From f7853c34241807bb97673a5e97719123be39a09e Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 7 Jul 2023 16:19:09 +0200 Subject: locking/rtmutex: Fix task->pi_waiters integrity Henry reported that rt_mutex_adjust_prio_check() has an ordering problem and puts the lie to the comment in [7]. Sharing the sort key between lock->waiters and owner->pi_waiters *does* create problems, since unlike what the comment claims, holding [L] is insufficient. Notably, consider: A / \ M1 M2 | | B C That is, task A owns both M1 and M2, B and C block on them. In this case a concurrent chain walk (B & C) will modify their resp. sort keys in [7] while holding M1->wait_lock and M2->wait_lock. So holding [L] is meaningless, they're different Ls. This then gives rise to a race condition between [7] and [11], where the requeue of pi_waiters will observe an inconsistent tree order. B C (holds M1->wait_lock, (holds M2->wait_lock, holds B->pi_lock) holds A->pi_lock) [7] waiter_update_prio(); ... [8] raw_spin_unlock(B->pi_lock); ... [10] raw_spin_lock(A->pi_lock); [11] rt_mutex_enqueue_pi(); // observes inconsistent A->pi_waiters // tree order Fixing this means either extending the range of the owner lock from [10-13] to [6-13], with the immediate problem that this means [6-8] hold both blocked and owner locks, or duplicating the sort key. Since the locking in chain walk is horrible enough without having to consider pi_lock nesting rules, duplicate the sort key instead. By giving each tree their own sort key, the above race becomes harmless, if C sees B at the old location, then B will correct things (if they need correcting) when it walks up the chain and reaches A. Fixes: fb00aca47440 ("rtmutex: Turn the plist into an rb-tree") Reported-by: Henry Wu Signed-off-by: Peter Zijlstra (Intel) Acked-by: Thomas Gleixner Tested-by: Henry Wu Link: https://lkml.kernel.org/r/20230707161052.GF2883469%40hirez.programming.kicks-ass.net --- kernel/locking/rtmutex.c | 170 +++++++++++++++++++++++++++------------- kernel/locking/rtmutex_api.c | 2 +- kernel/locking/rtmutex_common.h | 47 ++++++++--- kernel/locking/ww_mutex.h | 12 +-- 4 files changed, 155 insertions(+), 76 deletions(-) (limited to 'kernel') diff --git a/kernel/locking/rtmutex.c b/kernel/locking/rtmutex.c index 728f434de2bb..21db0df0eb00 100644 --- a/kernel/locking/rtmutex.c +++ b/kernel/locking/rtmutex.c @@ -333,21 +333,43 @@ static __always_inline int __waiter_prio(struct task_struct *task) return prio; } +/* + * Update the waiter->tree copy of the sort keys. + */ static __always_inline void waiter_update_prio(struct rt_mutex_waiter *waiter, struct task_struct *task) { - waiter->prio = __waiter_prio(task); - waiter->deadline = task->dl.deadline; + lockdep_assert_held(&waiter->lock->wait_lock); + lockdep_assert(RB_EMPTY_NODE(&waiter->tree.entry)); + + waiter->tree.prio = __waiter_prio(task); + waiter->tree.deadline = task->dl.deadline; +} + +/* + * Update the waiter->pi_tree copy of the sort keys (from the tree copy). + */ +static __always_inline void +waiter_clone_prio(struct rt_mutex_waiter *waiter, struct task_struct *task) +{ + lockdep_assert_held(&waiter->lock->wait_lock); + lockdep_assert_held(&task->pi_lock); + lockdep_assert(RB_EMPTY_NODE(&waiter->pi_tree.entry)); + + waiter->pi_tree.prio = waiter->tree.prio; + waiter->pi_tree.deadline = waiter->tree.deadline; } /* - * Only use with rt_mutex_waiter_{less,equal}() + * Only use with rt_waiter_node_{less,equal}() */ +#define task_to_waiter_node(p) \ + &(struct rt_waiter_node){ .prio = __waiter_prio(p), .deadline = (p)->dl.deadline } #define task_to_waiter(p) \ - &(struct rt_mutex_waiter){ .prio = __waiter_prio(p), .deadline = (p)->dl.deadline } + &(struct rt_mutex_waiter){ .tree = *task_to_waiter_node(p) } -static __always_inline int rt_mutex_waiter_less(struct rt_mutex_waiter *left, - struct rt_mutex_waiter *right) +static __always_inline int rt_waiter_node_less(struct rt_waiter_node *left, + struct rt_waiter_node *right) { if (left->prio < right->prio) return 1; @@ -364,8 +386,8 @@ static __always_inline int rt_mutex_waiter_less(struct rt_mutex_waiter *left, return 0; } -static __always_inline int rt_mutex_waiter_equal(struct rt_mutex_waiter *left, - struct rt_mutex_waiter *right) +static __always_inline int rt_waiter_node_equal(struct rt_waiter_node *left, + struct rt_waiter_node *right) { if (left->prio != right->prio) return 0; @@ -385,7 +407,7 @@ static __always_inline int rt_mutex_waiter_equal(struct rt_mutex_waiter *left, static inline bool rt_mutex_steal(struct rt_mutex_waiter *waiter, struct rt_mutex_waiter *top_waiter) { - if (rt_mutex_waiter_less(waiter, top_waiter)) + if (rt_waiter_node_less(&waiter->tree, &top_waiter->tree)) return true; #ifdef RT_MUTEX_BUILD_SPINLOCKS @@ -393,30 +415,30 @@ static inline bool rt_mutex_steal(struct rt_mutex_waiter *waiter, * Note that RT tasks are excluded from same priority (lateral) * steals to prevent the introduction of an unbounded latency. */ - if (rt_prio(waiter->prio) || dl_prio(waiter->prio)) + if (rt_prio(waiter->tree.prio) || dl_prio(waiter->tree.prio)) return false; - return rt_mutex_waiter_equal(waiter, top_waiter); + return rt_waiter_node_equal(&waiter->tree, &top_waiter->tree); #else return false; #endif } #define __node_2_waiter(node) \ - rb_entry((node), struct rt_mutex_waiter, tree_entry) + rb_entry((node), struct rt_mutex_waiter, tree.entry) static __always_inline bool __waiter_less(struct rb_node *a, const struct rb_node *b) { struct rt_mutex_waiter *aw = __node_2_waiter(a); struct rt_mutex_waiter *bw = __node_2_waiter(b); - if (rt_mutex_waiter_less(aw, bw)) + if (rt_waiter_node_less(&aw->tree, &bw->tree)) return 1; if (!build_ww_mutex()) return 0; - if (rt_mutex_waiter_less(bw, aw)) + if (rt_waiter_node_less(&bw->tree, &aw->tree)) return 0; /* NOTE: relies on waiter->ww_ctx being set before insertion */ @@ -434,48 +456,58 @@ static __always_inline bool __waiter_less(struct rb_node *a, const struct rb_nod static __always_inline void rt_mutex_enqueue(struct rt_mutex_base *lock, struct rt_mutex_waiter *waiter) { - rb_add_cached(&waiter->tree_entry, &lock->waiters, __waiter_less); + lockdep_assert_held(&lock->wait_lock); + + rb_add_cached(&waiter->tree.entry, &lock->waiters, __waiter_less); } static __always_inline void rt_mutex_dequeue(struct rt_mutex_base *lock, struct rt_mutex_waiter *waiter) { - if (RB_EMPTY_NODE(&waiter->tree_entry)) + lockdep_assert_held(&lock->wait_lock); + + if (RB_EMPTY_NODE(&waiter->tree.entry)) return; - rb_erase_cached(&waiter->tree_entry, &lock->waiters); - RB_CLEAR_NODE(&waiter->tree_entry); + rb_erase_cached(&waiter->tree.entry, &lock->waiters); + RB_CLEAR_NODE(&waiter->tree.entry); } -#define __node_2_pi_waiter(node) \ - rb_entry((node), struct rt_mutex_waiter, pi_tree_entry) +#define __node_2_rt_node(node) \ + rb_entry((node), struct rt_waiter_node, entry) -static __always_inline bool -__pi_waiter_less(struct rb_node *a, const struct rb_node *b) +static __always_inline bool __pi_waiter_less(struct rb_node *a, const struct rb_node *b) { - return rt_mutex_waiter_less(__node_2_pi_waiter(a), __node_2_pi_waiter(b)); + return rt_waiter_node_less(__node_2_rt_node(a), __node_2_rt_node(b)); } static __always_inline void rt_mutex_enqueue_pi(struct task_struct *task, struct rt_mutex_waiter *waiter) { - rb_add_cached(&waiter->pi_tree_entry, &task->pi_waiters, __pi_waiter_less); + lockdep_assert_held(&task->pi_lock); + + rb_add_cached(&waiter->pi_tree.entry, &task->pi_waiters, __pi_waiter_less); } static __always_inline void rt_mutex_dequeue_pi(struct task_struct *task, struct rt_mutex_waiter *waiter) { - if (RB_EMPTY_NODE(&waiter->pi_tree_entry)) + lockdep_assert_held(&task->pi_lock); + + if (RB_EMPTY_NODE(&waiter->pi_tree.entry)) return; - rb_erase_cached(&waiter->pi_tree_entry, &task->pi_waiters); - RB_CLEAR_NODE(&waiter->pi_tree_entry); + rb_erase_cached(&waiter->pi_tree.entry, &task->pi_waiters); + RB_CLEAR_NODE(&waiter->pi_tree.entry); } -static __always_inline void rt_mutex_adjust_prio(struct task_struct *p) +static __always_inline void rt_mutex_adjust_prio(struct rt_mutex_base *lock, + struct task_struct *p) { struct task_struct *pi_task = NULL; + lockdep_assert_held(&lock->wait_lock); + lockdep_assert(rt_mutex_owner(lock) == p); lockdep_assert_held(&p->pi_lock); if (task_has_pi_waiters(p)) @@ -571,9 +603,14 @@ static __always_inline struct rt_mutex_base *task_blocked_on_lock(struct task_st * Chain walk basics and protection scope * * [R] refcount on task - * [P] task->pi_lock held + * [Pn] task->pi_lock held * [L] rtmutex->wait_lock held * + * Normal locking order: + * + * rtmutex->wait_lock + * task->pi_lock + * * Step Description Protected by * function arguments: * @task [R] @@ -588,27 +625,32 @@ static __always_inline struct rt_mutex_base *task_blocked_on_lock(struct task_st * again: * loop_sanity_check(); * retry: - * [1] lock(task->pi_lock); [R] acquire [P] - * [2] waiter = task->pi_blocked_on; [P] - * [3] check_exit_conditions_1(); [P] - * [4] lock = waiter->lock; [P] - * [5] if (!try_lock(lock->wait_lock)) { [P] try to acquire [L] - * unlock(task->pi_lock); release [P] + * [1] lock(task->pi_lock); [R] acquire [P1] + * [2] waiter = task->pi_blocked_on; [P1] + * [3] check_exit_conditions_1(); [P1] + * [4] lock = waiter->lock; [P1] + * [5] if (!try_lock(lock->wait_lock)) { [P1] try to acquire [L] + * unlock(task->pi_lock); release [P1] * goto retry; * } - * [6] check_exit_conditions_2(); [P] + [L] - * [7] requeue_lock_waiter(lock, waiter); [P] + [L] - * [8] unlock(task->pi_lock); release [P] + * [6] check_exit_conditions_2(); [P1] + [L] + * [7] requeue_lock_waiter(lock, waiter); [P1] + [L] + * [8] unlock(task->pi_lock); release [P1] * put_task_struct(task); release [R] * [9] check_exit_conditions_3(); [L] * [10] task = owner(lock); [L] * get_task_struct(task); [L] acquire [R] - * lock(task->pi_lock); [L] acquire [P] - * [11] requeue_pi_waiter(tsk, waiters(lock));[P] + [L] - * [12] check_exit_conditions_4(); [P] + [L] - * [13] unlock(task->pi_lock); release [P] + * lock(task->pi_lock); [L] acquire [P2] + * [11] requeue_pi_waiter(tsk, waiters(lock));[P2] + [L] + * [12] check_exit_conditions_4(); [P2] + [L] + * [13] unlock(task->pi_lock); release [P2] * unlock(lock->wait_lock); release [L] * goto again; + * + * Where P1 is the blocking task and P2 is the lock owner; going up one step + * the owner becomes the next blocked task etc.. + * +* */ static int __sched rt_mutex_adjust_prio_chain(struct task_struct *task, enum rtmutex_chainwalk chwalk, @@ -756,7 +798,7 @@ static int __sched rt_mutex_adjust_prio_chain(struct task_struct *task, * enabled we continue, but stop the requeueing in the chain * walk. */ - if (rt_mutex_waiter_equal(waiter, task_to_waiter(task))) { + if (rt_waiter_node_equal(&waiter->tree, task_to_waiter_node(task))) { if (!detect_deadlock) goto out_unlock_pi; else @@ -764,13 +806,18 @@ static int __sched rt_mutex_adjust_prio_chain(struct task_struct *task, } /* - * [4] Get the next lock + * [4] Get the next lock; per holding task->pi_lock we can't unblock + * and guarantee @lock's existence. */ lock = waiter->lock; /* * [5] We need to trylock here as we are holding task->pi_lock, * which is the reverse lock order versus the other rtmutex * operations. + * + * Per the above, holding task->pi_lock guarantees lock exists, so + * inverting this lock order is infeasible from a life-time + * perspective. */ if (!raw_spin_trylock(&lock->wait_lock)) { raw_spin_unlock_irq(&task->pi_lock); @@ -874,17 +921,18 @@ static int __sched rt_mutex_adjust_prio_chain(struct task_struct *task, * or * * DL CBS enforcement advancing the effective deadline. - * - * Even though pi_waiters also uses these fields, and that tree is only - * updated in [11], we can do this here, since we hold [L], which - * serializes all pi_waiters access and rb_erase() does not care about - * the values of the node being removed. */ waiter_update_prio(waiter, task); rt_mutex_enqueue(lock, waiter); - /* [8] Release the task */ + /* + * [8] Release the (blocking) task in preparation for + * taking the owner task in [10]. + * + * Since we hold lock->waiter_lock, task cannot unblock, even if we + * release task->pi_lock. + */ raw_spin_unlock(&task->pi_lock); put_task_struct(task); @@ -908,7 +956,12 @@ static int __sched rt_mutex_adjust_prio_chain(struct task_struct *task, return 0; } - /* [10] Grab the next task, i.e. the owner of @lock */ + /* + * [10] Grab the next task, i.e. the owner of @lock + * + * Per holding lock->wait_lock and checking for !owner above, there + * must be an owner and it cannot go away. + */ task = get_task_struct(rt_mutex_owner(lock)); raw_spin_lock(&task->pi_lock); @@ -921,8 +974,9 @@ static int __sched rt_mutex_adjust_prio_chain(struct task_struct *task, * and adjust the priority of the owner. */ rt_mutex_dequeue_pi(task, prerequeue_top_waiter); + waiter_clone_prio(waiter, task); rt_mutex_enqueue_pi(task, waiter); - rt_mutex_adjust_prio(task); + rt_mutex_adjust_prio(lock, task); } else if (prerequeue_top_waiter == waiter) { /* @@ -937,8 +991,9 @@ static int __sched rt_mutex_adjust_prio_chain(struct task_struct *task, */ rt_mutex_dequeue_pi(task, waiter); waiter = rt_mutex_top_waiter(lock); + waiter_clone_prio(waiter, task); rt_mutex_enqueue_pi(task, waiter); - rt_mutex_adjust_prio(task); + rt_mutex_adjust_prio(lock, task); } else { /* * Nothing changed. No need to do any priority @@ -1154,6 +1209,7 @@ static int __sched task_blocks_on_rt_mutex(struct rt_mutex_base *lock, waiter->task = task; waiter->lock = lock; waiter_update_prio(waiter, task); + waiter_clone_prio(waiter, task); /* Get the top priority waiter on the lock */ if (rt_mutex_has_waiters(lock)) @@ -1187,7 +1243,7 @@ static int __sched task_blocks_on_rt_mutex(struct rt_mutex_base *lock, rt_mutex_dequeue_pi(owner, top_waiter); rt_mutex_enqueue_pi(owner, waiter); - rt_mutex_adjust_prio(owner); + rt_mutex_adjust_prio(lock, owner); if (owner->pi_blocked_on) chain_walk = 1; } else if (rt_mutex_cond_detect_deadlock(waiter, chwalk)) { @@ -1234,6 +1290,8 @@ static void __sched mark_wakeup_next_waiter(struct rt_wake_q_head *wqh, { struct rt_mutex_waiter *waiter; + lockdep_assert_held(&lock->wait_lock); + raw_spin_lock(¤t->pi_lock); waiter = rt_mutex_top_waiter(lock); @@ -1246,7 +1304,7 @@ static void __sched mark_wakeup_next_waiter(struct rt_wake_q_head *wqh, * task unblocks. */ rt_mutex_dequeue_pi(current, waiter); - rt_mutex_adjust_prio(current); + rt_mutex_adjust_prio(lock, current); /* * As we are waking up the top waiter, and the waiter stays @@ -1482,7 +1540,7 @@ static void __sched remove_waiter(struct rt_mutex_base *lock, if (rt_mutex_has_waiters(lock)) rt_mutex_enqueue_pi(owner, rt_mutex_top_waiter(lock)); - rt_mutex_adjust_prio(owner); + rt_mutex_adjust_prio(lock, owner); /* Store the lock on which owner is blocked or NULL */ next_lock = task_blocked_on_lock(owner); diff --git a/kernel/locking/rtmutex_api.c b/kernel/locking/rtmutex_api.c index cb9fdff76a8a..a6974d044593 100644 --- a/kernel/locking/rtmutex_api.c +++ b/kernel/locking/rtmutex_api.c @@ -459,7 +459,7 @@ void __sched rt_mutex_adjust_pi(struct task_struct *task) raw_spin_lock_irqsave(&task->pi_lock, flags); waiter = task->pi_blocked_on; - if (!waiter || rt_mutex_waiter_equal(waiter, task_to_waiter(task))) { + if (!waiter || rt_waiter_node_equal(&waiter->tree, task_to_waiter_node(task))) { raw_spin_unlock_irqrestore(&task->pi_lock, flags); return; } diff --git a/kernel/locking/rtmutex_common.h b/kernel/locking/rtmutex_common.h index c47e8361bfb5..1162e07cdaea 100644 --- a/kernel/locking/rtmutex_common.h +++ b/kernel/locking/rtmutex_common.h @@ -17,27 +17,44 @@ #include #include + +/* + * This is a helper for the struct rt_mutex_waiter below. A waiter goes in two + * separate trees and they need their own copy of the sort keys because of + * different locking requirements. + * + * @entry: rbtree node to enqueue into the waiters tree + * @prio: Priority of the waiter + * @deadline: Deadline of the waiter if applicable + * + * See rt_waiter_node_less() and waiter_*_prio(). + */ +struct rt_waiter_node { + struct rb_node entry; + int prio; + u64 deadline; +}; + /* * This is the control structure for tasks blocked on a rt_mutex, * which is allocated on the kernel stack on of the blocked task. * - * @tree_entry: pi node to enqueue into the mutex waiters tree - * @pi_tree_entry: pi node to enqueue into the mutex owner waiters tree + * @tree: node to enqueue into the mutex waiters tree + * @pi_tree: node to enqueue into the mutex owner waiters tree * @task: task reference to the blocked task * @lock: Pointer to the rt_mutex on which the waiter blocks * @wake_state: Wakeup state to use (TASK_NORMAL or TASK_RTLOCK_WAIT) - * @prio: Priority of the waiter - * @deadline: Deadline of the waiter if applicable * @ww_ctx: WW context pointer + * + * @tree is ordered by @lock->wait_lock + * @pi_tree is ordered by rt_mutex_owner(@lock)->pi_lock */ struct rt_mutex_waiter { - struct rb_node tree_entry; - struct rb_node pi_tree_entry; + struct rt_waiter_node tree; + struct rt_waiter_node pi_tree; struct task_struct *task; struct rt_mutex_base *lock; unsigned int wake_state; - int prio; - u64 deadline; struct ww_acquire_ctx *ww_ctx; }; @@ -105,7 +122,7 @@ static inline bool rt_mutex_waiter_is_top_waiter(struct rt_mutex_base *lock, { struct rb_node *leftmost = rb_first_cached(&lock->waiters); - return rb_entry(leftmost, struct rt_mutex_waiter, tree_entry) == waiter; + return rb_entry(leftmost, struct rt_mutex_waiter, tree.entry) == waiter; } static inline struct rt_mutex_waiter *rt_mutex_top_waiter(struct rt_mutex_base *lock) @@ -113,8 +130,10 @@ static inline struct rt_mutex_waiter *rt_mutex_top_waiter(struct rt_mutex_base * struct rb_node *leftmost = rb_first_cached(&lock->waiters); struct rt_mutex_waiter *w = NULL; + lockdep_assert_held(&lock->wait_lock); + if (leftmost) { - w = rb_entry(leftmost, struct rt_mutex_waiter, tree_entry); + w = rb_entry(leftmost, struct rt_mutex_waiter, tree.entry); BUG_ON(w->lock != lock); } return w; @@ -127,8 +146,10 @@ static inline int task_has_pi_waiters(struct task_struct *p) static inline struct rt_mutex_waiter *task_top_pi_waiter(struct task_struct *p) { + lockdep_assert_held(&p->pi_lock); + return rb_entry(p->pi_waiters.rb_leftmost, struct rt_mutex_waiter, - pi_tree_entry); + pi_tree.entry); } #define RT_MUTEX_HAS_WAITERS 1UL @@ -190,8 +211,8 @@ static inline void debug_rt_mutex_free_waiter(struct rt_mutex_waiter *waiter) static inline void rt_mutex_init_waiter(struct rt_mutex_waiter *waiter) { debug_rt_mutex_init_waiter(waiter); - RB_CLEAR_NODE(&waiter->pi_tree_entry); - RB_CLEAR_NODE(&waiter->tree_entry); + RB_CLEAR_NODE(&waiter->pi_tree.entry); + RB_CLEAR_NODE(&waiter->tree.entry); waiter->wake_state = TASK_NORMAL; waiter->task = NULL; } diff --git a/kernel/locking/ww_mutex.h b/kernel/locking/ww_mutex.h index 56f139201f24..3ad2cc4823e5 100644 --- a/kernel/locking/ww_mutex.h +++ b/kernel/locking/ww_mutex.h @@ -96,25 +96,25 @@ __ww_waiter_first(struct rt_mutex *lock) struct rb_node *n = rb_first(&lock->rtmutex.waiters.rb_root); if (!n) return NULL; - return rb_entry(n, struct rt_mutex_waiter, tree_entry); + return rb_entry(n, struct rt_mutex_waiter, tree.entry); } static inline struct rt_mutex_waiter * __ww_waiter_next(struct rt_mutex *lock, struct rt_mutex_waiter *w) { - struct rb_node *n = rb_next(&w->tree_entry); + struct rb_node *n = rb_next(&w->tree.entry); if (!n) return NULL; - return rb_entry(n, struct rt_mutex_waiter, tree_entry); + return rb_entry(n, struct rt_mutex_waiter, tree.entry); } static inline struct rt_mutex_waiter * __ww_waiter_prev(struct rt_mutex *lock, struct rt_mutex_waiter *w) { - struct rb_node *n = rb_prev(&w->tree_entry); + struct rb_node *n = rb_prev(&w->tree.entry); if (!n) return NULL; - return rb_entry(n, struct rt_mutex_waiter, tree_entry); + return rb_entry(n, struct rt_mutex_waiter, tree.entry); } static inline struct rt_mutex_waiter * @@ -123,7 +123,7 @@ __ww_waiter_last(struct rt_mutex *lock) struct rb_node *n = rb_last(&lock->rtmutex.waiters.rb_root); if (!n) return NULL; - return rb_entry(n, struct rt_mutex_waiter, tree_entry); + return rb_entry(n, struct rt_mutex_waiter, tree.entry); } static inline void -- cgit From aa6fde93f3a49e42c0fe0490d7f3711bac0d162e Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 17 Jul 2023 12:50:02 -1000 Subject: workqueue: Scale up wq_cpu_intensive_thresh_us if BogoMIPS is below 4000 wq_cpu_intensive_thresh_us is used to detect CPU-hogging per-cpu work items. Once detected, they're excluded from concurrency management to prevent them from blocking other per-cpu work items. If CONFIG_WQ_CPU_INTENSIVE_REPORT is enabled, repeat offenders are also reported so that the code can be updated. The default threshold is 10ms which is long enough to do fair bit of work on modern CPUs while short enough to be usually not noticeable. This unfortunately leads to a lot of, arguable spurious, detections on very slow CPUs. Using the same threshold across CPUs whose performance levels may be apart by multiple levels of magnitude doesn't make whole lot of sense. This patch scales up wq_cpu_intensive_thresh_us upto 1 second when BogoMIPS is below 4000. This is obviously very inaccurate but it doesn't have to be accurate to be useful. The mechanism is still useful when the threshold is fully scaled up and the benefits of reports are usually shared with everyone regardless of who's reporting, so as long as there are sufficient number of fast machines reporting, we don't lose much. Some (or is it all?) ARM CPUs systemtically report significantly lower BogoMIPS. While this doesn't break anything, given how widespread ARM CPUs are, it's at least a missed opportunity and it probably would be a good idea to teach workqueue about it. Signed-off-by: Tejun Heo Reported-and-Tested-by: Geert Uytterhoeven --- kernel/workqueue.c | 43 ++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 