From 966d2b04e070bc040319aaebfec09e0144dc3341 Mon Sep 17 00:00:00 2001 From: Douglas Miller Date: Sat, 28 Jan 2017 06:42:20 -0600 Subject: percpu-refcount: fix reference leak during percpu-atomic transition percpu_ref_tryget() and percpu_ref_tryget_live() should return "true" IFF they acquire a reference. But the return value from atomic_long_inc_not_zero() is a long and may have high bits set, e.g. PERCPU_COUNT_BIAS, and the return value of the tryget routines is bool so the reference may actually be acquired but the routines return "false" which results in a reference leak since the caller assumes it does not need to do a corresponding percpu_ref_put(). This was seen when performing CPU hotplug during I/O, as hangs in blk_mq_freeze_queue_wait where percpu_ref_kill (blk_mq_freeze_queue_start) raced with percpu_ref_tryget (blk_mq_timeout_work). Sample stack trace: __switch_to+0x2c0/0x450 __schedule+0x2f8/0x970 schedule+0x48/0xc0 blk_mq_freeze_queue_wait+0x94/0x120 blk_mq_queue_reinit_work+0xb8/0x180 blk_mq_queue_reinit_prepare+0x84/0xa0 cpuhp_invoke_callback+0x17c/0x600 cpuhp_up_callbacks+0x58/0x150 _cpu_up+0xf0/0x1c0 do_cpu_up+0x120/0x150 cpu_subsys_online+0x64/0xe0 device_online+0xb4/0x120 online_store+0xb4/0xc0 dev_attr_store+0x68/0xa0 sysfs_kf_write+0x80/0xb0 kernfs_fop_write+0x17c/0x250 __vfs_write+0x6c/0x1e0 vfs_write+0xd0/0x270 SyS_write+0x6c/0x110 system_call+0x38/0xe0 Examination of the queue showed a single reference (no PERCPU_COUNT_BIAS, and __PERCPU_REF_DEAD, __PERCPU_REF_ATOMIC set) and no requests. However, conditions at the time of the race are count of PERCPU_COUNT_BIAS + 0 and __PERCPU_REF_DEAD and __PERCPU_REF_ATOMIC set. The fix is to make the tryget routines use an actual boolean internally instead of the atomic long result truncated to a int. Fixes: e625305b3907 percpu-refcount: make percpu_ref based on longs instead of ints Link: https://bugzilla.kernel.org/show_bug.cgi?id=190751 Signed-off-by: Douglas Miller Reviewed-by: Jens Axboe Signed-off-by: Tejun Heo Fixes: e625305b3907 ("percpu-refcount: make percpu_ref based on longs instead of ints") Cc: stable@vger.kernel.org # v3.18+ --- include/linux/percpu-refcount.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/percpu-refcount.h b/include/linux/percpu-refcount.h index 1c7eec09e5eb..3a481a49546e 100644 --- a/include/linux/percpu-refcount.h +++ b/include/linux/percpu-refcount.h @@ -204,7 +204,7 @@ static inline void percpu_ref_get(struct percpu_ref *ref) static inline bool percpu_ref_tryget(struct percpu_ref *ref) { unsigned long __percpu *percpu_count; - int ret; + bool ret; rcu_read_lock_sched(); @@ -238,7 +238,7 @@ static inline bool percpu_ref_tryget(struct percpu_ref *ref) static inline bool percpu_ref_tryget_live(struct percpu_ref *ref) { unsigned long __percpu *percpu_count; - int ret = false; + bool ret = false; rcu_read_lock_sched(); -- cgit v1.3-14-g43fede From f1712c73714088a7252d276a57126d56c7d37e64 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Fri, 27 Jan 2017 08:11:44 -0800 Subject: can: Fix kernel panic at security_sock_rcv_skb Zhang Yanmin reported crashes [1] and provided a patch adding a synchronize_rcu() call in can_rx_unregister() The main problem seems that the sockets themselves are not RCU protected. If CAN uses RCU for delivery, then sockets should be freed only after one RCU grace period. Recent kernels could use sock_set_flag(sk, SOCK_RCU_FREE), but let's ease stable backports with the following fix instead. [1] BUG: unable to handle kernel NULL pointer dereference at (null) IP: [] selinux_socket_sock_rcv_skb+0x65/0x2a0 Call Trace: [] security_sock_rcv_skb+0x4c/0x60 [] sk_filter+0x41/0x210 [] sock_queue_rcv_skb+0x53/0x3a0 [] raw_rcv+0x2a3/0x3c0 [] can_rcv_filter+0x12b/0x370 [] can_receive+0xd9/0x120 [] can_rcv+0xab/0x100 [] __netif_receive_skb_core+0xd8c/0x11f0 [] __netif_receive_skb+0x24/0xb0 [] process_backlog+0x127/0x280 [] net_rx_action+0x33b/0x4f0 [] __do_softirq+0x184/0x440 [] do_softirq_own_stack+0x1c/0x30 [] do_softirq.part.18+0x3b/0x40 [] do_softirq+0x1d/0x20 [] netif_rx_ni+0xe5/0x110 [] slcan_receive_buf+0x507/0x520 [] flush_to_ldisc+0x21c/0x230 [] process_one_work+0x24f/0x670 [] worker_thread+0x9d/0x6f0 [] ? rescuer_thread+0x480/0x480 [] kthread+0x12c/0x150 [] ret_from_fork+0x3f/0x70 Reported-by: Zhang Yanmin Signed-off-by: Eric Dumazet Acked-by: Oliver Hartkopp Signed-off-by: David S. Miller --- include/linux/can/core.h | 7 +++---- net/can/af_can.c | 12 ++++++++++-- net/can/af_can.h | 3 ++- net/can/bcm.c | 4 ++-- net/can/gw.c | 2 +- net/can/raw.c | 4 ++-- 6 files changed, 20 insertions(+), 12 deletions(-) (limited to 'include/linux') diff --git a/include/linux/can/core.h b/include/linux/can/core.h index a0875001b13c..df08a41d5be5 100644 --- a/include/linux/can/core.h +++ b/include/linux/can/core.h @@ -45,10 +45,9 @@ struct can_proto { extern int can_proto_register(const struct can_proto *cp); extern void can_proto_unregister(const struct can_proto *cp); -extern int can_rx_register(struct net_device *dev, canid_t can_id, - canid_t mask, - void (*func)(struct sk_buff *, void *), - void *data, char *ident); +int can_rx_register(struct net_device *dev, canid_t can_id, canid_t mask, + void (*func)(struct sk_buff *, void *), + void *data, char *ident, struct sock *sk); extern void can_rx_unregister(struct net_device *dev, canid_t can_id, canid_t mask, diff --git a/net/can/af_can.c b/net/can/af_can.c index 1108079d934f..5488e4a6ccd0 100644 --- a/net/can/af_can.c +++ b/net/can/af_can.c @@ -445,6 +445,7 @@ static struct hlist_head *find_rcv_list(canid_t *can_id, canid_t *mask, * @func: callback function on filter match * @data: returned parameter for callback function * @ident: string for calling module identification + * @sk: socket pointer (might be NULL) * * Description: * Invokes the callback function with the received sk_buff and the given @@ -468,7 +469,7 @@ static struct hlist_head *find_rcv_list(canid_t *can_id, canid_t *mask, */ int can_rx_register(struct net_device *dev, canid_t can_id, canid_t mask, void (*func)(struct sk_buff *, void *), void *data, - char *ident) + char *ident, struct sock *sk) { struct receiver *r; struct hlist_head *rl; @@ -496,6 +497,7 @@ int can_rx_register(struct net_device *dev, canid_t can_id, canid_t mask, r->func = func; r->data = data; r->ident = ident; + r->sk = sk; hlist_add_head_rcu(&r->list, rl); d->entries++; @@ -520,8 +522,11 @@ EXPORT_SYMBOL(can_rx_register); static void can_rx_delete_receiver(struct rcu_head *rp) { struct receiver *r = container_of(rp, struct receiver, rcu); + struct sock *sk = r->sk; kmem_cache_free(rcv_cache, r); + if (sk) + sock_put(sk); } /** @@ -596,8 +601,11 @@ void can_rx_unregister(struct net_device *dev, canid_t can_id, canid_t mask, spin_unlock(&can_rcvlists_lock); /* schedule the receiver item for deletion */ - if (r) + if (r) { + if (r->sk) + sock_hold(r->sk); call_rcu(&r->rcu, can_rx_delete_receiver); + } } EXPORT_SYMBOL(can_rx_unregister); diff --git a/net/can/af_can.h b/net/can/af_can.h index fca0fe9fc45a..b86f5129e838 100644 --- a/net/can/af_can.h +++ b/net/can/af_can.h @@ -50,13 +50,14 @@ struct receiver { struct hlist_node list; - struct rcu_head rcu; canid_t can_id; canid_t mask; unsigned long matches; void (*func)(struct sk_buff *, void *); void *data; char *ident; + struct sock *sk; + struct rcu_head rcu; }; #define CAN_SFF_RCV_ARRAY_SZ (1 << CAN_SFF_ID_BITS) diff --git a/net/can/bcm.c b/net/can/bcm.c index 21ac75390e3d..5c9407181918 100644 --- a/net/can/bcm.c +++ b/net/can/bcm.c @@ -1216,7 +1216,7 @@ static int bcm_rx_setup(struct bcm_msg_head *msg_head, struct msghdr *msg, err = can_rx_register(dev, op->can_id, REGMASK(op->can_id), bcm_rx_handler, op, - "bcm"); + "bcm", sk); op->rx_reg_dev = dev; dev_put(dev); @@ -1225,7 +1225,7 @@ static int bcm_rx_setup(struct bcm_msg_head *msg_head, struct msghdr *msg, } else err = can_rx_register(NULL, op->can_id, REGMASK(op->can_id), - bcm_rx_handler, op, "bcm"); + bcm_rx_handler, op, "bcm", sk); if (err) { /* this bcm rx op is broken -> remove it */ list_del(&op->list); diff --git a/net/can/gw.c b/net/can/gw.c index a54ab0c82104..7056a1a2bb70 100644 --- a/net/can/gw.c +++ b/net/can/gw.c @@ -442,7 +442,7 @@ static inline int cgw_register_filter(struct cgw_job *gwj) { return can_rx_register(gwj->src.dev, gwj->ccgw.filter.can_id, gwj->ccgw.filter.can_mask, can_can_gw_rcv, - gwj, "gw"); + gwj, "gw", NULL); } static inline void cgw_unregister_filter(struct cgw_job *gwj) diff --git a/net/can/raw.c b/net/can/raw.c index b075f028d7e2..6dc546a06673 100644 --- a/net/can/raw.c +++ b/net/can/raw.c @@ -190,7 +190,7 @@ static int raw_enable_filters(struct net_device *dev, struct sock *sk, for (i = 0; i < count; i++) { err = can_rx_register(dev, filter[i].can_id, filter[i].can_mask, - raw_rcv, sk, "raw"); + raw_rcv, sk, "raw", sk); if (err) { /* clean up successfully registered filters */ while (--i >= 0) @@ -211,7 +211,7 @@ static int raw_enable_errfilter(struct net_device *dev, struct sock *sk, if (err_mask) err = can_rx_register(dev, 0, err_mask | CAN_ERR_FLAG, - raw_rcv, sk, "raw"); + raw_rcv, sk, "raw", sk); return err; } -- cgit v1.3-14-g43fede From e26bfebdfc0d212d366de9990a096665d5c0209a Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 31 Jan 2017 09:45:28 +0000 Subject: fscache: Fix dead object requeue Under some circumstances, an fscache object can become queued such that it fscache_object_work_func() can be called once the object is in the OBJECT_DEAD state. This results in the kernel oopsing when it tries to invoke the handler for the state (which is hard coded to 0x2). The way this comes about is something like the following: (1) The object dispatcher is processing a work state for an object. This is done in workqueue context. (2) An out-of-band event comes in that isn't masked, causing the object to be queued, say EV_KILL. (3) The object dispatcher finishes processing the current work state on that object and then sees there's another event to process, so, without returning to the workqueue core, it processes that event too. It then follows the chain of events that initiates until we reach OBJECT_DEAD without going through a wait state (such as WAIT_FOR_CLEARANCE). At this point, object->events may be 0, object->event_mask will be 0 and oob_event_mask will be 0. (4) The object dispatcher returns to the workqueue processor, and in due course, this sees that the object's work item is still queued and invokes it again. (5) The current state is a work state (OBJECT_DEAD), so the dispatcher jumps to it - resulting in an OOPS. When I'm seeing this, the work state in (1) appears to have been either LOOK_UP_OBJECT or CREATE_OBJECT (object->oob_table is fscache_osm_lookup_oob). The window for (2) is very small: (A) object->event_mask is cleared whilst the event dispatch process is underway - though there's no memory barrier to force this to the top of the function. The window, therefore is from the time the object was selected by the workqueue processor and made requeueable to the time the mask was cleared. (B) fscache_raise_event() will only queue the object if it manages to set the event bit and the corresponding event_mask bit was set. The enqueuement is then deferred slightly whilst we get a ref on the object and get the per-CPU variable for workqueue congestion. This slight deferral slightly increases the probability by allowing extra time for the workqueue to make the item requeueable. Handle this by giving the dead state a processor function and checking the for the dead state address rather than seeing if the processor function is address 0x2. The dead state processor function can then set a flag to indicate that it's occurred and give a warning if it occurs more than once per object. If this race occurs, an oops similar to the following is seen (note the RIP value): BUG: unable to handle kernel NULL pointer dereference at 0000000000000002 IP: [<0000000000000002>] 0x1 PGD 0 Oops: 0010 [#1] SMP Modules linked in: ... CPU: 17 PID: 16077 Comm: kworker/u48:9 Not tainted 3.10.0-327.18.2.el7.x86_64 #1 Hardware name: HP ProLiant DL380 Gen9/ProLiant DL380 Gen9, BIOS P89 12/27/2015 Workqueue: fscache_object fscache_object_work_func [fscache] task: ffff880302b63980 ti: ffff880717544000 task.ti: ffff880717544000 RIP: 0010:[<0000000000000002>] [<0000000000000002>] 0x1 RSP: 0018:ffff880717547df8 EFLAGS: 00010202 RAX: ffffffffa0368640 RBX: ffff880edf7a4480 RCX: dead000000200200 RDX: 0000000000000002 RSI: 00000000ffffffff RDI: ffff880edf7a4480 RBP: ffff880717547e18 R08: 0000000000000000 R09: dfc40a25cb3a4510 R10: dfc40a25cb3a4510 R11: 0000000000000400 R12: 0000000000000000 R13: ffff880edf7a4510 R14: ffff8817f6153400 R15: 0000000000000600 FS: 0000000000000000(0000) GS:ffff88181f420000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000002 CR3: 000000000194a000 CR4: 00000000001407e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Stack: ffffffffa0363695 ffff880edf7a4510 ffff88093f16f900 ffff8817faa4ec00 ffff880717547e60 ffffffff8109d5db 00000000faa4ec18 0000000000000000 ffff8817faa4ec18 ffff88093f16f930 ffff880302b63980 ffff88093f16f900 Call Trace: [] ? fscache_object_work_func+0xa5/0x200 [fscache] [] process_one_work+0x17b/0x470 [] worker_thread+0x21c/0x400 [] ? rescuer_thread+0x400/0x400 [] kthread+0xcf/0xe0 [] ? kthread_create_on_node+0x140/0x140 [] ret_from_fork+0x58/0x90 [] ? kthread_create_on_node+0x140/0x140 Signed-off-by: David Howells Acked-by: Jeremy McNicoll Tested-by: Frank Sorenson Tested-by: Benjamin Coddington Reviewed-by: Benjamin Coddington Signed-off-by: Al Viro --- fs/fscache/object.c | 26 ++++++++++++++++++++++++-- include/linux/fscache-cache.h | 1 + 2 files changed, 25 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/fs/fscache/object.c b/fs/fscache/object.c index be02a086ed9b..7a182c87f378 100644 --- a/fs/fscache/object.c +++ b/fs/fscache/object.c @@ -30,6 +30,7 @@ static const struct fscache_state *fscache_look_up_object(struct fscache_object static const struct fscache_state *fscache_object_available(struct fscache_object *, int); static const struct fscache_state *fscache_parent_ready(struct fscache_object *, int); static const struct fscache_state *fscache_update_object(struct fscache_object *, int); +static const struct fscache_state *fscache_object_dead(struct fscache_object *, int); #define __STATE_NAME(n) fscache_osm_##n #define STATE(n) (&__STATE_NAME(n)) @@ -91,7 +92,7 @@ static WORK_STATE(LOOKUP_FAILURE, "LCFL", fscache_lookup_failure); static WORK_STATE(KILL_OBJECT, "KILL", fscache_kill_object); static WORK_STATE(KILL_DEPENDENTS, "KDEP", fscache_kill_dependents); static WORK_STATE(DROP_OBJECT, "DROP", fscache_drop_object); -static WORK_STATE(OBJECT_DEAD, "DEAD", (void*)2UL); +static WORK_STATE(OBJECT_DEAD, "DEAD", fscache_object_dead); static WAIT_STATE(WAIT_FOR_INIT, "?INI", TRANSIT_TO(INIT_OBJECT, 1 << FSCACHE_OBJECT_EV_NEW_CHILD)); @@ -229,6 +230,10 @@ execute_work_state: event = -1; if (new_state == NO_TRANSIT) { _debug("{OBJ%x} %s notrans", object->debug_id, state->name); + if (unlikely(state == STATE(OBJECT_DEAD))) { + _leave(" [dead]"); + return; + } fscache_enqueue_object(object); event_mask = object->oob_event_mask; goto unmask_events; @@ -239,7 +244,7 @@ execute_work_state: object->state = state = new_state; if (state->work) { - if (unlikely(state->work == ((void *)2UL))) { + if (unlikely(state == STATE(OBJECT_DEAD))) { _leave(" [dead]"); return; } @@ -1083,3 +1088,20 @@ void fscache_object_mark_killed(struct fscache_object *object, } } EXPORT_SYMBOL(fscache_object_mark_killed); + +/* + * The object is dead. We can get here if an object gets queued by an event + * that would lead to its death (such as EV_KILL) when the dispatcher is + * already running (and so can be requeued) but hasn't yet cleared the event + * mask. + */ +static const struct fscache_state *fscache_object_dead(struct fscache_object *object, + int event) +{ + if (!test_and_set_bit(FSCACHE_OBJECT_RUN_AFTER_DEAD, + &object->flags)) + return NO_TRANSIT; + + WARN(true, "FS-Cache object redispatched after death"); + return NO_TRANSIT; +} diff --git a/include/linux/fscache-cache.h b/include/linux/fscache-cache.h index 13ba552e6c09..4c467ef50159 100644 --- a/include/linux/fscache-cache.h +++ b/include/linux/fscache-cache.h @@ -360,6 +360,7 @@ struct fscache_object { #define FSCACHE_OBJECT_IS_AVAILABLE 5 /* T if object has become active */ #define FSCACHE_OBJECT_RETIRED 6 /* T if object was retired on relinquishment */ #define FSCACHE_OBJECT_KILLED_BY_CACHE 7 /* T if object was killed by the cache */ +#define FSCACHE_OBJECT_RUN_AFTER_DEAD 8 /* T if object has been dispatched after death */ struct list_head cache_link; /* link in cache->object_list */ struct hlist_node cookie_link; /* link in cookie->backing_objects */ -- cgit v1.3-14-g43fede From dd86e373e09fb16b83e8adf5c48c421a4ca76468 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 31 Jan 2017 23:58:38 +0100 Subject: perf/x86/intel/rapl: Make package handling more robust The package management code in RAPL relies on package mapping being available before a CPU is started. This changed with: 9d85eb9119f4 ("x86/smpboot: Make logical package management more