From 62695a84eb8f2e718bf4dfb21700afaa7a08e0ea Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Sat, 18 Oct 2008 20:26:09 -0700 Subject: vmscan: move isolate_lru_page() to vmscan.c On large memory systems, the VM can spend way too much time scanning through pages that it cannot (or should not) evict from memory. Not only does it use up CPU time, but it also provokes lock contention and can leave large systems under memory presure in a catatonic state. This patch series improves VM scalability by: 1) putting filesystem backed, swap backed and unevictable pages onto their own LRUs, so the system only scans the pages that it can/should evict from memory 2) switching to two handed clock replacement for the anonymous LRUs, so the number of pages that need to be scanned when the system starts swapping is bound to a reasonable number 3) keeping unevictable pages off the LRU completely, so the VM does not waste CPU time scanning them. ramfs, ramdisk, SHM_LOCKED shared memory segments and mlock()ed VMA pages are keept on the unevictable list. This patch: isolate_lru_page logically belongs to be in vmscan.c than migrate.c. It is tough, because we don't need that function without memory migration so there is a valid argument to have it in migrate.c. However a subsequent patch needs to make use of it in the core mm, so we can happily move it to vmscan.c. Also, make the function a little more generic by not requiring that it adds an isolated page to a given list. Callers can do that. Note that we now have '__isolate_lru_page()', that does something quite different, visible outside of vmscan.c for use with memory controller. Methinks we need to rationalize these names/purposes. --lts [akpm@linux-foundation.org: fix mm/memory_hotplug.c build] Signed-off-by: Nick Piggin Signed-off-by: Rik van Riel Signed-off-by: Lee Schermerhorn Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/vmscan.c | 45 +++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 45 insertions(+) (limited to 'mm/vmscan.c') diff --git a/mm/vmscan.c b/mm/vmscan.c index 1ff1a58e7c1..1fd4912a596 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -844,6 +844,51 @@ static unsigned long clear_active_flags(struct list_head *page_list) return nr_active; } +/** + * isolate_lru_page - tries to isolate a page from its LRU list + * @page: page to isolate from its LRU list + * + * Isolates a @page from an LRU list, clears PageLRU and adjusts the + * vmstat statistic corresponding to whatever LRU list the page was on. + * + * Returns 0 if the page was removed from an LRU list. + * Returns -EBUSY if the page was not on an LRU list. + * + * The returned page will have PageLRU() cleared. If it was found on + * the active list, it will have PageActive set. That flag may need + * to be cleared by the caller before letting the page go. + * + * The vmstat statistic corresponding to the list on which the page was + * found will be decremented. + * + * Restrictions: + * (1) Must be called with an elevated refcount on the page. This is a + * fundamentnal difference from isolate_lru_pages (which is called + * without a stable reference). + * (2) the lru_lock must not be held. + * (3) interrupts must be enabled. + */ +int isolate_lru_page(struct page *page) +{ + int ret = -EBUSY; + + if (PageLRU(page)) { + struct zone *zone = page_zone(page); + + spin_lock_irq(&zone->lru_lock); + if (PageLRU(page) && get_page_unless_zero(page)) { + ret = 0; + ClearPageLRU(page); + if (PageActive(page)) + del_page_from_active_list(zone, page); + else + del_page_from_inactive_list(zone, page); + } + spin_unlock_irq(&zone->lru_lock); + } + return ret; +} + /* * shrink_inactive_list() is a helper for shrink_zone(). It returns the number * of reclaimed pages -- cgit v1.2.3 From b69408e88bd86b98feb7b9a38fd865e1ddb29827 Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Sat, 18 Oct 2008 20:26:14 -0700 Subject: vmscan: Use an indexed array for LRU variables Currently we are defining explicit variables for the inactive and active list. An indexed array can be more generic and avoid repeating similar code in several places in the reclaim code. We are saving a few bytes in terms of code size: Before: text data bss dec hex filename 4097753 573120 4092484 8763357 85b7dd vmlinux After: text data bss dec hex filename 4097729 573120 4092484 8763333 85b7c5 vmlinux Having an easy way to add new lru lists may ease future work on the reclaim code. Signed-off-by: Rik van Riel Signed-off-by: Lee Schermerhorn Signed-off-by: Christoph Lameter Signed-off-by: KOSAKI Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/vmscan.c | 120 +++++++++++++++++++++++++++++------------------------------- 1 file changed, 57 insertions(+), 63 deletions(-) (limited to 'mm/vmscan.c') diff --git a/mm/vmscan.c b/mm/vmscan.c index 1fd4912a596..46fdaa546b8 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -819,10 +819,10 @@ static unsigned long isolate_pages_global(unsigned long nr, int active) { if (active) - return isolate_lru_pages(nr, &z->active_list, dst, + return isolate_lru_pages(nr, &z->lru[LRU_ACTIVE].list, dst, scanned, order, mode); else - return isolate_lru_pages(nr, &z->inactive_list, dst, + return isolate_lru_pages(nr, &z->lru[LRU_INACTIVE].list, dst, scanned, order, mode); } @@ -973,10 +973,7 @@ static unsigned long shrink_inactive_list(unsigned long max_scan, VM_BUG_ON(PageLRU(page)); SetPageLRU(page); list_del(&page->lru); - if (PageActive(page)) - add_page_to_active_list(zone, page); - else - add_page_to_inactive_list(zone, page); + add_page_to_lru_list(zone, page, page_lru(page)); if (!pagevec_add(&pvec, page)) { spin_unlock_irq(&zone->lru_lock); __pagevec_release(&pvec); @@ -1144,8 +1141,8 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, int pgdeactivate = 0; unsigned long pgscanned; LIST_HEAD(l_hold); /* The pages which were snipped off */ - LIST_HEAD(l_inactive); /* Pages to go onto the inactive_list */ - LIST_HEAD(l_active); /* Pages to go onto the active_list */ + LIST_HEAD(l_active); + LIST_HEAD(l_inactive); struct page *page; struct pagevec pvec; int reclaim_mapped = 0; @@ -1194,7 +1191,7 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, VM_BUG_ON(!PageActive(page)); ClearPageActive(page); - list_move(&page->lru, &zone->inactive_list); + list_move(&page->lru, &zone->lru[LRU_INACTIVE].list); mem_cgroup_move_lists(page, false); pgmoved++; if (!pagevec_add(&pvec, page)) { @@ -1224,7 +1221,7 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, SetPageLRU(page); VM_BUG_ON(!PageActive(page)); - list_move(&page->lru, &zone->active_list); + list_move(&page->lru, &zone->lru[LRU_ACTIVE].list); mem_cgroup_move_lists(page, true); pgmoved++; if (!pagevec_add(&pvec, page)) { @@ -1244,65 +1241,64 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, pagevec_release(&pvec); } +static unsigned long shrink_list(enum lru_list l, unsigned long nr_to_scan, + struct zone *zone, struct scan_control *sc, int priority) +{ + if (l == LRU_ACTIVE) { + shrink_active_list(nr_to_scan, zone, sc, priority); + return 0; + } + return shrink_inactive_list(nr_to_scan, zone, sc); +} + /* * This is a basic per-zone page freer. Used by both kswapd and direct reclaim. */ static unsigned long shrink_zone(int priority, struct zone *zone, struct scan_control *sc) { - unsigned long nr_active; - unsigned long nr_inactive; + unsigned long nr[NR_LRU_LISTS]; unsigned long nr_to_scan; unsigned long nr_reclaimed = 0; + enum lru_list l; if (scan_global_lru(sc)) { /* * Add one to nr_to_scan just to make sure that the kernel * will slowly sift through the active list. */ - zone->nr_scan_active += - (zone_page_state(zone, NR_ACTIVE) >> priority) + 1; - nr_active = zone->nr_scan_active; - zone->nr_scan_inactive += - (zone_page_state(zone, NR_INACTIVE) >> priority) + 1; - nr_inactive = zone->nr_scan_inactive; - if (nr_inactive >= sc->swap_cluster_max) - zone->nr_scan_inactive = 0; - else - nr_inactive = 0; - - if (nr_active >= sc->swap_cluster_max) - zone->nr_scan_active = 0; - else - nr_active = 0; + for_each_lru(l) { + zone->lru[l].nr_scan += (zone_page_state(zone, + NR_LRU_BASE + l) >> priority) + 1; + nr[l] = zone->lru[l].nr_scan; + if (nr[l] >= sc->swap_cluster_max) + zone->lru[l].nr_scan = 0; + else + nr[l] = 0; + } } else { /* * This reclaim occurs not because zone memory shortage but * because memory controller hits its limit. * Then, don't modify zone reclaim related data. */ - nr_active = mem_cgroup_calc_reclaim_active(sc->mem_cgroup, - zone, priority); + nr[LRU_ACTIVE] = mem_cgroup_calc_reclaim(sc->mem_cgroup, + zone, priority, LRU_ACTIVE); - nr_inactive = mem_cgroup_calc_reclaim_inactive(sc->mem_cgroup, - zone, priority); + nr[LRU_INACTIVE] = mem_cgroup_calc_reclaim(sc->mem_cgroup, + zone, priority, LRU_INACTIVE); } - - while (nr_active || nr_inactive) { - if (nr_active) { - nr_to_scan = min(nr_active, + while (nr[LRU_ACTIVE] || nr[LRU_INACTIVE]) { + for_each_lru(l) { + if (nr[l]) { + nr_to_scan = min(nr[l], (unsigned long)sc->swap_cluster_max); - nr_active -= nr_to_scan; - shrink_active_list(nr_to_scan, zone, sc, priority); - } + nr[l] -= nr_to_scan; - if (nr_inactive) { - nr_to_scan = min(nr_inactive, - (unsigned long)sc->swap_cluster_max); - nr_inactive -= nr_to_scan; - nr_reclaimed += shrink_inactive_list(nr_to_scan, zone, - sc); + nr_reclaimed += shrink_list(l, nr_to_scan, + zone, sc, priority); + } } } @@ -1819,6 +1815,7 @@ static unsigned long shrink_all_zones(unsigned long nr_pages, int prio, { struct zone *zone; unsigned long nr_to_scan, ret = 0; + enum lru_list l; for_each_zone(zone) { @@ -1828,28 +1825,25 @@ static unsigned long shrink_all_zones(unsigned long nr_pages, int prio, if (zone_is_all_unreclaimable(zone) && prio != DEF_PRIORITY) continue; - /* For pass = 0 we don't shrink the active list */ - if (pass > 0) { - zone->nr_scan_active += - (zone_page_state(zone, NR_ACTIVE) >> prio) + 1; - if (zone->nr_scan_active >= nr_pages || pass > 3) { - zone->nr_scan_active = 0; + for_each_lru(l) { + /* For pass = 0 we don't shrink the active list */ + if (pass == 0 && l == LRU_ACTIVE) + continue; + + zone->lru[l].nr_scan += + (zone_page_state(zone, NR_LRU_BASE + l) + >> prio) + 1; + if (zone->lru[l].nr_scan >= nr_pages || pass > 3) { + zone->lru[l].nr_scan = 0; nr_to_scan = min(nr_pages, - zone_page_state(zone, NR_ACTIVE)); - shrink_active_list(nr_to_scan, zone, sc, prio); + zone_page_state(zone, + NR_LRU_BASE + l)); + ret += shrink_list(l, nr_to_scan, zone, + sc, prio); + if (ret >= nr_pages) + return ret; } } - - zone->nr_scan_inactive += - (zone_page_state(zone, NR_INACTIVE) >> prio) + 1; - if (zone->nr_scan_inactive >= nr_pages || pass > 3) { - zone->nr_scan_inactive = 0; - nr_to_scan = min(nr_pages, - zone_page_state(zone, NR_INACTIVE)); - ret += shrink_inactive_list(nr_to_scan, zone, sc); - if (ret >= nr_pages) - return ret; - } } return ret; -- cgit v1.2.3 From 68a22394c286a2daf06ee8d65d8835f738faefa5 Mon Sep 17 00:00:00 2001 From: Rik van Riel Date: Sat, 18 Oct 2008 20:26:23 -0700 Subject: vmscan: free swap space on swap-in/activation If vm_swap_full() (swap space more than 50% full), the system will free swap space at swapin time. With this patch, the system will also free the swap space in the pageout code, when we decide that the page is not a candidate for swapout (and just wasting swap space). Signed-off-by: Rik van Riel Signed-off-by: Lee Schermerhorn Signed-off-by: MinChan Kim Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/vmscan.