From f5d6d2da0d9098a4aa0ebcc187aa0fc167045d6b Mon Sep 17 00:00:00 2001 From: Tobias Klauser Date: Wed, 26 Oct 2016 13:37:04 +0200 Subject: sched/fair: Remove unused but set variable 'rq' Since commit: 8663e24d56dc ("sched/fair: Reorder cgroup creation code") ... the variable 'rq' in alloc_fair_sched_group() is set but no longer used. Remove it to fix the following GCC warning when building with 'W=1': kernel/sched/fair.c:8842:13: warning: variable ‘rq’ set but not used [-Wunused-but-set-variable] Signed-off-by: Tobias Klauser Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/20161026113704.8981-1-tklauser@distanz.ch Signed-off-by: Ingo Molnar --- kernel/sched/fair.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'kernel/sched') diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index d941c97dfbc3..c242944f5cbd 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -8839,7 +8839,6 @@ int alloc_fair_sched_group(struct task_group *tg, struct task_group *parent) { struct sched_entity *se; struct cfs_rq *cfs_rq; - struct rq *rq; int i; tg->cfs_rq = kzalloc(sizeof(cfs_rq) * nr_cpu_ids, GFP_KERNEL); @@ -8854,8 +8853,6 @@ int alloc_fair_sched_group(struct task_group *tg, struct task_group *parent) init_cfs_bandwidth(tg_cfs_bandwidth(tg)); for_each_possible_cpu(i) { - rq = cpu_rq(i); - cfs_rq = kzalloc_node(sizeof(struct cfs_rq), GFP_KERNEL, cpu_to_node(i)); if (!cfs_rq) -- cgit v1.2.3 From 9dcb8b685fc30813b35ab4b4bf39244430753190 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Wed, 26 Oct 2016 10:15:30 -0700 Subject: mm: remove per-zone hashtable of bitlock waitqueues The per-zone waitqueues exist because of a scalability issue with the page waitqueues on some NUMA machines, but it turns out that they hurt normal loads, and now with the vmalloced stacks they also end up breaking gfs2 that uses a bit_wait on a stack object: wait_on_bit(&gh->gh_iflags, HIF_WAIT, TASK_UNINTERRUPTIBLE) where 'gh' can be a reference to the local variable 'mount_gh' on the stack of fill_super(). The reason the per-zone hash table breaks for this case is that there is no "zone" for virtual allocations, and trying to look up the physical page to get at it will fail (with a BUG_ON()). It turns out that I actually complained to the mm people about the per-zone hash table for another reason just a month ago: the zone lookup also hurts the regular use of "unlock_page()" a lot, because the zone lookup ends up forcing several unnecessary cache misses and generates horrible code. As part of that earlier discussion, we had a much better solution for the NUMA scalability issue - by just making the page lock have a separate contention bit, the waitqueue doesn't even have to be looked at for the normal case. Peter Zijlstra already has a patch for that, but let's see if anybody even notices. In the meantime, let's fix the actual gfs2 breakage by simplifying the bitlock waitqueues and removing the per-zone issue. Reported-by: Andreas Gruenbacher Tested-by: Bob Peterson Acked-by: Mel Gorman Cc: Peter Zijlstra Cc: Andy Lutomirski Cc: Steven Whitehouse Signed-off-by: Linus Torvalds --- include/linux/mmzone.h | 30 +------------ kernel/sched/core.c | 16 +++++++ kernel/sched/wait.c | 10 ----- mm/filemap.c | 4 +- mm/memory_hotplug.c | 28 ------------ mm/page_alloc.c | 115 +------------------------------------------------ 6 files changed, 21 insertions(+), 182 deletions(-) (limited to 'kernel/sched') diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 7f2ae99e5daf..0f088f3a2fed 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -440,33 +440,7 @@ struct zone { seqlock_t span_seqlock; #endif - /* - * wait_table -- the array holding the hash table - * wait_table_hash_nr_entries -- the size of the hash table array - * wait_table_bits -- wait_table_size == (1 << wait_table_bits) - * - * The purpose of all these is to keep track of the people - * waiting for a page to become available and make them - * runnable again when possible. The trouble is that this - * consumes a lot of space, especially when so few things - * wait on pages at a given time. So instead of using - * per-page waitqueues, we