merging. Will +* get ignored with page shuffling enabled. +*/ +#define FOP_TO_TAIL ((__force fop_t)BIT(1)) +/* prevent >1 _updater_ of zone percpu pageset ->high and ->batch fields */ static DEFINE_MUTEX (pcp_batch_high_lock);#define MIN_PERCPU_PAGELIST_FRACTION (8)@@ -1040,6 +1046,8 @@ static inline void __free_one_page(struct page *page, unsigned long pfn,if (is_shuffle_order (order)) to_tail =shuffle_pick_tail (); +else if (fop_flags &FOP_TO_TAIL ) + to_tail = true;else to_tail =buddy_merge_likely (pfn, buddy_pfn, page, order);@@ -3289,7 +3297,7 @@ void __putback_isolated_page(struct page *page, unsigned int order, int mt)/* Return isolated page to tail of freelist. */ __free_one_page (page,page_to_pfn (page), zone, order, mt, -FOP_SKIP_REPORT_NOTIFY ); +FOP_SKIP_REPORT_NOTIFY |FOP_TO_TAIL ); }/* Signature · 1 line
2.26.2
Signature · 2 lines
Wei Yang Help you, Help me
Re: [PATCH RFC 2/4] mm/page_alloc: place pages to tail in __putback_isolated_page()
In reply to David HildenbrandQuoted 64 lines · David Hildenbrand
__putback_isolated_page()already documents that pages will be placed to the tail of the freelist - this is, however, not the case for "order >=MAX_ORDER- 2" (seebuddy_merge_likely()) - which should be the case for all existing users.This change affects two users:
- free page reporting
- page isolation, when undoing the isolation.This behavior is desireable for pages that haven't really been touched lately, so exactly the two users that don't actually read/write page content, but rather move untouched pages.
The new behavior is especially desirable for memory onlining, where we allow allocation of newly onlined pages via
undo_isolate_page_range()inonline_pages(). Right now, we always place them to the head of the free list, resulting in undesireable behavior: Assume we add individual memory chunks viaadd_memory()and online them right away to the NORMAL zone. We create a dependency chain of unmovable allocations e.g., via the memmap. The memmap of the next chunk will be placed onto previous chunks - if the last block cannot get offlined+removed, all dependent ones cannot get offlined+removed. While this can already be observed with individual DIMMs, it's more of an issue for virtio-mem (and I suspect also ppc DLPAR).Note: If we observe a degradation due to the changed page isolation behavior (which I doubt), we can always make this configurable by the instance triggering undo of isolation (e.g.,
alloc_contig_range(), memory onlining, memory offlining).Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Cc Scott Cheloha <[email protected]>Cc Michael Ellerman <[email protected]>Signed-off-by David Hildenbrand <[email protected]>---
mm/page_alloc.c| 10 +++++++++-1 file changed, 9 insertions(+), 1 deletion(-)mm/page_alloc.c +6
-0@@ -89,6 +89,12 @@ typedef int __bitwise fop_t;*/ #define FOP_SKIP_REPORT_NOTIFY ((__force fop_t)BIT(0)) +/* +* Place the freed page to the tail of the freelist after buddy merging. Will +* get ignored with page shuffling enabled. +*/ +#define FOP_TO_TAIL ((__force fop_t)BIT(1)) +/* prevent >1 _updater_ of zone percpu pageset ->high and ->batch fields */ static DEFINE_MUTEX (pcp_batch_high_lock);#define MIN_PERCPU_PAGELIST_FRACTION (8)
mm/page_alloc.c static inline void __free_one_page(struct page *page, unsigned long pfn, · line 1050
Quoted 5 lines
mm/page_alloc.c +2
-0@@ -1040,6 +1046,8 @@ static inline void __free_one_page(struct page *page, unsigned long pfn,if (is_shuffle_order (order)) to_tail =shuffle_pick_tail (); +else if (fop_flags &FOP_TO_TAIL ) + to_tail = true;
Take another look into this part. Maybe we can move this check at top?
For online_page case, currently we have following call flow:
online_page
online_pages_range
shuffle_zone
This means we would always shuffle the newly added pages. Maybe we don't need to do the shuffle when adding them to the free_list?
Quoted 14 lines
mm/page_alloc.c +1
-1else to_tail =buddy_merge_likely (pfn, buddy_pfn, page, order);@@ -3289,7 +3297,7 @@ void __putback_isolated_page(struct page *page, unsigned int order, int mt)/* Return isolated page to tail of freelist. */ __free_one_page (page,page_to_pfn (page), zone, order, mt, -FOP_SKIP_REPORT_NOTIFY ); +FOP_SKIP_REPORT_NOTIFY |FOP_TO_TAIL ); }/* Signature · 1 line
2.26.2
Signature · 2 lines
Wei Yang Help you, Help me
Re: [PATCH RFC 3/4] mm/page_alloc: always move pages to the tail of the freelist in unset_migratetype_isolate()
In reply to David HildenbrandQuoted 9 lines · David Hildenbrand
Page isolation doesn't actually touch the pages, it simply isolates pageblocks and moves all free pages to the
MIGRATE_ISOLATEfreelist.We already place pages to the tail of the freelists when undoing isolation via
__putback_isolated_page(), let's do it in any case (e.g., if order ==pageblock_order) and document the behavior.This change results in all pages getting onlined via
online_pages()to be placed to the tail of the freelist.
I am sorry to not follow again. unset_migratetype_isolate() is used in __offline_pages if my understanding is correct. How does it contribute on online_pages?
Quoted 137 lines
Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Cc Scott Cheloha <[email protected]>Cc Michael Ellerman <[email protected]>Signed-off-by David Hildenbrand <[email protected]>---
include/linux/page-isolation.h | 2 ++ mm/page_alloc.c | 36 +++++++++++++++++++++++++++++----- mm/page_isolation.c | 8 ++++++-- 3 files changed, 39 insertions(+), 7 deletions(-)include/linux/page-isolation.h +2
-0@@ -38,6 +38,8 @@ struct page *has_unmovable_pages(struct zone *zone, struct page *page,void set_pageblock_migratetype (struct page *page,int migratetype);int move_freepages_block (struct zone *zone,struct page *page,int migratetype,int *num_movable); +int move_freepages_block_tail (struct zone *zone,struct page *page, +int migratetype);/* * Changes migrate type in [start_pfn, end_pfn) to be MIGRATE_ISOLATE. mm/page_alloc.c +31
-5@@ -899,6 +899,15 @@ static inline void move_to_free_list(struct page *page, struct zone *zone,list_move (&page->lru, &area->free_list[migratetype]); } +/* Used for pages which are on another list */ +static inline void move_to_free_list_tail (struct page *page,struct zone *zone, +unsigned int order,int migratetype) +{ +struct free_area *area = &zone->free_area[order]; + +list_move_tail (&page->lru, &area->free_list[migratetype]); +} +static inline void del_page_from_free_list (struct page *page,struct zone *zone,unsigned int order) {@@ -2323,7 +2332,7 @@ static inline struct page *__rmqueue_cma_fallback(struct zone *zone,*/ static int move_freepages (struct zone *zone,struct page *start_page,struct page *end_page, -int migratetype,int *num_movable) +int migratetype,int *num_movable,bool to_tail) {struct page *page;unsigned int order;@@ -2354,7 +2363,10 @@ static int move_freepages(struct zone *zone,VM_BUG_ON_PAGE (page_zone (page) != zone, page); order =page_order (page); -move_to_free_list (page, zone, order, migratetype); +if (to_tail) +move_to_free_list_tail (page, zone, order, migratetype); +else +move_to_free_list (page, zone, order, migratetype); page +=1 << order; pages_moved +=1 << order; }@@ -2362,8 +2374,9 @@ static int move_freepages(struct zone *zone,return pages_moved; } -int move_freepages_block (struct zone *zone,struct page *page, -int migratetype,int *num_movable) +static int __move_freepages_block (struct zone *zone,struct page *page, +int migratetype,int *num_movable, +bool to_tail) {unsigned long start_pfn, end_pfn;struct page *start_page, *end_page;@@ -2384,7 +2397,20 @@ int move_freepages_block(struct zone *zone, struct page *page,return 0 ;return move_freepages (zone, start_page, end_page, migratetype, - num_movable); + num_movable, to_tail); +} + +int move_freepages_block (struct zone *zone,struct page *page, +int migratetype,int *num_movable) +{ +return __move_freepages_block (zone, page, migratetype, num_movable, + false); +} + +int move_freepages_block_tail (struct zone *zone,struct page *page, +int migratetype) +{ +return __move_freepages_block (zone, page, migratetype,NULL , true); } static void change_pageblock_range(struct page *pageblock_page,mm/page_isolation.c +6
-2@@ -83,7 +83,7 @@ static void unset_migratetype_isolate(struct page *page, unsigned migratetype)* Because freepage with more than pageblock_order on isolated * pageblock is restricted to merge due to freepage counting problem, * it is possible that there is free buddy page. -* move_freepages_block() doesn't care of merge so we need other +* move_freepages_block*() don't care about merging, so we need another * approach in order to merge them. Isolation and free will make * these pages to be merged. */ @@ -106,9 +106,13 @@ static void unset_migratetype_isolate(struct page *page, unsigned migratetype)* If we isolate freepage with more than pageblock_order, there * should be no freepage in the range, so we could avoid costly * pageblock scanning for freepage moving. +* +* We didn't actually touch any of the isolated pages, so place them +* to the tail of the freelists. This is especially relevant during +* memory onlining. */ if (!isolated_page) { - nr_pages =move_freepages_block (zone, page, migratetype,NULL ); + nr_pages =move_freepages_block_tail (zone, page, migratetype);__mod_zone_freepage_state (zone, nr_pages, migratetype); }set_pageblock_migratetype (page, migratetype);Signature · 1 line
2.26.2
Signature · 2 lines
Wei Yang Help you, Help me
Quoted 36 lines · David Hildenbrand
Quoted 27 lines
Am 16.09.2020 um 20:50 schrieb [email protected]:
Quoted 13 lines · David Hildenbrand
When adding separate memory blocks via add_memory*() and onlining them immediately, the metadata (especially the memmap) of the next block will be placed onto one of the just added+onlined block. This creates a chain of unmovable allocations: If the last memory block cannot get offlined+removed() so will all dependant ones. We directly have unmovable allocations all over the place.
This can be observed quite easily using virtio-mem, however, it can also be observed when using DIMMs. The freshly onlined pages will usually be placed to the head of the freelists, meaning they will be allocated next, turning the just-added memory usually immediately un-removable. The fresh pages are cold, prefering to allocate others (that might be hot) also feels to be the natural thing to do.
It also applies to the hyper-v balloon xen-balloon, and ppc64 dlpar: whenQuoted 8 lines
adding separate, successive memory blocks, each memory block will have unmovable allocations on them - for example gigantic pages will fail to allocate.
While theZONE_NORMALdoesn't provide any guarantees that memory can get offlined+removed again (any kind of fragmentation with unmovable allocations is possible), there are many scenarios (hotplugging a lot of memory, running workload, hotunplug some memory/as much as possible) where we can offline+remove quite a lot with this patchset.Quoted 2 lines
Hi David,
Hi Oscar.
Quoted 1 line
I did not read through the patchset yet, so sorry if the question is nonsense, but is this not trying to fix the same issue the vmemmap patches did? [1]
Not nonesense at all. It only helps to some degree, though. It solves the dependencies due to the memmap. However, it‘s not completely ideal, especially for single memory blocks.
Quoted 7 lines
With single memory blocks (virtio-mem, xen-balloon, hv balloon, ppc dlpar) you still have unmovable (vmemmap chunks) all over the physical address space. Consider the gigantic page example after hotplug. You directly fragmented all hotplugged memory.
Of course, there might be (less extreme) dependencies due page tables for the identity mapping, extended struct pages and similar.
Having that said, there are other benefits when preferring other memory over just hotplugged memory. Think about adding+onlining memory during boot (dimms under QEMU, virtio-mem), once the system is up you will have most (all) of that memory completely untouched.
So while vmemmap on hotplugged memory would tackle some part of the issue, there are cases where this approach is better, and there are even benefits when combining both.
While everything changes with shuffle.
Quoted 10 lines
Thanks!
David
Quoted 5 lines
I was about to give it a new respin now that thw hwpoison stuff has been settled.