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'mm/vmscan.c') diff --git a/mm/vmscan.c b/mm/vmscan.c index 46fdaa546b8..e656035d340 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -647,6 +647,9 @@ free_it: continue; activate_locked: + /* Not a candidate for swapping, so reclaim swap space. */ + if (PageSwapCache(page) && vm_swap_full()) + remove_exclusive_swap_page_ref(page); SetPageActive(page); pgactivate++; keep_locked: @@ -1228,6 +1231,8 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, __mod_zone_page_state(zone, NR_ACTIVE, pgmoved); pgmoved = 0; spin_unlock_irq(&zone->lru_lock); + if (vm_swap_full()) + pagevec_swap_free(&pvec); __pagevec_release(&pvec); spin_lock_irq(&zone->lru_lock); } @@ -1237,6 +1242,8 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, __count_zone_vm_events(PGREFILL, zone, pgscanned); __count_vm_events(PGDEACTIVATE, pgdeactivate); spin_unlock_irq(&zone->lru_lock); + if (vm_swap_full()) + pagevec_swap_free(&pvec); pagevec_release(&pvec); } -- cgit v1.2.3 From 4f98a2fee8acdb4ac84545df98cccecfd130f8db Mon Sep 17 00:00:00 2001 From: Rik van Riel Date: Sat, 18 Oct 2008 20:26:32 -0700 Subject: vmscan: split LRU lists into anon & file sets Split the LRU lists in two, one set for pages that are backed by real file systems ("file") and one for pages that are backed by memory and swap ("anon"). The latter includes tmpfs. The advantage of doing this is that the VM will not have to scan over lots of anonymous pages (which we generally do not want to swap out), just to find the page cache pages that it should evict. This patch has the infrastructure and a basic policy to balance how much we scan the anon lists and how much we scan the file lists. The big policy changes are in separate patches. [lee.schermerhorn@hp.com: collect lru meminfo statistics from correct offset] [kosaki.motohiro@jp.fujitsu.com: prevent incorrect oom under split_lru] [kosaki.motohiro@jp.fujitsu.com: fix pagevec_move_tail() doesn't treat unevictable page] [hugh@veritas.com: memcg swapbacked pages active] [hugh@veritas.com: splitlru: BDI_CAP_SWAP_BACKED] [akpm@linux-foundation.org: fix /proc/vmstat units] [nishimura@mxp.nes.nec.co.jp: memcg: fix handling of shmem migration] [kosaki.motohiro@jp.fujitsu.com: adjust Quicklists field of /proc/meminfo] [kosaki.motohiro@jp.fujitsu.com: fix style issue of get_scan_ratio()] Signed-off-by: Rik van Riel Signed-off-by: Lee Schermerhorn Signed-off-by: KOSAKI Motohiro Signed-off-by: Hugh Dickins Signed-off-by: Daisuke Nishimura Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/vmscan.c | 416 +++++++++++++++++++++++++++++++----------------------------- 1 file changed, 216 insertions(+), 200 deletions(-) (limited to 'mm/vmscan.c') diff --git a/mm/vmscan.c b/mm/vmscan.c index e656035d340..d10d2f9a33f 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -78,7 +78,7 @@ struct scan_control { unsigned long (*isolate_pages)(unsigned long nr, struct list_head *dst, unsigned long *scanned, int order, int mode, struct zone *z, struct mem_cgroup *mem_cont, - int active); + int active, int file); }; #define lru_to_page(_head) (list_entry((_head)->prev, struct page, lru)) @@ -680,7 +680,7 @@ keep: * * returns 0 on success, -ve errno on failure. */ -int __isolate_lru_page(struct page *page, int mode) +int __isolate_lru_page(struct page *page, int mode, int file) { int ret = -EINVAL; @@ -696,6 +696,9 @@ int __isolate_lru_page(struct page *page, int mode) if (mode != ISOLATE_BOTH && (!PageActive(page) != !mode)) return ret; + if (mode != ISOLATE_BOTH && (!page_is_file_cache(page) != !file)) + return ret; + ret = -EBUSY; if (likely(get_page_unless_zero(page))) { /* @@ -726,12 +729,13 @@ int __isolate_lru_page(struct page *page, int mode) * @scanned: The number of pages that were scanned. * @order: The caller's attempted allocation order * @mode: One of the LRU isolation modes + * @file: True [1] if isolating file [!anon] pages * * returns how many pages were moved onto *@dst. */ static unsigned long isolate_lru_pages(unsigned long nr_to_scan, struct list_head *src, struct list_head *dst, - unsigned long *scanned, int order, int mode) + unsigned long *scanned, int order, int mode, int file) { unsigned long nr_taken = 0; unsigned long scan; @@ -748,7 +752,7 @@ static unsigned long isolate_lru_pages(unsigned long nr_to_scan, VM_BUG_ON(!PageLRU(page)); - switch (__isolate_lru_page(page, mode)) { + switch (__isolate_lru_page(page, mode, file)) { case 0: list_move(&page->lru, dst); nr_taken++; @@ -791,10 +795,11 @@ static unsigned long isolate_lru_pages(unsigned long nr_to_scan, break; cursor_page = pfn_to_page(pfn); + /* Check that we have not crossed a zone boundary. */ if (unlikely(page_zone_id(cursor_page) != zone_id)) continue; - switch (__isolate_lru_page(cursor_page, mode)) { + switch (__isolate_lru_page(cursor_page, mode, file)) { case 0: list_move(&cursor_page->lru, dst); nr_taken++; @@ -819,30 +824,37 @@ static unsigned long isolate_pages_global(unsigned long nr, unsigned long *scanned, int order, int mode, struct zone *z, struct mem_cgroup *mem_cont, - int active) + int active, int file) { + int lru = LRU_BASE; if (active) - return isolate_lru_pages(nr, &z->lru[LRU_ACTIVE].list, dst, - scanned, order, mode); - else - return isolate_lru_pages(nr, &z->lru[LRU_INACTIVE].list, dst, - scanned, order, mode); + lru += LRU_ACTIVE; + if (file) + lru += LRU_FILE; + return isolate_lru_pages(nr, &z->lru[lru].list, dst, scanned, order, + mode, !!file); } /* * clear_active_flags() is a helper for shrink_active_list(), clearing * any active bits from the pages in the list. */ -static unsigned long clear_active_flags(struct list_head *page_list) +static unsigned long clear_active_flags(struct list_head *page_list, + unsigned int *count) { int nr_active = 0; + int lru; struct page *page; - list_for_each_entry(page, page_list, lru) + list_for_each_entry(page, page_list, lru) { + lru = page_is_file_cache(page); if (PageActive(page)) { + lru += LRU_ACTIVE; ClearPageActive(page); nr_active++; } + count[lru]++; + } return nr_active; } @@ -880,12 +892,12 @@ int isolate_lru_page(struct page *page) spin_lock_irq(&zone->lru_lock); if (PageLRU(page) && get_page_unless_zero(page)) { + int lru = LRU_BASE; ret = 0; ClearPageLRU(page); - if (PageActive(page)) - del_page_from_active_list(zone, page); - else - del_page_from_inactive_list(zone, page); + + lru += page_is_file_cache(page) + !!PageActive(page); + del_page_from_lru_list(zone, page, lru); } spin_unlock_irq(&zone->lru_lock); } @@ -897,7 +909,7 @@ int isolate_lru_page(struct page *page) * of reclaimed pages */ static unsigned long shrink_inactive_list(unsigned long max_scan, - struct zone *zone, struct scan_control *sc) + struct zone *zone, struct scan_control *sc, int file) { LIST_HEAD(page_list); struct pagevec pvec; @@ -914,20 +926,32 @@ static unsigned long shrink_inactive_list(unsigned long max_scan, unsigned long nr_scan; unsigned long nr_freed; unsigned long nr_active; + unsigned int count[NR_LRU_LISTS] = { 0, }; + int mode = (sc->order > PAGE_ALLOC_COSTLY_ORDER) ? + ISOLATE_BOTH : ISOLATE_INACTIVE; nr_taken = sc->isolate_pages(sc->swap_cluster_max, - &page_list, &nr_scan, sc->order, - (sc->order > PAGE_ALLOC_COSTLY_ORDER)? - ISOLATE_BOTH : ISOLATE_INACTIVE, - zone, sc->mem_cgroup, 0); - nr_active = clear_active_flags(&page_list); + &page_list, &nr_scan, sc->order, mode, + zone, sc->mem_cgroup, 0, file); + nr_active = clear_active_flags(&page_list, count); __count_vm_events(PGDEACTIVATE, nr_active); - __mod_zone_page_state(zone, NR_ACTIVE, -nr_active); - __mod_zone_page_state(zone, NR_INACTIVE, - -(nr_taken - nr_active)); - if (scan_global_lru(sc)) + __mod_zone_page_state(zone, NR_ACTIVE_FILE, + -count[LRU_ACTIVE_FILE]); + __mod_zone_page_state(zone, NR_INACTIVE_FILE, + -count[LRU_INACTIVE_FILE]); + __mod_zone_page_state(zone, NR_ACTIVE_ANON, + -count[LRU_ACTIVE_ANON]); + __mod_zone_page_state(zone, NR_INACTIVE_ANON, + -count[LRU_INACTIVE_ANON]); + + if (scan_global_lru(sc)) { zone->pages_scanned += nr_scan; + zone->recent_scanned[0] += count[LRU_INACTIVE_ANON]; + zone->recent_scanned[0] += count[LRU_ACTIVE_ANON]; + zone->recent_scanned[1] += count[LRU_INACTIVE_FILE]; + zone->recent_scanned[1] += count[LRU_ACTIVE_FILE]; + } spin_unlock_irq(&zone->lru_lock); nr_scanned += nr_scan; @@ -947,7 +971,7 @@ static unsigned long shrink_inactive_list(unsigned long max_scan, * The attempt at page out may have made some * of the pages active, mark them inactive again. */ - nr_active = clear_active_flags(&page_list); + nr_active = clear_active_flags(&page_list, count); count_vm_events(PGDEACTIVATE, nr_active); nr_freed += shrink_page_list(&page_list, sc, @@ -977,6 +1001,10 @@ static unsigned long shrink_inactive_list(unsigned long max_scan, SetPageLRU(page); list_del(&page->lru); add_page_to_lru_list(zone, page, page_lru(page)); + if (PageActive(page) && scan_global_lru(sc)) { + int file = !!page_is_file_cache(page); + zone->recent_rotated[file]++; + } if (!pagevec_add(&pvec, page)) { spin_unlock_irq(&zone->lru_lock); __pagevec_release(&pvec); @@ -1007,115 +1035,7 @@ static inline void note_zone_scanning_priority(struct zone *zone, int priority) static inline int zone_is_near_oom(struct zone *zone) { - return zone->pages_scanned >= (zone_page_state(zone, NR_ACTIVE) - + zone_page_state(zone, NR_INACTIVE))*3; -} - -/* - * Determine we should try to reclaim mapped pages. - * This is called only when sc->mem_cgroup is NULL. - */ -static int calc_reclaim_mapped(struct scan_control *sc, struct zone *zone, - int priority) -{ - long mapped_ratio; - long distress; - long swap_tendency; - long imbalance; - int reclaim_mapped = 0; - int prev_priority; - - if (scan_global_lru(sc) && zone_is_near_oom(zone)) - return 1; - /* - * `distress' is a measure of how much trouble we're having - * reclaiming pages. 