From 81d68a96a39844853b37f20cc8282d9b65b78ef3 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 12 May 2008 21:20:42 +0200 Subject: ftrace: trace irq disabled critical timings This patch adds latency tracing for critical timings (how long interrupts are disabled for). "irqsoff" is added to /debugfs/tracing/available_tracers Note: tracing_max_latency also holds the max latency for irqsoff (in usecs). (default to large number so one must start latency tracing) tracing_thresh threshold (in usecs) to always print out if irqs off is detected to be longer than stated here. If irq_thresh is non-zero, then max_irq_latency is ignored. Here's an example of a trace with ftrace_enabled = 0 ======= preemption latency trace v1.1.5 on 2.6.24-rc7 Signed-off-by: Ingo Molnar -------------------------------------------------------------------- latency: 100 us, #3/3, CPU#1 | (M:rt VP:0, KP:0, SP:0 HP:0 #P:2) ----------------- | task: swapper-0 (uid:0 nice:0 policy:0 rt_prio:0) ----------------- => started at: _spin_lock_irqsave+0x2a/0xb7 => ended at: _spin_unlock_irqrestore+0x32/0x5f _------=> CPU# / _-----=> irqs-off | / _----=> need-resched || / _---=> hardirq/softirq ||| / _--=> preempt-depth |||| / ||||| delay cmd pid ||||| time | caller \ / ||||| \ | / swapper-0 1d.s3 0us+: _spin_lock_irqsave+0x2a/0xb7 (e1000_update_stats+0x47/0x64c [e1000]) swapper-0 1d.s3 100us : _spin_unlock_irqrestore+0x32/0x5f (e1000_update_stats+0x641/0x64c [e1000]) swapper-0 1d.s3 100us : trace_hardirqs_on_caller+0x75/0x89 (_spin_unlock_irqrestore+0x32/0x5f) vim:ft=help ======= And this is a trace with ftrace_enabled == 1 ======= preemption latency trace v1.1.5 on 2.6.24-rc7 -------------------------------------------------------------------- latency: 102 us, #12/12, CPU#1 | (M:rt VP:0, KP:0, SP:0 HP:0 #P:2) ----------------- | task: swapper-0 (uid:0 nice:0 policy:0 rt_prio:0) ----------------- => started at: _spin_lock_irqsave+0x2a/0xb7 => ended at: _spin_unlock_irqrestore+0x32/0x5f _------=> CPU# / _-----=> irqs-off | / _----=> need-resched || / _---=> hardirq/softirq ||| / _--=> preempt-depth |||| / ||||| delay cmd pid ||||| time | caller \ / ||||| \ | / swapper-0 1dNs3 0us+: _spin_lock_irqsave+0x2a/0xb7 (e1000_update_stats+0x47/0x64c [e1000]) swapper-0 1dNs3 46us : e1000_read_phy_reg+0x16/0x225 [e1000] (e1000_update_stats+0x5e2/0x64c [e1000]) swapper-0 1dNs3 46us : e1000_swfw_sync_acquire+0x10/0x99 [e1000] (e1000_read_phy_reg+0x49/0x225 [e1000]) swapper-0 1dNs3 46us : e1000_get_hw_eeprom_semaphore+0x12/0xa6 [e1000] (e1000_swfw_sync_acquire+0x36/0x99 [e1000]) swapper-0 1dNs3 47us : __const_udelay+0x9/0x47 (e1000_read_phy_reg+0x116/0x225 [e1000]) swapper-0 1dNs3 47us+: __delay+0x9/0x50 (__const_udelay+0x45/0x47) swapper-0 1dNs3 97us : preempt_schedule+0xc/0x84 (__delay+0x4e/0x50) swapper-0 1dNs3 98us : e1000_swfw_sync_release+0xc/0x55 [e1000] (e1000_read_phy_reg+0x211/0x225 [e1000]) swapper-0 1dNs3 99us+: e1000_put_hw_eeprom_semaphore+0x9/0x35 [e1000] (e1000_swfw_sync_release+0x50/0x55 [e1000]) swapper-0 1dNs3 101us : _spin_unlock_irqrestore+0xe/0x5f (e1000_update_stats+0x641/0x64c [e1000]) swapper-0 1dNs3 102us : _spin_unlock_irqrestore+0x32/0x5f (e1000_update_stats+0x641/0x64c [e1000]) swapper-0 1dNs3 102us : trace_hardirqs_on_caller+0x75/0x89 (_spin_unlock_irqrestore+0x32/0x5f) vim:ft=help ======= Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- kernel/fork.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/fork.c') diff --git a/kernel/fork.c b/kernel/fork.c index 19908b26cf8..d66d676dc36 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -909,7 +909,7 @@ static struct task_struct *copy_process(unsigned long clone_flags, rt_mutex_init_task(p); -#ifdef CONFIG_TRACE_IRQFLAGS +#if defined(CONFIG_TRACE_IRQFLAGS) && defined(CONFIG_LOCKDEP) DEBUG_LOCKS_WARN_ON(!p->hardirqs_enabled); DEBUG_LOCKS_WARN_ON(!p->softirqs_enabled); #endif -- cgit v1.2.3-70-g09d2 From da9cbc87395308a21465bd25441297bbba0477e1 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Mon, 30 Jun 2008 20:42:08 +0200 Subject: block: blkdev.h cleanup, move iocontext stuff to iocontext.h Signed-off-by: Jens Axboe --- include/linux/blkdev.h | 17 ----------------- include/linux/iocontext.h | 18 ++++++++++++++++++ kernel/exit.c | 1 + kernel/fork.c | 1 + 4 files changed, 20 insertions(+), 17 deletions(-) (limited to 'kernel/fork.c') diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 4a9ed45270f..443df75d2cd 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -33,12 +33,6 @@ struct sg_io_hdr; #define BLKDEV_MIN_RQ 4 #define BLKDEV_MAX_RQ 128 /* Default maximum */ -int put_io_context(struct