42 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 02a8f402eeb5..800b4208dba9 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -52,6 +52,7 @@ #include #include #include +#include #include "workqueue_internal.h" @@ -338,8 +339,10 @@ static cpumask_var_t *wq_numa_possible_cpumask; * Per-cpu work items which run for longer than the following threshold are * automatically considered CPU intensive and excluded from concurrency * management to prevent them from noticeably delaying other per-cpu work items. + * ULONG_MAX indicates that the user hasn't overridden it with a boot parameter. + * The actual value is initialized in wq_cpu_intensive_thresh_init(). */ -static unsigned long wq_cpu_intensive_thresh_us = 10000; +static unsigned long wq_cpu_intensive_thresh_us = ULONG_MAX; module_param_named(cpu_intensive_thresh_us, wq_cpu_intensive_thresh_us, ulong, 0644); static bool wq_disable_numa; @@ -6513,6 +6516,42 @@ void __init workqueue_init_early(void) !system_freezable_power_efficient_wq); } +static void __init wq_cpu_intensive_thresh_init(void) +{ + unsigned long thresh; + unsigned long bogo; + + /* if the user set it to a specific value, keep it */ + if (wq_cpu_intensive_thresh_us != ULONG_MAX) + return; + + /* + * The default of 10ms is derived from the fact that most modern (as of + * 2023) processors can do a lot in 10ms and that it's just below what + * most consider human-perceivable. However, the kernel also runs on a + * lot slower CPUs including microcontrollers where the threshold is way + * too low. + * + * Let's scale up the threshold upto 1 second if BogoMips is below 4000. + * This is by no means accurate but it doesn't have to be. The mechanism + * is still useful even when the threshold is fully scaled up. Also, as + * the reports would usually be applicable to everyone, some machines + * operating on longer thresholds won't significantly diminish their + * usefulness. + */ + thresh = 10 * USEC_PER_MSEC; + + /* see init/calibrate.c for lpj -> BogoMIPS calculation */ + bogo = max_t(unsigned long, loops_per_jiffy / 500000 * HZ, 1); + if (bogo < 4000) + thresh = min_t(unsigned long, thresh * 4000 / bogo, USEC_PER_SEC); + + pr_debug("wq_cpu_intensive_thresh: lpj=%lu BogoMIPS=%lu thresh_us=%lu\n", + loops_per_jiffy, bogo, thresh); + + wq_cpu_intensive_thresh_us = thresh; +} + /** * workqueue_init - bring workqueue subsystem fully online * @@ -6528,6 +6567,8 @@ void __init workqueue_init(void) struct worker_pool *pool; int cpu, bkt; + wq_cpu_intensive_thresh_init(); + /* * It'd be simpler to initialize NUMA in workqueue_init_early() but * CPU to node mapping may not be available that early on some -- cgit From f2c67a3e60d1071b65848efaa8c3b66c363dd025 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Tue, 25 Jul 2023 10:42:05 +0200 Subject: bpf: Disable preemption in bpf_perf_event_output The nesting protection in bpf_perf_event_output relies on disabled preemption, which is guaranteed for kprobes and tracepoints. However bpf_perf_event_output can be also called from uprobes context through bpf_prog_run_array_sleepable function which disables migration, but keeps preemption enabled. This can cause task to be preempted by another one inside the nesting protection and lead eventually to two tasks using same perf_sample_data buffer and cause crashes like: kernel tried to execute NX-protected page - exploit attempt? (uid: 0) BUG: unable to handle page fault for address: ffffffff82be3eea ... Call Trace: ? __die+0x1f/0x70 ? page_fault_oops+0x176/0x4d0 ? exc_page_fault+0x132/0x230 ? asm_exc_page_fault+0x22/0x30 ? perf_output_sample+0x12b/0x910 ? perf_event_output+0xd0/0x1d0 ? bpf_perf_event_output+0x162/0x1d0 ? bpf_prog_c6271286d9a4c938_krava1+0x76/0x87 ? __uprobe_perf_func+0x12b/0x540 ? uprobe_dispatcher+0x2c4/0x430 ? uprobe_notify_resume+0x2da/0xce0 ? atomic_notifier_call_chain+0x7b/0x110 ? exit_to_user_mode_prepare+0x13e/0x290 ? irqentry_exit_to_user_mode+0x5/0x30 ? asm_exc_int3+0x35/0x40 Fixing this by disabling preemption in bpf_perf_event_output. Cc: stable@vger.kernel.org Fixes: 8c7dcb84e3b7 ("bpf: implement sleepable uprobes by chaining gps") Acked-by: Hou Tao Signed-off-by: Jiri Olsa Link: https://lore.kernel.org/r/20230725084206.580930-2-jolsa@kernel.org Signed-off-by: Alexei Starovoitov --- kernel/trace/bpf_trace.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c index 5f2dcabad202..bf18a53731b0 100644 --- a/kernel/trace/bpf_trace.c +++ b/kernel/trace/bpf_trace.c @@ -661,8 +661,7 @@ static DEFINE_PER_CPU(int, bpf_trace_nest_level); BPF_CALL_5(bpf_perf_event_output, struct pt_regs *, regs, struct bpf_map *, map, u64, flags, void *, data, u64, size) { - struct bpf_trace_sample_data *sds = this_cpu_ptr(&bpf_trace_sds); - int nest_level = this_cpu_inc_return(bpf_trace_nest_level); + struct bpf_trace_sample_data *sds; struct perf_raw_record raw = { .frag = { .size = size, @@ -670,7 +669,11 @@ BPF_CALL_5(bpf_perf_event_output, struct pt_regs *, regs, struct bpf_map *, map, }, }; struct perf_sample_data *sd; - int err; + int nest_level, err; + + preempt_disable(); + sds = this_cpu_ptr(&bpf_trace_sds); + nest_level = this_cpu_inc_return(bpf_trace_nest_level); if (WARN_ON_ONCE(nest_level > ARRAY_SIZE(sds->sds))) { err = -EBUSY; @@ -688,9 +691,9 @@ BPF_CALL_5(bpf_perf_event_output, struct pt_regs *, regs, struct bpf_map *, map, perf_sample_save_raw_data(sd, &raw); err = __bpf_perf_event_output(regs, map, flags, sd); - out: this_cpu_dec(bpf_trace_nest_level); + preempt_enable(); return err; } -- cgit From d62cc390c2e99ae267ffe4b8d7e2e08b6c758c32 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Tue, 25 Jul 2023 10:42:06 +0200 Subject: bpf: Disable preemption in bpf_event_output We received report [1] of kernel crash, which is caused by using nesting protection without disabled preemption. The bpf_event_output can be called by programs executed by bpf_prog_run_array_cg function that disabled migration but keeps preemption enabled. This can cause task to be preempted by another one inside the nesting protection and lead eventually to two tasks using same perf_sample_data buffer and cause crashes like: BUG: kernel NULL pointer dereference, address: 0000000000000001 #PF: supervisor instruction fetch in kernel mode #PF: error_code(0x0010) - not-present page ... ? perf_output_sample+0x12a/0x9a0 ? finish_task_switch.isra.0+0x81/0x280 ? perf_event_output+0x66/0xa0 ? bpf_event_output+0x13a/0x190 ? bpf_event_output_data+0x22/0x40 ? bpf_prog_dfc84bbde731b257_cil_sock4_connect+0x40a/0xacb ? xa_load+0x87/0xe0 ? __cgroup_bpf_run_filter_sock_addr+0xc1/0x1a0 ? release_sock+0x3e/0x90 ? sk_setsockopt+0x1a1/0x12f0 ? udp_pre_connect+0x36/0x50 ? inet_dgram_connect+0x93/0xa0 ? __sys_connect+0xb4/0xe0 ? udp_setsockopt+0x27/0x40 ? __pfx_udp_push_pending_frames+0x10/0x10 ? __sys_setsockopt+0xdf/0x1a0 ? __x64_sys_connect+0xf/0x20 ? do_syscall_64+0x3a/0x90 ? entry_SYSCALL_64_after_hwframe+0x72/0xdc Fixing this by disabling preemption in bpf_event_output. [1] https://github.com/cilium/cilium/issues/26756 Cc: stable@vger.kernel.org Reported-by: Oleg "livelace" Popov Closes: https://github.com/cilium/cilium/issues/26756 Fixes: 2a916f2f546c ("bpf: Use migrate_disable/enable in array macros and cgroup/lirc code.") Acked-by: Hou Tao Signed-off-by: Jiri Olsa Link: https://lore.kernel.org/r/20230725084206.580930-3-jolsa@kernel.org Signed-off-by: Alexei Starovoitov --- kernel/trace/bpf_trace.