Signature · 2 lines
Wei Yang Help you, Help me
Re: [PATCH RFC 1/4] mm/page_alloc: convert "report" flag of __free_one_page() to a proper flag
In reply to Wei YangQuoted 39 lines · Wei Yang
Quoted 25 lines · David Hildenbrand
Let's prepare for additional flags and avoid long parameter lists of bools. Follow-up patches will also make use of the flags in
__free_pages_ok(), however, I wasn't able to come up with a better name for the type - should be good enough for internal purposes.Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Signed-off-by David Hildenbrand <[email protected]>---
mm/page_alloc.c| 28 ++++++++++++++++++++--------1 file changed, 20 insertions(+), 8 deletions(-)mm/page_alloc.c +0
-0@@ -77,6 +77,18 @@#include "shuffle.h" mm/page_alloc.c line 87
Quoted 10 lines
mm/page_alloc.c +8
-0#include "page_reporting.h" +/* Free One Page flags: for internal, non-pcp variants of free_pages(). */ +typedef int __bitwise fop_t ; + +/* No special request */ +#define FOP_NONE ((__force fop_t)0) + +/* +* Skip free page reporting notification after buddy merging (will *not* mark
__free_one_page()may not merge buddy when its buddy is not available.
Quoted 2 lines
Would this comment be a little confusing?
I rather meant the process than if it's actually happening.
"Skip free page reporting notification for the (possibly merged) page."
Thanks!
Signature · 3 lines
Thanks, David / dhildenb
Re: [PATCH RFC 2/4] mm/page_alloc: place pages to tail in __putback_isolated_page()
In reply to Wei YangQuoted 14 lines · Wei Yang
Quoted 13 lines · David Hildenbrand
__putback_isolated_page()already documents that pages will be placed to the tail of the freelist - this is, however, not the case for "order >=MAX_ORDER- 2" (seebuddy_merge_likely()) - which should be the case for all existing users.This change affects two users:
- free page reporting
- page isolation, when undoing the isolation.This behavior is desireable for pages that haven't really been touched lately, so exactly the two users that don't actually read/write page content, but rather move untouched pages.
Quoted 8 lines
Quoted 3 lines
The new behavior is especially desirable for memory onlining, where we allow allocation of newly onlined pages via
undo_isolate_page_range()inonline_pages(). Right now, we always place them to the head of theThe code looks good, while I don't fully understand the log here.
undo_isolate_page_range()is used in__offline_pagesandalloc_contig_range. I don't connect them withonline_pages(). Do I miss something?
Yeah, please look at -mm / -next instead. See
https://lkml.kernel.org/r/[email protected]
Signature · 3 lines
Thanks, David / dhildenb
Re: [PATCH RFC 2/4] mm/page_alloc: place pages to tail in __putback_isolated_page()
In reply to Wei YangQuoted 73 lines · Wei Yang
Quoted 64 lines · David Hildenbrand
__putback_isolated_page()already documents that pages will be placed to the tail of the freelist - this is, however, not the case for "order >=MAX_ORDER- 2" (seebuddy_merge_likely()) - which should be the case for all existing users.This change affects two users:
- free page reporting
- page isolation, when undoing the isolation.This behavior is desireable for pages that haven't really been touched lately, so exactly the two users that don't actually read/write page content, but rather move untouched pages.
The new behavior is especially desirable for memory onlining, where we allow allocation of newly onlined pages via
undo_isolate_page_range()inonline_pages(). Right now, we always place them to the head of the free list, resulting in undesireable behavior: Assume we add individual memory chunks viaadd_memory()and online them right away to the NORMAL zone. We create a dependency chain of unmovable allocations e.g., via the memmap. The memmap of the next chunk will be placed onto previous chunks - if the last block cannot get offlined+removed, all dependent ones cannot get offlined+removed. While this can already be observed with individual DIMMs, it's more of an issue for virtio-mem (and I suspect also ppc DLPAR).Note: If we observe a degradation due to the changed page isolation behavior (which I doubt), we can always make this configurable by the instance triggering undo of isolation (e.g.,
alloc_contig_range(), memory onlining, memory offlining).Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Cc Scott Cheloha <[email protected]>Cc Michael Ellerman <[email protected]>Signed-off-by David Hildenbrand <[email protected]>---
mm/page_alloc.c| 10 +++++++++-1 file changed, 9 insertions(+), 1 deletion(-)mm/page_alloc.c +6
-0@@ -89,6 +89,12 @@ typedef int __bitwise fop_t;*/ #define FOP_SKIP_REPORT_NOTIFY ((__force fop_t)BIT(0)) +/* +* Place the freed page to the tail of the freelist after buddy merging. Will +* get ignored with page shuffling enabled. +*/ +#define FOP_TO_TAIL ((__force fop_t)BIT(1)) +/* prevent >1 _updater_ of zone percpu pageset ->high and ->batch fields */ static DEFINE_MUTEX (pcp_batch_high_lock);#define MIN_PERCPU_PAGELIST_FRACTION (8)mm/page_alloc.c static inline void __free_one_page(struct page *page, unsigned long pfn, · line 1050
Quoted 5 lines
mm/page_alloc.c +2
-0@@ -1040,6 +1046,8 @@ static inline void __free_one_page(struct page *page, unsigned long pfn,if (is_shuffle_order (order)) to_tail =shuffle_pick_tail (); +else if (fop_flags &FOP_TO_TAIL ) + to_tail = true;Take another look into this part. Maybe we can move this check at top?
Quoted 8 lines
For
online_pagecase, currently we have following call flow:
online_page
online_pages_range
shuffle_zoneThis means we would always shuffle the newly added pages. Maybe we don't need to do the shuffle when adding them to the
free_list?
Yeah we don't, but it doesn't really buy us too much as the call paths I am touching are used by other mechanisms as well that need it (especially undoing page isolation).
Signature · 3 lines
Thanks, David / dhildenb
Re: [PATCH RFC 3/4] mm/page_alloc: always move pages to the tail of the freelist in unset_migratetype_isolate()
In reply to Wei YangQuoted 8 lines · Wei Yang
Quoted 7 lines · David Hildenbrand
Page isolation doesn't actually touch the pages, it simply isolates pageblocks and moves all free pages to the
MIGRATE_ISOLATEfreelist.We already place pages to the tail of the freelists when undoing isolation via
__putback_isolated_page(), let's do it in any case (e.g., if order ==pageblock_order) and document the behavior.
Quoted 6 lines
Quoted 2 lines
This change results in all pages getting onlined via
online_pages()to be placed to the tail of the freelist.I am sorry to not follow again.
unset_migratetype_isolate()is used in__offline_pagesif my understanding is correct. How does it contribute ononline_pages?
See -next / -mm, that should make it clearer.
Signature · 3 lines
Thanks, David / dhildenb
Quoted 39 lines · Wei Yang
Quoted 38 lines · David Hildenbrand
Quoted 27 lines
Am 16.09.2020 um 20:50 schrieb [email protected]:
Quoted 13 lines · David Hildenbrand
When adding separate memory blocks via add_memory*() and onlining them immediately, the metadata (especially the memmap) of the next block will be placed onto one of the just added+onlined block. This creates a chain of unmovable allocations: If the last memory block cannot get offlined+removed() so will all dependant ones. We directly have unmovable allocations all over the place.
This can be observed quite easily using virtio-mem, however, it can also be observed when using DIMMs. The freshly onlined pages will usually be placed to the head of the freelists, meaning they will be allocated next, turning the just-added memory usually immediately un-removable. The fresh pages are cold, prefering to allocate others (that might be hot) also feels to be the natural thing to do.
It also applies to the hyper-v balloon xen-balloon, and ppc64 dlpar: whenQuoted 8 lines
adding separate, successive memory blocks, each memory block will have unmovable allocations on them - for example gigantic pages will fail to allocate.
While theZONE_NORMALdoesn't provide any guarantees that memory can get offlined+removed again (any kind of fragmentation with unmovable allocations is possible), there are many scenarios (hotplugging a lot of memory, running workload, hotunplug some memory/as much as possible) where we can offline+remove quite a lot with this patchset.Quoted 2 lines
Hi David,
Hi Oscar.
Quoted 1 line
I did not read through the patchset yet, so sorry if the question is nonsense, but is this not trying to fix the same issue the vmemmap patches did? [1]
Not nonesense at all. It only helps to some degree, though. It solves the dependencies due to the memmap. However, it‘s not completely ideal, especially for single memory blocks.
With single memory blocks (virtio-mem, xen-balloon, hv balloon, ppc dlpar) you still have unmovable (vmemmap chunks) all over the physical address space. Consider the gigantic page example after hotplug. You directly fragmented all hotplugged memory.
Quoted 8 lines
Quoted 5 lines
Of course, there might be (less extreme) dependencies due page tables for the identity mapping, extended struct pages and similar.
Having that said, there are other benefits when preferring other memory over just hotplugged memory. Think about adding+onlining memory during boot (dimms under QEMU, virtio-mem), once the system is up you will have most (all) of that memory completely untouched.
So while vmemmap on hotplugged memory would tackle some part of the issue, there are cases where this approach is better, and there are even benefits when combining both.
While everything changes with shuffle.
Right. Shuffling would naturally try to break the dependencies.
Shuffling is quite rare though, it has to be enabled explicitly on the cmdline and might not be of too much help in virtualized environments.
Signature · 3 lines
Thanks, David / dhildenb
Re: [PATCH RFC 2/4] mm/page_alloc: place pages to tail in __putback_isolated_page()
In reply to David HildenbrandQuoted 19 lines · David Hildenbrand
Quoted 18 lines · Wei Yang
Quoted 16 lines · David Hildenbrand
__putback_isolated_page()already documents that pages will be placed to the tail of the freelist - this is, however, not the case for "order >=MAX_ORDER- 2" (seebuddy_merge_likely()) - which should be the case for all existing users.This change affects two users:
- free page reporting
- page isolation, when undoing the isolation.This behavior is desireable for pages that haven't really been touched lately, so exactly the two users that don't actually read/write page content, but rather move untouched pages.
The new behavior is especially desirable for memory onlining, where we allow allocation of newly onlined pages via
undo_isolate_page_range()inonline_pages(). Right now, we always place them to the head of the
Quoted 9 lines
Quoted 4 lines
The code looks good, while I don't fully understand the log here.
undo_isolate_page_range()is used in__offline_pagesandalloc_contig_range. I don't connect them withonline_pages(). Do I miss something?Yeah, please look at -mm / -next instead. See
Thanks, I get the point.
Quoted 5 lines
Signature · 4 lines
-- Thanks, David / dhildenb
Signature · 2 lines
Wei Yang Help you, Help me
Quoted 36 lines · David Hildenbrand
Quoted 27 lines
Am 16.09.2020 um 20:50 schrieb [email protected]:
Quoted 13 lines · David Hildenbrand
When adding separate memory blocks via add_memory*() and onlining them immediately, the metadata (especially the memmap) of the next block will be placed onto one of the just added+onlined block. This creates a chain of unmovable allocations: If the last memory block cannot get offlined+removed() so will all dependant ones. We directly have unmovable allocations all over the place.
This can be observed quite easily using virtio-mem, however, it can also be observed when using DIMMs. The freshly onlined pages will usually be placed to the head of the freelists, meaning they will be allocated next, turning the just-added memory usually immediately un-removable. The fresh pages are cold, prefering to allocate others (that might be hot) also feels to be the natural thing to do.
It also applies to the hyper-v balloon xen-balloon, and ppc64 dlpar: whenQuoted 8 lines
adding separate, successive memory blocks, each memory block will have unmovable allocations on them - for example gigantic pages will fail to allocate.
While theZONE_NORMALdoesn't provide any guarantees that memory can get offlined+removed again (any kind of fragmentation with unmovable allocations is possible), there are many scenarios (hotplugging a lot of memory, running workload, hotunplug some memory/as much as possible) where we can offline+remove quite a lot with this patchset.Quoted 2 lines
Hi David,
Hi Oscar.
Quoted 1 line
I did not read through the patchset yet, so sorry if the question is nonsense, but is this not trying to fix the same issue the vmemmap patches did? [1]
Not nonesense at all. It only helps to some degree, though. It solves the dependencies due to the memmap. However, it‘s not completely ideal, especially for single memory blocks.
Quoted 7 lines
With single memory blocks (virtio-mem, xen-balloon, hv balloon, ppc dlpar) you still have unmovable (vmemmap chunks) all over the physical address space. Consider the gigantic page example after hotplug. You directly fragmented all hotplugged memory.
Of course, there might be (less extreme) dependencies due page tables for the identity mapping, extended struct pages and similar.
Having that said, there are other benefits when preferring other memory over just hotplugged memory. Think about adding+onlining memory during boot (dimms under QEMU, virtio-mem), once the system is up you will have most (all) of that memory completely untouched.
So while vmemmap on hotplugged memory would tackle some part of the issue, there are cases where this approach is better, and there are even benefits when combining both.
