X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=mm%2Foom_kill.c;h=e9654b7ad11bf1b77da5ad2c683e20cb0b006c4c;hb=refs%2Fheads%2Fvserver;hp=f9c115a6291d7dc6cc2e7cadd5a627515992e8ff;hpb=43bc926fffd92024b46cafaf7350d669ba9ca884;p=linux-2.6.git diff --git a/mm/oom_kill.c b/mm/oom_kill.c index f9c115a62..e9654b7ad 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -15,17 +15,22 @@ * kernel subsystems and hints as to where to find out what things do. */ +#include #include #include #include #include #include #include +#include +#include +#include +int sysctl_panic_on_oom; /* #define DEBUG */ /** - * oom_badness - calculate a numeric value for how bad this task has been + * badness - calculate a numeric value for how bad this task has been * @p: task struct of which task we should calculate * @uptime: current uptime in seconds * @@ -61,12 +66,22 @@ unsigned long badness(struct task_struct *p, unsigned long uptime) */ points = mm->total_vm; + /* + * add points for context badness + */ + + points += vx_badness(p, mm); + /* * After this unlock we can no longer dereference local variable `mm' */ task_unlock(p); - /* FIXME: add vserver badness ;) */ + /* + * swapoff can easily use up all memory, so kill those first. + */ + if (p->flags & PF_SWAPOFF) + return ULONG_MAX; /* * Processes which fork a lot of child processes are likely @@ -127,6 +142,14 @@ unsigned long badness(struct task_struct *p, unsigned long uptime) if (cap_t(p->cap_effective) & CAP_TO_MASK(CAP_SYS_RAWIO)) points /= 4; + /* + * If p's nodes don't overlap ours, it may still help to kill p + * because p may have allocated or otherwise mapped memory on + * this node before. However it will be less likely. + */ + if (!cpuset_excl_nodes_overlap(p)) + points /= 8; + /* * Adjust the score by oomkilladj. */ @@ -138,8 +161,8 @@ unsigned long badness(struct task_struct *p, unsigned long uptime) } #ifdef DEBUG - printk(KERN_DEBUG "OOMkill: task %d (%s) got %d points\n", - p->pid, p->comm, points); + printk(KERN_DEBUG "OOMkill: task %d:#%u (%s) got %d points\n", + p->pid, p->xid, p->comm, points); #endif return points; } @@ -158,12 +181,18 @@ static inline int constrained_alloc(struct zonelist *zonelist, gfp_t gfp_mask) { #ifdef CONFIG_NUMA struct zone **z; - nodemask_t nodes = node_online_map; + nodemask_t nodes; + int node; + + nodes_clear(nodes); + /* node has memory ? */ + for_each_online_node(node) + if (NODE_DATA(node)->node_present_pages) + node_set(node, nodes); for (z = zonelist->zones; *z; z++) - if (cpuset_zone_allowed(*z, gfp_mask)) - node_clear((*z)->zone_pgdat->node_id, - nodes); + if (cpuset_zone_allowed_softwall(*z, gfp_mask)) + node_clear(zone_to_nid(*z), nodes); else return CONSTRAINT_CPUSET; @@ -190,27 +219,49 @@ static struct task_struct *select_bad_process(unsigned long *ppoints) do_posix_clock_monotonic_gettime(&uptime); do_each_thread(g, p) { unsigned long points; - int releasing; - /* skip the init task with pid == 1 */ - if (p->pid == 1) - continue; - if (p->oomkilladj == OOM_DISABLE) + /* + * skip kernel threads and tasks which have already released + * their mm. + */ + if (!p->mm) continue; - /* If p's nodes don't overlap ours, it won't help to kill p. */ - if (!cpuset_excl_nodes_overlap(p)) + /* skip the init task */ + if (is_init(p)) continue; /* - * This is in the process of releasing memory so for wait it - * to finish before killing some other task by mistake. + * This task already has access to memory reserves and is + * being killed. Don't allow any other task access to the + * memory reserve. + * + * Note: this may have a chance of deadlock if it gets + * blocked waiting for another task which itself is waiting + * for memory. Is there a better alternative? */ - releasing = test_tsk_thread_flag(p, TIF_MEMDIE) || - p->flags & PF_EXITING; - if (releasing && !(p->flags & PF_DEAD)) + if (test_tsk_thread_flag(p, TIF_MEMDIE)) return ERR_PTR(-1UL); - if (p->flags & PF_SWAPOFF) - return p; + + /* + * This is in the process of releasing memory so wait for it + * to finish before killing some other task by mistake. + * + * However, if p is the current task, we allow the 'kill' to + * go ahead if it is exiting: this will simply set TIF_MEMDIE, + * which will allow it to gain access to memory reserves in + * the process of exiting and releasing its resources. + * Otherwise we could get an easy OOM deadlock. + */ + if (p->flags & PF_EXITING) { + if (p != current) + return ERR_PTR(-1UL); + + chosen = p; + *ppoints = ULONG_MAX; + } + + if (p->oomkilladj == OOM_DISABLE) + continue; points = badness(p, uptime.tv_sec); if (points > *ppoints || !chosen) { @@ -218,32 +269,32 @@ static struct task_struct *select_bad_process(unsigned long *ppoints) *ppoints = points; } } while_each_thread(g, p); + return chosen; } /** - * We must be careful though to never send SIGKILL a process with - * CAP_SYS_RAW_IO set, send SIGTERM instead (but it's unlikely that - * we select a process with CAP_SYS_RAW_IO set). + * Send SIGKILL to the selected process irrespective of CAP_SYS_RAW_IO + * flag though it's unlikely that we select a process with CAP_SYS_RAW_IO + * set. */ -static void __oom_kill_task(task_t *p, const char *message) +static void __oom_kill_task(struct task_struct *p, int verbose) { - if (p->pid == 1) { + if (is_init(p)) { WARN_ON(1); printk(KERN_WARNING "tried to kill init!