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c index bf18a53731b0..bd1a42b23f3f 100644 --- a/kernel/trace/bpf_trace.c +++ b/kernel/trace/bpf_trace.c @@ -718,7 +718,6 @@ static DEFINE_PER_CPU(struct bpf_trace_sample_data, bpf_misc_sds); u64 bpf_event_output(struct bpf_map *map, u64 flags, void *meta, u64 meta_size, void *ctx, u64 ctx_size, bpf_ctx_copy_t ctx_copy) { - int nest_level = this_cpu_inc_return(bpf_event_output_nest_level); struct perf_raw_frag frag = { .copy = ctx_copy, .size = ctx_size, @@ -735,8 +734,12 @@ u64 bpf_event_output(struct bpf_map *map, u64 flags, void *meta, u64 meta_size, }; struct perf_sample_data *sd; struct pt_regs *regs; + int nest_level; u64 ret; + preempt_disable(); + nest_level = this_cpu_inc_return(bpf_event_output_nest_level); + if (WARN_ON_ONCE(nest_level > ARRAY_SIZE(bpf_misc_sds.sds))) { ret = -EBUSY; goto out; @@ -751,6 +754,7 @@ u64 bpf_event_output(struct bpf_map *map, u64 flags, void *meta, u64 meta_size, ret = __bpf_perf_event_output(regs, map, flags, sd); out: this_cpu_dec(bpf_event_output_nest_level); + preempt_enable(); return ret; } -- cgit From 1f9f4f4777e7958e5c1fbdfd9ddf4207dc00a40f Mon Sep 17 00:00:00 2001 From: "Masami Hiramatsu (Google)" Date: Tue, 25 Jul 2023 10:37:19 +0900 Subject: tracing/probes: Fix to add NULL check for BTF APIs Since find_btf_func_param() abd btf_type_by_id() can return NULL, the caller must check the return value correctly. Link: https://lore.kernel.org/all/169024903951.395371.11361556840733470934.stgit@devnote2/ Fixes: b576e09701c7 ("tracing/probes: Support function parameters if BTF is available") Signed-off-by: Masami Hiramatsu (Google) Reviewed-by: Steven Rostedt (Google) --- kernel/trace/trace_probe.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_probe.c b/kernel/trace/trace_probe.c index b2b726bea1f9..c68a72707852 100644 --- a/kernel/trace/trace_probe.c +++ b/kernel/trace/trace_probe.c @@ -386,12 +386,12 @@ static const struct btf_type *find_btf_func_proto(const char *funcname) /* Get BTF_KIND_FUNC type */ t = btf_type_by_id(btf, id); - if (!btf_type_is_func(t)) + if (!t || !btf_type_is_func(t)) return ERR_PTR(-ENOENT); /* The type of BTF_KIND_FUNC is BTF_KIND_FUNC_PROTO */ t = btf_type_by_id(btf, t->type); - if (!btf_type_is_func_proto(t)) + if (!t || !btf_type_is_func_proto(t)) return ERR_PTR(-ENOENT); return t; @@ -443,7 +443,7 @@ static int parse_btf_arg(const char *varname, struct fetch_insn *code, if (!ctx->params) { params = find_btf_func_param(ctx->funcname, &ctx->nr_params, ctx->flags & TPARG_FL_TPOINT); - if (IS_ERR(params)) { + if (IS_ERR_OR_NULL(params)) { trace_probe_log_err(ctx->offset, NO_BTF_ENTRY); return PTR_ERR(params); } @@ -1273,7 +1273,7 @@ const char **traceprobe_expand_meta_args(int argc, const char *argv[], params = find_btf_func_param(ctx->funcname, &nr_params, ctx->flags & TPARG_FL_TPOINT); - if (IS_ERR(params)) { + if (IS_ERR_OR_NULL(params)) { if (args_idx != -1) { /* $arg* requires BTF info */ trace_probe_log_err(0, NOSUP_BTFARG); -- cgit From 5f0bc0b042fc77ff70e14c790abdec960cde4ec1 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Tue, 25 Jul 2023 09:38:32 -0700 Subject: mm: suppress mm fault logging if fatal signal already pending Commit eda0047296a1 ("mm: make the page fault mmap locking killable") intentionally made it much easier to trigger the "page fault fails because a fatal signal is pending" situation, by having the mmap locking fail early in that case. We have long aborted page faults in other fatal cases when the actual IO for a page is interrupted by SIGKILL - which is particularly useful for the traditional case of NFS hanging due to network issues, but local filesystems could cause it too if you happened to get the SIGKILL while waiting for a page to be faulted in (eg lock_folio_maybe_drop_mmap()). So aborting the page fault wasn't a new condition - but it now triggers earlier, before we even get to 'handle_mm_fault()'. And as a result the error doesn't go through our 'fault_signal_pending()' logic, and doesn't get filtered away there. Normally you'd never even notice, because if a fatal signal is pending, the new SIGSEGV we send ends up being ignored anyway. But it turns out that there is one very noticeable exception: if you enable 'show_unhandled_signals', the aborted page fault will be logged in the kernel messages, and you'll get a scary line looking something like this in your logs: pverados[2183248]: segfault at 55e5a00f9ae0 ip 000055e5a00f9ae0 sp 00007ffc0720bea8 error 14 in perl[55e5a00d4000+195000] likely on CPU 10 (core 4, socket 0) which is rather misleading. It's not really a segfault at all, it's just "the thread was killed before the page fault completed, so we aborted the page fault". Fix this by just making it clear that a pending fatal signal means that any new signal coming in after that is implicitly handled. This will avoid the misleading logging, since now the signal isn't 'unhandled' any more. Reported-and-tested-by: Fiona Ebner Tested-by: Thomas Lamprecht Link: https://lore.kernel.org/lkml/8d063a26-43f5-0bb7-3203-c6a04dc159f8@proxmox.com/ Acked-by: Oleg Nesterov Fixes: eda0047296a1 ("mm: make the page fault mmap locking killable") Signed-off-by: Linus Torvalds --- kernel/signal.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'kernel') diff --git a/kernel/signal.c b/kernel/signal.c index b5370fe5c198..128e9bb3d1a2 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -562,6 +562,10 @@ bool unhandled_signal(struct task_struct *tsk, int sig) if (handler != SIG_IGN && handler != SIG_DFL) return false; + /* If dying, we handle all new signals by ignoring them */ + if (fatal_signal_pending(tsk)) + return false; + /* if ptraced, let the tracer determine */ return !tsk->ptrace; } -- cgit From 2d093282b0d4357373497f65db6a05eb0c28b7c8 Mon Sep 17 00:00:00 2001 From: Zheng Yejian Date: Mon, 24 Jul 2023 13:40:40 +0800 Subject: ring-buffer: Fix wrong stat of cpu_buffer->read When pages are removed in rb_remove_pages(), 'cpu_buffer->read' is set to 0 in order to make sure any read iterators reset themselves. However, this will mess 'entries' stating, see following steps: # cd /sys/kernel/tracing/ # 1. Enlarge ring buffer prepare for later reducing: # echo 20 > per_cpu/cpu0/buffer_size_kb # 2. Write a log into ring buffer of cpu0: # taskset -c 0 echo "hello1" > trace_marker # 3. Read the log: # cat per_cpu/cpu0/trace_pipe <...>-332 [000] ..... 