0 -> no problems. 100 -> great trouble. - */ - if (scan_global_lru(sc)) - prev_priority = zone->prev_priority; - else - prev_priority = mem_cgroup_get_reclaim_priority(sc->mem_cgroup); - - distress = 100 >> min(prev_priority, priority); - - /* - * The point of this algorithm is to decide when to start - * reclaiming mapped memory instead of just pagecache. Work out - * how much memory - * is mapped. - */ - if (scan_global_lru(sc)) - mapped_ratio = ((global_page_state(NR_FILE_MAPPED) + - global_page_state(NR_ANON_PAGES)) * 100) / - vm_total_pages; - else - mapped_ratio = mem_cgroup_calc_mapped_ratio(sc->mem_cgroup); - - /* - * Now decide how much we really want to unmap some pages. The - * mapped ratio is downgraded - just because there's a lot of - * mapped memory doesn't necessarily mean that page reclaim - * isn't succeeding. - * - * The distress ratio is important - we don't want to start - * going oom. - * - * A 100% value of vm_swappiness overrides this algorithm - * altogether. - */ - swap_tendency = mapped_ratio / 2 + distress + sc->swappiness; - - /* - * If there's huge imbalance between active and inactive - * (think active 100 times larger than inactive) we should - * become more permissive, or the system will take too much - * cpu before it start swapping during memory pressure. - * Distress is about avoiding early-oom, this is about - * making swappiness graceful despite setting it to low - * values. - * - * Avoid div by zero with nr_inactive+1, and max resulting - * value is vm_total_pages. - */ - if (scan_global_lru(sc)) { - imbalance = zone_page_state(zone, NR_ACTIVE); - imbalance /= zone_page_state(zone, NR_INACTIVE) + 1; - } else - imbalance = mem_cgroup_reclaim_imbalance(sc->mem_cgroup); - - /* - * Reduce the effect of imbalance if swappiness is low, - * this means for a swappiness very low, the imbalance - * must be much higher than 100 for this logic to make - * the difference. - * - * Max temporary value is vm_total_pages*100. - */ - imbalance *= (vm_swappiness + 1); - imbalance /= 100; - - /* - * If not much of the ram is mapped, makes the imbalance - * less relevant, it's high priority we refill the inactive - * list with mapped pages only in presence of high ratio of - * mapped pages. - * - * Max temporary value is vm_total_pages*100. - */ - imbalance *= mapped_ratio; - imbalance /= 100; - - /* apply imbalance feedback to swap_tendency */ - swap_tendency += imbalance; - - /* - * Now use this metric to decide whether to start moving mapped - * memory onto the inactive list. - */ - if (swap_tendency >= 100) - reclaim_mapped = 1; - - return reclaim_mapped; + return zone->pages_scanned >= (zone_lru_pages(zone) * 3); } /* @@ -1138,7 +1058,7 @@ static int calc_reclaim_mapped(struct scan_control *sc, struct zone *zone, static void shrink_active_list(unsigned long nr_pages, struct zone *zone, - struct scan_control *sc, int priority) + struct scan_control *sc, int priority, int file) { unsigned long pgmoved; int pgdeactivate = 0; @@ -1148,43 +1068,42 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, LIST_HEAD(l_inactive); struct page *page; struct pagevec pvec; - int reclaim_mapped = 0; - - if (sc->may_swap) - reclaim_mapped = calc_reclaim_mapped(sc, zone, priority); + enum lru_list lru; lru_add_drain(); spin_lock_irq(&zone->lru_lock); pgmoved = sc->isolate_pages(nr_pages, &l_hold, &pgscanned, sc->order, ISOLATE_ACTIVE, zone, - sc->mem_cgroup, 1); + sc->mem_cgroup, 1, file); /* * zone->pages_scanned is used for detect zone's oom * mem_cgroup remembers nr_scan by itself. */ - if (scan_global_lru(sc)) + if (scan_global_lru(sc)) { zone->pages_scanned += pgscanned; + zone->recent_scanned[!!file] += pgmoved; + } - __mod_zone_page_state(zone, NR_ACTIVE, -pgmoved); + if (file) + __mod_zone_page_state(zone, NR_ACTIVE_FILE, -pgmoved); + else + __mod_zone_page_state(zone, NR_ACTIVE_ANON, -pgmoved); spin_unlock_irq(&zone->lru_lock); while (!list_empty(&l_hold)) { cond_resched(); page = lru_to_page(&l_hold); list_del(&page->lru); - if (page_mapped(page)) { - if (!reclaim_mapped || - (total_swap_pages == 0 && PageAnon(page)) || - page_referenced(page, 0, sc->mem_cgroup)) { - list_add(&page->lru, &l_active); - continue; - } - } list_add(&page->lru, &l_inactive); } + /* + * Now put the pages back on the appropriate [file or anon] inactive + * and active lists. + */ pagevec_init(&pvec, 1); pgmoved = 0; + lru = LRU_BASE + file * LRU_FILE; spin_lock_irq(&zone->lru_lock); while (!list_empty(&l_inactive)) { page = lru_to_page(&l_inactive); @@ -1194,11 +1113,11 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, VM_BUG_ON(!PageActive(page)); ClearPageActive(page); - list_move(&page->lru, &zone->lru[LRU_INACTIVE].list); + list_move(&page->lru, &zone->lru[lru].list); mem_cgroup_move_lists(page, false); pgmoved++; if (!pagevec_add(&pvec, page)) { - __mod_zone_page_state(zone, NR_INACTIVE, pgmoved); + __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved); spin_unlock_irq(&zone->lru_lock); pgdeactivate += pgmoved; pgmoved = 0; @@ -1208,7 +1127,7 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, spin_lock_irq(&zone->lru_lock); } } - __mod_zone_page_state(zone, NR_INACTIVE, pgmoved); + __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved); pgdeactivate += pgmoved; if (buffer_heads_over_limit) { spin_unlock_irq(&zone->lru_lock); @@ -1217,6 +1136,7 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, } pgmoved = 0; + lru = LRU_ACTIVE + file * LRU_FILE; while (!list_empty(&l_active)) { page = lru_to_page(&l_active); prefetchw_prev_lru_page(page, &l_active, flags); @@ -1224,11 +1144,11 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, SetPageLRU(page); VM_BUG_ON(!PageActive(page)); - list_move(&page->lru, &zone->lru[LRU_ACTIVE].list); + list_move(&page->lru, &zone->lru[lru].list); mem_cgroup_move_lists(page, true); pgmoved++; if (!pagevec_add(&pvec, page)) { - __mod_zone_page_state(zone, NR_ACTIVE, pgmoved); + __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved); pgmoved = 0; spin_unlock_irq(&zone->lru_lock); if (vm_swap_full()) @@ -1237,7 +1157,8 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, spin_lock_irq(&zone->lru_lock); } } - __mod_zone_page_state(zone, NR_ACTIVE, pgmoved); + __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved); + zone->recent_rotated[!!file] += pgmoved; __count_zone_vm_events(PGREFILL, zone, pgscanned); __count_vm_events(PGDEACTIVATE, pgdeactivate); @@ -1248,16 +1169,103 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, pagevec_release(&pvec); } -static unsigned long shrink_list(enum lru_list l, unsigned long nr_to_scan, +static unsigned long shrink_list(enum lru_list lru, unsigned long nr_to_scan, struct zone *zone, struct scan_control *sc, int priority) { - if (l == LRU_ACTIVE) { - shrink_active_list(nr_to_scan, zone, sc, priority); + int file = is_file_lru(lru); + + if (lru == LRU_ACTIVE_ANON || lru == LRU_ACTIVE_FILE) { + shrink_active_list(nr_to_scan, zone, sc, priority, file); return 0; } - return shrink_inactive_list(nr_to_scan, zone, sc); + return shrink_inactive_list(nr_to_scan, zone, sc, file); +} + +/* + * Determine how aggressively the anon and file LRU lists should be + * scanned. The relative value of each set of LRU lists is determined + * by looking at the fraction of the pages scanned we did rotate back + * onto the active list instead of evict. + * + * percent[0] specifies how much pressure to put on ram/swap backed + * memory, while percent[1] determines pressure on the file LRUs. + */ +static void get_scan_ratio(struct zone *zone, struct scan_control *sc, + unsigned long *percent) +{ + unsigned long anon, file, free; + unsigned long anon_prio, file_prio; + unsigned long ap, fp; + + anon = zone_page_state(zone, NR_ACTIVE_ANON) + + zone_page_state(zone, NR_INACTIVE_ANON); + file = zone_page_state(zone, NR_ACTIVE_FILE) + + zone_page_state(zone, NR_INACTIVE_FILE); + free = zone_page_state(zone, NR_FREE_PAGES); + + /* If we have no swap space, do not bother scanning anon pages. */ + if (nr_swap_pages <= 0) { + percent[0] = 0; + percent[1] = 100; + return; + } + + /* If we have very few page cache pages, force-scan anon pages. */ + if (unlikely(file + free <= zone->pages_high)) { + percent[0] = 100; + percent[1] = 0; + return; + } + + /* + * OK, so we have swap space and a fair amount of page cache + * pages. We use the recently rotated / recently scanned + * ratios to determine how valuable each cache is. + * + * Because workloads change over time (and to avoid overflow) + * we keep these statistics as a floating average, which ends + * up weighing recent references more than old ones. + * + * anon in [0], file in [1] + */ + if (unlikely(zone->recent_scanned[0] > anon / 4)) { + spin_lock_irq(&zone->lru_lock); + zone->recent_scanned[0] /= 2; + zone->recent_rotated[0] /= 2; + spin_unlock_irq(&zone->lru_lock); + } + + if (unlikely(zone->recent_scanned[1] > file / 