I see the point, but I don't think the head/tail mechanism is great for this. It might sort of work, but with other interfering activity there are no guarantees and it relies on a subtle implementation detail. There are better mechanisms possible I think, such as preparing a larger MIGRATE_UNMOVABLE area in the existing memory before we allocate those long-term management structures. Or onlining a bunch of blocks as zone_movable first and only later convert to zone_normal in a controlled way when existing normal zone becomes depeted?
I guess it's an issue that the e.g. 128M block onlines are so disconnected from each other it's hard to employ a strategy that works best for e.g. a whole bunch of GB onlined at once. But I noticed some effort towards new API, so maybe that will be solved there too?
Quoted 10 lines
Thanks!
David
Quoted 5 lines
I was about to give it a new respin now that thw hwpoison stuff has been settled.
Quoted 37 lines · Vlastimil Babka
Quoted 36 lines · David Hildenbrand
Quoted 27 lines
Am 16.09.2020 um 20:50 schrieb [email protected]:
Quoted 13 lines · David Hildenbrand
When adding separate memory blocks via add_memory*() and onlining them immediately, the metadata (especially the memmap) of the next block will be placed onto one of the just added+onlined block. This creates a chain of unmovable allocations: If the last memory block cannot get offlined+removed() so will all dependant ones. We directly have unmovable allocations all over the place.
This can be observed quite easily using virtio-mem, however, it can also be observed when using DIMMs. The freshly onlined pages will usually be placed to the head of the freelists, meaning they will be allocated next, turning the just-added memory usually immediately un-removable. The fresh pages are cold, prefering to allocate others (that might be hot) also feels to be the natural thing to do.
It also applies to the hyper-v balloon xen-balloon, and ppc64 dlpar: whenQuoted 8 lines
adding separate, successive memory blocks, each memory block will have unmovable allocations on them - for example gigantic pages will fail to allocate.
While theZONE_NORMALdoesn't provide any guarantees that memory can get offlined+removed again (any kind of fragmentation with unmovable allocations is possible), there are many scenarios (hotplugging a lot of memory, running workload, hotunplug some memory/as much as possible) where we can offline+remove quite a lot with this patchset.Quoted 2 lines
Hi David,
Hi Oscar.
Quoted 1 line
I did not read through the patchset yet, so sorry if the question is nonsense, but is this not trying to fix the same issue the vmemmap patches did? [1]
Not nonesense at all. It only helps to some degree, though. It solves the dependencies due to the memmap. However, it‘s not completely ideal, especially for single memory blocks.
Quoted 8 lines
Quoted 7 lines
With single memory blocks (virtio-mem, xen-balloon, hv balloon, ppc dlpar) you still have unmovable (vmemmap chunks) all over the physical address space. Consider the gigantic page example after hotplug. You directly fragmented all hotplugged memory.
Of course, there might be (less extreme) dependencies due page tables for the identity mapping, extended struct pages and similar.
Having that said, there are other benefits when preferring other memory over just hotplugged memory. Think about adding+onlining memory during boot (dimms under QEMU, virtio-mem), once the system is up you will have most (all) of that memory completely untouched.
So while vmemmap on hotplugged memory would tackle some part of the issue, there are cases where this approach is better, and there are even benefits when combining both.
Hi Vlastimil,
Quoted 3 lines
I see the point, but I don't think the head/tail mechanism is great for this. It might sort of work, but with other interfering activity there are no guarantees and it relies on a subtle implementation detail. There are better mechanisms
For the specified use case of adding+onlining a whole bunch of memory this works just fine. We don't care too much about "other interfering activity" as you mention here, or about guarantees - this is a pure optimization that seems to work just fine in practice.
I'm not sure about the "subtle implementation detail" - buddy merging, and head/tail of buddy lists are a basic concept of our page allocator. If that would ever change, the optimization here would be lost and we would have to think of something else. Nothing would actually break - and it's all kept directly in page_alloc.c
I'd like to stress that what I propose here is both simple and powerful.
Quoted 4 lines
possible I think, such as preparing a larger
MIGRATE_UNMOVABLEarea in the existing memory before we allocate those long-term management structures. Or onlining a bunch of blocks aszone_movablefirst and only later convert tozone_normalin a controlled way when existing normal zone becomes depeted?
I see the following (more or less complicated) alternatives
1) Having a larger MIGRATE_UNMOVABLE area
a) Sizing it is difficult. I mean you would have to plan ahead for all memory you might eventually hotplug later - and that could even be impossible if you hotplug quite a lot of memory to a smaller machine. (I've seen people in the vm/container world trying to hotplug 128GB DIMMs to 2GB VMs ... and failing for obvious reasons)
b) not really desired. You usually want to have most memory movable, not the opposite (just because you might hotplug memory in small chunks later).
2) smarter onlining
I have prototype patches for better auto-onlining (which I'll share at some point), where I balance between ZONE_NORMAL and ZONE_MOVABLE in a defined ratio. Assuming something very simple, adding separate memory blocks and onlining them based on the current zone ratio (assuming a 1:4 normal:movable target ratio) would (without some other policies I have in place) result in something like this for hotplugged memory (via virtio-mem):
[N][M][M][M][M][N][M][M][M][M][N][M][M][M][M]...
(note: layout is suboptimal, just a simple example)
But even here, all [N] memory blocks would immediately be use for allocations for the memmap of successive blocks. It doesn't solve the dependency issues.
Now assume we would want to group [N] in a way to allow for gigantic pages, like
[N][N][N][N][N][N][N][N][M][M][M][M] ....
we would, once again, never be able to allocate a gigantic page because all [N] would contain a memmap.
3) conversion from MOVABLE -> NORMAL
While a conversion from MOVABLE to NORMAL would be interesting to see, it's going to be a challenging task to actually implement (people expect that page_zone() remains stable). Without any hacks, we'd have to
1. offline the selected (MOVABLE) memory block/chunk
2. online the selected memory block/chunk to the NORMAL zone
This is not something we can do out of random context (for example, we need both, the device hotplug lock and the memory hotplug lock, as we might race with user space) - so there might still be a chance of corner-case OOMs.
(I assume there could also be quite a negative performance impact when always relying on the conversion, and not properly planning ahead as in 2.)
Quoted 5 lines
I guess it's an issue that the e.g. 128M block onlines are so disconnected from each other it's hard to employ a strategy that works best for e.g. a whole bunch of GB onlined at once. But I noticed some effort towards new API, so maybe that will be solved there too?
While new interfaces might make it easier to identify boundaries of separate DIMMs (e.g., to online a single DIMM either movable or unmovable - which can partially be done right now when going via memory resource boundaries), it doesn't help for the use case of adding separate memory blocks.
So while having an automatic conversion from MOVABLE -> NORMAL would be interesting, I doubt we'll see it in the foreseeable future. Are there any similarly simple alternatives to optimize this?
Thanks!
Signature · 3 lines
Thanks, David / dhildenb
Quoted 45 lines · Vlastimil Babka
Quoted 36 lines · David Hildenbrand
Quoted 27 lines
Am 16.09.2020 um 20:50 schrieb [email protected]:
Quoted 13 lines · David Hildenbrand
When adding separate memory blocks via add_memory*() and onlining them immediately, the metadata (especially the memmap) of the next block will be placed onto one of the just added+onlined block. This creates a chain of unmovable allocations: If the last memory block cannot get offlined+removed() so will all dependant ones. We directly have unmovable allocations all over the place.
This can be observed quite easily using virtio-mem, however, it can also be observed when using DIMMs. The freshly onlined pages will usually be placed to the head of the freelists, meaning they will be allocated next, turning the just-added memory usually immediately un-removable. The fresh pages are cold, prefering to allocate others (that might be hot) also feels to be the natural thing to do.
It also applies to the hyper-v balloon xen-balloon, and ppc64 dlpar: whenQuoted 8 lines
adding separate, successive memory blocks, each memory block will have unmovable allocations on them - for example gigantic pages will fail to allocate.
While theZONE_NORMALdoesn't provide any guarantees that memory can get offlined+removed again (any kind of fragmentation with unmovable allocations is possible), there are many scenarios (hotplugging a lot of memory, running workload, hotunplug some memory/as much as possible) where we can offline+remove quite a lot with this patchset.Quoted 2 lines
Hi David,
Hi Oscar.
Quoted 1 line
I did not read through the patchset yet, so sorry if the question is nonsense, but is this not trying to fix the same issue the vmemmap patches did? [1]
Not nonesense at all. It only helps to some degree, though. It solves the dependencies due to the memmap. However, it‘s not completely ideal, especially for single memory blocks.
Quoted 7 lines
With single memory blocks (virtio-mem, xen-balloon, hv balloon, ppc dlpar) you still have unmovable (vmemmap chunks) all over the physical address space. Consider the gigantic page example after hotplug. You directly fragmented all hotplugged memory.
Of course, there might be (less extreme) dependencies due page tables for the identity mapping, extended struct pages and similar.
Having that said, there are other benefits when preferring other memory over just hotplugged memory. Think about adding+onlining memory during boot (dimms under QEMU, virtio-mem), once the system is up you will have most (all) of that memory completely untouched.
So while vmemmap on hotplugged memory would tackle some part of the issue, there are cases where this approach is better, and there are even benefits when combining both.
Quoted 8 lines
I see the point, but I don't think the head/tail mechanism is great for this. It might sort of work, but with other interfering activity there are no guarantees and it relies on a subtle implementation detail. There are better mechanisms possible I think, such as preparing a larger
MIGRATE_UNMOVABLEarea in the existing memory before we allocate those long-term management structures. Or onlining a bunch of blocks aszone_movablefirst and only later convert tozone_normalin a controlled way when existing normal zone becomes depeted?
To be honest, David's approach is easy to understand for me.
And I don't see some negative effect.
Quoted 15 lines
I guess it's an issue that the e.g. 128M block onlines are so disconnected from each other it's hard to employ a strategy that works best for e.g. a whole bunch of GB onlined at once. But I noticed some effort towards new API, so maybe that will be solved there too?
Quoted 10 lines
Thanks!
David
Quoted 5 lines
I was about to give it a new respin now that thw hwpoison stuff has been settled.
Signature · 2 lines
Wei Yang Help you, Help me
Quoted 52 lines · David Hildenbrand
Quoted 32 lines
Quoted 32 lines
Quoted 23 lines
Quoted 13 lines · David Hildenbrand
When adding separate memory blocks via add_memory*() and onlining them immediately, the metadata (especially the memmap) of the next block will be placed onto one of the just added+onlined block. This creates a chain of unmovable allocations: If the last memory block cannot get offlined+removed() so will all dependant ones. We directly have unmovable allocations all over the place.
This can be observed quite easily using virtio-mem, however, it can also be observed when using DIMMs. The freshly onlined pages will usually be placed to the head of the freelists, meaning they will be allocated next, turning the just-added memory usually immediately un-removable. The fresh pages are cold, prefering to allocate others (that might be hot) also feels to be the natural thing to do.
It also applies to the hyper-v balloon xen-balloon, and ppc64 dlpar: whenQuoted 8 lines
adding separate, successive memory blocks, each memory block will have unmovable allocations on them - for example gigantic pages will fail to allocate.
While theZONE_NORMALdoesn't provide any guarantees that memory can get offlined+removed again (any kind of fragmentation with unmovable allocations is possible), there are many scenarios (hotplugging a lot of memory, running workload, hotunplug some memory/as much as possible) where we can offline+remove quite a lot with this patchset.Quoted 2 lines
Hi David,
Hi Oscar.
Quoted 1 line
I did not read through the patchset yet, so sorry if the question is nonsense, but is this not trying to fix the same issue the vmemmap patches did? [1]
Not nonesense at all. It only helps to some degree, though. It solves the dependencies due to the memmap. However, it???s not completely ideal, especially for single memory blocks.
Quoted 8 lines
Quoted 7 lines
With single memory blocks (virtio-mem, xen-balloon, hv balloon, ppc dlpar) you still have unmovable (vmemmap chunks) all over the physical address space. Consider the gigantic page example after hotplug. You directly fragmented all hotplugged memory.
Of course, there might be (less extreme) dependencies due page tables for the identity mapping, extended struct pages and similar.
Having that said, there are other benefits when preferring other memory over just hotplugged memory. Think about adding+onlining memory during boot (dimms under QEMU, virtio-mem), once the system is up you will have most (all) of that memory completely untouched.
So while vmemmap on hotplugged memory would tackle some part of the issue, there are cases where this approach is better, and there are even benefits when combining both.