use a waitqueue hash table. - * - * The bucket discipline is to sleep on the same queue when - * colliding and wake all in that wait queue when removing. - * When something wakes, it must check to be sure its page is - * truly available, a la thundering herd. The cost of a - * collision is great, but given the expected load of the - * table, they should be so rare as to be outweighed by the - * benefits from the saved space. - * - * __wait_on_page_locked() and unlock_page() in mm/filemap.c, are the - * primary users of these fields, and in mm/page_alloc.c - * free_area_init_core() performs the initialization of them. - */ - wait_queue_head_t *wait_table; - unsigned long wait_table_hash_nr_entries; - unsigned long wait_table_bits; + int initialized; /* Write-intensive fields used from the page allocator */ ZONE_PADDING(_pad1_) @@ -546,7 +520,7 @@ static inline bool zone_spans_pfn(const struct zone *zone, unsigned long pfn) static inline bool zone_is_initialized(struct zone *zone) { - return !!zone->wait_table; + return zone->initialized; } static inline bool zone_is_empty(struct zone *zone) diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 94732d1ab00a..42d4027f9e26 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -7515,11 +7515,27 @@ static struct kmem_cache *task_group_cache __read_mostly; DECLARE_PER_CPU(cpumask_var_t, load_balance_mask); DECLARE_PER_CPU(cpumask_var_t, select_idle_mask); +#define WAIT_TABLE_BITS 8 +#define WAIT_TABLE_SIZE (1 << WAIT_TABLE_BITS) +static wait_queue_head_t bit_wait_table[WAIT_TABLE_SIZE] __cacheline_aligned; + +wait_queue_head_t *bit_waitqueue(void *word, int bit) +{ + const int shift = BITS_PER_LONG == 32 ? 5 : 6; + unsigned long val = (unsigned long)word << shift | bit; + + return bit_wait_table + hash_long(val, WAIT_TABLE_BITS); +} +EXPORT_SYMBOL(bit_waitqueue); + void __init sched_init(void) { int i, j; unsigned long alloc_size = 0, ptr; + for (i = 0; i < WAIT_TABLE_SIZE; i++) + init_waitqueue_head(bit_wait_table + i); + #ifdef CONFIG_FAIR_GROUP_SCHED alloc_size += 2 * nr_cpu_ids * sizeof(void **); #endif diff --git a/kernel/sched/wait.c b/kernel/sched/wait.c index 4f7053579fe3..9453efe9b25a 100644 --- a/kernel/sched/wait.c +++ b/kernel/sched/wait.c @@ -480,16 +480,6 @@ void wake_up_bit(void *word, int bit) } EXPORT_SYMBOL(wake_up_bit); -wait_queue_head_t *bit_waitqueue(void *word, int bit) -{ - const int shift = BITS_PER_LONG == 32 ? 5 : 6; - const struct zone *zone = page_zone(virt_to_page(word)); - unsigned long val = (unsigned long)word << shift | bit; - - return &zone->wait_table[hash_long(val, zone->wait_table_bits)]; -} -EXPORT_SYMBOL(bit_waitqueue); - /* * Manipulate the atomic_t address to produce a better bit waitqueue table hash * index (we're keying off bit -1, but that would produce a horrible hash diff --git a/mm/filemap.c b/mm/filemap.c index 849f459ad078..c7fe2f16503f 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -790,9 +790,7 @@ EXPORT_SYMBOL(__page_cache_alloc); */ wait_queue_head_t *page_waitqueue(struct page *page) { - const struct zone *zone = page_zone(page); - - return &zone->wait_table[hash_ptr(page, zone->wait_table_bits)]; + return bit_waitqueue(page, 0); } EXPORT_SYMBOL(page_waitqueue); diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c index 962927309b6e..b18dab401be6 100644 --- a/mm/memory_hotplug.c +++ b/mm/memory_hotplug.c @@ -268,7 +268,6 @@ void __init register_page_bootmem_info_node(struct pglist_data *pgdat) unsigned long i, pfn, end_pfn, nr_pages; int node = pgdat->node_id; struct page *page; - struct zone *zone; nr_pages = PAGE_ALIGN(sizeof(struct pglist_data)) >> PAGE_SHIFT; page = virt_to_page(pgdat); @@ -276,19 +275,6 @@ void __init register_page_bootmem_info_node(struct pglist_data *pgdat) for (i = 0; i < nr_pages; i++, page++) get_page_bootmem(node, page, NODE_INFO); - zone = &pgdat->node_zones[0]; - for (; zone < pgdat->node_zones + MAX_NR_ZONES - 1; zone++) { - if (zone_is_initialized(zone)) { - nr_pages = zone->wait_table_hash_nr_entries - * sizeof(wait_queue_head_t); - nr_pages = PAGE_ALIGN(nr_pages) >> PAGE_SHIFT; - page = virt_to_page(zone->wait_table); - - for (i = 0; i < nr_pages; i++, page++) - get_page_bootmem(node, page, NODE_INFO); - } - } - pfn = pgdat->node_start_pfn; end_pfn = pgdat_end_pfn(pgdat); @@ -2158,20 +2144,6 @@ void try_offline_node(int nid) */ node_set_offline(nid); unregister_one_node(nid); - - /* free waittable in each zone */ - for (i = 0; i < MAX_NR_ZONES; i++) { - struct zone *zone = pgdat->node_zones + i; - - /* - * wait_table may be allocated from boot memory, - * here only free if it's allocated by vmalloc. - */ - if (is_vmalloc_addr(zone->wait_table)) { - vfree(zone->wait_table); - zone->wait_table = NULL; - } - } } EXPORT_SYMBOL(try_offline_node); diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 2b3bf6767d54..de7c6e43b1c9 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -4976,72 +4976,6 @@ void __ref build_all_zonelists(pg_data_t *pgdat, struct zone *zone) #endif } -/* - * Helper functions to size the waitqueue hash table. - * Essentially these want to choose hash table sizes sufficiently - * large so that collisions trying to wait on pages are rare. - * But in fact, the number of active page waitqueues on typical - * systems is ridiculously low, less than 200. So this is even - * conservative, even though it seems large. - * - * The constant PAGES_PER_WAITQUEUE specifies the ratio of pages to - * waitqueues, i.e. the size of the waitq table given the number of pages. - */ -#define PAGES_PER_WAITQUEUE 256 - -#ifndef CONFIG_MEMORY_HOTPLUG -static inline unsigned long wait_table_hash_nr_entries(unsigned long pages) -{ - unsigned long size = 1; - - pages /= PAGES_PER_WAITQUEUE; - - while (size < pages) - size <<= 1; - - /* - * Once we have dozens or even hundreds of threads sleeping - * on IO we've got bigger problems than wait queue collision. - * Limit the size of the wait table to a reasonable size. - */ - size = min(size, 4096UL); - - return max(size, 4UL); -} -#else -/* - * A zone's size might be changed by hot-add, so it is not possible to determine - * a suitable size for its wait_table. So we use the maximum size now. - * - * The max wait table size = 4096 x sizeof(wait_queue_head_t). ie: - * - * i386 (preemption config) : 4096 x 16 = 64Kbyte. - * ia64, x86-64 (no preemption): 4096 x 20 = 80Kbyte. - * ia64, x86-64 (preemption) : 4096 x 24 = 96Kbyte. - * - * The maximum entries are prepared when a zone's memory is (512K + 256) pages - * or more by the traditional way. (See above). It equals: - * - * i386, x86-64, powerpc(4K page size) : = ( 2G + 1M)byte. - * ia64(16K page size) : = ( 8G + 4M)byte. - * powerpc (64K page size) : = (32G +16M)byte. - */ -static inline unsigned long wait_table_hash_nr_entries(unsigned long pages) -{ - return 4096UL; -} -#endif - -/* - * This is an integer logarithm so that shifts can be used later - * to extract the more random high bits from the multiplicative - * hash function before the remainder is taken. - */ -static inline unsigned long wait_table_bits(unsigned long size) -{ - return ffz(~size); -} - /* * Initially all pages are reserved - free ones are freed * up by free_all_bootmem() once the early boot process is @@ -5304,49 +5238,6 @@ void __init setup_per_cpu_pageset(void) alloc_percpu(struct per_cpu_nodestat); } -static noinline __ref -int zone_wait_table_init(struct zone *zone, unsigned long zone_size_pages) -{ - int i; - size_t alloc_size; - - /* - * The per-page waitqueue mechanism uses hashed waitqueues - * per zone. - */ - zone->wait_table_hash_nr_entries = - wait_table_hash_nr_entries(zone_size_pages); - zone->wait_table_bits = - wait_table_bits(zone->wait_table_hash_nr_entries); - alloc_size = zone->wait_table_hash_nr_entries - * sizeof(wait_queue_head_t); - - if (!slab_is_available()) { - zone->wait_table = (wait_queue_head_t *) - memblock_virt_alloc_node_nopanic( - alloc_size, zone->zone_pgdat->node_id); - } else { - /* - * This case means that a zone whose size was 0 gets new memory - * via memory hot-add. - * But it may be the case that a new node was hot-added. In - * this case vmalloc() will not be able to use this new node's - * memory - this wait_table must be initialized to use this