io_context *ioc); -void exit_io_context(void); -struct io_context *get_io_context(gfp_t gfp_flags, int node); -struct io_context *alloc_io_context(gfp_t gfp_flags, int node); -void copy_io_context(struct io_context **pdst, struct io_context **psrc); - struct request; typedef void (rq_end_io_fn)(struct request *, int); @@ -981,17 +975,6 @@ static inline long nr_blockdev_pages(void) return 0; } -static inline void exit_io_context(void) -{ -} - -struct io_context; -static inline int put_io_context(struct io_context *ioc) -{ - return 1; -} - - #endif /* CONFIG_BLOCK */ #endif diff --git a/include/linux/iocontext.h b/include/linux/iocontext.h index 2b7a1187cb2..08b987bccf8 100644 --- a/include/linux/iocontext.h +++ b/include/linux/iocontext.h @@ -99,4 +99,22 @@ static inline struct io_context *ioc_task_link(struct io_context *ioc) return NULL; } +#ifdef CONFIG_BLOCK +int put_io_context(struct io_context *ioc); +void exit_io_context(void); +struct io_context *get_io_context(gfp_t gfp_flags, int node); +struct io_context *alloc_io_context(gfp_t gfp_flags, int node); +void copy_io_context(struct io_context **pdst, struct io_context **psrc); +#else +static inline void exit_io_context(void) +{ +} + +struct io_context; +static inline int put_io_context(struct io_context *ioc) +{ + return 1; +} +#endif + #endif diff --git a/kernel/exit.c b/kernel/exit.c index 8f6185e69b6..ceb25878283 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -13,6 +13,7 @@ #include #include #include +#include #include #include #include diff --git a/kernel/fork.c b/kernel/fork.c index 19908b26cf8..b71ccd09fc8 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -23,6 +23,7 @@ #include #include #include +#include #include #include #include -- cgit v1.2.3-70-g09d2 From d12c1a37925a8ec386994169605fe99217295199 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Mon, 14 Jul 2008 12:09:28 +0200 Subject: lockdep: fix kernel/fork.c warning fix: [ 0.184011] ------------[ cut here ]------------ [ 0.188011] WARNING: at kernel/fork.c:918 copy_process+0x1c0/0x1084() [ 0.192011] Pid: 0, comm: swapper Not tainted 2.6.26-tip-00351-g01d4a50-dirty #14521 [ 0.196011] [] warn_on_slowpath+0x3c/0x60 [ 0.200012] [] ? __alloc_pages_internal+0x92/0x36b [ 0.208012] [] ? __spin_lock_init+0x24/0x4a [ 0.212012] [] copy_process+0x1c0/0x1084 [ 0.216013] [] do_fork+0xb8/0x1ad [ 0.220013] [] ? acpi_os_release_lock+0x8/0xa [ 0.228013] [] ? acpi_os_vprintf+0x20/0x24 [ 0.232014] [] kernel_thread+0x75/0x7d [ 0.236014] [] ? kernel_init+0x0/0x24a [ 0.240014] [] ? kernel_init+0x0/0x24a [ 0.244014] [] ? kernel_thread_helper+0x0/0x10 [ 0.252015] [] rest_init+0x14/0x50 [ 0.256015] [] start_kernel+0x2b9/0x2c0 [ 0.260015] [] __init_begin+0x4f/0x57 [ 0.264016] ======================= [ 0.268016] ---[ end trace 4eaa2a86a8e2da22 ]--- [ 0.272016] enabled ExtINT on CPU#0 which occurs if CONFIG_TRACE_IRQFLAGS=y, CONFIG_DEBUG_LOCKDEP=y, but CONFIG_PROVE_LOCKING is disabled. Signed-off-by: Ingo Molnar --- kernel/fork.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/fork.c') diff --git a/kernel/fork.c b/kernel/fork.c index 19908b26cf8..cdb1f82d3bd 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -909,7 +909,7 @@ static struct task_struct *copy_process(unsigned long clone_flags, rt_mutex_init_task(p); -#ifdef CONFIG_TRACE_IRQFLAGS +#ifdef CONFIG_PROVE_LOCKING DEBUG_LOCKS_WARN_ON(!p->hardirqs_enabled); DEBUG_LOCKS_WARN_ON(!p->softirqs_enabled); #endif -- cgit v1.2.3-70-g09d2 From f470021adb9190819c03d6d8c5c860a17480aa6d Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Mon, 24 Mar 2008 18:36:23 -0700 Subject: ptrace children revamp ptrace no longer fiddles with the children/sibling links, and the old ptrace_children list is gone. Now ptrace, whether of one's own children or another's via PTRACE_ATTACH, just uses the new ptraced list instead. There should be no user-visible difference that matters. The only change is the order in which do_wait() sees multiple stopped children and stopped ptrace attachees. Since wait_task_stopped() was changed earlier so it no longer reorders the children list, we already know this won't cause any new problems. Signed-off-by: Roland McGrath --- include/linux/init_task.h | 4 +- include/linux/sched.h | 26 +++--- kernel/exit.c | 226 ++++++++++++++++++++++++---------------------- kernel/fork.c | 6 +- kernel/ptrace.c | 37 +++++--- 5 files changed, 160 insertions(+), 139 deletions(-) (limited to 'kernel/fork.c') diff --git a/include/linux/init_task.h b/include/linux/init_task.h index 9927a88674a..93c45acf249 100644 --- a/include/linux/init_task.h +++ b/include/linux/init_task.h @@ -140,8 +140,8 @@ extern struct group_info init_groups; .nr_cpus_allowed = NR_CPUS, \ }, \ .tasks = LIST_HEAD_INIT(tsk.tasks), \ - .ptrace_children= LIST_HEAD_INIT(tsk.ptrace_children), \ - .ptrace_list = LIST_HEAD_INIT(tsk.ptrace_list), \ + .ptraced = LIST_HEAD_INIT(tsk.ptraced), \ + .ptrace_entry = LIST_HEAD_INIT(tsk.ptrace_entry), \ .real_parent = &tsk, \ .parent = &tsk, \ .children = LIST_HEAD_INIT(tsk.children), \ diff --git a/include/linux/sched.h b/include/linux/sched.h index ba2f859c6e4..1941d8b5cf1 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1062,12 +1062,6 @@ struct task_struct { #endif struct list_head tasks; - /* - * ptrace_list/ptrace_children forms the list of my children - * that were stolen by a ptracer. - */ - struct list_head ptrace_children; - struct list_head ptrace_list; struct mm_struct *mm, *active_mm; @@ -1089,18 +1083,25 @@ struct task_struct { /* * pointers to (original) parent process, youngest child, younger sibling, * older sibling, respectively. (p->father can be replaced with - * p->parent->pid) + * p->real_parent->pid) */ - struct task_struct *real_parent; /* real parent process (when being debugged) */ - struct task_struct *parent; /* parent process */ + struct task_struct *real_parent; /* real parent process */ + struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */ /* - * children/sibling forms the list of my children plus the - * tasks I'm ptracing. + * children/sibling forms the list of my natural children */ struct list_head children; /* list of my children */ struct list_head sibling; /* linkage in my parent's children list */ struct task_struct *group_leader; /* threadgroup leader */ + /* + * ptraced is the list of tasks this task is using ptrace on. + * This includes both natural children and PTRACE_ATTACH targets. + * p->ptrace_entry is p's link on the p->parent->ptraced list. + */ + struct list_head ptraced; + struct list_head ptrace_entry; + /* PID/PID hash table linkage. */ struct pid_link pids[PIDTYPE_MAX]; struct list_head thread_group; @@ -1876,9 +1877,6 @@ extern void wait_task_inactive(struct task_struct * p); #define wait_task_inactive(p) do { } while (0) #endif -#define remove_parent(p) list_del_init(&(p)->sibling) -#define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children) - #define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks) #define for_each_process(p) \ diff --git a/kernel/exit.c b/kernel/exit.c index 7453356a961..1e909826a80 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -71,7 +71,7 @@ static void __unhash_process(struct task_struct *p) __get_cpu_var(process_counts)--; } list_del_rcu(&p->thread_group); - remove_parent(p); + list_del_init(&p->sibling); } /* @@ -152,6 +152,18 @@ static void delayed_put_task_struct(struct rcu_head *rhp) put_task_struct(container_of(rhp, struct task_struct, rcu)); } +/* + * Do final ptrace-related cleanup of a zombie being reaped. + * + * Called with write_lock(&tasklist_lock) held. + */ +static void ptrace_release_task(struct task_struct *p) +{ + BUG_ON(!list_empty(&p->ptraced)); + ptrace_unlink(p); + BUG_ON(!list_empty(&p->ptrace_entry)); +} + void release_task(struct task_struct * p) { struct task_struct *leader; @@ -160,8 +172,7 @@ repeat: atomic_dec(&p->user->processes); proc_flush_task(p); write_lock_irq(&tasklist_lock); - ptrace_unlink(p); - BUG_ON(!list_empty(&p->ptrace_list) || !list_empty(&p->ptrace_children)); + ptrace_release_task(p); __exit_signal(p); /* @@ -315,9 +326,8 @@ static void reparent_to_kthreadd(void) ptrace_unlink(current); /* Reparent to init */ - remove_parent(current); current->real_parent = current->parent = kthreadd_task; - add_parent(current); + list_move_tail(¤t->sibling, ¤t->real_parent->children); /* Set the exit signal to SIGCHLD so we signal init on exit */ current->exit_signal = SIGCHLD; @@ -692,37 +702,71 @@ static void exit_mm(struct task_struct * tsk) mmput(mm); } -static void -reparent_thread(struct task_struct *p, struct task_struct *father, int traced) +/* + * Detach all tasks we were using ptrace on. + * Any that need to be release_task'd are put on the @dead list. + * + * Called with write_lock(&tasklist_lock) held. + */ +static void ptrace_exit(struct task_struct *parent, struct list_head *dead) { - if (p->pdeath_signal) - /* We already hold the tasklist_lock here. */ - group_send_sig_info(p->pdeath_signal, SEND_SIG_NOINFO, p); + struct task_struct *p, *n; - /* Move the child from its dying parent to the new one. */ - if (unlikely(traced)) { - /* Preserve ptrace links if someone else is tracing this child. */ - list_del_init(&p->ptrace_list); - if (ptrace_reparented(p)) - list_add(&p->ptrace_list, &p->real_parent->ptrace_children); - } else { - /* If this child is being traced, then we're the one tracing it - * anyway, so let go of it. + list_for_each_entry_safe(p, n, &parent->ptraced, ptrace_entry) { + __ptrace_unlink(p); + + if (p->exit_state != EXIT_ZOMBIE) + continue; + + /* + * If it's a zombie, our attachedness prevented normal + * parent notification or self-reaping. Do notification + * now if it would have happened earlier. If it should + * reap itself, add it to the @dead list. We can't call + * release_task() here because we already hold tasklist_lock. + * + * If it's our own child, there is no notification to do. */ - p->ptrace = 0; - remove_parent(p); - p->parent = p->real_parent; - add_parent(p); + if (!task_detached(p) && thread_group_empty(p)) { + if (!same_thread_group(p->real_parent, parent)) + do_notify_parent(p, p->exit_signal); + } - if (task_is_traced(p)) { + if (task_detached(p)) { /* - * If it was at a trace stop, turn it into - * a normal stop since it's no longer being - * traced. + * Mark it as in the process of being reaped. */ - ptrace_untrace(p); + p->exit_state = EXIT_DEAD; + list_add(&p->ptrace_entry, dead); } } +} + +/* + * Finish up exit-time ptrace cleanup. + * + * Called without locks. + */ +static void ptrace_exit_finish(struct task_struct *parent, + struct list_head *dead) +{ + struct task_struct *p, *n; + + BUG_ON(!list_empty(&parent->ptraced)); + + list_for_each_entry_safe(p, n, dead, ptrace_entry) { + list_del_init(&p->ptrace_entry); + release_task(p); + } +} + +static void reparent_thread(struct task_struct *p, struct task_struct *father) +{ + if (p->pdeath_signal) + /* We already hold the tasklist_lock here. */ + group_send_sig_info(p->pdeath_signal, SEND_SIG_NOINFO, p); + + list_move_tail(&p->sibling, &p->real_parent->children); /* If this is a threaded reparent there is no need to * notify anyone anything has happened. @@ -737,7 +781,8 @@ reparent_thread(struct task_struct *p, struct task_struct *father, int traced) /* If we'd notified the old parent about this child's death, * also notify the new parent. */ - if (!traced && p->exit_state == EXIT_ZOMBIE && + if (!ptrace_reparented(p) && + p->exit_state == EXIT_ZOMBIE && !task_detached(p) && thread_group_empty(p)) do_notify_parent(p, p->exit_signal); @@ -754,12 +799,15 @@ reparent_thread(struct task_struct *p, struct task_struct *father, int traced) static void forget_original_parent(struct task_struct *father) { struct task_struct *p, *n, *reaper = father; - struct list_head ptrace_dead; - - INIT_LIST_HEAD(&ptrace_dead); + LIST_HEAD(ptrace_dead); write_lock_irq(&tasklist_lock); + /* + * First clean up ptrace if we were using it. + */ + ptrace_exit(father, &ptrace_dead); + do { reaper = next_thread(reaper); if (reaper == father) { @@ -768,58 +816,19 @@ static void forget_original_parent(struct task_struct *father) } } while (reaper->flags & PF_EXITING); - /* - * There are only two places where our children can be: - * - * - in our child list - * - in our ptraced child list - * - * Search them and reparent children. - */ list_for_each_entry_safe(p, n, &father->children, sibling) { - int ptrace; - - ptrace = p->ptrace; - - /* if father isn't the real parent, then ptrace must be enabled */ - BUG_ON(father != p->real_parent && !ptrace); - - if (father == p->real_parent) { - /* reparent with a reaper, real father it's us */ - p->real_parent = reaper; - reparent_thread(p, father, 0); - } else { - /* reparent ptraced task to its real parent */ - __ptrace_unlink (p); - if (p->exit_state == EXIT_ZOMBIE && !task_detached(p) && - thread_group_empty(p)) - do_notify_parent(p, p->exit_signal); - } - - /* - * if the ptraced child is a detached zombie we must collect - * it before we exit, or it will remain zombie forever since - * we prevented it from self-reap itself while it was being - * traced by us, to be able to see it in wait4. - */ - if (unlikely(ptrace && p->exit_state == EXIT_ZOMBIE && task_detached(p))) - list_add(&p->ptrace_list, &ptrace_dead); - } - - list_for_each_entry_safe(p, n, &father->ptrace_children, ptrace_list) { p->real_parent = reaper; - reparent_thread(p, father, 1); + if (p->parent == father) { + BUG_ON(p->ptrace); + p->parent = p->real_parent; + } + reparent_thread(p, father); } write_unlock_irq(&tasklist_lock); BUG_ON(!list_empty(&father->children)); - BUG_ON(!list_empty(&father->ptrace_children)); - - list_for_each_entry_safe(p, n, &ptrace_dead, ptrace_list) { - list_del_init(&p->ptrace_list); - release_task(p); - } + ptrace_exit_finish(father, &ptrace_dead); } /* @@ -1180,13 +1189,6 @@ static int eligible_child(enum pid_type type, struct pid *pid, int options, return 0; } - /* - * Do not consider detached threads that are - * not ptraced: - */ - if (task_detached(p) && !p->ptrace) - return 0; - /* Wait for all children (clone and not) if __WALL is set; * otherwise, wait for clone children *only* if __WCLONE is * set; otherwise, wait for non-clone children *only*. (Note: @@ -1399,7 +1401,7 @@ static int wait_task_zombie(struct task_struct *p, int options, * the lock and this task is uninteresting. If we return nonzero, we have * released the lock and the system call should return. */ -static int wait_task_stopped(struct task_struct *p, +static int wait_task_stopped(int ptrace, struct task_struct *p, int options, struct siginfo __user *infop, int __user *stat_addr, struct rusage __user *ru) { @@ -1407,7 +1409,7 @@ static int wait_task_stopped(struct task_struct *p, uid_t uid = 0; /* unneeded, required by compiler */ pid_t pid; - if (!(p->ptrace & PT_PTRACED) && !(options & WUNTRACED)) + if (!(options & WUNTRACED)) return 0; exit_code = 0; @@ -1416,7 +1418,7 @@ static int wait_task_stopped(struct task_struct *p, if (unlikely(!task_is_stopped_or_traced(p))) goto unlock_sig; - if (!(p->ptrace & PT_PTRACED) && p->signal->group_stop_count > 0) + if (!ptrace && p->signal->group_stop_count > 0) /* * A group stop is in progress and this is the group leader. * We won't report until all threads have stopped. @@ -1445,7 +1447,7 @@ unlock_sig: */ get_task_struct(p); pid = task_pid_vnr(p); - why = (p->ptrace & PT_PTRACED) ? CLD_TRAPPED : CLD_STOPPED; + why = ptrace ? CLD_TRAPPED : CLD_STOPPED; read_unlock(&tasklist_lock); if (unlikely(options & WNOWAIT)) @@ -1536,7 +1538,7 @@ static int wait_task_continued(struct task_struct *p, int options, * Returns zero if the search for a child should continue; * then *@notask_error is 0 if @p is an eligible child, or still -ECHILD. */ -static int wait_consider_task(struct task_struct *parent, +static int wait_consider_task(struct task_struct *parent, int ptrace, struct task_struct *p, int *notask_error, enum pid_type type, struct pid *pid, int options, struct siginfo __user *infop, @@ -1546,6 +1548,15 @@ static int wait_consider_task(struct task_struct *parent, if (ret <= 0) return ret; + if (likely(!ptrace) && unlikely(p->ptrace)) { + /* + * This child is hidden by ptrace. + * We aren't allowed to see it now, but eventually we will. + */ + *notask_error = 0; + return 0; + } + if (p->exit_state == EXIT_DEAD) return 0; @@ -1562,7 +1573,8 @@ static int wait_consider_task(struct task_struct *parent, *notask_error = 0; if (task_is_stopped_or_traced(p)) - return wait_task_stopped(p, options, infop, stat_addr, ru); + return wait_task_stopped(ptrace, p, options, + infop, stat_addr, ru); return wait_task_continued(p, options, infop, stat_addr, ru); } @@ -1583,11 +1595,16 @@ static int do_wait_thread(struct task_struct *tsk, int *notask_error, struct task_struct *p; list_for_each_entry(p, &tsk->children, sibling) { - int ret = wait_consider_task(tsk, p, notask_error, - type, pid, options, - infop, stat_addr, ru); - if (ret) - return ret; + /* + * Do not consider detached threads. + */ + if (!task_detached(p)) { + int ret = wait_consider_task(tsk, 0, p, notask_error, + type, pid, options, + infop, stat_addr, ru); + if (ret) + return ret; + } } return 0; @@ -1601,21 +1618,16 @@ static int ptrace_do_wait(struct task_struct *tsk, int *notask_error, struct task_struct *p; /* - * If we never saw an eligile child, check for children stolen by - * ptrace. We don't leave -ECHILD in *@notask_error if there are any, - * because we will eventually be allowed to wait for them again. + * Traditionally we see ptrace'd stopped tasks regardless of options. */ - if (!*notask_error) - return 0; + options |= WUNTRACED; - list_for_each_entry(p, &tsk->ptrace_children, ptrace_list) { - int ret = eligible_child(type, pid, options, p); - if (unlikely(ret < 0)) + list_for_each_entry(p, &tsk->ptraced, ptrace_entry) { + int ret = wait_consider_task(tsk, 1, p, notask_error, + type, pid, options, + infop, stat_addr, ru); + if (ret) return ret; - if (ret) { - *notask_error = 0; - return 0; - } } return 0; diff --git a/kernel/fork.c b/kernel/fork.c index 4bd2f516401..adefc1131f2 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1125,8 +1125,8 @@ static struct task_struct *copy_process(unsigned long clone_flags, */ p->group_leader = p; INIT_LIST_HEAD(&p->thread_group); - INIT_LIST_HEAD(&p->ptrace_children); - INIT_LIST_HEAD(&p->ptrace_list); + INIT_LIST_HEAD(&p->ptrace_entry); + INIT_LIST_HEAD(&p->ptraced); /* Now