Hi Vlastimil,
Quoted 3 lines
I see the point, but I don't think the head/tail mechanism is great for this. It might sort of work, but with other interfering activity there are no guarantees and it relies on a subtle implementation detail. There are better mechanisms
For the specified use case of adding+onlining a whole bunch of memory this works just fine. We don't care too much about "other interfering activity" as you mention here, or about guarantees - this is a pure optimization that seems to work just fine in practice.
Quoted 6 lines
I'm not sure about the "subtle implementation detail" - buddy merging, and head/tail of buddy lists are a basic concept of our page allocator. If that would ever change, the optimization here would be lost and we would have to think of something else. Nothing would actually break - and it's all kept directly in
page_alloc.c
It's somewhat subtle because it's relying heavily on the exact ordering of how pages are pulled from the free lists at the moment. Lets say for example that someone was brave enough to tackle the problem of the giant zone lock and split the zone into allocation arenas (like what glibc does to split the lock). Depending on the exact ordering of how pages are added and removed from the list would break your approach. I'm wary of anything that relies on the ordering of freelists for correctness becauuse it limits the ability to fix the zone lock (which has been overdue for fixing for years now and getting worse as node sizes increase).
To be robust, you'd need to do something like early memory bring-up whereby pages are directly allocated from one part of the DIMM (presumably the start) and use that for the metadata -- potentially all the metadata that would be necessary to plug/unplug the entire DIMM. This would effectively be unmovable but if you want to guarantee that all the memory except the metadata can be unplugged, you do not have much alteratives. Playing games with the ordering of the freelists will simply end up as "sometimes works, sometimes does not".
In terms of forcing ranges to be UNMOVABLE or MOVABLE (either via zones or by implementing "sticky" pageblocks which hits complex reclaim-related problems), you start running into problems similar to lowmem starvation where a page cache allocation fails because unmovable metadata cannot be allocated.
I suggest you keep it simple -- statically allocate the potential metadata needed in the future even though it limits the maximum amount of memory that can be unplugged. The alternative is unpredictable plug/unplug success rates.
Signature · 2 lines
Mel Gorman SUSE Labs
Quoted 53 lines · Mel Gorman
Quoted 52 lines · David Hildenbrand
Quoted 32 lines
Quoted 32 lines
Quoted 23 lines
Quoted 13 lines · David Hildenbrand
When adding separate memory blocks via add_memory*() and onlining them immediately, the metadata (especially the memmap) of the next block will be placed onto one of the just added+onlined block. This creates a chain of unmovable allocations: If the last memory block cannot get offlined+removed() so will all dependant ones. We directly have unmovable allocations all over the place.
This can be observed quite easily using virtio-mem, however, it can also be observed when using DIMMs. The freshly onlined pages will usually be placed to the head of the freelists, meaning they will be allocated next, turning the just-added memory usually immediately un-removable. The fresh pages are cold, prefering to allocate others (that might be hot) also feels to be the natural thing to do.
It also applies to the hyper-v balloon xen-balloon, and ppc64 dlpar: whenQuoted 8 lines
adding separate, successive memory blocks, each memory block will have unmovable allocations on them - for example gigantic pages will fail to allocate.
While theZONE_NORMALdoesn't provide any guarantees that memory can get offlined+removed again (any kind of fragmentation with unmovable allocations is possible), there are many scenarios (hotplugging a lot of memory, running workload, hotunplug some memory/as much as possible) where we can offline+remove quite a lot with this patchset.Quoted 2 lines
Hi David,
Hi Oscar.
Quoted 1 line
I did not read through the patchset yet, so sorry if the question is nonsense, but is this not trying to fix the same issue the vmemmap patches did? [1]
Not nonesense at all. It only helps to some degree, though. It solves the dependencies due to the memmap. However, it???s not completely ideal, especially for single memory blocks.
Quoted 8 lines
Quoted 7 lines
With single memory blocks (virtio-mem, xen-balloon, hv balloon, ppc dlpar) you still have unmovable (vmemmap chunks) all over the physical address space. Consider the gigantic page example after hotplug. You directly fragmented all hotplugged memory.
Of course, there might be (less extreme) dependencies due page tables for the identity mapping, extended struct pages and similar.
Having that said, there are other benefits when preferring other memory over just hotplugged memory. Think about adding+onlining memory during boot (dimms under QEMU, virtio-mem), once the system is up you will have most (all) of that memory completely untouched.
So while vmemmap on hotplugged memory would tackle some part of the issue, there are cases where this approach is better, and there are even benefits when combining both.
Hi Vlastimil,
Quoted 3 lines
I see the point, but I don't think the head/tail mechanism is great for this. It might sort of work, but with other interfering activity there are no guarantees and it relies on a subtle implementation detail. There are better mechanisms
For the specified use case of adding+onlining a whole bunch of memory this works just fine. We don't care too much about "other interfering activity" as you mention here, or about guarantees - this is a pure optimization that seems to work just fine in practice.
Quoted 6 lines
Quoted 6 lines
I'm not sure about the "subtle implementation detail" - buddy merging, and head/tail of buddy lists are a basic concept of our page allocator. If that would ever change, the optimization here would be lost and we would have to think of something else. Nothing would actually break - and it's all kept directly in
page_alloc.c
Hi Mel,
thanks for your reply.
Quoted 7 lines
It's somewhat subtle because it's relying heavily on the exact ordering of how pages are pulled from the free lists at the moment. Lets say for example that someone was brave enough to tackle the problem of the giant zone lock and split the zone into allocation arenas (like what glibc does to split the lock). Depending on the exact ordering of how pages are added and removed from the list would break your approach. I'm wary of
First of all, it would not break it (as I already said). The optimization would be lost. Totally acceptable.
However, I assume we would apply the same technique (optimized buddy merging - placing to head/tail, page shuffling) on these allocation arenas. So the optimization would still mostly apply, just in different granularity - which would be fine.
Quoted 3 lines
anything that relies on the ordering of freelists for correctness becauuse it limits the ability to fix the zone lock (which has been overdue for fixing for years now and getting worse as node sizes increase).
"for correctness" - no, this is an optimization. As I said, there are no guarantees. Please keep that in mind.
(also, page shuffling relies on the ordering of freelists right now ... for correctness)
Quoted 9 lines
To be robust, you'd need to do something like early memory bring-up whereby pages are directly allocated from one part of the DIMM (presumably the start) and use that for the metadata -- potentially all the metadata that would be necessary to plug/unplug the entire DIMM. This would effectively be unmovable but if you want to guarantee that all the memory except the metadata can be unplugged, you do not have much alteratives. Playing games with the ordering of the freelists will simply end up as "sometimes works, sometimes does not".
As answered to Oscar already, while something like that might be feasible for DIMMs in the future (and there are still quite some issues to be sorted out), it isn't always desired adding separate (small - e.g., 128MB) memory blocks. You - again- have unmovable allocations all over the place that won't allow you to allocate any gigantic page.
Quoted 6 lines
In terms of forcing ranges to be UNMOVABLE or MOVABLE (either via zones or by implementing "sticky" pageblocks which hits complex reclaim-related problems), you start running into problems similar to lowmem starvation where a page cache allocation fails because unmovable metadata cannot be allocated.
Exactly.
Quoted 6 lines
I suggest you keep it simple -- statically allocate the potential metadata needed in the future even though it limits the maximum amount of memory that can be unplugged. The alternative is unpredictable plug/unplug success rates.
I'm sorry I can't follow. How is this "simple"? Or even "simpler" than what I suggest?
And as I said, it doesn't always work. Assume I hotplug 128GB to a 2GB machine via virtio-mem (which works just fine, as we add+online memory in small chunks compared to a single, huge DIMM), I would have to pre-allocate 2GB just for the memmap - which obviously doesn't work.
Again, I'd like to stress that this is a pure optimization that I am proposing - nothing would "break" when ripping it out again, except that we lose the optimizations I mentioned.
Signature · 3 lines
Thanks, David / dhildenb
Re: [PATCH RFC 1/4] mm/page_alloc: convert "report" flag of __free_one_page() to a proper flag
In reply to David HildenbrandQuoted 5 lines · David Hildenbrand
Let's prepare for additional flags and avoid long parameter lists of bools. Follow-up patches will also make use of the flags in
__free_pages_ok(), however, I wasn't able to come up with a better name for the type - should be good enough for internal purposes.
Quoted 10 lines
Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Signed-off-by David Hildenbrand <[email protected]>
Quoted 79 lines
--- mm/page_alloc.c | 28 ++++++++++++++++++++-------- 1 file changed, 20 insertions(+), 8 deletions(-)mm/page_alloc.c +20
-8@@ -77,6 +77,18 @@#include "shuffle.h" #include "page_reporting.h" +/* Free One Page flags: for internal, non-pcp variants of free_pages(). */ +typedef int __bitwise fop_t ; + +/* No special request */ +#define FOP_NONE ((__force fop_t)0) + +/* +* Skip free page reporting notification after buddy merging (will *not* mark +* the page reported, only skip the notification). +*/ +#define FOP_SKIP_REPORT_NOTIFY ((__force fop_t)BIT(0)) +/* prevent >1 _updater_ of zone percpu pageset ->high and ->batch fields */ static DEFINE_MUTEX (pcp_batch_high_lock);#define MIN_PERCPU_PAGELIST_FRACTION (8)@@ -948,10 +960,9 @@ buddy_merge_likely(unsigned long pfn, unsigned long buddy_pfn,* -- nyc */ -static inline void __free_one_page (struct page *page, -unsigned long pfn, -struct zone *zone,unsigned int order, -int migratetype,bool report) +static inline void __free_one_page (struct page *page,unsigned long pfn, +struct zone *zone,unsigned int order, +int migratetype,fop_t fop_flags) {struct capture_control *capc =task_capc (zone);unsigned long buddy_pfn;@@ -1038,7 +1049,7 @@ static inline void __free_one_page(struct page *page,add_to_free_list (page, zone, order, migratetype);/* Notify page reporting subsystem of freed page */ -if (report) +if (!(fop_flags &FOP_SKIP_REPORT_NOTIFY ))page_reporting_notify_free (order); }@@ -1368,7 +1379,7 @@ static void free_pcppages_bulk(struct zone *zone, int count,if (unlikely (isolated_pageblocks)) mt =get_pageblock_migratetype (page); -__free_one_page (page,page_to_pfn (page), zone,0 , mt, true); +__free_one_page (page,page_to_pfn (page), zone,0 , mt,FOP_NONE );trace_mm_page_pcpu_drain (page,0 , mt); }spin_unlock (&zone->lock);@@ -1384,7 +1395,7 @@ static void free_one_page(struct zone *zone, is_migrate_isolate(migratetype))) { migratetype =get_pfnblock_migratetype (page, pfn); } -__free_one_page (page, pfn, zone, order, migratetype, true); +__free_one_page (page, pfn, zone, order, migratetype,FOP_NONE );spin_unlock (&zone->lock); }@@ -3277,7 +3288,8 @@ void __putback_isolated_page(struct page *page, unsigned int order, int mt)lockdep_assert_held (&zone->lock);/* Return isolated page to tail of freelist. */ -__free_one_page (page,page_to_pfn (page), zone, order, mt, false); +__free_one_page (page,page_to_pfn (page), zone, order, mt, +FOP_SKIP_REPORT_NOTIFY ); }/*
Re: [PATCH RFC 2/4] mm/page_alloc: place pages to tail in __putback_isolated_page()
In reply to David HildenbrandQuoted 4 lines · David Hildenbrand
__putback_isolated_page()already documents that pages will be placed to the tail of the freelist - this is, however, not the case for "order >=MAX_ORDER- 2" (seebuddy_merge_likely()) - which should be the case for all existing users.
I think here should be a sentence saying something along "Thus this patch introduces a FOP_TO_TAIL flag to really ensure moving pages to tail."
Quoted 59 lines
This change affects two users:
- free page reporting
- page isolation, when undoing the isolation.This behavior is desireable for pages that haven't really been touched lately, so exactly the two users that don't actually read/write page content, but rather move untouched pages.
