From 19965460e31c73a934d2c19c152f876a75bdff3e Mon Sep 17 00:00:00 2001 From: David Rientjes Date: Tue, 11 Dec 2012 16:00:26 -0800 Subject: mm, memcg: make mem_cgroup_out_of_memory() static mem_cgroup_out_of_memory() is only referenced from within file scope, so it can be marked static. Signed-off-by: David Rientjes Acked-by: KAMEZAWA Hiroyuki Acked-by: Michal Hocko Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/oom.h | 2 -- 1 file changed, 2 deletions(-) (limited to 'include/linux/oom.h') diff --git a/include/linux/oom.h b/include/linux/oom.h index fb9826847b89..4a4188da4570 100644 --- a/include/linux/oom.h +++ b/include/linux/oom.h @@ -49,8 +49,6 @@ extern void check_panic_on_oom(enum oom_constraint constraint, gfp_t gfp_mask, extern enum oom_scan_t oom_scan_process_thread(struct task_struct *task, unsigned long totalpages, const nodemask_t *nodemask, bool force_kill); -extern void mem_cgroup_out_of_memory(struct mem_cgroup *memcg, gfp_t gfp_mask, - int order); extern void out_of_memory(struct zonelist *zonelist, gfp_t gfp_mask, int order, nodemask_t *mask, bool force_kill); -- cgit v1.3-8-gc7d7 From a9c58b907dbc6821533dfc295b63caf111ff1f16 Mon Sep 17 00:00:00 2001 From: David Rientjes Date: Tue, 11 Dec 2012 16:02:54 -0800 Subject: mm, oom: change type of oom_score_adj to short The maximum oom_score_adj is 1000 and the minimum oom_score_adj is -1000, so this range can be represented by the signed short type with no functional change. The extra space this frees up in struct signal_struct will be used for per-thread oom kill flags in the next patch. Signed-off-by: David Rientjes Cc: KAMEZAWA Hiroyuki Cc: KOSAKI Motohiro Reviewed-by: Michal Hocko Cc: Anton Vorontsov Cc: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/staging/android/lowmemorykiller.c | 16 ++++++++-------- fs/proc/base.c | 10 +++++----- include/linux/oom.h | 4 ++-- include/linux/sched.h | 6 +++--- include/trace/events/oom.h | 4 ++-- include/trace/events/task.h | 8 ++++---- mm/ksm.c | 2 +- mm/oom_kill.c | 10 +++++----- mm/swapfile.c | 2 +- 9 files changed, 31 insertions(+), 31 deletions(-) (limited to 'include/linux/oom.h') diff --git a/drivers/staging/android/lowmemorykiller.c b/drivers/staging/android/lowmemorykiller.c index b91e4bc332a7..3b91b0fd4de3 100644 --- a/drivers/staging/android/lowmemorykiller.c +++ b/drivers/staging/android/lowmemorykiller.c @@ -40,7 +40,7 @@ #include static uint32_t lowmem_debug_level = 2; -static int lowmem_adj[6] = { +static short lowmem_adj[6] = { 0, 1, 6, @@ -70,9 +70,9 @@ static int lowmem_shrink(struct shrinker *s, struct shrink_control *sc) int rem = 0; int tasksize; int i; - int min_score_adj = OOM_SCORE_ADJ_MAX + 1; + short min_score_adj = OOM_SCORE_ADJ_MAX + 1; int selected_tasksize = 0; - int selected_oom_score_adj; + short selected_oom_score_adj; int array_size = ARRAY_SIZE(lowmem_adj); int other_free = global_page_state(NR_FREE_PAGES); int other_file = global_page_state(NR_FILE_PAGES) - @@ -90,7 +90,7 @@ static int lowmem_shrink(struct shrinker *s, struct shrink_control *sc) } } if (sc->nr_to_scan > 0) - lowmem_print(3, "lowmem_shrink %lu, %x, ofree %d %d, ma %d\n", + lowmem_print(3, "lowmem_shrink %lu, %x, ofree %d %d, ma %hd\n", sc->nr_to_scan, sc->gfp_mask, other_free, other_file, min_score_adj); rem = global_page_state(NR_ACTIVE_ANON) + @@ -107,7 +107,7 @@ static int lowmem_shrink(struct shrinker *s, struct shrink_control *sc) rcu_read_lock(); for_each_process(tsk) { struct task_struct *p; - int oom_score_adj; + short oom_score_adj; if (tsk->flags & PF_KTHREAD) continue; @@ -141,11 +141,11 @@ static int lowmem_shrink(struct shrinker *s, struct shrink_control *sc) selected = p; selected_tasksize = tasksize; selected_oom_score_adj = oom_score_adj; - lowmem_print(2, "select %d (%s), adj %d, size %d, to kill\n", + lowmem_print(2, "select %d (%s), adj %hd, size %d, to kill\n", p->pid, p->comm, oom_score_adj, tasksize); } if (selected) { - lowmem_print(1, "send sigkill to %d (%s), adj %d, size %d\n", + lowmem_print(1, "send sigkill to %d (%s), adj %hd, size %d\n", selected->pid, selected->comm, selected_oom_score_adj, selected_tasksize); lowmem_deathpending_timeout = jiffies + HZ; @@ -176,7 +176,7 @@ static void __exit lowmem_exit(void) } module_param_named(cost, lowmem_shrinker.seeks, int, S_IRUGO | S_IWUSR); -module_param_array_named(adj, lowmem_adj, int, &lowmem_adj_size, +module_param_array_named(adj, lowmem_adj, short, &lowmem_adj_size, S_IRUGO | S_IWUSR); module_param_array_named(minfree, lowmem_minfree, uint, &lowmem_minfree_size, S_IRUGO | S_IWUSR); diff --git a/fs/proc/base.c b/fs/proc/base.c index 9e28356a959a..aa63d25157b8 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -985,7 +985,7 @@ static ssize_t oom_score_adj_read(struct file *file, char __user *buf, { struct task_struct *task = get_proc_task(file->f_path.dentry->d_inode); char buffer[PROC_NUMBUF]; - int oom_score_adj = OOM_SCORE_ADJ_MIN; + short oom_score_adj = OOM_SCORE_ADJ_MIN; unsigned long flags; size_t len; @@ -996,7 +996,7 @@ static ssize_t oom_score_adj_read(struct file *file, char __user *buf, unlock_task_sighand(task, &flags); } put_task_struct(task); - len = snprintf(buffer, sizeof(buffer), "%d\n", oom_score_adj); + len = snprintf(buffer, sizeof(buffer), "%hd\n", oom_score_adj); return simple_read_from_buffer(buf, count, ppos, buffer, len); } @@ -1043,15 +1043,15 @@ static ssize_t oom_score_adj_write(struct file *file, const char __user *buf, goto err_task_lock; } - if (oom_score_adj < task->signal->oom_score_adj_min && + if ((short)oom_score_adj < task->signal->oom_score_adj_min && !capable(CAP_SYS_RESOURCE)) { err = -EACCES; goto err_sighand; } - task->signal->oom_score_adj = oom_score_adj; + task->signal->oom_score_adj = (short)oom_score_adj; if (has_capability_noaudit(current, CAP_SYS_RESOURCE)) - task->signal->oom_score_adj_min = oom_score_adj; + task->signal->oom_score_adj_min = (short)oom_score_adj; trace_oom_score_adj_update(task); err_sighand: diff --git a/include/linux/oom.h b/include/linux/oom.h index 4a4188da4570..922dab164eb0 100644 --- a/include/linux/oom.h +++ b/include/linux/oom.h @@ -29,8 +29,8 @@ enum oom_scan_t { OOM_SCAN_SELECT, /* always select this thread first */ }; -extern void compare_swap_oom_score_adj(int old_val, int new_val); -extern int test_set_oom_score_adj(int new_val); +extern void compare_swap_oom_score_adj(short old_val, short new_val); +extern short test_set_oom_score_adj(short new_val); extern unsigned long oom_badness(struct task_struct *p, struct mem_cgroup *memcg, const nodemask_t *nodemask, diff --git a/include/linux/sched.h b/include/linux/sched.h index 0dd42a02df2e..ed30456152da 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -631,9 +631,9 @@ struct signal_struct { struct rw_semaphore group_rwsem; #endif - int oom_score_adj; /* OOM kill score adjustment */ - int oom_score_adj_min; /* OOM kill score adjustment minimum value. - * Only settable by CAP_SYS_RESOURCE. */ + short oom_score_adj; /* OOM kill score adjustment */ + short oom_score_adj_min; /* OOM kill score adjustment min value. + * Only settable by CAP_SYS_RESOURCE. */ struct mutex cred_guard_mutex; /* guard against foreign influences on * credential calculations diff --git a/include/trace/events/oom.h b/include/trace/events/oom.h index dd4ba3b92002..1e974983757e 100644 --- a/include/trace/events/oom.h +++ b/include/trace/events/oom.h @@ -14,7 +14,7 @@ TRACE_EVENT(oom_score_adj_update, TP_STRUCT__entry( __field( pid_t, pid) __array( char, comm, TASK_COMM_LEN ) - __field( int, oom_score_adj) + __field( short, oom_score_adj) ), TP_fast_assign( @@ -23,7 +23,7 @@ TRACE_EVENT(oom_score_adj_update, __entry->oom_score_adj = task->signal->oom_score_adj; ), - TP_printk("pid=%d comm=%s oom_score_adj=%d", + TP_printk("pid=%d comm=%s oom_score_adj=%hd", __entry->pid, __entry->comm, __entry->oom_score_adj) ); diff --git a/include/trace/events/task.h b/include/trace/events/task.h index b53add02e929..102a646e1996 100644 --- a/include/trace/events/task.h +++ b/include/trace/events/task.h @@ -15,7 +15,7 @@ TRACE_EVENT(task_newtask, __field( pid_t, pid) __array( char, comm, TASK_COMM_LEN) __field( unsigned long, clone_flags) - __field( int, oom_score_adj) + __field( short, oom_score_adj) ), TP_fast_assign( @@ -25,7 +25,7 @@ TRACE_EVENT(task_newtask, __entry->oom_score_adj = task->signal->oom_score_adj; ), - TP_printk("pid=%d comm=%s clone_flags=%lx oom_score_adj=%d", + TP_printk("pid=%d comm=%s clone_flags=%lx oom_score_adj=%hd", __entry->pid, __entry->comm, __entry->clone_flags, __entry->oom_score_adj) ); @@ -40,7 +40,7 @@ TRACE_EVENT(task_rename, __field( pid_t, pid) __array( char, oldcomm, TASK_COMM_LEN) __array( char, newcomm, TASK_COMM_LEN) - __field( int, oom_score_adj) + __field( short, oom_score_adj) ), TP_fast_assign( @@ -50,7 +50,7 @@ TRACE_EVENT(task_rename, __entry->oom_score_adj = task->signal->oom_score_adj; ), - TP_printk("pid=%d oldcomm=%s newcomm=%s oom_score_adj=%d", + TP_printk("pid=%d oldcomm=%s newcomm=%s oom_score_adj=%hd", __entry->pid, __entry->oldcomm, __entry->newcomm, __entry->oom_score_adj) ); diff --git a/mm/ksm.c b/mm/ksm.c index 31ae5ea1eac0..b4d5a9deb17f 100644 --- a/mm/ksm.c +++ b/mm/ksm.c @@ -1919,7 +1919,7 @@ static ssize_t run_store(struct kobject *kobj, struct kobj_attribute *attr, if (ksm_run != flags) { ksm_run = flags; if (flags & KSM_RUN_UNMERGE) { - int oom_score_adj; + short oom_score_adj; oom_score_adj = test_set_oom_score_adj(OOM_SCORE_ADJ_MAX); err = unmerge_and_remove_all_rmap_items(); diff --git a/mm/oom_kill.c b/mm/oom_kill.c index 7e9e9113bd05..37ab4c5ab6e8 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -53,7 +53,7 @@ static DEFINE_SPINLOCK(zone_scan_lock); * @old_val. Usually used to reinstate a previous value to prevent racing with * userspacing tuning the value in the interim. */ -void compare_swap_oom_score_adj(int old_val, int new_val) +void compare_swap_oom_score_adj(short old_val, short new_val) { struct sighand_struct *sighand = current->sighand; @@ -72,7 +72,7 @@ void compare_swap_oom_score_adj(int old_val, int new_val) * synchronization and returns the old value. Usually used to temporarily * set a value, save the old value in the caller, and then reinstate it later. */ -int test_set_oom_score_adj(int new_val) +short test_set_oom_score_adj(short new_val) { struct sighand_struct *sighand = current->sighand; int old_val; @@ -193,7 +193,7 @@ unsigned long oom_badness(struct task_struct *p, struct mem_cgroup *memcg, if (!p) return 0; - adj = p->signal->oom_score_adj; + adj = (long)p->signal->oom_score_adj; if (adj == OOM_SCORE_ADJ_MIN) { task_unlock(p); return 0; @@ -399,7 +399,7 @@ static void dump_tasks(const struct mem_cgroup *memcg, const nodemask_t *nodemas continue; } - pr_info("[%5d] %5d %5d %8lu %8lu %7lu %8lu %5d %s\n", + pr_info("[%5d] %5d %5d %8lu %8lu %7lu %8lu %5hd %s\n", task->pid, from_kuid(&init_user_ns, task_uid(task)), task->tgid, task->mm->total_vm, get_mm_rss(task->mm), task->mm->nr_ptes, @@ -415,7 +415,7 @@ static void dump_header(struct task_struct *p, gfp_t gfp_mask, int order, { task_lock(current); pr_warning("%s invoked oom-killer: gfp_mask=0x%x, order=%d, " - "oom_score_adj=%d\n", + "oom_score_adj=%hd\n", current->comm, gfp_mask, order, current->signal->oom_score_adj); cpuset_print_task_mems_allowed(current); diff --git a/mm/swapfile.c b/mm/swapfile.c index 0fbb45283c66..bb6f6a04e92d 100644 --- a/mm/swapfile.c +++ b/mm/swapfile.c @@ -1498,7 +1498,7 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile) struct address_space *mapping; struct inode *inode; struct filename *pathname; - int oom_score_adj; + short oom_score_adj; int i, type, prev; int err; -- cgit v1.3-8-gc7d7 From e1e12d2f3104be886073ac6c5c4678f30b1b9e51 Mon Sep 17 00:00:00 2001 From: David Rientjes Date: Tue, 11 Dec 2012 16:02:56 -0800 Subject: mm, oom: fix race when specifying a thread as the oom origin test_set_oom_score_adj() and compare_swap_oom_score_adj() are used to specify that current should be killed first if an oom condition occurs in between the two calls. The usage is short oom_score_adj = test_set_oom_score_adj(OOM_SCORE_ADJ_MAX); ... compare_swap_oom_score_adj(OOM_SCORE_ADJ_MAX, oom_score_adj); to store the thread's oom_score_adj, temporarily change it to the maximum score possible, and then restore the old value if it is still the same. This happens to still be racy, however, if the user writes OOM_SCORE_ADJ_MAX to /proc/pid/oom_score_adj in between the two calls. The compare_swap_oom_score_adj() will then incorrectly reset the old value prior to the write of OOM_SCORE_ADJ_MAX. To fix this, introduce a new oom_flags_t member in struct signal_struct that will be used for per-thread oom killer flags. KSM and swapoff can now use a bit in this member to specify that threads should be killed first in oom conditions without playing around with oom_score_adj. This also allows the correct oom_score_adj to always be shown when reading /proc/pid/oom_score. Signed-off-by: David Rientjes Cc: KAMEZAWA Hiroyuki Cc: KOSAKI Motohiro Reviewed-by: Michal Hocko Cc: Anton Vorontsov Cc: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/oom.h | 19 +++++++++++++++++-- include/linux/sched.h | 1 + include/linux/types.h | 1 + mm/ksm.c | 7 ++----- mm/oom_kill.c | 49 +++++++------------------------------------------ mm/swapfile.c | 5 ++--- 6 files changed, 30 insertions(+), 52 deletions(-) (limited to 'include/linux/oom.h') diff --git a/include/linux/oom.h b/include/linux/oom.h index 922dab164eb0..da60007075b5 100644 --- a/include/linux/oom.h +++ b/include/linux/oom.h @@ -29,8 +29,23 @@ enum oom_scan_t { OOM_SCAN_SELECT, /* always select this thread first */ }; -extern void compare_swap_oom_score_adj(short old_val, short new_val); -extern short test_set_oom_score_adj(short new_val); +/* Thread is the potential origin of an oom condition; kill first on oom */ +#define OOM_FLAG_ORIGIN ((__force oom_flags_t)0x1) + +static inline void set_current_oom_origin(void) +{ + current->signal->oom_flags |= OOM_FLAG_ORIGIN; +} + +static inline void clear_current_oom_origin(void) +{ + current->signal->oom_flags &= ~OOM_FLAG_ORIGIN; +} + +static inline bool oom_task_origin(const struct task_struct *p) +{ + return !!(p->signal->oom_flags & OOM_FLAG_ORIGIN); +} extern unsigned long oom_badness(struct task_struct *p, struct mem_cgroup *memcg, const nodemask_t *nodemask, diff --git a/include/linux/sched.h b/include/linux/sched.h index ed30456152da..3e387df065fc 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -631,6 +631,7 @@ struct signal_struct { struct rw_semaphore group_rwsem; #endif + oom_flags_t oom_flags; short oom_score_adj; /* OOM kill score adjustment */ short oom_score_adj_min; /* OOM kill score adjustment min value. * Only settable by CAP_SYS_RESOURCE. */ diff --git a/include/linux/types.h b/include/linux/types.h index 1cc0e4b9a048..4d118ba11349 100644 --- a/include/linux/types.h +++ b/include/linux/types.h @@ -156,6 +156,7 @@ typedef u32 