new - * node itself as well. - * To use this new node's memory, further consideration will be - * necessary. - */ - zone->wait_table = vmalloc(alloc_size); - } - if (!zone->wait_table) - return -ENOMEM; - - for (i = 0; i < zone->wait_table_hash_nr_entries; ++i) - init_waitqueue_head(zone->wait_table + i); - - return 0; -} - static __meminit void zone_pcp_init(struct zone *zone) { /* @@ -5367,10 +5258,7 @@ int __meminit init_currently_empty_zone(struct zone *zone, unsigned long size) { struct pglist_data *pgdat = zone->zone_pgdat; - int ret; - ret = zone_wait_table_init(zone, size); - if (ret) - return ret; + pgdat->nr_zones = zone_idx(zone) + 1; zone->zone_start_pfn = zone_start_pfn; @@ -5382,6 +5270,7 @@ int __meminit init_currently_empty_zone(struct zone *zone, zone_start_pfn, (zone_start_pfn + size)); zone_init_free_lists(zone); + zone->initialized = 1; return 0; } -- cgit v1.2.3 From 382005027fedc50b28d40ae64ef1461cca38953e Mon Sep 17 00:00:00 2001 From: Tetsuo Handa Date: Wed, 2 Nov 2016 19:50:29 +0900 Subject: sched/core: Fix oops in sched_show_task() When CONFIG_THREAD_INFO_IN_TASK=y, it is possible that an exited thread remains in the task list after its stack pointer was already set to NULL. Therefore, thread_saved_pc() and stack_not_used() in sched_show_task() will trigger NULL pointer dereference if an attempt to dump such thread's traces (e.g. SysRq-t, khungtaskd) is made. Since show_stack() in sched_show_task() calls try_get_task_stack() and sched_show_task() is called from interrupt context, calling try_get_task_stack() from sched_show_task() will be safe as well. Signed-off-by: Tetsuo Handa Acked-by: Andy Lutomirski Acked-by: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: bp@alien8.de Cc: brgerst@gmail.com Cc: jann@thejh.net Cc: keescook@chromium.org Cc: linux-api@vger.kernel.org Cc: tycho.andersen@canonical.com Link: http://lkml.kernel.org/r/201611021950.FEJ34368.HFFJOOMLtQOVSF@I-love.SAKURA.ne.jp Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel/sched') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 42d4027f9e26..9abf66b6c271 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5192,6 +5192,8 @@ void sched_show_task(struct task_struct *p) int ppid; unsigned long state = p->state; + if (!try_get_task_stack(p)) + return; if (state) state = __ffs(state) + 1; printk(KERN_INFO "%-15.15s %c", p->comm, @@ -5221,6 +5223,7 @@ void sched_show_task(struct task_struct *p) print_worker_info(KERN_INFO, p); show_stack(p, NULL); + put_task_stack(p); } void show_state_filter(unsigned long state_filter) -- cgit v1.2.3 From 8243d5597793b5e85143c9a935e1b971c59740a9 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Tue, 1 Nov 2016 17:47:18 -0600 Subject: sched/core: Remove pointless printout in sched_show_task() In sched_show_task() we print out a useless hex number, not even a symbol, and there's a big question mark whether this even makes sense anyway, I suspect we should just remove it all. Signed-off-by: Linus Torvalds Acked-by: Andy Lutomirski Cc: Peter Zijlstra Cc: Tetsuo Handa Cc: Thomas Gleixner Cc: bp@alien8.de Cc: brgerst@gmail.com Cc: jann@thejh.net Cc: keescook@chromium.org Cc: linux-api@vger.kernel.org Cc: tycho.andersen@canonical.com Link: http://lkml.kernel.org/r/CA+55aFzphURPFzAvU4z6Moy7ZmimcwPuUdYU8bj9z0J+S8X1rw@mail.gmail.com Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 9 --------- 1 file changed, 9 deletions(-) (limited to 'kernel/sched') diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 9abf66b6c271..154fd689fe02 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5198,17 +5198,8 @@ void sched_show_task(struct task_struct *p) state = __ffs(state) + 1; printk(KERN_INFO "%-15.15s %c", p->comm, state < sizeof(stat_nam) - 1 ? stat_nam[state] : '?'); -#if BITS_PER_LONG == 32 - if (state == TASK_RUNNING) - printk(KERN_CONT " running "); - else - printk(KERN_CONT " %08lx ", thread_saved_pc(p)); -#else if (state == TASK_RUNNING) printk(KERN_CONT " running task "); - else - printk(KERN_CONT " %016lx ", thread_saved_pc(p)); -#endif #ifdef CONFIG_DEBUG_STACK_USAGE free = stack_not_used(p); #endif -- cgit v1.2.3