The new behavior is especially desirable for memory onlining, where we allow allocation of newly onlined pages via
undo_isolate_page_range()inonline_pages(). Right now, we always place them to the head of the free list, resulting in undesireable behavior: Assume we add individual memory chunks viaadd_memory()and online them right away to the NORMAL zone. We create a dependency chain of unmovable allocations e.g., via the memmap. The memmap of the next chunk will be placed onto previous chunks - if the last block cannot get offlined+removed, all dependent ones cannot get offlined+removed. While this can already be observed with individual DIMMs, it's more of an issue for virtio-mem (and I suspect also ppc DLPAR).Note: If we observe a degradation due to the changed page isolation behavior (which I doubt), we can always make this configurable by the instance triggering undo of isolation (e.g.,
alloc_contig_range(), memory onlining, memory offlining).Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Cc Scott Cheloha <[email protected]>Cc Michael Ellerman <[email protected]>Signed-off-by David Hildenbrand <[email protected]>--- mm/page_alloc.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-)mm/page_alloc.c +6
-0@@ -89,6 +89,12 @@ typedef int __bitwise fop_t;*/ #define FOP_SKIP_REPORT_NOTIFY ((__force fop_t)BIT(0)) +/* +* Place the freed page to the tail of the freelist after buddy merging. Will +* get ignored with page shuffling enabled. +*/ +#define FOP_TO_TAIL ((__force fop_t)BIT(1)) +/* prevent >1 _updater_ of zone percpu pageset ->high and ->batch fields */ static DEFINE_MUTEX (pcp_batch_high_lock);#define MIN_PERCPU_PAGELIST_FRACTION (8)
mm/page_alloc.c static inline void __free_one_page(struct page *page, unsigned long pfn, · line 1050
Quoted 5 lines
mm/page_alloc.c +2
-0@@ -1040,6 +1046,8 @@ static inline void __free_one_page(struct page *page, unsigned long pfn,if (is_shuffle_order (order)) to_tail =shuffle_pick_tail (); +else if (fop_flags &FOP_TO_TAIL ) + to_tail = true;
Should we really let random shuffling decision have a larger priority than explicit FOP_TO_TAIL request? Wei Yang mentioned that there's a call to shuffle_zone() anyway to process a freshly added memory, so we don't need to do that also during the process of addition itself? Might help with your goal of reducing dependencies even on systems that do have shuffling enabled?
Thanks,
Vlastimil
Quoted 13 lines
mm/page_alloc.c +1
-1else to_tail =buddy_merge_likely (pfn, buddy_pfn, page, order);@@ -3289,7 +3297,7 @@ void __putback_isolated_page(struct page *page, unsigned int order, int mt)/* Return isolated page to tail of freelist. */ __free_one_page (page,page_to_pfn (page), zone, order, mt, -FOP_SKIP_REPORT_NOTIFY ); +FOP_SKIP_REPORT_NOTIFY |FOP_TO_TAIL ); }/*
Re: [PATCH RFC 3/4] mm/page_alloc: always move pages to the tail of the freelist in unset_migratetype_isolate()
In reply to David HildenbrandQuoted 47 lines · David Hildenbrand
Page isolation doesn't actually touch the pages, it simply isolates pageblocks and moves all free pages to the
MIGRATE_ISOLATEfreelist.We already place pages to the tail of the freelists when undoing isolation via
__putback_isolated_page(), let's do it in any case (e.g., if order ==pageblock_order) and document the behavior.This change results in all pages getting onlined via
online_pages()to be placed to the tail of the freelist.Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Cc Scott Cheloha <[email protected]>Cc Michael Ellerman <[email protected]>Signed-off-by David Hildenbrand <[email protected]>--- include/linux/page-isolation.h | 2 ++ mm/page_alloc.c | 36 +++++++++++++++++++++++++++++----- mm/page_isolation.c | 8 ++++++-- 3 files changed, 39 insertions(+), 7 deletions(-)include/linux/page-isolation.h +2
-0@@ -38,6 +38,8 @@ struct page *has_unmovable_pages(struct zone *zone, struct page *page,void set_pageblock_migratetype (struct page *page,int migratetype);int move_freepages_block (struct zone *zone,struct page *page,int migratetype,int *num_movable); +int move_freepages_block_tail (struct zone *zone,struct page *page, +int migratetype);/* * Changes migrate type in [start_pfn, end_pfn) to be MIGRATE_ISOLATE. mm/page_alloc.c +0
-0@@ -899,6 +899,15 @@ static inline void move_to_free_list(struct page *page, struct zone *zone,list_move (&page->lru, &area->free_list[migratetype]);
mm/page_alloc.c static inline void move_to_free_list(struct page *page, struct zone *zone, · line 909
Quoted 10 lines
mm/page_alloc.c +8
-0} +/* Used for pages which are on another list */ +static inline void move_to_free_list_tail (struct page *page,struct zone *zone, +unsigned int order,int migratetype) +{ +struct free_area *area = &zone->free_area[order]; + +list_move_tail (&page->lru, &area->free_list[migratetype]); +}
There are just 3 callers of move_to_free_list() before this patch, I would just add the to_tail parameter there instead of new wrapper. For callers with constant parameter, the inline will eliminate it anyway.
Quoted 46 lines
mm/page_alloc.c +13
-5static inline void del_page_from_free_list (struct page *page,struct zone *zone,unsigned int order) {@@ -2323,7 +2332,7 @@ static inline struct page *__rmqueue_cma_fallback(struct zone *zone,*/ static int move_freepages (struct zone *zone,struct page *start_page,struct page *end_page, -int migratetype,int *num_movable) +int migratetype,int *num_movable,bool to_tail) {struct page *page;unsigned int order;@@ -2354,7 +2363,10 @@ static int move_freepages(struct zone *zone,VM_BUG_ON_PAGE (page_zone (page) != zone, page); order =page_order (page); -move_to_free_list (page, zone, order, migratetype); +if (to_tail) +move_to_free_list_tail (page, zone, order, migratetype); +else +move_to_free_list (page, zone, order, migratetype); page +=1 << order; pages_moved +=1 << order; }@@ -2362,8 +2374,9 @@ static int move_freepages(struct zone *zone,return pages_moved; } -int move_freepages_block (struct zone *zone,struct page *page, -int migratetype,int *num_movable) +static int __move_freepages_block (struct zone *zone,struct page *page, +int migratetype,int *num_movable, +bool to_tail) {unsigned long start_pfn, end_pfn;struct page *start_page, *end_page;@@ -2384,7 +2397,20 @@ int move_freepages_block(struct zone *zone, struct page *page,return 0 ;return move_freepages (zone, start_page, end_page, migratetype, - num_movable); + num_movable, to_tail); +} + +int move_freepages_block (struct zone *zone,struct page *page, +int migratetype,int *num_movable)
mm/page_alloc.c int move_freepages_block(struct zone *zone, struct page *page, · line 2414
Quoted 10 lines
mm/page_alloc.c +9
-0+{ +return __move_freepages_block (zone, page, migratetype, num_movable, + false); +} + +int move_freepages_block_tail (struct zone *zone,struct page *page, +int migratetype) +{ +return __move_freepages_block (zone, page, migratetype,NULL , true); }
Likewise, just 5 callers of move_freepages_block(), all in the files you're already changing, so no need for this wrappers IMHO.
Thanks,
Vlastimil
Quoted 30 lines
mm/page_alloc.c +0
-0static void change_pageblock_range(struct page *pageblock_page,mm/page_isolation.c +6
-2@@ -83,7 +83,7 @@ static void unset_migratetype_isolate(struct page *page, unsigned migratetype)* Because freepage with more than pageblock_order on isolated * pageblock is restricted to merge due to freepage counting problem, * it is possible that there is free buddy page. -* move_freepages_block() doesn't care of merge so we need other +* move_freepages_block*() don't care about merging, so we need another * approach in order to merge them. Isolation and free will make * these pages to be merged. */ @@ -106,9 +106,13 @@ static void unset_migratetype_isolate(struct page *page, unsigned migratetype)* If we isolate freepage with more than pageblock_order, there * should be no freepage in the range, so we could avoid costly * pageblock scanning for freepage moving. +* +* We didn't actually touch any of the isolated pages, so place them +* to the tail of the freelists. This is especially relevant during +* memory onlining. */ if (!isolated_page) { - nr_pages =move_freepages_block (zone, page, migratetype,NULL ); + nr_pages =move_freepages_block_tail (zone, page, migratetype);__mod_zone_freepage_state (zone, nr_pages, migratetype); }set_pageblock_migratetype (page, migratetype);
Re: [PATCH RFC 4/4] mm/page_alloc: place pages to tail in __free_pages_core()
In reply to David HildenbrandQuoted 27 lines · David Hildenbrand
__free_pages_core()is used when exposing fresh memory to the buddy during system boot and when onlining memory ingeneric_online_page().
generic_online_page()is used in two cases:1. Direct memory onlining in
online_pages().
2. Deferred memory onlining in memory-ballooning-like mechanisms (HyperV balloon and virtio-mem), when parts of a section are kept fake-offline to be fake-onlined later on.In 1, we already place pages to the tail of the freelist. Pages will be freed to
MIGRATE_ISOLATElists first and moved to the tail of the freelists viaundo_isolate_page_range().In 2, we currently don't implement a proper rule. In case of virtio-mem, where we currently always online
MAX_ORDER- 1 pages, the pages will be placed to the HEAD of the freelist - undesireable. While the hyper-v balloon callsgeneric_online_page()with single pages, usually it will call it on successive single pages in a larger block.The pages are fresh, so place them to the tail of the freelists and avoid the PCP.
Note: If we detect that the new behavior is undesireable for
__free_pages_core()during boot, we can let the caller specify the behavior.
Quoted 14 lines
Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Cc "K. Y. Srinivasan" <[email protected]>Cc Haiyang Zhang <[email protected]>Cc Stephen Hemminger <[email protected]>Cc Wei Liu <[email protected]>Signed-off-by David Hildenbrand <[email protected]>
Quoted 103 lines
--- mm/page_alloc.c | 32 ++++++++++++++++++++------------ 1 file changed, 20 insertions(+), 12 deletions(-)mm/page_alloc.c +20
-12@@ -264,7 +264,8 @@ bool pm_suspended_storage(void)unsigned int pageblock_order__read_mostly ;#endif -static void __free_pages_ok (struct page *page,unsigned int order); +static void __free_pages_ok (struct page *page,unsigned int order, +fop_t fop_flags);/* * results with 256, 32 in the lowmem_reserve sysctl: @@ -676,7 +677,7 @@ static void bad_page(struct page *page, const char *reason)void free_compound_page (struct page *page) {mem_cgroup_uncharge (page); -__free_pages_ok (page,compound_order (page)); +__free_pages_ok (page,compound_order (page),FOP_NONE ); }void prep_compound_page (struct page *page,unsigned int order)@@ -1402,17 +1403,15 @@ static void free_pcppages_bulk(struct zone *zone, int count,spin_unlock (&zone->lock); } -static void free_one_page (struct zone *zone, -struct page *page,unsigned long pfn, -unsigned int order, -int migratetype) +static void free_one_page (struct zone *zone,struct page *page,unsigned long pfn, +unsigned int order,int migratetype,fop_t fop_flags) {spin_lock (&zone->lock);if (unlikely (has_isolate_pageblock (zone) ||is_migrate_isolate (migratetype))) { migratetype =get_pfnblock_migratetype (page, pfn); } -__free_one_page (page, pfn, zone, order, migratetype,FOP_NONE ); +__free_one_page (page, pfn, zone, order, migratetype, fop_flags);spin_unlock (&zone->lock); }@@ -1490,7 +1489,8 @@ void __meminit reserve_bootmem_region(phys_addr_t start, phys_addr_t end) } } -static void __free_pages_ok (struct page *page,unsigned int order) +static void __free_pages_ok (struct page *page,unsigned int order, +fop_t fop_flags) {unsigned long flags;int migratetype;@@ -1502,7 +1502,8 @@ static void __free_pages_ok(struct page *page, unsigned int order) migratetype =get_pfnblock_migratetype (page, pfn);local_irq_save (flags);__count_vm_events (PGFREE ,1 << order); -free_one_page (page_zone (page), page, pfn, order, migratetype); +free_one_page (page_zone (page), page, pfn, order, migratetype, + fop_flags);local_irq_restore (flags); }@@ -1523,7 +1524,13 @@ void __free_pages_core(struct page *page, unsigned int order)atomic_long_add (nr_pages, &page_zone (page)->managed_pages);set_page_refcounted (page); -__free_pages (page, order); + +/* +* Bypass PCP and place fresh pages right to the tail, primarily +* relevant for memory onlining. +*/ +page_ref_dec (page); +__free_pages_ok (page, order,FOP_TO_TAIL ); }#ifdef CONFIG_NEED_MULTIPLE_NODES @@ -3167,7 +3174,8 @@ static void free_unref_page_commit(struct page *page, unsigned long pfn)*/ if (migratetype >=MIGRATE_PCPTYPES ) {if (unlikely (is_migrate_isolate (migratetype))) { -free_one_page (zone, page, pfn,0 , migratetype); +free_one_page (zone, page, pfn,0 , migratetype, +FOP_NONE );return ; } migratetype =MIGRATE_MOVABLE ;@@ -4984,7 +4992,7 @@ static inline void free_the_page(struct page *page, unsigned int order)if (order ==0 )/* Via pcp? */ free_unref_page (page);else -__free_pages_ok (page, order); +__free_pages_ok (page, order,FOP_NONE ); }void __free_pages (struct page *page,unsigned int order)
Quoted 10 lines · David Hildenbrand
Quoted 2 lines · Vlastimil Babka
On 9/16/20 9:31 PM, David Hildenbrand wrote:
Hi Vlastimil,
Quoted 3 lines
I see the point, but I don't think the head/tail mechanism is great for this. It might sort of work, but with other interfering activity there are no guarantees and it relies on a subtle implementation detail. There are better mechanisms
Quoted 7 lines
For the specified use case of adding+onlining a whole bunch of memory this works just fine. We don't care too much about "other interfering activity" as you mention here, or about guarantees - this is a pure optimization that seems to work just fine in practice.