dma_addr_t; #endif typedef unsigned __bitwise__ gfp_t; typedef unsigned __bitwise__ fmode_t; +typedef unsigned __bitwise__ oom_flags_t; #ifdef CONFIG_PHYS_ADDR_T_64BIT typedef u64 phys_addr_t; diff --git a/mm/ksm.c b/mm/ksm.c index b4d5a9deb17f..382d930a0bf1 100644 --- a/mm/ksm.c +++ b/mm/ksm.c @@ -1919,12 +1919,9 @@ static ssize_t run_store(struct kobject *kobj, struct kobj_attribute *attr, if (ksm_run != flags) { ksm_run = flags; if (flags & KSM_RUN_UNMERGE) { - short oom_score_adj; - - oom_score_adj = test_set_oom_score_adj(OOM_SCORE_ADJ_MAX); + set_current_oom_origin(); err = unmerge_and_remove_all_rmap_items(); - compare_swap_oom_score_adj(OOM_SCORE_ADJ_MAX, - oom_score_adj); + clear_current_oom_origin(); if (err) { ksm_run = KSM_RUN_STOP; count = err; diff --git a/mm/oom_kill.c b/mm/oom_kill.c index 37ab4c5ab6e8..18f1ae2b45de 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -44,48 +44,6 @@ int sysctl_oom_kill_allocating_task; int sysctl_oom_dump_tasks = 1; static DEFINE_SPINLOCK(zone_scan_lock); -/* - * compare_swap_oom_score_adj() - compare and swap current's oom_score_adj - * @old_val: old oom_score_adj for compare - * @new_val: new oom_score_adj for swap - * - * Sets the oom_score_adj value for current to @new_val iff its present value is - * @old_val. Usually used to reinstate a previous value to prevent racing with - * userspacing tuning the value in the interim. - */ -void compare_swap_oom_score_adj(short old_val, short new_val) -{ - struct sighand_struct *sighand = current->sighand; - - spin_lock_irq(&sighand->siglock); - if (current->signal->oom_score_adj == old_val) - current->signal->oom_score_adj = new_val; - trace_oom_score_adj_update(current); - spin_unlock_irq(&sighand->siglock); -} - -/** - * test_set_oom_score_adj() - set current's oom_score_adj and return old value - * @new_val: new oom_score_adj value - * - * Sets the oom_score_adj value for current to @new_val with proper - * synchronization and returns the old value. Usually used to temporarily - * set a value, save the old value in the caller, and then reinstate it later. - */ -short test_set_oom_score_adj(short new_val) -{ - struct sighand_struct *sighand = current->sighand; - int old_val; - - spin_lock_irq(&sighand->siglock); - old_val = current->signal->oom_score_adj; - current->signal->oom_score_adj = new_val; - trace_oom_score_adj_update(current); - spin_unlock_irq(&sighand->siglock); - - return old_val; -} - #ifdef CONFIG_NUMA /** * has_intersects_mems_allowed() - check task eligiblity for kill @@ -310,6 +268,13 @@ enum oom_scan_t oom_scan_process_thread(struct task_struct *task, if (!task->mm) return OOM_SCAN_CONTINUE; + /* + * If task is allocating a lot of memory and has been marked to be + * killed first if it triggers an oom, then select it. + */ + if (oom_task_origin(task)) + return OOM_SCAN_SELECT; + if (task->flags & PF_EXITING && !force_kill) { /* * If this task is not being ptraced on exit, then wait for it diff --git a/mm/swapfile.c b/mm/swapfile.c index bb6f6a04e92d..e97a0e5aea91 100644 --- a/mm/swapfile.c +++ b/mm/swapfile.c @@ -1498,7 +1498,6 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile) struct address_space *mapping; struct inode *inode; struct filename *pathname; - short oom_score_adj; int i, type, prev; int err; @@ -1557,9 +1556,9 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile) p->flags &= ~SWP_WRITEOK; spin_unlock(&swap_lock); - oom_score_adj = test_set_oom_score_adj(OOM_SCORE_ADJ_MAX); + set_current_oom_origin(); err = try_to_unuse(type, false, 0); /* force all pages to be unused */ - compare_swap_oom_score_adj(OOM_SCORE_ADJ_MAX, oom_score_adj); + clear_current_oom_origin(); if (err) { /* re-insert swap space back into swap_list */ -- cgit v1.3-8-gc7d7