robust") because the ACPI/BIOS information turned out to be unreliable, but that left RAPL in broken state. This was not noticed because on a regular boot all CPUs are online before RAPL is initialized. A possible fix would be to reintroduce the mess which allocates a package data structure in CPU prepare and when it turns out to already exist in starting throw it away later in the CPU online callback. But that's a horrible hack and not required at all because RAPL becomes functional for perf only in the CPU online callback. That's correct because user space is not yet informed about the CPU being onlined, so nothing caan rely on RAPL being available on that particular CPU. Move the allocation to the CPU online callback and simplify the hotplug handling. At this point the package mapping is established and correct. This also adds a missing check for available package data in the event_init() function. Reported-by: Yasuaki Ishimatsu Signed-off-by: Thomas Gleixner Cc: Alexander Shishkin Cc: Arnaldo Carvalho de Melo Cc: Jiri Olsa Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Sebastian Siewior Cc: Stephane Eranian Cc: Vince Weaver Fixes: 9d85eb9119f4 ("x86/smpboot: Make logical package management more robust") Link: http://lkml.kernel.org/r/20170131230141.212593966@linutronix.de Signed-off-by: Ingo Molnar --- arch/x86/events/intel/rapl.c | 60 +++++++++++++++++++------------------------- include/linux/cpuhotplug.h | 1 - 2 files changed, 26 insertions(+), 35 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/events/intel/rapl.c b/arch/x86/events/intel/rapl.c index 17c3564d087a..22ef4f72cf32 100644 --- a/arch/x86/events/intel/rapl.c +++ b/arch/x86/events/intel/rapl.c @@ -161,7 +161,13 @@ static u64 rapl_timer_ms; static inline struct rapl_pmu *cpu_to_rapl_pmu(unsigned int cpu) { - return rapl_pmus->pmus[topology_logical_package_id(cpu)]; + unsigned int pkgid = topology_logical_package_id(cpu); + + /* + * The unsigned check also catches the '-1' return value for non + * existent mappings in the topology map. + */ + return pkgid < rapl_pmus->maxpkg ? rapl_pmus->pmus[pkgid] : NULL; } static inline u64 rapl_read_counter(struct perf_event *event) @@ -402,6 +408,8 @@ static int rapl_pmu_event_init(struct perf_event *event) /* must be done before validate_group */ pmu = cpu_to_rapl_pmu(event->cpu); + if (!pmu) + return -EINVAL; event->cpu = pmu->cpu; event->pmu_private = pmu; event->hw.event_base = msr; @@ -585,6 +593,20 @@ static int rapl_cpu_online(unsigned int cpu) struct rapl_pmu *pmu = cpu_to_rapl_pmu(cpu); int target; + if (!pmu) { + pmu = kzalloc_node(sizeof(*pmu), GFP_KERNEL, cpu_to_node(cpu)); + if (!pmu) + return -ENOMEM; + + raw_spin_lock_init(&pmu->lock); + INIT_LIST_HEAD(&pmu->active_list); + pmu->pmu = &rapl_pmus->pmu; + pmu->timer_interval = ms_to_ktime(rapl_timer_ms); + rapl_hrtimer_init(pmu); + + rapl_pmus->pmus[topology_logical_package_id(cpu)] = pmu; + } + /* * Check if there is an online cpu in the package which collects rapl * events already. @@ -598,27 +620,6 @@ static int rapl_cpu_online(unsigned int cpu) return 0; } -static int rapl_cpu_prepare(unsigned int cpu) -{ - struct rapl_pmu *pmu = cpu_to_rapl_pmu(cpu); - - if (pmu) - return 0; - - pmu = kzalloc_node(sizeof(*pmu), GFP_KERNEL, cpu_to_node(cpu)); - if (!pmu) - return -ENOMEM; - - raw_spin_lock_init(&pmu->lock); - INIT_LIST_HEAD(&pmu->active_list); - pmu->pmu = &rapl_pmus->pmu; - pmu->timer_interval = ms_to_ktime(rapl_timer_ms); - pmu->cpu = -1; - rapl_hrtimer_init(pmu); - rapl_pmus->pmus[topology_logical_package_id(cpu)] = pmu; - return 0; -} - static int rapl_check_hw_unit(bool apply_quirk) { u64 msr_rapl_power_unit_bits; @@ -803,29 +804,21 @@ static int __init rapl_pmu_init(void) /* * Install callbacks. Core will call them for each online cpu. */ - - ret = cpuhp_setup_state(CPUHP_PERF_X86_RAPL_PREP, "perf/x86/rapl:prepare", - rapl_cpu_prepare, NULL); - if (ret) - goto out; - ret = cpuhp_setup_state(CPUHP_AP_PERF_X86_RAPL_ONLINE, "perf/x86/rapl:online", rapl_cpu_online, rapl_cpu_offline); if (ret) - goto out1; + goto out; ret = perf_pmu_register(&rapl_pmus->pmu, "power", -1); if (ret) - goto out2; + goto out1; rapl_advertise(); return 0; -out2: - cpuhp_remove_state(CPUHP_AP_PERF_X86_RAPL_ONLINE); out1: - cpuhp_remove_state(CPUHP_PERF_X86_RAPL_PREP); + cpuhp_remove_state(CPUHP_AP_PERF_X86_RAPL_ONLINE); out: pr_warn("Initialization failed (%d), disabled\n", ret); cleanup_rapl_pmus(); @@ -836,7 +829,6 @@ module_init(rapl_pmu_init); static void __exit intel_rapl_exit(void) { cpuhp_remove_state_nocalls(CPUHP_AP_PERF_X86_RAPL_ONLINE); - cpuhp_remove_state_nocalls(CPUHP_PERF_X86_RAPL_PREP); perf_pmu_unregister(&rapl_pmus->pmu); cleanup_rapl_pmus(); } diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h index d936a0021839..8329f3dc592c 100644 --- a/include/linux/cpuhotplug.h +++ b/include/linux/cpuhotplug.h @@ -10,7 +10,6 @@ enum cpuhp_state { CPUHP_PERF_X86_PREPARE, CPUHP_PERF_X86_UNCORE_PREP, CPUHP_PERF_X86_AMD_UNCORE_PREP, - CPUHP_PERF_X86_RAPL_PREP, CPUHP_PERF_BFIN, CPUHP_PERF_POWER, CPUHP_PERF_SUPERH, -- cgit v1.3-14-g43fede From fff4b87e594ad3d2e4f51e8d3d86a6f9d3d8b654 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 31 Jan 2017 23:58:40 +0100 Subject: perf/x86/intel/uncore: Make package handling more robust The package management code in uncore relies on package mapping being available before a CPU is started. This changed with: 9d85eb9119f4 ("x86/smpboot: Make logical package management more robust") because the ACPI/BIOS information turned out to be unreliable, but that left uncore in broken state. This was not noticed because on a regular boot all CPUs are online before uncore is initialized. Move the allocation to the CPU online callback and simplify the hotplug handling. At this point the package mapping is established and correct. Signed-off-by: Thomas Gleixner Cc: Alexander Shishkin Cc: Arnaldo Carvalho de Melo Cc: Jiri Olsa Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Sebastian Siewior Cc: Stephane Eranian Cc: Vince Weaver Cc: Yasuaki Ishimatsu Fixes: 9d85eb9119f4 ("x86/smpboot: Make logical package management more robust") Link: http://lkml.kernel.org/r/20170131230141.377156255@linutronix.de Signed-off-by: Ingo Molnar --- arch/x86/events/intel/uncore.c | 196 +++++++++++++++++++---------------------- include/linux/cpuhotplug.h | 2 - 2 files changed, 91 insertions(+), 107 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/events/intel/uncore.c b/arch/x86/events/intel/uncore.c index 56c5235dcc29..1ab45976474d 100644 --- a/arch/x86/events/intel/uncore.c +++ b/arch/x86/events/intel/uncore.c @@ -100,7 +100,13 @@ ssize_t uncore_event_show(struct kobject *kobj, struct intel_uncore_box *uncore_pmu_to_box(struct intel_uncore_pmu *pmu, int cpu) { - return pmu->boxes[topology_logical_package_id(cpu)]; + unsigned int pkgid = topology_logical_package_id(cpu); + + /* + * The unsigned check also catches the '-1' return value for non + * existent mappings in the topology map. + */ + return pkgid < max_packages ? pmu->boxes[pkgid] : NULL; } u64 uncore_msr_read_counter(struct intel_uncore_box *box, struct perf_event *event) @@ -1034,76 +1040,6 @@ static void uncore_pci_exit(void) } } -static int uncore_cpu_dying(unsigned int cpu) -{ - struct intel_uncore_type *type, **types = uncore_msr_uncores; - struct intel_uncore_pmu *pmu; - struct intel_uncore_box *box; - int i, pkg; - - pkg = topology_logical_package_id(cpu); - for (; *types; types++) { - type = *types; - pmu = type->pmus; - for (i = 0; i < type->num_boxes; i++, pmu++) { - box = pmu->boxes[pkg]; - if (box && atomic_dec_return(&box->refcnt) == 0) - uncore_box_exit(box); - } - } - return 0; -} - -static int uncore_cpu_starting(unsigned int cpu) -{ - struct intel_uncore_type *type, **types = uncore_msr_uncores; - struct intel_uncore_pmu *pmu; - struct intel_uncore_box *box; - int i, pkg; - - pkg = topology_logical_package_id(cpu); - for (; *types; types++) { - type = *types; - pmu = type->pmus; - for (i = 0; i < type->num_boxes; i++, pmu++) { - box = pmu->boxes[pkg]; - if (!box) - continue; - /* The first cpu on a package activates the box */ - if (atomic_inc_return(&box->refcnt) == 1) - uncore_box_init(box); - } - } - - return 0; -} - -static int uncore_cpu_prepare(unsigned int cpu) -{ - struct intel_uncore_type *type, **types = uncore_msr_uncores; - struct intel_uncore_pmu *pmu; - struct intel_uncore_box *box; - int i, pkg; - - pkg = topology_logical_package_id(cpu); - for (; *types; types++) { - type = *types; - pmu = type->pmus; - for (i = 0; i < type->num_boxes; i++, pmu++) { - if (pmu->boxes[pkg]) - continue; - /* First cpu of a package allocates the box */ - box = uncore_alloc_box(type, cpu_to_node(cpu)); - if (!box) - return -ENOMEM; - box->pmu = pmu; - box->pkgid = pkg; - pmu->boxes[pkg] = box; - } - } - return 0; -} - static void uncore_change_type_ctx(struct intel_uncore_type *type, int old_cpu, int new_cpu) { @@ -1143,12 +1079,14 @@ static void uncore_change_context(struct intel_uncore_type **uncores, static int uncore_event_cpu_offline(unsigned int cpu) { - int target; + struct intel_uncore_type *type, **types = uncore_msr_uncores; + struct intel_uncore_pmu *pmu; + struct intel_uncore_box *box; + int i, pkg, target; /* Check if exiting cpu is used for collecting uncore events */ if (!cpumask_test_and_clear_cpu(cpu, &uncore_cpu_mask)) - return 0; - + goto unref; /* Find a new cpu to collect uncore events */ target = cpumask_any_but(topology_core_cpumask(cpu), cpu); @@ -1160,12 +1098,82 @@ static int uncore_event_cpu_offline(unsigned int cpu) uncore_change_context(uncore_msr_uncores, cpu, target); uncore_change_context(uncore_pci_uncores, cpu, target); + +unref: + /* Clear the references */ + pkg = topology_logical_package_id(cpu); + for (; *types; types++) { + type = *types; + pmu = type->pmus; + for (i = 0; i < type->num_boxes; i++, pmu++) { + box = pmu->boxes[pkg]; + if (box && atomic_dec_return(&box->refcnt) == 0) + uncore_box_exit(box); + } + } return 0; } +static int allocate_boxes(struct intel_uncore_type **types, + unsigned int pkg, unsigned int cpu) +{ + struct intel_uncore_box *box, *tmp; + struct intel_uncore_type *type; + struct intel_uncore_pmu *pmu; + LIST_HEAD(allocated); + int i; + + /* Try to allocate all required boxes */ + for (; *types; types++) { + type = *types; + pmu = type->pmus; + for (i = 0; i < type->num_boxes; i++, pmu++) { + if (pmu->boxes[pkg]) + continue; + box = uncore_alloc_box(type, cpu_to_node(cpu)); + if (!box) + goto cleanup; + box->pmu = pmu; + box->pkgid = pkg; + list_add(&box->active_list, &allocated); + } + } + /* Install them in the pmus */ + list_for_each_entry_safe(box, tmp, &allocated, active_list) { + list_del_init(&box->active_list); + box->pmu->boxes[pkg] = box; + } + return 0; + +cleanup: + list_for_each_entry_safe(box, tmp, &allocated, active_list) { + list_del_init(&box->active_list); + kfree(box); + } + return -ENOMEM; +} + static int uncore_event_cpu_online(unsigned int cpu) { - int target; + struct intel_uncore_type *type, **types = uncore_msr_uncores; + struct intel_uncore_pmu *pmu; + struct intel_uncore_box *box; + int i, ret, pkg, target; + + pkg = topology_logical_package_id(cpu); + ret = allocate_boxes(types, pkg, cpu); + if (ret) + return ret; + + for (; *types; types++) { + type = *types; + pmu = type->pmus; + for (i = 0; i < type->num_boxes; i++, pmu++) { + box = pmu->boxes[pkg]; + if (!box && atomic_inc_return(&box->refcnt) == 1) + uncore_box_init(box); + } + } /* * Check if there is an online cpu in the package @@ -1355,33 +1363,13 @@ static int __init intel_uncore_init(void) if (cret && pret) return -ENODEV; - /* - * Install callbacks. Core will call them for each online cpu. - * - * The first online cpu of each package allocates and takes - * the refcounts for all other online cpus in that package. - * If msrs are not enabled no allocation is required and - * uncore_cpu_prepare() is not called for each online cpu. - */ - if (!cret) { - ret = cpuhp_setup_state(CPUHP_PERF_X86_UNCORE_PREP, - "perf/x86/intel/uncore:prepare", - uncore_cpu_prepare, NULL); - if (ret) - goto err; - } else { - cpuhp_setup_state_nocalls(CPUHP_PERF_X86_UNCORE_PREP, - "perf/x86/intel/uncore:prepare", - uncore_cpu_prepare, NULL); - } - - cpuhp_setup_state(CPUHP_AP_PERF_X86_UNCORE_STARTING, - "perf/x86/uncore:starting", - uncore_cpu_starting, uncore_cpu_dying); - - cpuhp_setup_state(CPUHP_AP_PERF_X86_UNCORE_ONLINE, - "perf/x86/uncore:online", - uncore_event_cpu_online, uncore_event_cpu_offline); + /* Install hotplug callbacks to setup the targets for each package */ + ret = cpuhp_setup_state(CPUHP_AP_PERF_X86_UNCORE_ONLINE, + "perf/x86/intel/uncore:online", + uncore_event_cpu_online, + uncore_event_cpu_offline); + if (ret) + goto err; return 0; err: @@ -1393,9 +1381,7 @@ module_init(intel_uncore_init); static void __exit