I'm not sure about the "subtle implementation detail" - buddy merging, and head/tail of buddy lists are a basic concept of our page allocator.
Mel already explained that, so I won't repeat.
Quoted 3 lines
If that would ever change, the optimization here would be lost and we would have to think of something else. Nothing would actually break - and it's all kept directly in
page_alloc.c
Sure, but then it can become a pointless code churn.
Quoted 7 lines
I'd like to stress that what I propose here is both simple and powerful.
Quoted 4 lines
possible I think, such as preparing a larger
MIGRATE_UNMOVABLEarea in the existing memory before we allocate those long-term management structures. Or onlining a bunch of blocks aszone_movablefirst and only later convert tozone_normalin a controlled way when existing normal zone becomes depeted?
Quoted 6 lines
I see the following (more or less complicated) alternatives
1) Having a larger
MIGRATE_UNMOVABLEareaa) Sizing it is difficult. I mean you would have to plan ahead for all memory you might eventually hotplug later - and that could even be
Yeah, hence my worry about existing interfaces that work on 128MB blocks individually without a larger strategy.
Quoted 3 lines
impossible if you hotplug quite a lot of memory to a smaller machine. (I've seen people in the vm/container world trying to hotplug 128GB DIMMs to 2GB VMs ... and failing for obvious reasons)
Some planning should still be possible to maximize the contiguous area without unmovable allocations.
Quoted 21 lines
b) not really desired. You usually want to have most memory movable, not the opposite (just because you might hotplug memory in small chunks later).
2) smarter onlining
I have prototype patches for better auto-onlining (which I'll share at some point), where I balance between
ZONE_NORMALandZONE_MOVABLEin a defined ratio. Assuming something very simple, adding separate memory blocks and onlining them based on the current zone ratio (assuming a 1:4 normal:movable target ratio) would (without some other policies I have in place) result in something like this for hotplugged memory (via virtio-mem):[N][M][M][M][M][N][M][M][M][M][N][M][M][M][M]...
(note: layout is suboptimal, just a simple example)
But even here, all [N] memory blocks would immediately be use for allocations for the memmap of successive blocks. It doesn't solve the dependency issues.
Quoted 7 lines
Now assume we would want to group [N] in a way to allow for gigantic pages, like
[N][N][N][N][N][N][N][N][M][M][M][M] ....
we would, once again, never be able to allocate a gigantic page because all [N] would contain a memmap.
The second approach should work, if you know how much you are going to online, and plan the size the N group accordingly, and if the onlined amount is several gigabytes, then only the first one (or first X) will be unusable for a gigantic page, but the rest would be? Can't get much better than that.
Quoted 6 lines
3) conversion from MOVABLE -> NORMAL
While a conversion from MOVABLE to NORMAL would be interesting to see, it's going to be a challenging task to actually implement (people expect that
page_zone()remains stable). Without any hacks, we'd have to
Quoted 7 lines
1. offline the selected (MOVABLE) memory block/chunk
2. online the selected memory block/chunk to the NORMAL zoneThis is not something we can do out of random context (for example, we need both, the device hotplug lock and the memory hotplug lock, as we might race with user space) - so there might still be a chance of corner-case OOMs.
Right, it's trickier than I thought.
Quoted 15 lines
(I assume there could also be quite a negative performance impact when always relying on the conversion, and not properly planning ahead as in 2.)
Quoted 5 lines
I guess it's an issue that the e.g. 128M block onlines are so disconnected from each other it's hard to employ a strategy that works best for e.g. a whole bunch of GB onlined at once. But I noticed some effort towards new API, so maybe that will be solved there too?
While new interfaces might make it easier to identify boundaries of separate DIMMs (e.g., to online a single DIMM either movable or unmovable - which can partially be done right now when going via memory resource boundaries), it doesn't help for the use case of adding separate memory blocks.
Quoted 3 lines
So while having an automatic conversion from MOVABLE -> NORMAL would be interesting, I doubt we'll see it in the foreseeable future. Are there any similarly simple alternatives to optimize this?
I've reviewed the series and I won't block it - yes it's an optimistic approach that can break and leave us with code churn. But at least it's not that much code and the extra test in __free_one_page() shouldn't make this hotpath too worse. But I still hope we can achieve a more robust solution one day.
Quoted 2 lines
Thanks!
Quoted 5 lines
Quoted 3 lines
If that would ever change, the optimization here would be lost and we would have to think of something else. Nothing would actually break - and it's all kept directly in
page_alloc.cSure, but then it can become a pointless code churn.
Indeed, and if there are valid concerns that this will happen in the near future (e.g., < 1 year), I agree that we should look into alternatives right from the start. Otherwise it's good enough until some of the other things I mentioned below become real (which could also take a while ...).
Quoted 10 lines
Quoted 10 lines
I'd like to stress that what I propose here is both simple and powerful.
Quoted 4 lines
possible I think, such as preparing a larger
MIGRATE_UNMOVABLEarea in the existing memory before we allocate those long-term management structures. Or onlining a bunch of blocks aszone_movablefirst and only later convert tozone_normalin a controlled way when existing normal zone becomes depeted?I see the following (more or less complicated) alternatives
Quoted 7 lines
Quoted 4 lines
1) Having a larger
MIGRATE_UNMOVABLEareaa) Sizing it is difficult. I mean you would have to plan ahead for all memory you might eventually hotplug later - and that could even be
Yeah, hence my worry about existing interfaces that work on 128MB blocks individually without a larger strategy.
Yes, in the works :)
Quoted 7 lines
Quoted 4 lines
impossible if you hotplug quite a lot of memory to a smaller machine. (I've seen people in the vm/container world trying to hotplug 128GB DIMMs to 2GB VMs ... and failing for obvious reasons)
Some planning should still be possible to maximize the contiguous area without unmovable allocations.
Indeed, optimizing that is very high on my list of things to look into ...
Quoted 8 lines
Quoted 3 lines
we would, once again, never be able to allocate a gigantic page because all [N] would contain a memmap.
The second approach should work, if you know how much you are going to online, and plan the size the N group accordingly, and if the onlined amount is several gigabytes, then only the first one (or first X) will be unusable for a gigantic page, but the rest would be? Can't get much better than that.
Indeed, it's the optimal case (assuming one can come up with a safe zone balance - which is usually possible, but unfortunately, there are exceptions one at least has to identify).
[...]
Quoted 3 lines
I've reviewed the series and I won't block it - yes it's an optimistic approach that can break and leave us with code churn. But at least it's not that much
Thanks.
I'll try to document somewhere that the behavior of FOP_TO_TAIL is a pure optimization and might change in the future - along with the case it tried to optimize (so people know what the use case was).
Quoted 1 line
code and the extra test in
__free_one_page()shouldn't make this hotpath too
I assume the compiler is able to completely propagate constants and optimize that out - I haven't checked, though.
Quoted 1 line
worse. But I still hope we can achieve a more robust solution one day.
I definitely agree. I'd also prefer some kind of guarantees, but I learned that things always sound easier than they actually are when it comes to memory management in Linux ... and they take a lot of time (for example, Michal's/Oscar's attempts to implement vmemmap on hotadded memory).
Signature · 3 lines
Thanks, David / dhildenb
Re: [PATCH RFC 3/4] mm/page_alloc: always move pages to the tail of the freelist in unset_migratetype_isolate()
In reply to Vlastimil BabkaQuoted 50 lines · Vlastimil Babka
Quoted 49 lines · David Hildenbrand
Page isolation doesn't actually touch the pages, it simply isolates pageblocks and moves all free pages to the
MIGRATE_ISOLATEfreelist.We already place pages to the tail of the freelists when undoing isolation via
__putback_isolated_page(), let's do it in any case (e.g., if order ==pageblock_order) and document the behavior.This change results in all pages getting onlined via
online_pages()to be placed to the tail of the freelist.Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Cc Scott Cheloha <[email protected]>Cc Michael Ellerman <[email protected]>Signed-off-by David Hildenbrand <[email protected]>--- include/linux/page-isolation.h | 2 ++ mm/page_alloc.c | 36 +++++++++++++++++++++++++++++----- mm/page_isolation.c | 8 ++++++-- 3 files changed, 39 insertions(+), 7 deletions(-)include/linux/page-isolation.h +2
-0@@ -38,6 +38,8 @@ struct page *has_unmovable_pages(struct zone *zone, struct page *page,void set_pageblock_migratetype (struct page *page,int migratetype);int move_freepages_block (struct zone *zone,struct page *page,int migratetype,int *num_movable); +int move_freepages_block_tail (struct zone *zone,struct page *page, +int migratetype);/* * Changes migrate type in [start_pfn, end_pfn) to be MIGRATE_ISOLATE. mm/page_alloc.c +0
-0@@ -899,6 +899,15 @@ static inline void move_to_free_list(struct page *page, struct zone *zone,list_move (&page->lru, &area->free_list[migratetype]); }
Quoted 12 lines
mm/page_alloc.c static inline void move_to_free_list(struct page *page, struct zone *zone, · line 909
Quoted 8 lines
mm/page_alloc.c +8
-0+/* Used for pages which are on another list */ +static inline void move_to_free_list_tail (struct page *page,struct zone *zone, +unsigned int order,int migratetype) +{ +struct free_area *area = &zone->free_area[order]; + +list_move_tail (&page->lru, &area->free_list[migratetype]); +}There are just 3 callers of
move_to_free_list()before this patch, I would just add theto_tailparameter there instead of new wrapper. For callers with constant parameter, the inline will eliminate it anyway.