that the task is set up, run cgroup callbacks if * necessary. We need to run them before the task is visible @@ -1198,7 +1198,7 @@ static struct task_struct *copy_process(unsigned long clone_flags, } if (likely(p->pid)) { - add_parent(p); + list_add_tail(&p->sibling, &p->real_parent->children); if (unlikely(p->ptrace & PT_PTRACED)) __ptrace_link(p, current->parent); diff --git a/kernel/ptrace.c b/kernel/ptrace.c index e337390fce0..8392a9da645 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -33,13 +33,9 @@ */ void __ptrace_link(struct task_struct *child, struct task_struct *new_parent) { - BUG_ON(!list_empty(&child->ptrace_list)); - if (child->parent == new_parent) - return; - list_add(&child->ptrace_list, &child->parent->ptrace_children); - remove_parent(child); + BUG_ON(!list_empty(&child->ptrace_entry)); + list_add(&child->ptrace_entry, &new_parent->ptraced); child->parent = new_parent; - add_parent(child); } /* @@ -73,12 +69,8 @@ void __ptrace_unlink(struct task_struct *child) BUG_ON(!child->ptrace); child->ptrace = 0; - if (ptrace_reparented(child)) { - list_del_init(&child->ptrace_list); - remove_parent(child); - child->parent = child->real_parent; - add_parent(child); - } + child->parent = child->real_parent; + list_del_init(&child->ptrace_entry); if (task_is_traced(child)) ptrace_untrace(child); @@ -492,15 +484,34 @@ int ptrace_traceme(void) /* * Are we already being traced? */ +repeat: task_lock(current); if (!(current->ptrace & PT_PTRACED)) { + /* + * See ptrace_attach() comments about the locking here. + */ + unsigned long flags; + if (!write_trylock_irqsave(&tasklist_lock, flags)) { + task_unlock(current); + do { + cpu_relax(); + } while (!write_can_lock(&tasklist_lock)); + goto repeat; + } + ret = security_ptrace(current->parent, current, PTRACE_MODE_ATTACH); + /* * Set the ptrace bit in the process ptrace flags. + * Then link us on our parent's ptraced list. */ - if (!ret) + if (!ret) { current->ptrace |= PT_PTRACED; + __ptrace_link(current, current->real_parent); + } + + write_unlock_irqrestore(&tasklist_lock, flags); } task_unlock(current); return ret; -- cgit v1.2.3-70-g09d2 From a1e78772d72b2616ed20e54896e68e0e7044854e Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 23 Jul 2008 21:27:23 -0700 Subject: hugetlb: reserve huge pages for reliable MAP_PRIVATE hugetlbfs mappings until fork() This patch reserves huge pages at mmap() time for MAP_PRIVATE mappings in a similar manner to the reservations taken for MAP_SHARED mappings. The reserve count is accounted both globally and on a per-VMA basis for private mappings. This guarantees that a process that successfully calls mmap() will successfully fault all pages in the future unless fork() is called. The characteristics of private mappings of hugetlbfs files behaviour after this patch are; 1. The process calling mmap() is guaranteed to succeed all future faults until it forks(). 2. On fork(), the parent may die due to SIGKILL on writes to the private mapping if enough pages are not available for the COW. For reasonably reliable behaviour in the face of a small huge page pool, children of hugepage-aware processes should not reference the mappings; such as might occur when fork()ing to exec(). 3. On fork(), the child VMAs inherit no reserves. Reads on pages already faulted by the parent will succeed. Successful writes will depend on enough huge pages being free in the pool. 4. Quotas of the hugetlbfs mount are checked at reserve time for the mapper and at fault time otherwise. Before this patch, all reads or writes in the child potentially needs page allocations that can later lead to the death of the parent. This applies to reads and writes of uninstantiated pages as well as COW. After the patch it is only a write to an instantiated page that causes problems. Signed-off-by: Mel Gorman Acked-by: Adam Litke Cc: Andy Whitcroft Cc: William Lee Irwin III Cc: Hugh Dickins Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/hugetlbfs/inode.c | 8 +-- include/linux/hugetlb.h | 9 ++- kernel/fork.c | 9 +++ mm/hugetlb.c | 158 ++++++++++++++++++++++++++++++++++++------------ 4 files changed, 140 insertions(+), 44 deletions(-) (limited to 'kernel/fork.c') diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c index aeabf80f81a..1576bbecd08 100644 --- a/fs/hugetlbfs/inode.c +++ b/fs/hugetlbfs/inode.c @@ -103,9 +103,9 @@ static int hugetlbfs_file_mmap(struct file *file, struct vm_area_struct *vma) ret = -ENOMEM; len = vma_len + ((loff_t)vma->vm_pgoff << PAGE_SHIFT); - if (vma->vm_flags & VM_MAYSHARE && - hugetlb_reserve_pages(inode, vma->vm_pgoff >> (HPAGE_SHIFT-PAGE_SHIFT), - len >> HPAGE_SHIFT)) + if (hugetlb_reserve_pages(inode, + vma->vm_pgoff >> (HPAGE_SHIFT-PAGE_SHIFT), + len >> HPAGE_SHIFT, vma)) goto out; ret = 0; @@ -942,7 +942,7 @@ struct file *hugetlb_file_setup(const char *name, size_t size) goto out_dentry; error = -ENOMEM; - if (hugetlb_reserve_pages(inode, 0, size >> HPAGE_SHIFT)) + if (hugetlb_reserve_pages(inode, 0, size >> HPAGE_SHIFT, NULL)) goto out_inode; d_instantiate(dentry, inode); diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h index a79e80b689d..185b14c9f02 100644 --- a/include/linux/hugetlb.h +++ b/include/linux/hugetlb.h @@ -17,6 +17,7 @@ static inline int is_vm_hugetlb_page(struct vm_area_struct *vma) return vma->vm_flags & VM_HUGETLB; } +void reset_vma_resv_huge_pages(struct vm_area_struct *vma); int hugetlb_sysctl_handler(struct ctl_table *, int, struct file *, void __user *, size_t *, loff_t *); int hugetlb_overcommit_handler(struct ctl_table *, int, struct file *, void __user *, size_t *, loff_t *); int hugetlb_treat_movable_handler(struct ctl_table *, int, struct file *, void __user *, size_t *, loff_t *); @@ -30,7 +31,8 @@ int hugetlb_report_node_meminfo(int, char *); unsigned long hugetlb_total_pages(void); int hugetlb_fault(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, int write_access); -int hugetlb_reserve_pages(struct inode *inode, long from, long to); +int hugetlb_reserve_pages(struct inode *inode, long from, long to, + struct vm_area_struct *vma); void hugetlb_unreserve_pages(struct inode *inode, long offset, long freed); extern unsigned long max_huge_pages; @@ -58,6 +60,11 @@ static inline int is_vm_hugetlb_page(struct vm_area_struct *vma) { return 0; } + +static inline void reset_vma_resv_huge_pages(struct vm_area_struct *vma) +{ +} + static inline unsigned long hugetlb_total_pages(void) { return 0; diff --git a/kernel/fork.c b/kernel/fork.c index adefc1131f2..552c8d8e77a 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -33,6 +33,7 @@ #include #include #include +#include #include #include #include @@ -306,6 +307,14 @@ static int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm) spin_unlock(&file->f_mapping->i_mmap_lock); } + /* + * Clear hugetlb-related page reserves for children. This only + * affects MAP_PRIVATE mappings. Faults generated by the child + * are not guaranteed to succeed, even if read-only + */ + if (is_vm_hugetlb_page(tmp)) + reset_vma_resv_huge_pages(tmp); + /* * Link in the new vma and copy the page table entries. */ diff --git a/mm/hugetlb.c b/mm/hugetlb.c index a4dbba8965f..0af500db363 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -40,6 +40,69 @@ static int hugetlb_next_nid; */ static DEFINE_SPINLOCK(hugetlb_lock); +/* + * These helpers are used to track how many pages are reserved for + * faults in a MAP_PRIVATE mapping. Only the process that called mmap() + * is guaranteed to have their future faults succeed. + * + * With the exception of reset_vma_resv_huge_pages() which is called at fork(), + * the reserve counters are updated with the hugetlb_lock held. It is safe + * to reset the VMA at fork() time as it is not in use yet and there is no + * chance of the global counters getting corrupted as a result of the values. + */ +static unsigned long vma_resv_huge_pages(struct vm_area_struct *vma) +{ + VM_BUG_ON(!is_vm_hugetlb_page(vma)); + if (!(vma->vm_flags & VM_SHARED)) + return (unsigned long)vma->vm_private_data; + return 0; +} + +static void set_vma_resv_huge_pages(struct vm_area_struct *vma, + unsigned long reserve) +{ + VM_BUG_ON(!is_vm_hugetlb_page(vma)); + VM_BUG_ON(vma->vm_flags & VM_SHARED); + + vma->vm_private_data = (void *)reserve; +} + +/* Decrement the reserved pages in the hugepage pool by one */ +static void decrement_hugepage_resv_vma(struct vm_area_struct *vma) +{ + if (vma->vm_flags & VM_SHARED) { + /* Shared mappings always use reserves */ + resv_huge_pages--; + } else { + /* + * Only the process that called mmap() has reserves for + * private mappings. + */ + if (vma_resv_huge_pages(vma)) { + resv_huge_pages--; + reserve = (unsigned long)vma->vm_private_data - 1; + vma->vm_private_data = (void *)reserve; + } + } +} + +void reset_vma_resv_huge_pages(struct vm_area_struct *vma) +{ + VM_BUG_ON(!is_vm_hugetlb_page(vma)); + if (!