62.406844: tracing_mark_write: hello1 # 4. Stop reading and see the stats, now 0 entries, and 1 event readed: # cat per_cpu/cpu0/stats entries: 0 [...] read events: 1 # 5. Reduce the ring buffer # echo 7 > per_cpu/cpu0/buffer_size_kb # 6. Now entries became unexpected 1 because actually no entries!!! # cat per_cpu/cpu0/stats entries: 1 [...] read events: 0 To fix it, introduce 'page_removed' field to count total removed pages since last reset, then use it to let read iterators reset themselves instead of changing the 'read' pointer. Link: https://lore.kernel.org/linux-trace-kernel/20230724054040.3489499-1-zhengyejian1@huawei.com Cc: Cc: Fixes: 83f40318dab0 ("ring-buffer: Make removal of ring buffer pages atomic") Signed-off-by: Zheng Yejian Signed-off-by: Steven Rostedt (Google) --- kernel/trace/ring_buffer.c | 22 ++++++++++++---------- 1 file changed, 12 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index de061dd47313..46b4a3c7c3bf 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -523,6 +523,8 @@ struct ring_buffer_per_cpu { rb_time_t before_stamp; u64 event_stamp[MAX_NEST]; u64 read_stamp; + /* pages removed since last reset */ + unsigned long pages_removed; /* ring buffer pages to update, > 0 to add, < 0 to remove */ long nr_pages_to_update; struct list_head new_pages; /* new pages to add */ @@ -559,6 +561,7 @@ struct ring_buffer_iter { struct buffer_page *head_page; struct buffer_page *cache_reader_page; unsigned long cache_read; + unsigned long cache_pages_removed; u64 read_stamp; u64 page_stamp; struct ring_buffer_event *event; @@ -1957,6 +1960,8 @@ rb_remove_pages(struct ring_buffer_per_cpu *cpu_buffer, unsigned long nr_pages) to_remove = rb_list_head(to_remove)->next; head_bit |= (unsigned long)to_remove & RB_PAGE_HEAD; } + /* Read iterators need to reset themselves when some pages removed */ + cpu_buffer->pages_removed += nr_removed; next_page = rb_list_head(to_remove)->next; @@ -1978,12 +1983,6 @@ rb_remove_pages(struct ring_buffer_per_cpu *cpu_buffer, unsigned long nr_pages) cpu_buffer->head_page = list_entry(next_page, struct buffer_page, list); - /* - * change read pointer to make sure any read iterators reset - * themselves - */ - cpu_buffer->read = 0; - /* pages are removed, resume tracing and then free the pages */ atomic_dec(&cpu_buffer->record_disabled); raw_spin_unlock_irq(&cpu_buffer->reader_lock); @@ -4395,6 +4394,7 @@ static void rb_iter_reset(struct ring_buffer_iter *iter) iter->cache_reader_page = iter->head_page; iter->cache_read = cpu_buffer->read; + iter->cache_pages_removed = cpu_buffer->pages_removed; if (iter->head) { iter->read_stamp = cpu_buffer->read_stamp; @@ -4849,12 +4849,13 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) buffer = cpu_buffer->buffer; /* - * Check if someone performed a consuming read to - * the buffer. A consuming read invalidates the iterator - * and we need to reset the iterator in this case. + * Check if someone performed a consuming read to the buffer + * or removed some pages from the buffer. In these cases, + * iterator was invalidated and we need to reset it. */ if (unlikely(iter->cache_read != cpu_buffer->read || - iter->cache_reader_page != cpu_buffer->reader_page)) + iter->cache_reader_page != cpu_buffer->reader_page || + iter->cache_pages_removed != cpu_buffer->pages_removed)) rb_iter_reset(iter); again: @@ -5298,6 +5299,7 @@ rb_reset_cpu(struct ring_buffer_per_cpu *cpu_buffer) cpu_buffer->last_overrun = 0; rb_head_page_activate(cpu_buffer); + cpu_buffer->pages_removed = 0; } /* Must have disabled the cpu buffer then done a synchronize_rcu */ -- cgit From 151e34d1c6eb2f9b4b0b9d085d1fa9f02cc74f48 Mon Sep 17 00:00:00 2001 From: Gaosheng Cui Date: Mon, 24 Jul 2023 22:08:24 +0800 Subject: ring-buffer: Fix kernel-doc warnings in ring_buffer.c Fix kernel-doc warnings: kernel/trace/ring_buffer.c:954: warning: Function parameter or member 'cpu' not described in 'ring_buffer_wake_waiters' kernel/trace/ring_buffer.c:3383: warning: Excess function parameter 'event' description in 'ring_buffer_unlock_commit' kernel/trace/ring_buffer.c:5359: warning: Excess function parameter 'cpu' description in 'ring_buffer_reset_online_cpus' Link: https://lkml.kernel.org/r/20230724140827.1023266-2-cuigaosheng1@huawei.com Cc: Signed-off-by: Gaosheng Cui Signed-off-by: Steven Rostedt (Google) --- kernel/trace/ring_buffer.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 46b4a3c7c3bf..52dea5dd5362 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -950,6 +950,7 @@ static void rb_wake_up_waiters(struct irq_work *work) /** * ring_buffer_wake_waiters - wake up any waiters on this ring buffer * @buffer: The ring buffer to wake waiters on + * @cpu: The CPU buffer to wake waiters on * * In the case of a file that represents a ring buffer is closing, * it is prudent to wake up any waiters that are on this. @@ -3375,7 +3376,6 @@ void ring_buffer_nest_end(struct trace_buffer *buffer) /** * ring_buffer_unlock_commit - commit a reserved * @buffer: The buffer to commit to - * @event: The event pointer to commit. * * This commits the data to the ring buffer, and releases any locks held. * @@ -5358,7 +5358,6 @@ EXPORT_SYMBOL_GPL(ring_buffer_reset_cpu); /** * ring_buffer_reset_online_cpus - reset a ring buffer per CPU buffer * @buffer: The ring buffer to reset a per cpu buffer of - * @cpu: The CPU buffer to be reset */ void ring_buffer_reset_online_cpus(struct trace_buffer *buffer) { -- cgit From b32c789f7dbbba3e55f6295ad64923cb69e6f563 Mon Sep 17 00:00:00 2001 From: Gaosheng Cui Date: Mon, 24 Jul 2023 22:08:25 +0800 Subject: tracing/synthetic: Fix kernel-doc warnings in trace_events_synth.c Fix kernel-doc warning: kernel/trace/trace_events_synth.c:1257: warning: Function parameter or member 'mod' not described in 'synth_event_gen_cmd_array_start' Link: https://lkml.kernel.org/r/20230724140827.1023266-3-cuigaosheng1@huawei.com Cc: Signed-off-by: Gaosheng Cui Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace_events_synth.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/trace/trace_events_synth.c b/kernel/trace/trace_events_synth.c index d6a70aff2410..dd398afc8e25 100644 --- a/kernel/trace/trace_events_synth.c +++ b/kernel/trace/trace_events_synth.c @@ -1230,6 +1230,7 @@ EXPORT_SYMBOL_GPL(__synth_event_gen_cmd_start); * synth_event_gen_cmd_array_start - Start synthetic event command from an array * @cmd: A pointer to the dynevent_cmd struct representing the new event * @name: The name of the synthetic event + * @mod: The module creating the event, NULL if not created from a module * @fields: An array of type/name field descriptions * @n_fields: The number of field descriptions contained in the fields array * -- cgit From bd7217f30c7f08fc20f745aa1e2174eb3ff6d82a Mon Sep 17 00:00:00 2001 From: Gaosheng Cui Date: Mon, 24 Jul 2023 22:08:26 +0800 Subject: tracing: Fix kernel-doc warnings in trace_events_trigger.c Fix kernel-doc warnings: kernel/trace/trace_events_trigger.c:59: warning: Function parameter or member 'buffer' not described in 'event_triggers_call' kernel/trace/trace_events_trigger.c:59: warning: Function parameter or member 'event' not described in 'event_triggers_call' Link: https://lkml.kernel.org/r/20230724140827.1023266-4-cuigaosheng1@huawei.com Cc: Signed-off-by: Gaosheng Cui Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace_events_trigger.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'kernel') diff --git a/kernel/trace/trace_events_trigger.c b/kernel/trace/trace_events_trigger.c index e535959939d3..46439e3bcec4 100644 --- a/kernel/trace/trace_events_trigger.c +++ b/kernel/trace/trace_events_trigger.c @@ -31,7 +31,9 @@ void trigger_data_free(struct event_trigger_data *data) /** * event_triggers_call - Call triggers associated with a trace event * @file: The trace_event_file associated with the event + * @buffer: The ring buffer that the event is being written to * @rec: The trace entry for the event, NULL for unconditional invocation + * @event: The event meta data in the ring buffer * * For each trigger associated with an event, invoke the trigger * function registered with the associated trigger command. If rec is -- cgit From 6c95d71bad61410e3717afcf7fd1837a8e03edf2 Mon Sep 17 00:00:00 2001 From: Gaosheng Cui Date: Mon, 24 Jul 2023 22:08:27 +0800 Subject: tracing: Fix kernel-doc warnings in trace_seq.c Fix kernel-doc warning: kernel/trace/trace_seq.c:142: warning: Function parameter or member 'args' not described in 'trace_seq_vprintf' Link: https://lkml.kernel.org/r/20230724140827.1023266-5-cuigaosheng1@huawei.com Cc: Signed-off-by: Gaosheng Cui Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace_seq.