Got the same feeling :-)
Quoted 94 lines
Quoted 47 lines
mm/page_alloc.c +13
-5static inline void del_page_from_free_list (struct page *page,struct zone *zone,unsigned int order) {@@ -2323,7 +2332,7 @@ static inline struct page *__rmqueue_cma_fallback(struct zone *zone,*/ static int move_freepages (struct zone *zone,struct page *start_page,struct page *end_page, -int migratetype,int *num_movable) +int migratetype,int *num_movable,bool to_tail) {struct page *page;unsigned int order;@@ -2354,7 +2363,10 @@ static int move_freepages(struct zone *zone,VM_BUG_ON_PAGE (page_zone (page) != zone, page); order =page_order (page); -move_to_free_list (page, zone, order, migratetype); +if (to_tail) +move_to_free_list_tail (page, zone, order, migratetype); +else +move_to_free_list (page, zone, order, migratetype); page +=1 << order; pages_moved +=1 << order; }@@ -2362,8 +2374,9 @@ static int move_freepages(struct zone *zone,return pages_moved; } -int move_freepages_block (struct zone *zone,struct page *page, -int migratetype,int *num_movable) +static int __move_freepages_block (struct zone *zone,struct page *page, +int migratetype,int *num_movable, +bool to_tail) {unsigned long start_pfn, end_pfn;struct page *start_page, *end_page;@@ -2384,7 +2397,20 @@ int move_freepages_block(struct zone *zone, struct page *page,return 0 ;return move_freepages (zone, start_page, end_page, migratetype, - num_movable); + num_movable, to_tail); +} + +int move_freepages_block (struct zone *zone,struct page *page, +int migratetype,int *num_movable)mm/page_alloc.c int move_freepages_block(struct zone *zone, struct page *page, · line 2414
Quoted 10 lines
mm/page_alloc.c +9
-0+{ +return __move_freepages_block (zone, page, migratetype, num_movable, + false); +} + +int move_freepages_block_tail (struct zone *zone,struct page *page, +int migratetype) +{ +return __move_freepages_block (zone, page, migratetype,NULL , true); }Likewise, just 5 callers of
move_freepages_block(), all in the files you're already changing, so no need for this wrappers IMHO.Thanks,
VlastimilQuoted 30 lines
mm/page_alloc.c +0
-0static void change_pageblock_range(struct page *pageblock_page,mm/page_isolation.c +6
-2@@ -83,7 +83,7 @@ static void unset_migratetype_isolate(struct page *page, unsigned migratetype)* Because freepage with more than pageblock_order on isolated * pageblock is restricted to merge due to freepage counting problem, * it is possible that there is free buddy page. -* move_freepages_block() doesn't care of merge so we need other +* move_freepages_block*() don't care about merging, so we need another * approach in order to merge them. Isolation and free will make * these pages to be merged. */ @@ -106,9 +106,13 @@ static void unset_migratetype_isolate(struct page *page, unsigned migratetype)* If we isolate freepage with more than pageblock_order, there * should be no freepage in the range, so we could avoid costly * pageblock scanning for freepage moving. +* +* We didn't actually touch any of the isolated pages, so place them +* to the tail of the freelists. This is especially relevant during +* memory onlining. */ if (!isolated_page) { - nr_pages =move_freepages_block (zone, page, migratetype,NULL ); + nr_pages =move_freepages_block_tail (zone, page, migratetype);__mod_zone_freepage_state (zone, nr_pages, migratetype); }set_pageblock_migratetype (page, migratetype);
Signature · 2 lines
Wei Yang Help you, Help me
Re: [PATCH RFC 3/4] mm/page_alloc: always move pages to the tail of the freelist in unset_migratetype_isolate()
Quoted 64 lines · Wei Yang
Quoted 50 lines · Vlastimil Babka
Quoted 49 lines · David Hildenbrand
Page isolation doesn't actually touch the pages, it simply isolates pageblocks and moves all free pages to the
MIGRATE_ISOLATEfreelist.We already place pages to the tail of the freelists when undoing isolation via
__putback_isolated_page(), let's do it in any case (e.g., if order ==pageblock_order) and document the behavior.This change results in all pages getting onlined via
online_pages()to be placed to the tail of the freelist.Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Cc Scott Cheloha <[email protected]>Cc Michael Ellerman <[email protected]>Signed-off-by David Hildenbrand <[email protected]>--- include/linux/page-isolation.h | 2 ++ mm/page_alloc.c | 36 +++++++++++++++++++++++++++++----- mm/page_isolation.c | 8 ++++++-- 3 files changed, 39 insertions(+), 7 deletions(-)include/linux/page-isolation.h +2
-0@@ -38,6 +38,8 @@ struct page *has_unmovable_pages(struct zone *zone, struct page *page,void set_pageblock_migratetype (struct page *page,int migratetype);int move_freepages_block (struct zone *zone,struct page *page,int migratetype,int *num_movable); +int move_freepages_block_tail (struct zone *zone,struct page *page, +int migratetype);/* * Changes migrate type in [start_pfn, end_pfn) to be MIGRATE_ISOLATE. mm/page_alloc.c +0
-0@@ -899,6 +899,15 @@ static inline void move_to_free_list(struct page *page, struct zone *zone,list_move (&page->lru, &area->free_list[migratetype]); }Quoted 12 lines
mm/page_alloc.c static inline void move_to_free_list(struct page *page, struct zone *zone, · line 909
Quoted 8 lines
mm/page_alloc.c +8
-0+/* Used for pages which are on another list */ +static inline void move_to_free_list_tail (struct page *page,struct zone *zone, +unsigned int order,int migratetype) +{ +struct free_area *area = &zone->free_area[order]; + +list_move_tail (&page->lru, &area->free_list[migratetype]); +}There are just 3 callers of
move_to_free_list()before this patch, I would just add theto_tailparameter there instead of new wrapper. For callers with constant parameter, the inline will eliminate it anyway.
Quoted 1 line
Got the same feeling :-)
I once was told boolean parameters are the root of all evil, so I tried to keep them file-local :)
One thing to be aware of is, that inline optimizations won't help as long as this function is in mm/page_alloc.c, see below.
Quoted 53 lines
Quoted 53 lines
Quoted 53 lines
mm/page_alloc.c +18
-5static inline void del_page_from_free_list (struct page *page,struct zone *zone,unsigned int order) {@@ -2323,7 +2332,7 @@ static inline struct page *__rmqueue_cma_fallback(struct zone *zone,*/ static int move_freepages (struct zone *zone,struct page *start_page,struct page *end_page, -int migratetype,int *num_movable) +int migratetype,int *num_movable,bool to_tail) {struct page *page;unsigned int order;@@ -2354,7 +2363,10 @@ static int move_freepages(struct zone *zone,VM_BUG_ON_PAGE (page_zone (page) != zone, page); order =page_order (page); -move_to_free_list (page, zone, order, migratetype); +if (to_tail) +move_to_free_list_tail (page, zone, order, migratetype); +else +move_to_free_list (page, zone, order, migratetype); page +=1 << order; pages_moved +=1 << order; }@@ -2362,8 +2374,9 @@ static int move_freepages(struct zone *zone,return pages_moved; } -int move_freepages_block (struct zone *zone,struct page *page, -int migratetype,int *num_movable) +static int __move_freepages_block (struct zone *zone,struct page *page, +int migratetype,int *num_movable, +bool to_tail) {unsigned long start_pfn, end_pfn;struct page *start_page, *end_page;@@ -2384,7 +2397,20 @@ int move_freepages_block(struct zone *zone, struct page *page,return 0 ;return move_freepages (zone, start_page, end_page, migratetype, - num_movable); + num_movable, to_tail); +} + +int move_freepages_block (struct zone *zone,struct page *page, +int migratetype,int *num_movable) +{ +return __move_freepages_block (zone, page, migratetype, num_movable, + false); +} +
Quoted 8 lines
Quoted 8 lines
mm/page_alloc.c int move_freepages_block(struct zone *zone, struct page *page, · line 2414
Quoted 5 lines
mm/page_alloc.c +4
-0+int move_freepages_block_tail (struct zone *zone,struct page *page, +int migratetype) +{ +return __move_freepages_block (zone, page, migratetype,NULL , true); }Likewise, just 5 callers of
move_freepages_block(), all in the files you're already changing, so no need for this wrappers IMHO.
As long as we don't want to move the implementation to the header, we'll need it for the constant propagation to work at compile time (we don't really have link-time optimizations). Or am I missing something?
Thanks!
Signature · 3 lines
Thanks, David / dhildenb
Re: [PATCH RFC 2/4] mm/page_alloc: place pages to tail in __putback_isolated_page()
In reply to Vlastimil BabkaQuoted 8 lines · Vlastimil Babka
Quoted 4 lines · David Hildenbrand
__putback_isolated_page()already documents that pages will be placed to the tail of the freelist - this is, however, not the case for "order >=MAX_ORDER- 2" (seebuddy_merge_likely()) - which should be the case for all existing users.I think here should be a sentence saying something along "Thus this patch introduces a
FOP_TO_TAILflag to really ensure moving pages to tail."
Agreed, thanks!
Quoted 61 lines
Quoted 61 lines
This change affects two users:
- free page reporting
- page isolation, when undoing the isolation.This behavior is desireable for pages that haven't really been touched lately, so exactly the two users that don't actually read/write page content, but rather move untouched pages.
The new behavior is especially desirable for memory onlining, where we allow allocation of newly onlined pages via
undo_isolate_page_range()inonline_pages(). Right now, we always place them to the head of the free list, resulting in undesireable behavior: Assume we add individual memory chunks viaadd_memory()and online them right away to the NORMAL zone. We create a dependency chain of unmovable allocations e.g., via the memmap. The memmap of the next chunk will be placed onto previous chunks - if the last block cannot get offlined+removed, all dependent ones cannot get offlined+removed. While this can already be observed with individual DIMMs, it's more of an issue for virtio-mem (and I suspect also ppc DLPAR).Note: If we observe a degradation due to the changed page isolation behavior (which I doubt), we can always make this configurable by the instance triggering undo of isolation (e.g.,
alloc_contig_range(), memory onlining, memory offlining).Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Cc Scott Cheloha <[email protected]>Cc Michael Ellerman <[email protected]>Signed-off-by David Hildenbrand <[email protected]>--- mm/page_alloc.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-)mm/page_alloc.c +6
-0@@ -89,6 +89,12 @@ typedef int __bitwise fop_t;*/ #define FOP_SKIP_REPORT_NOTIFY ((__force fop_t)BIT(0)) +/* +* Place the freed page to the tail of the freelist after buddy merging. Will +* get ignored with page shuffling enabled. +*/ +#define FOP_TO_TAIL ((__force fop_t)BIT(1)) +/* prevent >1 _updater_ of zone percpu pageset ->high and ->batch fields */ static DEFINE_MUTEX (pcp_batch_high_lock);#define MIN_PERCPU_PAGELIST_FRACTION (8)@@ -1040,6 +1046,8 @@ static inline void __free_one_page(struct page *page, unsigned long pfn,
Quoted 10 lines
mm/page_alloc.c static inline void __free_one_page(struct page *page, unsigned long pfn, · line 1050
Quoted 4 lines
mm/page_alloc.c +2
-0if (is_shuffle_order (order)) to_tail =shuffle_pick_tail (); +else if (fop_flags &FOP_TO_TAIL ) + to_tail = true;Should we really let random shuffling decision have a larger priority than explicit
FOP_TO_TAILrequest? Wei Yang mentioned that there's a call toshuffle_zone()anyway to process a freshly added memory, so we don't need to do that also during the process of addition itself? Might help with your goal of reducing dependencies even on systems that do have shuffling enabled?
So, we do have cases where generic_online_page() -> __free_pages_core() isn't called (see patch #4):
generic_online_page() is used in two cases:
1. Direct memory onlining in online_pages(). Here, we call shuffle_zone().
2. Deferred memory onlining in memory-ballooning-like mechanisms (HyperV balloon and virtio-mem), when parts of a section are kept fake-offline to be fake-onlined later on.
While we shuffle in the fist instance the whole zone, we wouldn't shuffle in the second case.
But maybe this should be tackled (just like when alloc_contig_free() a large contiguous range, memory offlining failing, alloc_contig_range() failing) by manually shuffling the zone again. That would be cleaner, and the right thing to do when exposing large, contiguous ranges again to the buddy.
Thanks!
Signature · 3 lines
Thanks, David / dhildenb
Re: [PATCH RFC 3/4] mm/page_alloc: always move pages to the tail of the freelist in unset_migratetype_isolate()
In reply to David HildenbrandQuoted 60 lines · David Hildenbrand
Quoted 51 lines
Quoted 51 lines
Quoted 51 lines
static inline void del_page_from_free_list(struct page *page, struct zone *zone,
mm/page_alloc.c +18
-5unsigned int order) {@@ -2323,7 +2332,7 @@ static inline struct page *__rmqueue_cma_fallback(struct zone *zone,*/ static int move_freepages (struct zone *zone,struct page *start_page,struct page *end_page, -int migratetype,int *num_movable) +int migratetype,int *num_movable,bool to_tail) {struct page *page;unsigned int order;@@ -2354,7 +2363,10 @@ static int move_freepages(struct zone *zone,VM_BUG_ON_PAGE (page_zone (page) != zone, page); order =page_order (page); -move_to_free_list (page, zone, order, migratetype); +if (to_tail) +move_to_free_list_tail (page, zone, order, migratetype); +else +move_to_free_list (page, zone, order, migratetype); page +=1 << order; pages_moved +=1 << order; }@@ -2362,8 +2374,9 @@ static int move_freepages(struct zone *zone,return pages_moved; } -int move_freepages_block (struct zone *zone,struct page *page, -int migratetype,int *num_movable) +static int __move_freepages_block (struct zone *zone,struct page *page, +int migratetype,int *num_movable, +bool to_tail) {unsigned long start_pfn, end_pfn;struct page *start_page, *end_page;@@ -2384,7 +2397,20 @@ int move_freepages_block(struct zone *zone, struct page *page,return 0 ;return move_freepages (zone, start_page, end_page, migratetype, - num_movable); + num_movable, to_tail); +} + +int move_freepages_block (struct zone *zone,struct page *page, +int migratetype,int *num_movable) +{ +return __move_freepages_block (zone, page, migratetype, num_movable, + false); +} +Quoted 8 lines
Quoted 8 lines
mm/page_alloc.c int move_freepages_block(struct zone *zone, struct page *page, · line 2414
Quoted 5 lines
mm/page_alloc.c +4
-0+int move_freepages_block_tail (struct zone *zone,struct page *page, +int migratetype) +{ +return __move_freepages_block (zone, page, migratetype,NULL , true); }Likewise, just 5 callers of
move_freepages_block(), all in the files you're already changing, so no need for this wrappers IMHO.