4)) { + spin_lock_irq(&zone->lru_lock); + zone->recent_scanned[1] /= 2; + zone->recent_rotated[1] /= 2; + spin_unlock_irq(&zone->lru_lock); + } + + /* + * With swappiness at 100, anonymous and file have the same priority. + * This scanning priority is essentially the inverse of IO cost. + */ + anon_prio = sc->swappiness; + file_prio = 200 - sc->swappiness; + + /* + * anon recent_rotated[0] + * %anon = 100 * ----------- / ----------------- * IO cost + * anon + file rotate_sum + */ + ap = (anon_prio + 1) * (zone->recent_scanned[0] + 1); + ap /= zone->recent_rotated[0] + 1; + + fp = (file_prio + 1) * (zone->recent_scanned[1] + 1); + fp /= zone->recent_rotated[1] + 1; + + /* Normalize to percentages */ + percent[0] = 100 * ap / (ap + fp + 1); + percent[1] = 100 - percent[0]; } + /* * This is a basic per-zone page freer. Used by both kswapd and direct reclaim. */ @@ -1267,36 +1275,43 @@ static unsigned long shrink_zone(int priority, struct zone *zone, unsigned long nr[NR_LRU_LISTS]; unsigned long nr_to_scan; unsigned long nr_reclaimed = 0; + unsigned long percent[2]; /* anon @ 0; file @ 1 */ enum lru_list l; - if (scan_global_lru(sc)) { - /* - * Add one to nr_to_scan just to make sure that the kernel - * will slowly sift through the active list. - */ - for_each_lru(l) { - zone->lru[l].nr_scan += (zone_page_state(zone, - NR_LRU_BASE + l) >> priority) + 1; + get_scan_ratio(zone, sc, percent); + + for_each_lru(l) { + if (scan_global_lru(sc)) { + int file = is_file_lru(l); + int scan; + /* + * Add one to nr_to_scan just to make sure that the + * kernel will slowly sift through each list. + */ + scan = zone_page_state(zone, NR_LRU_BASE + l); + if (priority) { + scan >>= priority; + scan = (scan * percent[file]) / 100; + } + zone->lru[l].nr_scan += scan + 1; nr[l] = zone->lru[l].nr_scan; if (nr[l] >= sc->swap_cluster_max) zone->lru[l].nr_scan = 0; else nr[l] = 0; + } else { + /* + * This reclaim occurs not because zone memory shortage + * but because memory controller hits its limit. + * Don't modify zone reclaim related data. + */ + nr[l] = mem_cgroup_calc_reclaim(sc->mem_cgroup, zone, + priority, l); } - } else { - /* - * This reclaim occurs not because zone memory shortage but - * because memory controller hits its limit. - * Then, don't modify zone reclaim related data. - */ - nr[LRU_ACTIVE] = mem_cgroup_calc_reclaim(sc->mem_cgroup, - zone, priority, LRU_ACTIVE); - - nr[LRU_INACTIVE] = mem_cgroup_calc_reclaim(sc->mem_cgroup, - zone, priority, LRU_INACTIVE); } - while (nr[LRU_ACTIVE] || nr[LRU_INACTIVE]) { + while (nr[LRU_ACTIVE_ANON] || nr[LRU_INACTIVE_ANON] || + nr[LRU_ACTIVE_FILE] || nr[LRU_INACTIVE_FILE]) { for_each_lru(l) { if (nr[l]) { nr_to_scan = min(nr[l], @@ -1369,7 +1384,7 @@ static unsigned long shrink_zones(int priority, struct zonelist *zonelist, return nr_reclaimed; } - + /* * This is the main entry point to direct page reclaim. * @@ -1412,8 +1427,7 @@ static unsigned long do_try_to_free_pages(struct zonelist *zonelist, if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) continue; - lru_pages += zone_page_state(zone, NR_ACTIVE) - + zone_page_state(zone, NR_INACTIVE); + lru_pages += zone_lru_pages(zone); } } @@ -1615,8 +1629,7 @@ loop_again: for (i = 0; i <= end_zone; i++) { struct zone *zone = pgdat->node_zones + i; - lru_pages += zone_page_state(zone, NR_ACTIVE) - + zone_page_state(zone, NR_INACTIVE); + lru_pages += zone_lru_pages(zone); } /* @@ -1660,8 +1673,7 @@ loop_again: if (zone_is_all_unreclaimable(zone)) continue; if (nr_slab == 0 && zone->pages_scanned >= - (zone_page_state(zone, NR_ACTIVE) - + zone_page_state(zone, NR_INACTIVE)) * 6) + (zone_lru_pages(zone) * 6)) zone_set_flag(zone, ZONE_ALL_UNRECLAIMABLE); /* @@ -1715,7 +1727,7 @@ out: /* * The background pageout daemon, started as a kernel thread - * from the init process. + * from the init process. * * This basically trickles out pages so that we have _some_ * free memory available even if there is no other activity @@ -1809,6 +1821,14 @@ void wakeup_kswapd(struct zone *zone, int order) wake_up_interruptible(&pgdat->kswapd_wait); } +unsigned long global_lru_pages(void) +{ + return global_page_state(NR_ACTIVE_ANON) + + global_page_state(NR_ACTIVE_FILE) + + global_page_state(NR_INACTIVE_ANON) + + global_page_state(NR_INACTIVE_FILE); +} + #ifdef CONFIG_PM /* * Helper function for shrink_all_memory(). Tries to reclaim 'nr_pages' pages @@ -1834,7 +1854,8 @@ static unsigned long shrink_all_zones(unsigned long nr_pages, int prio, for_each_lru(l) { /* For pass = 0 we don't shrink the active list */ - if (pass == 0 && l == LRU_ACTIVE) + if (pass == 0 && + (l == LRU_ACTIVE || l == LRU_ACTIVE_FILE)) continue; zone->lru[l].nr_scan += @@ -1856,11 +1877,6 @@ static unsigned long shrink_all_zones(unsigned long nr_pages, int prio, return ret; } -static unsigned long count_lru_pages(void) -{ - return global_page_state(NR_ACTIVE) + global_page_state(NR_INACTIVE); -} - /* * Try to free `nr_pages' of memory, system-wide, and return the number of * freed pages. @@ -1886,7 +1902,7 @@ unsigned long shrink_all_memory(unsigned long nr_pages) current->reclaim_state = &reclaim_state; - lru_pages = count_lru_pages(); + lru_pages = global_lru_pages(); nr_slab = global_page_state(NR_SLAB_RECLAIMABLE); /* If slab caches are huge, it's better to hit them first */ while (nr_slab >= lru_pages) { @@ -1929,7 +1945,7 @@ unsigned long shrink_all_memory(unsigned long nr_pages) reclaim_state.reclaimed_slab = 0; shrink_slab(sc.nr_scanned, sc.gfp_mask, - count_lru_pages()); + global_lru_pages()); ret += reclaim_state.reclaimed_slab; if (ret >= nr_pages) goto out; @@ -1946,7 +1962,7 @@ unsigned long shrink_all_memory(unsigned long nr_pages) if (!ret) { do { reclaim_state.reclaimed_slab = 0; - shrink_slab(nr_pages, sc.gfp_mask, count_lru_pages()); + shrink_slab(nr_pages, sc.gfp_mask, global_lru_pages()); ret += reclaim_state.reclaimed_slab; } while (ret < nr_pages && reclaim_state.reclaimed_slab > 0); } -- cgit v1.2.3 From 556adecba110bf5f1db6c6b56416cfab5bcab698 Mon Sep 17 00:00:00 2001 From: Rik van Riel Date: Sat, 18 Oct 2008 20:26:34 -0700 Subject: vmscan: second chance replacement for anonymous pages We avoid evicting and scanning anonymous pages for the most part, but under some workloads we can end up with most of memory filled with anonymous pages. At that point, we suddenly need to clear the referenced bits on all of memory, which can take ages on very large memory systems. We can reduce the maximum number of pages that need to be scanned by not taking the referenced state into account when deactivating an anonymous page. After all, every anonymous page starts out referenced, so why check? If an anonymous page gets referenced again before it reaches the end of the inactive list, we move it back to the active list. To keep the maximum amount of necessary work reasonable, we scale the active to inactive ratio with the size of memory, using the formula active:inactive ratio = sqrt(memory in GB * 10). Kswapd CPU use now seems to scale by the amount of pageout bandwidth, instead of by the amount of memory present in the system. [kamezawa.hiroyu@jp.fujitsu.com: fix OOM with memcg] [kamezawa.hiroyu@jp.fujitsu.com: memcg: lru scan fix] Signed-off-by: Rik van Riel Signed-off-by: KOSAKI Motohiro Signed-off-by: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/vmscan.c | 38 ++++++++++++++++++++++++++++++++++---- 1 file changed, 34 insertions(+), 4 deletions(-) (limited to 'mm/vmscan.c') diff --git a/mm/vmscan.c b/mm/vmscan.c index d10d2f9a33f..c82ee9a33cf 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -1090,6 +1090,7 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, __mod_zone_page_state(zone, NR_ACTIVE_ANON, -pgmoved); spin_unlock_irq(&zone->lru_lock); + pgmoved = 0; while (!list_empty(&l_hold)) { cond_resched(); page = lru_to_page(&l_hold); @@ -1097,6 +1098,13 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, list_add(&page->lru, &l_inactive); } + /* + * Count the referenced pages as rotated, even when they are moved + * to the inactive list. This helps balance scan pressure between + * file and anonymous pages in get_scan_ratio. + */ + zone->recent_rotated[!!file] += pgmoved; + /* * Now put the pages back on the appropriate [file or anon] inactive * and active lists. @@ -1158,7 +1166,6 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, } } __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved); - zone->recent_rotated[!!file] += pgmoved; __count_zone_vm_events(PGREFILL, zone, pgscanned); __count_vm_events(PGDEACTIVATE, pgdeactivate); @@ -1174,7 +1181,13 @@ static unsigned long shrink_list(enum lru_list lru, unsigned long nr_to_scan, { int file = is_file_lru(lru); - if (lru == LRU_ACTIVE_ANON || lru == LRU_ACTIVE_FILE) { + if (lru == LRU_ACTIVE_FILE) { + shrink_active_list(nr_to_scan, zone, sc, priority, file); + return 0; + } + + if (lru == LRU_ACTIVE_ANON && + (!scan_global_lru(sc) || inactive_anon_is_low(zone))) { shrink_active_list(nr_to_scan, zone, sc, priority, file); return 0; } @@ -1310,8 +1323,8 @@ static unsigned long shrink_zone(int priority, struct zone *zone, } } - while (nr[LRU_ACTIVE_ANON] || nr[LRU_INACTIVE_ANON] || - nr[LRU_ACTIVE_FILE] || nr[LRU_INACTIVE_FILE]) { + while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] || + nr[LRU_INACTIVE_FILE]) { for_each_lru(l) { if (nr[l]) { nr_to_scan = min(nr[l], @@ -1324,6 +1337,15 @@ static unsigned long shrink_zone(int priority, struct zone *zone, } } + /* + * Even if we did not try to evict anon pages at all, we want to + * rebalance the anon lru active/inactive ratio. + */ + if (!scan_global_lru(sc) || inactive_anon_is_low(zone)) + shrink_active_list(SWAP_CLUSTER_MAX, zone, sc, priority, 0); + else if (!scan_global_lru(sc)) + shrink_active_list(SWAP_CLUSTER_MAX, zone, sc, priority, 0); + throttle_vm_writeout(sc->gfp_mask); return nr_reclaimed; } @@ -1617,6 +1639,14 @@ loop_again: priority != DEF_PRIORITY) continue; + /* + * Do some background aging of the anon list, to give + * pages a chance to be referenced before reclaiming. + */ + if (inactive_anon_is_low(zone)) + shrink_active_list(SWAP_CLUSTER_MAX, zone, + &sc, priority, 0); + if (!zone_watermark_ok(zone, order, zone->pages_high, 0, 0)) { end_zone = i; -- cgit v1.2.3 From 7e9cd484204f9e5b316ed35b241abf088d76e0af Mon Sep 17 00:00:00 2001 From: Rik van Riel Date: Sat, 18 Oct 2008 20:26:35 -0700 Subject: vmscan: fix pagecache reclaim referenced bit check Moving referenced pages back to the head of the active list creates a huge scalability problem, because by the time a large memory system finally runs out of free memory, every single page in the system will have been referenced. Not only do we not have the time to scan every single page on the active list, but since they have will all have the referenced bit set, that bit conveys no useful information. A more scalable solution is to just move every page that hits the end of the active list to the inactive list. We clear the referenced bit off of mapped pages, which need just one reference to be moved back onto the active list. Unmapped pages will be moved back to the active list after two references (see mark_page_accessed). We preserve the PG_referenced flag on unmapped pages to preserve accesses that were made while the page was on the active list. Signed-off-by: Rik van Riel Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/vmscan.c | 45 +++++++++++++-------------------------------- 1 file changed, 13 insertions(+), 32 deletions(-) (limited to 'mm/vmscan.c') diff --git a/mm/vmscan.c b/mm/vmscan.c index c82ee9a33cf..9588973849d 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -1064,7 +1064,6 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, int pgdeactivate = 0; unsigned long pgscanned; LIST_HEAD(l_hold); /* The pages which were snipped off */ - LIST_HEAD(l_active); LIST_HEAD(l_inactive); struct page *page; struct pagevec pvec; @@ -1095,21 +1094,28 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, cond_resched(); page = lru_to_page(&l_hold); list_del(&page->lru); + + /* page_referenced clears PageReferenced */ + if (page_mapping_inuse(page) && + page_referenced(page, 0, sc->mem_cgroup)) + pgmoved++; + list_add(&page->lru, &l_inactive); } /* - * Count the referenced pages as rotated, even when they are moved - * to the inactive list. This helps balance scan pressure between - * file and anonymous pages in get_scan_ratio. - */ + * Count referenced pages from currently used mappings as + * rotated, even though they are moved to the inactive list. + * This helps balance scan pressure between file and anonymous + * pages in get_scan_ratio. + */ zone->recent_rotated[!!file] += pgmoved; /* - * Now put the pages back on the appropriate [file or anon] inactive - * and active lists. + * Move the pages to the [file or anon] inactive list. */ pagevec_init(&pvec, 1); + pgmoved = 0; lru = LRU_BASE + file * LRU_FILE; spin_lock_irq(&zone->lru_lock); @@ -1142,31 +1148,6 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, pagevec_strip(&pvec); spin_lock_irq(&zone->lru_lock); } - - pgmoved = 0; - lru = LRU_ACTIVE + file * LRU_FILE; - while (!list_empty(&l_active)) { - page = lru_to_page(&l_active); - prefetchw_prev_lru_page(page, &l_active, flags); - VM_BUG_ON(PageLRU(page)); - SetPageLRU(page); - VM_BUG_ON(!PageActive(page)); - - list_move(&page->lru, &zone->lru[lru].list); - mem_cgroup_move_lists(page, true); - pgmoved++; - if (!pagevec_add(&pvec, page)) { - __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved); - pgmoved = 0; - spin_unlock_irq(&zone->lru_lock); - if (vm_swap_full()) - pagevec_swap_free(&pvec); - __pagevec_release(&pvec); - spin_lock_irq(&zone->lru_lock); - } - } - __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved); - __count_zone_vm_events(PGREFILL, zone, pgscanned); __count_vm_events(PGDEACTIVATE, pgdeactivate); spin_unlock_irq(&zone->lru_lock); -- cgit v1.2.3 From 33c120ed2843090e2bd316de1588b8bf8b96cbde Mon Sep 17 00:00:00 2001 From: Rik van Riel Date: Sat, 18 Oct 2008 20:26:36 -0700 Subject: more aggressively use lumpy reclaim During an AIM7 run on a 16GB system, fork started failing around 32000 threads, despite the system having plenty of free swap and 15GB of pageable memory. This was on x86-64, so 8k stacks. If a higher order allocation fails, we can either: - keep evicting pages off the end of the LRUs and hope that we eventually create a contiguous region; this is somewhat unlikely if the system is under enough stress by new allocations - after trying normal eviction for a bit, use lumpy reclaim This patch switches the system to lumpy reclaim if the VM is having trouble freeing enough pages, using the same threshold for detection as used by pageout congestion wait. Signed-off-by: Rik van Riel Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/vmscan.c | 20 ++++++++++++++++---- 1 file changed, 16 insertions(+), 4 deletions(-) (limited to 'mm/vmscan.c') diff --git a/mm/vmscan.c b/mm/vmscan.c index 9588973849d..a8347b677e7 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -909,7 +909,8 @@ int isolate_lru_page(struct page *page) * of reclaimed pages */ static unsigned long shrink_inactive_list(unsigned long max_scan, - struct zone *zone, struct scan_control *sc, int file) + struct zone *zone, struct scan_control *sc, + int priority, int file) { LIST_HEAD(page_list); struct pagevec pvec; @@ -927,8 +928,19 @@ static unsigned long shrink_inactive_list(unsigned long max_scan, unsigned long nr_freed; unsigned long nr_active; unsigned int count[NR_LRU_LISTS] = { 0, }; - int mode = (sc->order > PAGE_ALLOC_COSTLY_ORDER) ? - ISOLATE_BOTH : ISOLATE_INACTIVE; + int mode = ISOLATE_INACTIVE; + + /* + * If we need a large contiguous chunk of memory, or have + * trouble getting a small set of contiguous pages, we + * will reclaim both active and inactive pages. + * + * We use the same threshold as pageout congestion_wait below. + */ + if (sc->order > PAGE_ALLOC_COSTLY_ORDER) + mode = ISOLATE_BOTH; + else if (sc->order && priority < DEF_PRIORITY - 2) + mode = ISOLATE_BOTH; nr_taken = sc->isolate_pages(sc->swap_cluster_max, &page_list, &nr_scan, sc->order, mode, @@ -1172,7 +1184,7 @@ static unsigned long shrink_list(enum lru_list lru, unsigned long nr_to_scan, shrink_active_list(nr_to_scan, zone, sc, priority, file); return 0; } - return shrink_inactive_list(nr_to_scan, zone, sc, file); + return shrink_inactive_list(nr_to_scan, zone, sc, priority, file); } /* -- cgit v1.2.3 From 894bc310419ac95f4fa4142dc364401a7e607f65 Mon Sep 17 00:00:00 2001 From: Lee Schermerhorn Date: Sat, 18 Oct 2008 20:26:39 -0700 Subject: Unevictable LRU Infrastructure When the system contains lots of mlocked or otherwise unevictable pages, the pageout code (kswapd) can spend lots of time scanning over these pages. Worse still, the presence of lots of unevictable pages can confuse kswapd into thinking that more aggressive pageout modes are required, resulting in all kinds of bad behaviour. Infrastructure to manage pages excluded from reclaim--i.e., hidden from vmscan. Based on a patch by Larry Woodman of Red Hat. Reworked to maintain "unevictable" pages on a separate per-zone LRU list, to "hide" them from vmscan. Kosaki Motohiro added the support for the memory controller unevictable lru list. Pages on the unevictable list have both PG_unevictable and PG_lru set. Thus, PG_unevictable is analogous to and mutually exclusive with PG_active--it specifies which LRU list the page is on. The unevictable infrastructure is enabled by a new mm Kconfig option [CONFIG_]UNEVICTABLE_LRU. A new function 'page_evictable(page, vma)' in vmscan.c tests whether or not a page may be evictable. Subsequent patches will add the various !evictable tests. We'll want to keep these tests light-weight for use in shrink_active_list() and, possibly, the fault path. To avoid races between tasks putting pages [back] onto an LRU list and tasks that might be moving the page from non-evictable to evictable state, the new function 'putback_lru_page()' -- inverse to 'isolate_lru_page()' -- tests the "evictability" of a page after placing it on the LRU, before dropping the reference. If the page has become unevictable, putback_lru_page() will redo the 'putback', thus moving the page to the unevictable list. This way, we avoid "stranding" evictable pages on the unevictable list. [akpm@linux-foundation.org: fix fallout from out-of-order merge] [riel@redhat.com: fix UNEVICTABLE_LRU and !PROC_PAGE_MONITOR build] [nishimura@mxp.nes.nec.co.jp: remove redundant mapping check] [kosaki.motohiro@jp.fujitsu.com: unevictable-lru-infrastructure: putback_lru_page()/unevictable page handling rework] [kosaki.motohiro@jp.fujitsu.com: kill unnecessary lock_page() in vmscan.c] [kosaki.motohiro@jp.fujitsu.com: revert migration change of unevictable lru infrastructure] [kosaki.motohiro@jp.fujitsu.com: revert to unevictable-lru-infrastructure-kconfig-fix.patch] [kosaki.motohiro@jp.fujitsu.com: restore patch failure of vmstat-unevictable-and-mlocked-pages-vm-events.patch] Signed-off-by: Lee Schermerhorn Signed-off-by: Rik van Riel Signed-off-by: KOSAKI Motohiro Debugged-by: Benjamin Kidwell Signed-off-by: Daisuke Nishimura Signed-off-by: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/vmscan.c | 149 ++++++++++++++++++++++++++++++++++++++++++++++++++++++------ 1 file changed, 136 insertions(+), 13 deletions(-) (limited to 'mm/vmscan.c') diff --git a/mm/vmscan.c b/mm/vmscan.c index a8347b677e7..154b9b608da 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -470,6 +470,79 @@ int remove_mapping(struct address_space *mapping, struct page *page) return 0; } +/** + * putback_lru_page - put previously isolated page onto appropriate LRU list + * @page: page to be put back to appropriate lru list + * + * Add previously isolated @page to appropriate LRU list. + * Page may still be unevictable for other reasons. + * + * lru_lock must not be held, interrupts must be enabled. + */ +#ifdef CONFIG_UNEVICTABLE_LRU +void putback_lru_page(struct page *page) +{ + int lru; + int active = !!TestClearPageActive(page); + + VM_BUG_ON(PageLRU(page)); + +redo: + ClearPageUnevictable(page); + + if (page_evictable(page, NULL)) { + /* + * For evictable pages, we can use the cache. + * In event of a race, worst case is we end up with an + * unevictable page on [in]active list. + * We know how to handle that. + */ + lru = active + page_is_file_cache(page); + lru_cache_add_lru(page, lru); + } else { + /* + * Put unevictable pages directly on zone's unevictable + * list. + */ + lru = LRU_UNEVICTABLE; + add_page_to_unevictable_list(page); + } + mem_cgroup_move_lists(page, lru); + + /* + * page's status can change while we move it among lru. If an evictable + * page is on unevictable list, it never be freed. To avoid that, + * check after we added it to the list, again. + */ + if (lru == LRU_UNEVICTABLE && page_evictable(page, NULL)) { + if (!isolate_lru_page(page)) { + put_page(page); + goto redo; + } + /* This means someone else dropped this page from LRU + * So, it will be freed or putback to LRU again. There is + * nothing to do here. + */ + } + + put_page(page); /* drop ref from isolate */ +} + +#else /* CONFIG_UNEVICTABLE_LRU */ + +void putback_lru_page(struct page *page) +{ + int lru; + VM_BUG_ON(PageLRU(page)); + + lru = !!TestClearPageActive(page) + page_is_file_cache(page); + lru_cache_add_lru(page, lru); + mem_cgroup_move_lists(page, lru); + put_page(page); +} +#endif /* CONFIG_UNEVICTABLE_LRU */ + + /* * shrink_page_list() returns the number of reclaimed pages */ @@ -503,6 +576,12 @@ static unsigned long shrink_page_list(struct list_head *page_list, sc->nr_scanned++; + if (unlikely(!page_evictable(page, NULL))) { + unlock_page(page); + putback_lru_page(page); + continue; + } + if (!sc->may_swap && page_mapped(page)) goto keep_locked; @@ -602,7 +681,7 @@ static unsigned long shrink_page_list(struct list_head *page_list, * possible for a page to have PageDirty set, but it is actually * clean (all its buffers are clean). This happens if the * buffers were written out directly, with submit_bh(). ext3 - * will do this, as well as the blockdev mapping. + * will do this, as well as the blockdev mapping. * try_to_release_page() will discover that cleanness and will * drop the buffers and mark the page clean - it can be freed. * @@ -650,6 +729,7 @@ activate_locked: /* Not a candidate for swapping, so reclaim swap space. */ if (PageSwapCache(page) && vm_swap_full()) remove_exclusive_swap_page_ref(page); + VM_BUG_ON(PageActive(page)); SetPageActive(page); pgactivate++; keep_locked: @@ -699,6 +779,14 @@ int __isolate_lru_page(struct page *page, int mode, int file) if (mode != ISOLATE_BOTH && (!page_is_file_cache(page) != !file)) return ret; + /* + * When this function is being called for lumpy reclaim, we + * initially look into all LRU pages, active, inactive and + * unevictable; only give shrink_page_list evictable pages. + */ + if (PageUnevictable(page)) + return ret; + ret = -EBUSY; if (likely(get_page_unless_zero(page))) { /* @@ -810,7 +898,7 @@ static unsigned long isolate_lru_pages(unsigned long nr_to_scan, /* else it is being freed elsewhere */ list_move(&cursor_page->lru, src); default: - break; + break; /* ! on LRU or wrong list */ } } } @@ -870,8 +958,9 @@ static unsigned long clear_active_flags(struct list_head *page_list, * Returns -EBUSY if the page was not on an LRU list. * * The returned page will have PageLRU() cleared. If it was found on - * the active list, it will have PageActive set. That flag may need - * to be cleared by the caller before letting the page go. + * the active list, it will have PageActive set. If it was found on + * the unevictable list, it will have the PageUnevictable bit set. That flag + * may need to be cleared by the caller before letting the page go. * * The vmstat statistic corresponding to the list on which the page was * found will be decremented. @@ -892,11 +981,10 @@ int isolate_lru_page(struct page *page) spin_lock_irq(&zone->lru_lock); if (PageLRU(page) && get_page_unless_zero(page)) { - int lru = LRU_BASE; + int lru = page_lru(page); ret = 0; ClearPageLRU(page); - lru += page_is_file_cache(page) + !!PageActive(page); del_page_from_lru_list(zone, page, lru); } spin_unlock_irq(&zone->lru_lock); @@ -1008,11 +1096,20 @@ static unsigned long shrink_inactive_list(unsigned long max_scan, * Put back any unfreeable pages. */ while (!list_empty(&page_list)) { + int lru; page = lru_to_page(&page_list); VM_BUG_ON(PageLRU(page)); - SetPageLRU(page); list_del(&page->lru); - add_page_to_lru_list(zone, page, page_lru(page)); + if (unlikely(!page_evictable(page, NULL))) { + spin_unlock_irq(&zone->lru_lock); + putback_lru_page(page); + spin_lock_irq(&zone->lru_lock); + continue; + } + SetPageLRU(page); + lru = page_lru(page); + add_page_to_lru_list(zone, page, lru); + mem_cgroup_move_lists(page, lru); if (PageActive(page) && scan_global_lru(sc)) { int file = !!page_is_file_cache(page); zone->recent_rotated[file]++; @@ -1107,6 +1204,11 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, page = lru_to_page(&l_hold); list_del(&page->lru); + if (unlikely(!page_evictable(page, NULL))) { + putback_lru_page(page); + continue; + } + /* page_referenced clears PageReferenced */ if (page_mapping_inuse(page) && page_referenced(page, 0, sc->mem_cgroup)) @@ -1140,7 +1242,7 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, ClearPageActive(page); list_move(&page->lru, &zone->lru[lru].list); - mem_cgroup_move_lists(page, false); + mem_cgroup_move_lists(page, lru); pgmoved++; if (!pagevec_add(&pvec, page)) { __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved); @@ -1286,7 +1388,7 @@ static unsigned long shrink_zone(int priority, struct zone *zone, get_scan_ratio(zone, sc, percent); - for_each_lru(l) { + for_each_evictable_lru(l) { if (scan_global_lru(sc)) { int file = is_file_lru(l); int scan; @@ -1318,7 +1420,7 @@ static unsigned long shrink_zone(int priority, struct zone *zone, while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] || nr[LRU_INACTIVE_FILE]) { - for_each_lru(l) { + for_each_evictable_lru(l) { if (nr[l]) { nr_to_scan = min(nr[l], (unsigned long)sc->swap_cluster_max); @@ -1875,8 +1977,8 @@ static unsigned long shrink_all_zones(unsigned long nr_pages, int prio, if (zone_is_all_unreclaimable(zone) && prio != DEF_PRIORITY) continue; - for_each_lru(l) { - /* For pass = 0 we don't shrink the active list */ + for_each_evictable_lru(l) { + /* For pass = 0, we don't shrink the active list */ if (pass == 0 && (l == LRU_ACTIVE || l == LRU_ACTIVE_FILE)) continue; @@ -2213,3 +2315,24 @@ int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order) return ret; } #endif + +#ifdef CONFIG_UNEVICTABLE_LRU +/* + * page_evictable - test whether a page is evictable + * @page: the page to test + * @vma: the VMA in which the page is or will be mapped, may be NULL + * + * Test whether page is evictable--i.e., should be placed on active/inactive + * lists vs unevictable list. + * + * Reasons page might not be evictable: + * TODO - later patches + */ +int page_evictable(struct page *page, struct vm_area_struct *vma) +{ + + /* TODO: test page [!]evictable conditions */ + + return 1; +} +#endif -- cgit v1.2.3 From bbfd28eee9fbd73e780b19beb3dc562befbb94fa Mon Sep 17 00:00:00 2001 From: Lee Schermerhorn Date: Sat, 18 Oct 2008 20:26:40 -0700 Subject: unevictable lru: add event counting with statistics Fix to unevictable-lru-page-statistics.patch Add unevictable lru infrastructure vm events to the statistics patch. Rename the "NORECL_" and "noreclaim_" symbols and text strings to "UNEVICTABLE_" and "unevictable_", respectively. Currently, both the infrastructure and the mlocked pages event are added by a single patch later in the series. This makes it difficult to add or rework the incremental patches. The events actually "belong" with the stats, so pull them up to here. Also, restore the event counting to putback_lru_page(). This was removed from previous patch in series where it was "misplaced". The actual events weren't defined that early. Signed-off-by: Lee Schermerhorn Cc: Rik van Riel Reviewed-by: KOSAKI Motohiro Cc: KAMEZAWA Hiroyuki Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/vmscan.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'mm/vmscan.c') diff --git a/mm/vmscan.c b/mm/vmscan.c index 154b9b608da..2804d23e2da 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -484,6 +484,7 @@ void putback_lru_page(struct page *page) { int lru; int active = !!TestClearPageActive(page); + int was_unevictable = PageUnevictable(page); VM_BUG_ON(PageLRU(page)); @@ -525,6 +526,11 @@ redo: */ } + if (was_unevictable && lru != LRU_UNEVICTABLE) + count_vm_event(UNEVICTABLE_PGRESCUED); + else if (!was_unevictable && lru == LRU_UNEVICTABLE) + count_vm_event(UNEVICTABLE_PGCULLED); + put_page(page); /* drop ref from isolate */ } -- cgit v1.2.3 From ba9ddf49391645e6bb93219131a40446538a5e76 Mon Sep 17 00:00:00 2001 From: Lee Schermerhorn Date: Sat, 18 Oct 2008 20:26:42 -0700 Subject: Ramfs and Ram Disk pages are unevictable Christoph Lameter pointed out that ram disk pages also clutter the LRU lists. When vmscan finds them dirty and tries to clean them, the ram disk writeback function just redirties the page so that it goes back onto the active list. Round and round she goes... With the ram disk driver [rd.c] replaced by the newer 'brd.c', this is no longer the case, as ram disk pages are no longer maintained on the lru. [This makes them unmigratable for defrag or memory hot remove, but that can be addressed by a separate patch series.] However, the ramfs pages behave like ram disk pages used to, so: Define new address_space flag [shares address_space flags member with mapping's gfp mask] to indicate that the address space contains all unevictable pages. This will provide for efficient testing of ramfs pages in page_evictable(). Also provide wrapper functions to set/test the unevictable state to minimize #ifdefs in ramfs driver and any other users of this facility. Set the unevictable state on address_space structures for new ramfs inodes. Test the unevictable state in page_evictable() to cull unevictable pages. These changes depend on [CONFIG_]UNEVICTABLE_LRU. [riel@redhat.com: undo the brd.c part] Signed-off-by: Lee Schermerhorn Signed-off-by: Rik van Riel Debugged-by: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/vmscan.