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/trace/trace_seq.c b/kernel/trace/trace_seq.c index e5e299260d0c..bac06ee3b98b 100644 --- a/kernel/trace/trace_seq.c +++ b/kernel/trace/trace_seq.c @@ -131,6 +131,7 @@ EXPORT_SYMBOL_GPL(trace_seq_bitmask); * trace_seq_vprintf - sequence printing of trace information * @s: trace sequence descriptor * @fmt: printf format string + * @args: Arguments for the format string * * The tracer may use either sequence operations or its own * copy to user routines. To simplify formatting of a trace -- cgit From dea499781a1150d285c62b26659f62fb00824fce Mon Sep 17 00:00:00 2001 From: Zheng Yejian Date: Wed, 26 Jul 2023 17:58:04 +0800 Subject: tracing: Fix warning in trace_buffered_event_disable() Warning happened in trace_buffered_event_disable() at WARN_ON_ONCE(!trace_buffered_event_ref) Call Trace: ? __warn+0xa5/0x1b0 ? trace_buffered_event_disable+0x189/0x1b0 __ftrace_event_enable_disable+0x19e/0x3e0 free_probe_data+0x3b/0xa0 unregister_ftrace_function_probe_func+0x6b8/0x800 event_enable_func+0x2f0/0x3d0 ftrace_process_regex.isra.0+0x12d/0x1b0 ftrace_filter_write+0xe6/0x140 vfs_write+0x1c9/0x6f0 [...] The cause of the warning is in __ftrace_event_enable_disable(), trace_buffered_event_enable() was called once while trace_buffered_event_disable() was called twice. Reproduction script show as below, for analysis, see the comments: ``` #!/bin/bash cd /sys/kernel/tracing/ # 1. Register a 'disable_event' command, then: # 1) SOFT_DISABLED_BIT was set; # 2) trace_buffered_event_enable() was called first time; echo 'cmdline_proc_show:disable_event:initcall:initcall_finish' > \ set_ftrace_filter # 2. Enable the event registered, then: # 1) SOFT_DISABLED_BIT was cleared; # 2) trace_buffered_event_disable() was called first time; echo 1 > events/initcall/initcall_finish/enable # 3. Try to call into cmdline_proc_show(), then SOFT_DISABLED_BIT was # set again!!! cat /proc/cmdline # 4. Unregister the 'disable_event' command, then: # 1) SOFT_DISABLED_BIT was cleared again; # 2) trace_buffered_event_disable() was called second time!!! echo '!cmdline_proc_show:disable_event:initcall:initcall_finish' > \ set_ftrace_filter ``` To fix it, IIUC, we can change to call trace_buffered_event_enable() at fist time soft-mode enabled, and call trace_buffered_event_disable() at last time soft-mode disabled. Link: https://lore.kernel.org/linux-trace-kernel/20230726095804.920457-1-zhengyejian1@huawei.com Cc: Fixes: 0fc1b09ff1ff ("tracing: Use temp buffer when filtering events") Signed-off-by: Zheng Yejian Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace_events.c | 14 ++++---------- 1 file changed, 4 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 5d6ae4eae510..578f1f7d49a6 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -611,7 +611,6 @@ static int __ftrace_event_enable_disable(struct trace_event_file *file, { struct trace_event_call *call = file->event_call; struct trace_array *tr = file->tr; - unsigned long file_flags = file->flags; int ret = 0; int disable; @@ -635,6 +634,8 @@ static int __ftrace_event_enable_disable(struct trace_event_file *file, break; disable = file->flags & EVENT_FILE_FL_SOFT_DISABLED; clear_bit(EVENT_FILE_FL_SOFT_MODE_BIT, &file->flags); + /* Disable use of trace_buffered_event */ + trace_buffered_event_disable(); } else disable = !(file->flags & EVENT_FILE_FL_SOFT_MODE); @@ -673,6 +674,8 @@ static int __ftrace_event_enable_disable(struct trace_event_file *file, if (atomic_inc_return(&file->sm_ref) > 1) break; set_bit(EVENT_FILE_FL_SOFT_MODE_BIT, &file->flags); + /* Enable use of trace_buffered_event */ + trace_buffered_event_enable(); } if (!(file->flags & EVENT_FILE_FL_ENABLED)) { @@ -712,15 +715,6 @@ static int __ftrace_event_enable_disable(struct trace_event_file *file, break; } - /* Enable or disable use of trace_buffered_event */ - if ((file_flags & EVENT_FILE_FL_SOFT_DISABLED) != - (file->flags & EVENT_FILE_FL_SOFT_DISABLED)) { - if (file->flags & EVENT_FILE_FL_SOFT_DISABLED) - trace_buffered_event_enable(); - else - trace_buffered_event_disable(); - } - return ret; } -- cgit From de02f2ac5d8cfb311f44f2bf144cc20002f1fbbd Mon Sep 17 00:00:00 2001 From: "Masami Hiramatsu (Google)" Date: Tue, 11 Jul 2023 10:50:47 +0900 Subject: kprobes: Prohibit probing on CFI preamble symbol Do not allow to probe on "__cfi_" or "__pfx_" started symbol, because those are used for CFI and not executed. Probing it will break the CFI. Link: https://lore.kernel.org/all/168904024679.116016.18089228029322008512.stgit@devnote2/ Signed-off-by: Masami Hiramatsu (Google) Reviewed-by: Steven Rostedt (Google) --- kernel/kprobes.c | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 1fc6095d502d..ca385b61d546 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -1545,6 +1545,17 @@ static int check_ftrace_location(struct kprobe *p) return 0; } +static bool is_cfi_preamble_symbol(unsigned long addr) +{ + char symbuf[KSYM_NAME_LEN]; + + if (lookup_symbol_name(addr, symbuf)) + return false; + + return str_has_prefix("__cfi_", symbuf) || + str_has_prefix("__pfx_", symbuf); +} + static int check_kprobe_address_safe(struct kprobe *p, struct module **probed_mod) { @@ -1563,7 +1574,8 @@ static int check_kprobe_address_safe(struct kprobe *p, within_kprobe_blacklist((unsigned long) p->addr) || jump_label_text_reserved(p->addr, p->addr) || static_call_text_reserved(p->addr, p->addr) || - find_bug((unsigned long)p->addr)) { + find_bug((unsigned long)p->addr) || + is_cfi_preamble_symbol((unsigned long)p->addr)) { ret = -EINVAL; goto out; } -- cgit From 640a604585aa30f93e39b17d4d6ba69fcb1e66c9 Mon Sep 17 00:00:00 2001 From: Hou Tao Date: Sat, 29 Jul 2023 17:51:06 +0800 Subject: bpf, cpumap: Make sure kthread is running before map update returns The following warning was reported when running stress-mode enabled xdp_redirect_cpu with some RT threads: ------------[ cut here ]------------ WARNING: CPU: 4 PID: 65 at kernel/bpf/cpumap.c:135 CPU: 4 PID: 65 Comm: kworker/4:1 Not tainted 6.5.0-rc2+ #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996) Workqueue: events cpu_map_kthread_stop RIP: 0010:put_cpu_map_entry+0xda/0x220 ...... Call Trace: ? show_regs+0x65/0x70 ? __warn+0xa5/0x240 ...... ? put_cpu_map_entry+0xda/0x220 cpu_map_kthread_stop+0x41/0x60 process_one_work+0x6b0/0xb80 worker_thread+0x96/0x720 kthread+0x1a5/0x1f0 ret_from_fork+0x3a/0x70 ret_from_fork_asm+0x1b/0x30 The root cause is the same as commit 436901649731 ("bpf: cpumap: Fix memory leak in cpu_map_update_elem"). The kthread is stopped prematurely by kthread_stop() in cpu_map_kthread_stop(), and kthread() doesn't call cpu_map_kthread_run() at all but XDP program has already queued some frames or skbs into ptr_ring. So when __cpu_map_ring_cleanup() checks the ptr_ring, it will