\n"); return; } - task_lock(p); - if (!p->mm || p->mm == &init_mm) { + if (!p->mm) { WARN_ON(1); printk(KERN_WARNING "tried to kill an mm-less task!\n"); - task_unlock(p); return; } - task_unlock(p); - printk(KERN_ERR "%s: Killed process %d (%s).\n", - message, p->pid, p->comm); + + if (verbose) + printk(KERN_ERR "Killed process %d:#%u (%s)\n", + p->pid, p->xid, p->comm); /* * We give our sacrificial lamb high priority and access to @@ -256,10 +307,10 @@ static void __oom_kill_task(task_t *p, const char *message) force_sig(SIGKILL, p); } -static int oom_kill_task(task_t *p, const char *message) +static int oom_kill_task(struct task_struct *p) { struct mm_struct *mm; - task_t * g, * q; + struct task_struct *g, *q; mm = p->mm; @@ -272,18 +323,28 @@ static int oom_kill_task(task_t *p, const char *message) * However, this is of no concern to us. */ - if (mm == NULL || mm == &init_mm) + if (mm == NULL) return 1; - __oom_kill_task(p, message); + /* + * Don't kill the process if any threads are set to OOM_DISABLE + */ + do_each_thread(g, q) { + if (q->mm == mm && q->oomkilladj == OOM_DISABLE) + return 1; + } while_each_thread(g, q); + + __oom_kill_task(p, 1); + /* * kill all processes that share the ->mm (i.e. all threads), - * but are in a different thread group + * but are in a different thread group. Don't let them have access + * to memory reserves though, otherwise we might deplete all memory. */ - do_each_thread(g, q) + do_each_thread(g, q) { if (q->mm == mm && q->tgid != p->tgid) - __oom_kill_task(q, message); - while_each_thread(g, q); + force_sig(SIGKILL, q); + } while_each_thread(g, q); return 0; } @@ -294,21 +355,45 @@ static int oom_kill_process(struct task_struct *p, unsigned long points, struct task_struct *c; struct list_head *tsk; - printk(KERN_ERR "Out of Memory: Kill process %d (%s) score %li and " - "children.\n", p->pid, p->comm, points); + /* + * If the task is already exiting, don't alarm the sysadmin or kill + * its children or threads, just set TIF_MEMDIE so it can die quickly + */ + if (p->flags & PF_EXITING) { + __oom_kill_task(p, 0); + return 0; + } + + printk(KERN_ERR "%s: kill process %d:#%u (%s) score %li or a child\n", + message, p->pid, p->xid, p->comm, points); + /* Try to kill a child first */ list_for_each(tsk, &p->children) { c = list_entry(tsk, struct task_struct, sibling); if (c->mm == p->mm) continue; - if (!oom_kill_task(c, message)) + if (!oom_kill_task(c)) return 0; } - return oom_kill_task(p, message); + return oom_kill_task(p); +} + +static BLOCKING_NOTIFIER_HEAD(oom_notify_list); + +int register_oom_notifier(struct notifier_block *nb) +{ + return blocking_notifier_chain_register(&oom_notify_list, nb); +} +EXPORT_SYMBOL_GPL(register_oom_notifier); + +int unregister_oom_notifier(struct notifier_block *nb) +{ + return blocking_notifier_chain_unregister(&oom_notify_list, nb); } +EXPORT_SYMBOL_GPL(unregister_oom_notifier); /** - * oom_kill - kill the "best" process when we run out of memory + * out_of_memory - kill the "best" process when we run out of memory * * If we run out of memory, we have the choice between either * killing a random task (bad), letting the system crash (worse) @@ -317,12 +402,19 @@ static int oom_kill_process(struct task_struct *p, unsigned long points, */ void out_of_memory(struct zonelist *zonelist, gfp_t gfp_mask, int order) { - task_t *p; + struct task_struct *p; unsigned long points = 0; + unsigned long freed = 0; + + blocking_notifier_call_chain(&oom_notify_list, 0, &freed); + if (freed > 0) + /* Got some memory back in the last second. */ + return; if (printk_ratelimit()) { - printk("oom-killer: gfp_mask=0x%x, order=%d\n", - gfp_mask, order); + printk(KERN_WARNING "%s invoked oom-killer: " + "gfp_mask=0x%x, order=%d, oomkilladj=%d\n", + current->comm, gfp_mask, order, current->oomkilladj); dump_stack(); show_mem(); } @@ -346,6 +438,8 @@ void out_of_memory(struct zonelist *zonelist, gfp_t gfp_mask, int order) break; case CONSTRAINT_NONE: + if (sysctl_panic_on_oom) + panic("out of memory. panic_on_oom is selected\n"); retry: /* * Rambo mode: Shoot down a process and hope it solves whatever