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'mm/vmscan.c') diff --git a/mm/vmscan.c b/mm/vmscan.c index 2804d23e2da..9babfbc1ddc 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -2332,11 +2332,16 @@ int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order) * lists vs unevictable list. * * Reasons page might not be evictable: + * (1) page's mapping marked unevictable + * * TODO - later patches */ int page_evictable(struct page *page, struct vm_area_struct *vma) { + if (mapping_unevictable(page_mapping(page))) + return 0; + /* TODO: test page [!]evictable conditions */ return 1; -- cgit v1.2.3 From 89e004ea55abe201b29e2d6e35124101f1288ef7 Mon Sep 17 00:00:00 2001 From: Lee Schermerhorn Date: Sat, 18 Oct 2008 20:26:43 -0700 Subject: SHM_LOCKED pages are unevictable Shmem segments locked into memory via shmctl(SHM_LOCKED) should not be kept on the normal LRU, since scanning them is a waste of time and might throw off kswapd's balancing algorithms. Place them on the unevictable LRU list instead. Use the AS_UNEVICTABLE flag to mark address_space of SHM_LOCKed shared memory regions as unevictable. Then these pages will be culled off the normal LRU lists during vmscan. Add new wrapper function to clear the mapping's unevictable state when/if shared memory segment is munlocked. Add 'scan_mapping_unevictable_page()' to mm/vmscan.c to scan all pages in the shmem segment's mapping [struct address_space] for evictability now that they're no longer locked. If so, move them to the appropriate zone lru list. Changes depend on [CONFIG_]UNEVICTABLE_LRU. [kosaki.motohiro@jp.fujitsu.com: revert shm change] Signed-off-by: Lee Schermerhorn Signed-off-by: Rik van Riel Signed-off-by: Kosaki Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/vmscan.c | 89 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 89 insertions(+) (limited to 'mm/vmscan.c') diff --git a/mm/vmscan.c b/mm/vmscan.c index 9babfbc1ddc..dfb342e0db9 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -2346,4 +2346,93 @@ int page_evictable(struct page *page, struct vm_area_struct *vma) return 1; } + +/** + * check_move_unevictable_page - check page for evictability and move to appropriate zone lru list + * @page: page to check evictability and move to appropriate lru list + * @zone: zone page is in + * + * Checks a page for evictability and moves the page to the appropriate + * zone lru list. + * + * Restrictions: zone->lru_lock must be held, page must be on LRU and must + * have PageUnevictable set. + */ +static void check_move_unevictable_page(struct page *page, struct zone *zone) +{ + VM_BUG_ON(PageActive(page)); + +retry: + ClearPageUnevictable(page); + if (page_evictable(page, NULL)) { + enum lru_list l = LRU_INACTIVE_ANON + page_is_file_cache(page); + __dec_zone_state(zone, NR_UNEVICTABLE); + list_move(&page->lru, &zone->lru[l].list); + __inc_zone_state(zone, NR_INACTIVE_ANON + l); + __count_vm_event(UNEVICTABLE_PGRESCUED); + } else { + /* + * rotate unevictable list + */ + SetPageUnevictable(page); + list_move(&page->lru, &zone->lru[LRU_UNEVICTABLE].list); + if (page_evictable(page, NULL)) + goto retry; + } +} + +/** + * scan_mapping_unevictable_pages - scan an address space for evictable pages + * @mapping: struct address_space to scan for evictable pages + * + * Scan all pages in mapping. Check unevictable pages for + * evictability and move them to the appropriate zone lru list. + */ +void scan_mapping_unevictable_pages(struct address_space *mapping) +{ + pgoff_t next = 0; + pgoff_t end = (i_size_read(mapping->host) + PAGE_CACHE_SIZE - 1) >> + PAGE_CACHE_SHIFT; + struct zone *zone; + struct pagevec pvec; + + if (mapping->nrpages == 0) + return; + + pagevec_init(&pvec, 0); + while (next < end && + pagevec_lookup(&pvec, mapping, next, PAGEVEC_SIZE)) { + int i; + int pg_scanned = 0; + + zone = NULL; + + for (i = 0; i < pagevec_count(&pvec); i++) { + struct page *page = pvec.pages[i]; + pgoff_t page_index = page->index; + struct zone *pagezone = page_zone(page); + + pg_scanned++; + if (page_index > next) + next = page_index; + next++; + + if (pagezone != zone) { + if (zone) + spin_unlock_irq(&zone->lru_lock); + zone = pagezone; + spin_lock_irq(&zone->lru_lock); + } + + if (PageLRU(page) && PageUnevictable(page)) + check_move_unevictable_page(page, zone); + } + if (zone) + spin_unlock_irq(&zone->lru_lock); + pagevec_release(&pvec); + + count_vm_events(UNEVICTABLE_PGSCANNED, pg_scanned); + } + +} #endif -- cgit v1.2.3 From b291f000393f5a0b679012b39d79fbc85c018233 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Sat, 18 Oct 2008 20:26:44 -0700 Subject: mlock: mlocked pages are unevictable Make sure that mlocked pages also live on the unevictable LRU, so kswapd will not scan them over and over again. This is achieved through various strategies: 1) add yet another page flag--PG_mlocked--to indicate that the page is locked for efficient testing in vmscan and, optionally, fault path. This allows early culling of unevictable pages, preventing them from getting to page_referenced()/try_to_unmap(). Also allows separate accounting of mlock'd pages, as Nick's original patch did. Note: Nick's original mlock patch used a PG_mlocked flag. I had removed this in favor of the PG_unevictable flag + an mlock_count [new page struct member]. I restored the PG_mlocked flag to eliminate the new count field. 2) add the mlock/unevictable infrastructure to mm/mlock.c, with internal APIs in mm/internal.h. This is a rework of Nick's original patch to these files, taking into account that mlocked pages are now kept on unevictable LRU list. 3) update vmscan.c:page_evictable() to check PageMlocked() and, if vma passed in, the vm_flags. Note that the vma will only be passed in for new pages in the fault path; and then only if the "cull unevictable pages in fault path" patch is included. 4) add try_to_unlock() to rmap.c to walk a page's rmap and ClearPageMlocked() if no other vmas have it mlocked. Reuses as much of try_to_unmap() as possible. This effectively replaces the use of one of the lru list links as an mlock count. If this mechanism let's pages in mlocked vmas leak through w/o PG_mlocked set [I don't know that it does], we should catch them later in try_to_unmap(). One hopes this will be rare, as it will be relatively expensive. Original mm/internal.h, mm/rmap.c and mm/mlock.c changes: Signed-off-by: Nick Piggin splitlru: introduce __get_user_pages(): New munlock processing need to GUP_FLAGS_IGNORE_VMA_PERMISSIONS. because current get_user_pages() can't grab PROT_NONE pages theresore it cause PROT_NONE pages can't munlock. [akpm@linux-foundation.org: fix this for pagemap-pass-mm-into-pagewalkers.patch] [akpm@linux-foundation.org: untangle patch interdependencies] [akpm@linux-foundation.org: fix things after out-of-order merging] [hugh@veritas.com: fix page-flags mess] [lee.schermerhorn@hp.com: fix munlock page table walk - now requires 'mm'] [kosaki.motohiro@jp.fujitsu.com: build fix] [kosaki.motohiro@jp.fujitsu.com: fix truncate race and sevaral comments] [kosaki.motohiro@jp.fujitsu.com: splitlru: introduce __get_user_pages()] Signed-off-by: KOSAKI Motohiro Signed-off-by: Rik van Riel Signed-off-by: Lee Schermerhorn Cc: Nick Piggin Cc: Dave Hansen Cc: Matt Mackall Signed-off-by: Hugh Dickins Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/vmscan.c | 36 ++++++++++++++++++++++++++---------- 1 file changed, 26 insertions(+), 10 deletions(-) (limited to 'mm/vmscan.c') diff --git a/mm/vmscan.c b/mm/vmscan.c index dfb342e0db9..e5aaaad159e 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -582,11 +582,8 @@ static unsigned long shrink_page_list(struct list_head *page_list, sc->nr_scanned++; - if (unlikely(!page_evictable(page, NULL))) { - unlock_page(page); - putback_lru_page(page); - continue; - } + if (unlikely(!page_evictable(page, NULL))) + goto cull_mlocked; if (!sc->may_swap && page_mapped(page)) goto keep_locked; @@ -624,9 +621,19 @@ static unsigned long shrink_page_list(struct list_head *page_list, * Anonymous process memory has backing store? * Try to allocate it some swap space here. */ - if (PageAnon(page) && !PageSwapCache(page)) + if (PageAnon(page) && !PageSwapCache(page)) { + switch (try_to_munlock(page)) { + case SWAP_FAIL: /* shouldn't happen */ + case SWAP_AGAIN: + goto keep_locked; + case SWAP_MLOCK: + goto cull_mlocked; + case SWAP_SUCCESS: + ; /* fall thru'; add to swap cache */ + } if (!add_to_swap(page, GFP_ATOMIC)) goto activate_locked; + } #endif /* CONFIG_SWAP */ mapping = page_mapping(page); @@ -641,6 +648,8 @@ static unsigned long shrink_page_list(struct list_head *page_list, goto activate_locked; case SWAP_AGAIN: goto keep_locked; + case SWAP_MLOCK: + goto cull_mlocked; case SWAP_SUCCESS: ; /* try to free the page below */ } @@ -731,6 +740,11 @@ free_it: } continue; +cull_mlocked: + unlock_page(page); + putback_lru_page(page); + continue; + activate_locked: /* Not a candidate for swapping, so reclaim swap space. */ if (PageSwapCache(page) && vm_swap_full()) @@ -742,7 +756,7 @@ keep_locked: unlock_page(page); keep: list_add(&page->lru, &ret_pages); - VM_BUG_ON(PageLRU(page)); + VM_BUG_ON(PageLRU(page) || PageUnevictable(page)); } list_splice(&ret_pages, page_list); if (pagevec_count(&freed_pvec)) @@ -2329,12 +2343,13 @@ int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order) * @vma: the VMA in which the page is or will be mapped, may be NULL * * Test whether page is evictable--i.e., should be placed on active/inactive - * lists vs unevictable list. + * lists vs unevictable list. The vma argument is !NULL when called from the + * fault path to determine how to instantate a new page. * * Reasons page might not be evictable: * (1) page's mapping marked unevictable + * (2) page is part of an mlocked VMA * - * TODO - later patches */ int page_evictable(struct page *page, struct vm_area_struct *vma) { @@ -2342,7 +2357,8 @@ int