find it was not emptied and report a warning. An alternative fix is to use __cpu_map_ring_cleanup() to drop these pending frames or skbs when kthread_stop() returns -EINTR, but it may confuse the user, because these frames or skbs have been handled correctly by XDP program. So instead of dropping these frames or skbs, just make sure the per-cpu kthread is running before __cpu_map_entry_alloc() returns. After apply the fix, the error handle for kthread_stop() will be unnecessary because it will always return 0, so just remove it. Fixes: 6710e1126934 ("bpf: introduce new bpf cpu map type BPF_MAP_TYPE_CPUMAP") Signed-off-by: Hou Tao Reviewed-by: Pu Lehui Acked-by: Jesper Dangaard Brouer Link: https://lore.kernel.org/r/20230729095107.1722450-2-houtao@huaweicloud.com Signed-off-by: Martin KaFai Lau --- kernel/bpf/cpumap.c | 21 +++++++++++---------- 1 file changed, 11 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/bpf/cpumap.c b/kernel/bpf/cpumap.c index 6ae02be7a48e..7eeb20025164 100644 --- a/kernel/bpf/cpumap.c +++ b/kernel/bpf/cpumap.c @@ -28,6 +28,7 @@ #include #include #include +#include #include #include @@ -73,6 +74,7 @@ struct bpf_cpu_map_entry { struct rcu_head rcu; struct work_struct kthread_stop_wq; + struct completion kthread_running; }; struct bpf_cpu_map { @@ -153,7 +155,6 @@ static void put_cpu_map_entry(struct bpf_cpu_map_entry *rcpu) static void cpu_map_kthread_stop(struct work_struct *work) { struct bpf_cpu_map_entry *rcpu; - int err; rcpu = container_of(work, struct bpf_cpu_map_entry, kthread_stop_wq); @@ -163,14 +164,7 @@ static void cpu_map_kthread_stop(struct work_struct *work) rcu_barrier(); /* kthread_stop will wake_up_process and wait for it to complete */ - err = kthread_stop(rcpu->kthread); - if (err) { - /* kthread_stop may be called before cpu_map_kthread_run - * is executed, so we need to release the memory related - * to rcpu. - */ - put_cpu_map_entry(rcpu); - } + kthread_stop(rcpu->kthread); } static void cpu_map_bpf_prog_run_skb(struct bpf_cpu_map_entry *rcpu, @@ -298,11 +292,11 @@ static int cpu_map_bpf_prog_run(struct bpf_cpu_map_entry *rcpu, void **frames, return nframes; } - static int cpu_map_kthread_run(void *data) { struct bpf_cpu_map_entry *rcpu = data; + complete(&rcpu->kthread_running); set_current_state(TASK_INTERRUPTIBLE); /* When kthread gives stop order, then rcpu have been disconnected @@ -467,6 +461,7 @@ __cpu_map_entry_alloc(struct bpf_map *map, struct bpf_cpumap_val *value, goto free_ptr_ring; /* Setup kthread */ + init_completion(&rcpu->kthread_running); rcpu->kthread = kthread_create_on_node(cpu_map_kthread_run, rcpu, numa, "cpumap/%d/map:%d", cpu, map->id); @@ -480,6 +475,12 @@ __cpu_map_entry_alloc(struct bpf_map *map, struct bpf_cpumap_val *value, kthread_bind(rcpu->kthread, cpu); wake_up_process(rcpu->kthread); + /* Make sure kthread has been running, so kthread_stop() will not + * stop the kthread prematurely and all pending frames or skbs + * will be handled by the kthread before kthread_stop() returns. + */ + wait_for_completion(&rcpu->kthread_running); + return rcpu; free_prog: -- cgit From 7c62b75cd1a792e14b037fa4f61f9b18914e7de1 Mon Sep 17 00:00:00 2001 From: Hou Tao Date: Sat, 29 Jul 2023 17:51:07 +0800 Subject: bpf, cpumap: Handle skb as well when clean up ptr_ring The following warning was reported when running xdp_redirect_cpu with both skb-mode and stress-mode enabled: ------------[ cut here ]------------ Incorrect XDP memory type (-2128176192) usage WARNING: CPU: 7 PID: 1442 at net/core/xdp.c:405 Modules linked in: CPU: 7 PID: 1442 Comm: kworker/7:0 Tainted: G 6.5.0-rc2+ #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996) Workqueue: events __cpu_map_entry_free RIP: 0010:__xdp_return+0x1e4/0x4a0 ...... Call Trace: ? show_regs+0x65/0x70 ? __warn+0xa5/0x240 ? __xdp_return+0x1e4/0x4a0 ...... xdp_return_frame+0x4d/0x150 __cpu_map_entry_free+0xf9/0x230 process_one_work+0x6b0/0xb80 worker_thread+0x96/0x720 kthread+0x1a5/0x1f0 ret_from_fork+0x3a/0x70 ret_from_fork_asm+0x1b/0x30 The reason for the warning is twofold. One is due to the kthread cpu_map_kthread_run() is stopped prematurely. Another one is __cpu_map_ring_cleanup() doesn't handle skb mode and treats skbs in ptr_ring as XDP frames. Prematurely-stopped kthread will be fixed by the preceding patch and ptr_ring will be empty when __cpu_map_ring_cleanup() is called. But as the comments in __cpu_map_ring_cleanup() said, handling and freeing skbs in ptr_ring as well to "catch any broken behaviour gracefully". Fixes: 11941f8a8536 ("bpf: cpumap: Implement generic cpumap") Signed-off-by: Hou Tao Acked-by: Jesper Dangaard Brouer Link: https://lore.kernel.org/r/20230729095107.1722450-3-houtao@huaweicloud.com Signed-off-by: Martin KaFai Lau --- kernel/bpf/cpumap.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/bpf/cpumap.c b/kernel/bpf/cpumap.c index 7eeb20025164..286ab3db0fde 100644 --- a/kernel/bpf/cpumap.c +++ b/kernel/bpf/cpumap.c @@ -131,11 +131,17 @@ static void __cpu_map_ring_cleanup(struct ptr_ring *ring) * invoked cpu_map_kthread_stop(). Catch any broken behaviour * gracefully and warn once. */ - struct xdp_frame *xdpf; + void *ptr; - while ((xdpf = ptr_ring_consume(ring))) - if (WARN_ON_ONCE(xdpf)) - xdp_return_frame(xdpf); + while ((ptr = ptr_ring_consume(ring))) { + WARN_ON_ONCE(1); + if (unlikely(__ptr_test_bit(0, &ptr))) { + __ptr_clear_bit(0, &ptr); + kfree_skb(ptr); + continue; + } + xdp_return_frame(ptr); + } } static void put_cpu_map_entry(struct bpf_cpu_map_entry *rcpu) -- cgit From df2f7cde73cb58c0e6a60f97d1cd6037138a45cd Mon Sep 17 00:00:00 2001 From: Vlastimil Babka Date: Mon, 7 Aug 2023 10:33:57 +0200 Subject: PM: hibernate: fix resume_store() return value when hibernation not available On a laptop with hibernation set up but not actively used, and with secure boot and lockdown enabled kernel, 6.5-rc1 gets stuck on boot with the following repeated messages: A start job is running for Resume from hibernation using device /dev/system/swap (24s / no limit) lockdown_is_locked_down: 25311154 callbacks suppressed Lockdown: systemd-hiberna: hibernation is restricted; see man kernel_lockdown.7 ... Checking the resume code leads to commit cc89c63e2fe3 ("PM: hibernate: move finding the resume device out of software_resume") which inadvertently changed the return value from resume_store() to 0 when !hibernation_available(). This apparently translates to userspace write() returning 0 as in number of bytes written, and userspace looping indefinitely in the attempt to write the intended value. Fix this by returning the full number of bytes that were to be written, as that's what was done before the commit. Fixes: cc89c63e2fe3 ("PM: hibernate: move finding the resume device out of software_resume") Signed-off-by: Vlastimil Babka Reviewed-by: Christoph Hellwig Acked-by: Greg Kroah-Hartman Signed-off-by: Rafael J. Wysocki --- kernel/power/hibernate.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/power/hibernate.c b/kernel/power/hibernate.c index e1b4bfa938dd..2b4a946a6ff5 100644 --- a/kernel/power/hibernate.c +++ b/kernel/power/hibernate.c @@ -1166,7 +1166,7 @@ static ssize_t resume_store(struct kobject *kobj, struct kobj_attribute *attr, int error; if (!hibernation_available()) - return 0; + return n; if (len && buf[len-1] == '\n') len--; -- cgit