(vma->vm_flags & VM_SHARED)) + vma->vm_private_data = (void *)0; +} + +/* Returns true if the VMA has associated reserve pages */ +static int vma_has_private_reserves(struct vm_area_struct *vma) +{ + if (vma->vm_flags & VM_SHARED) + return 0; + if (!vma_resv_huge_pages(vma)) + return 0; + return 1; +} + static void clear_huge_page(struct page *page, unsigned long addr) { int i; @@ -101,6 +164,15 @@ static struct page *dequeue_huge_page_vma(struct vm_area_struct *vma, struct zone *zone; struct zoneref *z; + /* + * A child process with MAP_PRIVATE mappings created by their parent + * have no page reserves. This check ensures that reservations are + * not "stolen". The child may still get SIGKILLed + */ + if (!vma_has_private_reserves(vma) && + free_huge_pages - resv_huge_pages == 0) + return NULL; + for_each_zone_zonelist_nodemask(zone, z, zonelist, MAX_NR_ZONES - 1, nodemask) { nid = zone_to_nid(zone); @@ -111,8 +183,8 @@ static struct page *dequeue_huge_page_vma(struct vm_area_struct *vma, list_del(&page->lru); free_huge_pages--; free_huge_pages_node[nid]--; - if (vma && vma->vm_flags & VM_MAYSHARE) - resv_huge_pages--; + decrement_hugepage_resv_vma(vma); + break; } } @@ -461,55 +533,40 @@ static void return_unused_surplus_pages(unsigned long unused_resv_pages) } } - -static struct page *alloc_huge_page_shared(struct vm_area_struct *vma, - unsigned long addr) +static struct page *alloc_huge_page(struct vm_area_struct *vma, + unsigned long addr) { struct page *page; + struct address_space *mapping = vma->vm_file->f_mapping; + struct inode *inode = mapping->host; + unsigned int chg = 0; + + /* + * Processes that did not create the mapping will have no reserves and + * will not have accounted against quota. Check that the quota can be + * made before satisfying the allocation + */ + if (!vma_has_private_reserves(vma)) { + chg = 1; + if (hugetlb_get_quota(inode->i_mapping, chg)) + return ERR_PTR(-ENOSPC); + } spin_lock(&hugetlb_lock); page = dequeue_huge_page_vma(vma, addr); spin_unlock(&hugetlb_lock); - return page ? page : ERR_PTR(-VM_FAULT_OOM); -} -static struct page *alloc_huge_page_private(struct vm_area_struct *vma, - unsigned long addr) -{ - struct page *page = NULL; - - if (hugetlb_get_quota(vma->vm_file->f_mapping, 1)) - return ERR_PTR(-VM_FAULT_SIGBUS); - - spin_lock(&hugetlb_lock); - if (free_huge_pages > resv_huge_pages) - page = dequeue_huge_page_vma(vma, addr); - spin_unlock(&hugetlb_lock); if (!page) { page = alloc_buddy_huge_page(vma, addr); if (!page) { - hugetlb_put_quota(vma->vm_file->f_mapping, 1); + hugetlb_put_quota(inode->i_mapping, chg); return ERR_PTR(-VM_FAULT_OOM); } } - return page; -} -static struct page *alloc_huge_page(struct vm_area_struct *vma, - unsigned long addr) -{ - struct page *page; - struct address_space *mapping = vma->vm_file->f_mapping; - - if (vma->vm_flags & VM_MAYSHARE) - page = alloc_huge_page_shared(vma, addr); - else - page = alloc_huge_page_private(vma, addr); + set_page_refcounted(page); + set_page_private(page, (unsigned long) mapping); - if (!IS_ERR(page)) { - set_page_refcounted(page); - set_page_private(page, (unsigned long) mapping); - } return page; } @@ -757,6 +814,13 @@ out: return ret; } +static void hugetlb_vm_op_close(struct vm_area_struct *vma) +{ + unsigned long reserve = vma_resv_huge_pages(vma); + if (reserve) + hugetlb_acct_memory(-reserve); +} + /* * We cannot handle pagefaults against hugetlb pages at all. They cause * handle_mm_fault() to try to instantiate regular-sized pages in the @@ -771,6 +835,7 @@ static int hugetlb_vm_op_fault(struct vm_area_struct *vma, struct vm_fault *vmf) struct vm_operations_struct hugetlb_vm_ops = { .fault = hugetlb_vm_op_fault, + .close = hugetlb_vm_op_close, }; static pte_t make_huge_pte(struct vm_area_struct *vma, struct page *page, @@ -1289,11 +1354,25 @@ static long region_truncate(struct list_head *head, long end) return chg; } -int hugetlb_reserve_pages(struct inode *inode, long from, long to) +int hugetlb_reserve_pages(struct inode *inode, + long from, long to, + struct vm_area_struct *vma) { long ret, chg; - chg = region_chg(&inode->i_mapping->private_list, from, to); + /* + * Shared mappings base their reservation on the number of pages that + * are already allocated on behalf of the file. Private mappings need + * to reserve the full area even if read-only as mprotect() may be + * called to make the mapping read-write. Assume !vma is a shm mapping + */ + if (!vma || vma->vm_flags & VM_SHARED) + chg = region_chg(&inode->i_mapping->private_list, from, to); + else { + chg = to - from; + set_vma_resv_huge_pages(vma, chg); + } + if (chg < 0) return chg; @@ -1304,7 +1383,8 @@ int hugetlb_reserve_pages(struct inode *inode, long from, long to) hugetlb_put_quota(inode->i_mapping, chg); return ret; } - region_add(&inode->i_mapping->private_list, from, to); + if (!vma || vma->vm_flags & VM_SHARED) + region_add(&inode->i_mapping->private_list, from, to); return 0; } -- cgit v1.2.3-70-g09d2