page_evictable(struct page *page, struct vm_area_struct *vma) if (mapping_unevictable(page_mapping(page))) return 0; - /* TODO: test page [!]evictable conditions */ + if (PageMlocked(page) || (vma && is_mlocked_vma(vma, page))) + return 0; return 1; } -- cgit v1.2.3 From af936a1606246a10c145feac3770f6287f483f02 Mon Sep 17 00:00:00 2001 From: Lee Schermerhorn Date: Sat, 18 Oct 2008 20:26:53 -0700 Subject: vmscan: unevictable LRU scan sysctl This patch adds a function to scan individual or all zones' unevictable lists and move any pages that have become evictable onto the respective zone's inactive list, where shrink_inactive_list() will deal with them. Adds sysctl to scan all nodes, and per node attributes to individual nodes' zones. Kosaki: If evictable page found in unevictable lru when write /proc/sys/vm/scan_unevictable_pages, print filename and file offset of these pages. [akpm@linux-foundation.org: fix one CONFIG_MMU=n build error] [kosaki.motohiro@jp.fujitsu.com: adapt vmscan-unevictable-lru-scan-sysctl.patch to new sysfs API] Signed-off-by: Lee Schermerhorn Signed-off-by: Rik van Riel Signed-off-by: KOSAKI Motohiro Signed-off-by: KOSAKI Motohiro Signed-off-by: Hugh Dickins Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/vmscan.c | 166 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 166 insertions(+) (limited to 'mm/vmscan.c') diff --git a/mm/vmscan.c b/mm/vmscan.c index e5aaaad159e..ca64e3e0c51 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -39,6 +39,7 @@ #include #include #include +#include #include #include @@ -2363,6 +2364,39 @@ int page_evictable(struct page *page, struct vm_area_struct *vma) return 1; } +static void show_page_path(struct page *page) +{ + char buf[256]; + if (page_is_file_cache(page)) { + struct address_space *mapping = page->mapping; + struct dentry *dentry; + pgoff_t pgoff = page->index << (PAGE_CACHE_SHIFT - PAGE_SHIFT); + + spin_lock(&mapping->i_mmap_lock); + dentry = d_find_alias(mapping->host); + printk(KERN_INFO "rescued: %s %lu\n", + dentry_path(dentry, buf, 256), pgoff); + spin_unlock(&mapping->i_mmap_lock); + } else { +#if defined(CONFIG_MM_OWNER) && defined(CONFIG_MMU) + struct anon_vma *anon_vma; + struct vm_area_struct *vma; + + anon_vma = page_lock_anon_vma(page); + if (!anon_vma) + return; + + list_for_each_entry(vma, &anon_vma->head, anon_vma_node) { + printk(KERN_INFO "rescued: anon %s\n", + vma->vm_mm->owner->comm); + break; + } + page_unlock_anon_vma(anon_vma); +#endif + } +} + + /** * check_move_unevictable_page - check page for evictability and move to appropriate zone lru list * @page: page to check evictability and move to appropriate lru list @@ -2382,6 +2416,9 @@ retry: ClearPageUnevictable(page); if (page_evictable(page, NULL)) { enum lru_list l = LRU_INACTIVE_ANON + page_is_file_cache(page); + + show_page_path(page); + __dec_zone_state(zone, NR_UNEVICTABLE); list_move(&page->lru, &zone->lru[l].list); __inc_zone_state(zone, NR_INACTIVE_ANON + l); @@ -2451,4 +2488,133 @@ void scan_mapping_unevictable_pages(struct address_space *mapping) } } + +/** + * scan_zone_unevictable_pages - check unevictable list for evictable pages + * @zone - zone of which to scan the unevictable list + * + * Scan @zone's unevictable LRU lists to check for pages that have become + * evictable. Move those that have to @zone's inactive list where they + * become candidates for reclaim, unless shrink_inactive_zone() decides + * to reactivate them. Pages that are still unevictable are rotated + * back onto @zone's unevictable list. + */ +#define SCAN_UNEVICTABLE_BATCH_SIZE 16UL /* arbitrary lock hold batch size */ +void scan_zone_unevictable_pages(struct zone *zone) +{ + struct list_head *l_unevictable = &zone->lru[LRU_UNEVICTABLE].list; + unsigned long scan; + unsigned long nr_to_scan = zone_page_state(zone, NR_UNEVICTABLE); + + while (nr_to_scan > 0) { + unsigned long batch_size = min(nr_to_scan, + SCAN_UNEVICTABLE_BATCH_SIZE); + + spin_lock_irq(&zone->lru_lock); + for (scan = 0; scan < batch_size; scan++) { + struct page *page = lru_to_page(l_unevictable); + + if (!trylock_page(page)) + continue; + + prefetchw_prev_lru_page(page, l_unevictable, flags); + + if (likely(PageLRU(page) && PageUnevictable(page))) + check_move_unevictable_page(page, zone); + + unlock_page(page); + } + spin_unlock_irq(&zone->lru_lock); + + nr_to_scan -= batch_size; + } +} + + +/** + * scan_all_zones_unevictable_pages - scan all unevictable lists for evictable pages + * + * A really big hammer: scan all zones' unevictable LRU lists to check for + * pages that have become evictable. Move those back to the zones' + * inactive list where they become candidates for reclaim. + * This occurs when, e.g., we have unswappable pages on the unevictable lists, + * and we add swap to the system. As such, it runs in the context of a task + * that has possibly/probably made some previously unevictable pages + * evictable. + */ +void scan_all_zones_unevictable_pages(void) +{ + struct zone *zone; + + for_each_zone(zone) { + scan_zone_unevictable_pages(zone); + } +} + +/* + * scan_unevictable_pages [vm] sysctl handler. On demand re-scan of + * all nodes' unevictable lists for evictable pages + */ +unsigned long scan_unevictable_pages; + +int scan_unevictable_handler(struct ctl_table *table, int write, + struct file *file, void __user *buffer, + size_t *length, loff_t *ppos) +{ + proc_doulongvec_minmax(table, write, file, buffer, length, ppos); + + if (write && *(unsigned long *)table->data) + scan_all_zones_unevictable_pages(); + + scan_unevictable_pages = 0; + return 0; +} + +/* + * per node 'scan_unevictable_pages' attribute. On demand re-scan of + * a specified node's per zone unevictable lists for evictable pages. + */ + +static ssize_t read_scan_unevictable_node(struct sys_device *dev, + struct sysdev_attribute *attr, + char *buf) +{ + return sprintf(buf, "0\n"); /* always zero; should fit... */ +} + +static ssize_t write_scan_unevictable_node(struct sys_device *dev, + struct sysdev_attribute *attr, + const char *buf, size_t count) +{ + struct zone *node_zones = NODE_DATA(dev->id)->node_zones; + struct zone *zone; + unsigned long res; + unsigned long req = strict_strtoul(buf, 10, &res); + + if (!req) + return 1; /* zero is no-op */ + + for (zone = node_zones; zone - node_zones < MAX_NR_ZONES; ++zone) { + if (!populated_zone(zone)) + continue; + scan_zone_unevictable_pages(zone); + } + return 1; +} + + +static SYSDEV_ATTR(scan_unevictable_pages, S_IRUGO | S_IWUSR, + read_scan_unevictable_node, + write_scan_unevictable_node); + +int scan_unevictable_register_node(struct node *node) +{ + return sysdev_create_file(&node->sysdev, &attr_scan_unevictable_pages); +} + +void scan_unevictable_unregister_node(struct node *node) +{ + sysdev_remove_file(&node->sysdev, &attr_scan_unevictable_pages); +} + #endif -- cgit v1.2.3 From e0f79b8f1f3394bb344b7b83d6f121ac2af327de Mon Sep 17 00:00:00 2001 From: Johannes Weiner Date: Sat, 18 Oct 2008 20:26:55 -0700 Subject: vmscan: don't accumulate scan pressure on unrelated lists During each reclaim scan we accumulate scan pressure on unrelated lists which will result in bogus scans and unwanted reclaims eventually. Scanning lists with few reclaim candidates results in a lot of rotation and therefor also disturbs the list balancing, putting even more pressure on the wrong lists. In a test-case with much streaming IO, and therefor a crowded inactive file page list, swapping started because a) anon pages were reclaimed after swap_cluster_max reclaim invocations -- nr_scan of this list has just accumulated b) active file pages were scanned because *their* nr_scan has also accumulated through the same logic. And this in return created a lot of rotation for file pages and resulted in a decrease of file list priority, again increasing the pressure on anon pages. The result was an evicted working set of anon pages while there were tons of inactive file pages that should have been taken instead. Signed-off-by: Johannes Weiner Reviewed-by: Rik van Riel Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/vmscan.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) (limited to 'mm/vmscan.c') diff --git a/mm/vmscan.c b/mm/vmscan.c index ca64e3e0c51..412d7872fc7 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -1413,16 +1413,13 @@ static unsigned long shrink_zone(int priority, struct zone *zone, if (scan_global_lru(sc)) { int file = is_file_lru(l); int scan; - /* - * Add one to nr_to_scan just to make sure that the - * kernel will slowly sift through each list. - */ + scan = zone_page_state(zone, NR_LRU_BASE + l); if (priority) { scan >>= priority; scan = (scan * percent[file]) / 100; } - zone->lru[l].nr_scan += scan + 1; + zone->lru[l].nr_scan += scan; nr[l] = zone->lru[l].nr_scan; if (nr[l] >= sc->swap_cluster_max) zone->lru[l].nr_scan = 0; -- cgit v1.2.3 From a978d6f521063514812a7094dbe5036e056e4de3 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Sat, 18 Oct 2008 20:26:58 -0700 Subject: mm: unlockless reclaim unlock_page is fairly expensive. It can be avoided in page reclaim success path. By definition if we have any other references to the page it would be a bug anyway. Signed-off-by: Nick Piggin Reviewed-by: KOSAKI Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/vmscan.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'mm/vmscan.c') diff --git a/mm/vmscan.c b/mm/vmscan.c index 412d7872fc7..3b5860294bb 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -732,7 +732,14 @@ static unsigned long shrink_page_list(struct list_head *page_list, if (!mapping || !__remove_mapping(mapping, page)) goto keep_locked; - unlock_page(page); + /* + * At this point, we have no other references and there is + * no way to pick any more up (removed from LRU, removed + * from pagecache). Can use non-atomic bitops now (and + * we obviously don't have to worry about waking up a process + * waiting on the page lock, because there are no references. + */ + __clear_page_locked(page); free_it: nr_reclaimed++; if (!pagevec_add(&freed_pvec, page)) { -- cgit v1.2.3