intel_uncore_exit(void) { - cpuhp_remove_state_nocalls(CPUHP_AP_PERF_X86_UNCORE_ONLINE); - cpuhp_remove_state_nocalls(CPUHP_AP_PERF_X86_UNCORE_STARTING); - cpuhp_remove_state_nocalls(CPUHP_PERF_X86_UNCORE_PREP); + cpuhp_remove_state(CPUHP_AP_PERF_X86_UNCORE_ONLINE); uncore_types_exit(uncore_msr_uncores); uncore_pci_exit(); } diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h index 8329f3dc592c..921acaaa1601 100644 --- a/include/linux/cpuhotplug.h +++ b/include/linux/cpuhotplug.h @@ -8,7 +8,6 @@ enum cpuhp_state { CPUHP_CREATE_THREADS, CPUHP_PERF_PREPARE, CPUHP_PERF_X86_PREPARE, - CPUHP_PERF_X86_UNCORE_PREP, CPUHP_PERF_X86_AMD_UNCORE_PREP, CPUHP_PERF_BFIN, CPUHP_PERF_POWER, @@ -85,7 +84,6 @@ enum cpuhp_state { CPUHP_AP_IRQ_ARMADA_XP_STARTING, CPUHP_AP_IRQ_BCM2836_STARTING, CPUHP_AP_ARM_MVEBU_COHERENCY, - CPUHP_AP_PERF_X86_UNCORE_STARTING, CPUHP_AP_PERF_X86_AMD_UNCORE_STARTING, CPUHP_AP_PERF_X86_STARTING, CPUHP_AP_PERF_X86_AMD_IBS_STARTING, -- cgit v1.3-14-g43fede From 1a2a14444d32b89b28116daea86f63ced1716668 Mon Sep 17 00:00:00 2001 From: Dimitris Michailidis Date: Tue, 31 Jan 2017 16:03:13 -0800 Subject: net: fix ndo_features_check/ndo_fix_features comment ordering Commit cdba756f5803a2 ("net: move ndo_features_check() close to ndo_start_xmit()") inadvertently moved the doc comment for .ndo_fix_features instead of .ndo_features_check. Fix the comment ordering. Fixes: cdba756f5803a2 ("net: move ndo_features_check() close to ndo_start_xmit()") Signed-off-by: Dimitris Michailidis Acked-by: Eric Dumazet Signed-off-by: David S. Miller --- include/linux/netdevice.h | 29 +++++++++++++++-------------- 1 file changed, 15 insertions(+), 14 deletions(-) (limited to 'include/linux') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 9bde9558b596..70ad0291d517 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -866,11 +866,15 @@ struct netdev_xdp { * of useless work if you return NETDEV_TX_BUSY. * Required; cannot be NULL. * - * netdev_features_t (*ndo_fix_features)(struct net_device *dev, - * netdev_features_t features); - * Adjusts the requested feature flags according to device-specific - * constraints, and returns the resulting flags. Must not modify - * the device state. + * netdev_features_t (*ndo_features_check)(struct sk_buff *skb, + * struct net_device *dev + * netdev_features_t features); + * Called by core transmit path to determine if device is capable of + * performing offload operations on a given packet. This is to give + * the device an opportunity to implement any restrictions that cannot + * be otherwise expressed by feature flags. The check is called with + * the set of features that the stack has calculated and it returns + * those the driver believes to be appropriate. * * u16 (*ndo_select_queue)(struct net_device *dev, struct sk_buff *skb, * void *accel_priv, select_queue_fallback_t fallback); @@ -1028,6 +1032,12 @@ struct netdev_xdp { * Called to release previously enslaved netdev. * * Feature/offload setting functions. + * netdev_features_t (*ndo_fix_features)(struct net_device *dev, + * netdev_features_t features); + * Adjusts the requested feature flags according to device-specific + * constraints, and returns the resulting flags. Must not modify + * the device state. + * * int (*ndo_set_features)(struct net_device *dev, netdev_features_t features); * Called to update device configuration to new features. Passed * feature set might be less than what was returned by ndo_fix_features()). @@ -1100,15 +1110,6 @@ struct netdev_xdp { * Callback to use for xmit over the accelerated station. This * is used in place of ndo_start_xmit on accelerated net * devices. - * netdev_features_t (*ndo_features_check)(struct sk_buff *skb, - * struct net_device *dev - * netdev_features_t features); - * Called by core transmit path to determine if device is capable of - * performing offload operations on a given packet. This is to give - * the device an opportunity to implement any restrictions that cannot - * be otherwise expressed by feature flags. The check is called with - * the set of features that the stack has calculated and it returns - * those the driver believes to be appropriate. * int (*ndo_set_tx_maxrate)(struct net_device *dev, * int queue_index, u32 maxrate); * Called when a user wants to set a max-rate limitation of specific -- cgit v1.3-14-g43fede