Quoted 3 lines
As long as we don't want to move the implementation to the header, we'll need it for the constant propagation to work at compile time (we don't really have link-time optimizations). Or am I missing something?
I guess move_freepages_block() is not exactly fast path, so we could do without it.
Quoted 2 lines
Thanks!
Re: [PATCH RFC 1/4] mm/page_alloc: convert "report" flag of __free_one_page() to a proper flag
In reply to David HildenbrandQuoted 5 lines · David Hildenbrand
Let's prepare for additional flags and avoid long parameter lists of bools. Follow-up patches will also make use of the flags in
__free_pages_ok(), however, I wasn't able to come up with a better name for the type - should be good enough for internal purposes.
Quoted 10 lines
Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Signed-off-by David Hildenbrand <[email protected]>
Signature · 2 lines
Oscar Salvador SUSE L3
Re: [PATCH RFC 3/4] mm/page_alloc: always move pages to the tail of the freelist in unset_migratetype_isolate()
In reply to Vlastimil BabkaQuoted 50 lines · Vlastimil Babka
Quoted 49 lines · David Hildenbrand
Page isolation doesn't actually touch the pages, it simply isolates pageblocks and moves all free pages to the
MIGRATE_ISOLATEfreelist.We already place pages to the tail of the freelists when undoing isolation via
__putback_isolated_page(), let's do it in any case (e.g., if order ==pageblock_order) and document the behavior.This change results in all pages getting onlined via
online_pages()to be placed to the tail of the freelist.Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Cc Scott Cheloha <[email protected]>Cc Michael Ellerman <[email protected]>Signed-off-by David Hildenbrand <[email protected]>--- include/linux/page-isolation.h | 2 ++ mm/page_alloc.c | 36 +++++++++++++++++++++++++++++----- mm/page_isolation.c | 8 ++++++-- 3 files changed, 39 insertions(+), 7 deletions(-)include/linux/page-isolation.h +2
-0@@ -38,6 +38,8 @@ struct page *has_unmovable_pages(struct zone *zone, struct page *page,void set_pageblock_migratetype (struct page *page,int migratetype);int move_freepages_block (struct zone *zone,struct page *page,int migratetype,int *num_movable); +int move_freepages_block_tail (struct zone *zone,struct page *page, +int migratetype);/* * Changes migrate type in [start_pfn, end_pfn) to be MIGRATE_ISOLATE. mm/page_alloc.c +0
-0@@ -899,6 +899,15 @@ static inline void move_to_free_list(struct page *page, struct zone *zone,list_move (&page->lru, &area->free_list[migratetype]); }
Quoted 13 lines
mm/page_alloc.c static inline void move_to_free_list(struct page *page, struct zone *zone, · line 909
Quoted 8 lines
mm/page_alloc.c +8
-0+/* Used for pages which are on another list */ +static inline void move_to_free_list_tail (struct page *page,struct zone *zone, +unsigned int order,int migratetype) +{ +struct free_area *area = &zone->free_area[order]; + +list_move_tail (&page->lru, &area->free_list[migratetype]); +}There are just 3 callers of
move_to_free_list()before this patch, I would just add theto_tailparameter there instead of new wrapper. For callers with constant parameter, the inline will eliminate it anyway.
So, I'll leave this as is for now, it nicely pairs with add_to_free_list()/add_to_free_list_tail() and ...
[...]
Quoted 6 lines
mm/page_alloc.c int move_freepages_block(struct zone *zone, struct page *page, · line 2414
Quoted 5 lines
mm/page_alloc.c +4
-0+int move_freepages_block_tail (struct zone *zone,struct page *page, +int migratetype) +{ +return __move_freepages_block (zone, page, migratetype,NULL , true); }
Drop these wrappers, converting callers of move_freepages_block().
Thanks!
Signature · 3 lines
Thanks, David / dhildenb
Re: [PATCH RFC 2/4] mm/page_alloc: place pages to tail in __putback_isolated_page()
In reply to David HildenbrandQuoted 30 lines · David Hildenbrand
__putback_isolated_page()already documents that pages will be placed to the tail of the freelist - this is, however, not the case for "order >=MAX_ORDER- 2" (seebuddy_merge_likely()) - which should be the case for all existing users.This change affects two users:
- free page reporting
- page isolation, when undoing the isolation.This behavior is desireable for pages that haven't really been touched lately, so exactly the two users that don't actually read/write page content, but rather move untouched pages.
The new behavior is especially desirable for memory onlining, where we allow allocation of newly onlined pages via
undo_isolate_page_range()inonline_pages(). Right now, we always place them to the head of the free list, resulting in undesireable behavior: Assume we add individual memory chunks viaadd_memory()and online them right away to the NORMAL zone. We create a dependency chain of unmovable allocations e.g., via the memmap. The memmap of the next chunk will be placed onto previous chunks - if the last block cannot get offlined+removed, all dependent ones cannot get offlined+removed. While this can already be observed with individual DIMMs, it's more of an issue for virtio-mem (and I suspect also ppc DLPAR).Note: If we observe a degradation due to the changed page isolation behavior (which I doubt), we can always make this configurable by the instance triggering undo of isolation (e.g.,
alloc_contig_range(), memory onlining, memory offlining).
Quoted 12 lines
Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Cc Scott Cheloha <[email protected]>Cc Michael Ellerman <[email protected]>Signed-off-by David Hildenbrand <[email protected]>
LGTM, the only thing is the shuffe_zone topic that Wei and Vlastimil rose. Feels a bit odd that takes precedence over something we explicitily demanded.
With the comment Vlastimil suggested:
Signature · 2 lines
Oscar Salvador SUSE L3
Re: [PATCH RFC 2/4] mm/page_alloc: place pages to tail in __putback_isolated_page()
Quoted 44 lines · Oscar Salvador
Quoted 30 lines · David Hildenbrand
__putback_isolated_page()already documents that pages will be placed to the tail of the freelist - this is, however, not the case for "order >=MAX_ORDER- 2" (seebuddy_merge_likely()) - which should be the case for all existing users.This change affects two users:
- free page reporting
- page isolation, when undoing the isolation.This behavior is desireable for pages that haven't really been touched lately, so exactly the two users that don't actually read/write page content, but rather move untouched pages.
The new behavior is especially desirable for memory onlining, where we allow allocation of newly onlined pages via
undo_isolate_page_range()inonline_pages(). Right now, we always place them to the head of the free list, resulting in undesireable behavior: Assume we add individual memory chunks viaadd_memory()and online them right away to the NORMAL zone. We create a dependency chain of unmovable allocations e.g., via the memmap. The memmap of the next chunk will be placed onto previous chunks - if the last block cannot get offlined+removed, all dependent ones cannot get offlined+removed. While this can already be observed with individual DIMMs, it's more of an issue for virtio-mem (and I suspect also ppc DLPAR).Note: If we observe a degradation due to the changed page isolation behavior (which I doubt), we can always make this configurable by the instance triggering undo of isolation (e.g.,
alloc_contig_range(), memory onlining, memory offlining).Quoted 12 lines
Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Cc Scott Cheloha <[email protected]>Cc Michael Ellerman <[email protected]>Signed-off-by David Hildenbrand <[email protected]>
Quoted 3 lines
LGTM, the only thing is the
shuffe_zonetopic that Wei and Vlastimil rose. Feels a bit odd that takes precedence over something we explicitily demanded.
Thanks, yeah I'll be changing that.
Quoted 4 lines
With the comment Vlastimil suggested:
Reviewed-by Oscar Salvador <[email protected]>
Signature · 3 lines
Thanks, David / dhildenb
Re: [PATCH RFC 3/4] mm/page_alloc: always move pages to the tail of the freelist in unset_migratetype_isolate()
In reply to David HildenbrandQuoted 10 lines · David Hildenbrand
Page isolation doesn't actually touch the pages, it simply isolates pageblocks and moves all free pages to the
MIGRATE_ISOLATEfreelist.We already place pages to the tail of the freelists when undoing isolation via
__putback_isolated_page(), let's do it in any case (e.g., if order ==pageblock_order) and document the behavior.This change results in all pages getting onlined via
online_pages()to be placed to the tail of the freelist.
Quoted 12 lines
Cc Andrew Morton <[email protected]>Cc Alexander Duyck <[email protected]>Cc Mel Gorman <[email protected]>Cc Michal Hocko <[email protected]>Cc Dave Hansen <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Wei Yang <[email protected]>Cc Oscar Salvador <[email protected]>Cc Mike Rapoport <[email protected]>Cc Scott Cheloha <[email protected]>Cc Michael Ellerman <[email protected]>Signed-off-by David Hildenbrand <[email protected]>
LGTM.
Feel the same way about move_freepages_block_tail/move_freepages_block_tail wrappers, I think we are better off without them.
Thanks
Signature · 2 lines
Oscar Salvador SUSE L3
Re: [PATCH RFC 4/4] mm/page_alloc: place pages to tail in __free_pages_core()
In reply to David Hildenbrandmm/page_alloc.c void __free_pages_core(struct page *page, unsigned int order) · line 1533
Quoted 12 lines · David Hildenbrand
mm/page_alloc.c +7
-1@@ -1523,7 +1524,13 @@ void __free_pages_core(struct page *page, unsigned int order)atomic_long_add (nr_pages, &page_zone (page)->managed_pages);set_page_refcounted (page); -__free_pages (page, order); + +/* +* Bypass PCP and place fresh pages right to the tail, primarily +* relevant for memory onlining. +*/ +page_ref_dec (page); +__free_pages_ok (page, order,FOP_TO_TAIL );
Sorry, I must be missing something obvious here, but I am a bit confused here. I get the part of placing them at the tail so rmqueue_bulk() won't find them, but I do not get why we decrement page's refcount. IIUC, its refcount will be 0, but why do we want to do that?
Another thing a bit unrelated... we mess three times with page's refcount (two before this patch).
Why do we have this dance in place?
Thanks
Signature · 2 lines
Oscar Salvador SUSE L3
Re: [PATCH RFC 4/4] mm/page_alloc: place pages to tail in __free_pages_core()
Quoted 19 lines · Oscar Salvador
mm/page_alloc.c void __free_pages_core(struct page *page, unsigned int order) · line 1533
Quoted 12 lines · David Hildenbrand
mm/page_alloc.c +7
-1@@ -1523,7 +1524,13 @@ void __free_pages_core(struct page *page, unsigned int order)atomic_long_add (nr_pages, &page_zone (page)->managed_pages);set_page_refcounted (page); -__free_pages (page, order); + +/* +* Bypass PCP and place fresh pages right to the tail, primarily +* relevant for memory onlining. +*/ +page_ref_dec (page); +__free_pages_ok (page, order,FOP_TO_TAIL );Sorry, I must be missing something obvious here, but I am a bit confused here. I get the part of placing them at the tail so
rmqueue_bulk()won't find them, but I do not get why we decrement page's refcount. IIUC, its refcount will be 0, but why do we want to do that?
Quoted 3 lines
Another thing a bit unrelated... we mess three times with page's refcount (two before this patch).
Why do we have this dance in place?
Hi Oscar!
Old code:
set_page_refcounted(): sets the refcount to 1.__free_pages()
-> put_page_testzero(): sets it to 0
-> free_the_page()->__free_pages_ok()
New code:
set_page_refcounted(): sets the refcount to 1.page_ref_dec(page): sets it to 0__free_pages_ok():
We could skip the set_page_refcounted() + page_ref_dec(page) and lose a couple of sanity checks but we could simply use a VM_BUG_ON_PAGE(page_ref_count(page), page), which is what we really care about when onlining memory.
Signature · 3 lines
Thanks, David / dhildenb
Oscar Salvador
Reviewed-byRe: [PATCH RFC 4/4] mm/page_alloc: place pages to tail in __free_pages_core()
Quoted 1 line · David Hildenbrand
Hi Oscar!
Hi David :-)
Quoted 8 lines
Old code:
set_page_refcounted(): sets the refcount to 1.__free_pages()
->put_page_testzero(): sets it to 0
-> free_the_page()->__free_pages_ok()
Quoted 5 lines
New code:
set_page_refcounted(): sets the refcount to 1.page_ref_dec(page): sets it to 0__free_pages_ok():
bleh, I misread the patch, somehow I managed to not see that you replaced __free_pages with __free_pages_ok.
To be honest, now that we do not need the page's refcount to be 1 for the put_page_testzero to trigger (and since you are decrementing it anyways), I think it would be much clear for those two to be gone.
But not strong, so:
Signature · 2 lines
Oscar Salvador SUSE L3