From b845c898b2f1ea458d5453f0fa1da6e2dfce3bb4 Mon Sep 17 00:00:00 2001 From: Daniel Borkmann Date: Tue, 21 Aug 2018 15:55:00 +0200 Subject: bpf, sockmap: fix sock_hash_alloc and reject zero-sized keys Currently, it is possible to create a sock hash map with key size of 0 and have the kernel return a fd back to user space. This is invalid for hash maps (and kernel also hasn't been tested for zero key size support in general at this point). Thus, reject such configuration. Fixes: 81110384441a ("bpf: sockmap, add hash map support") Signed-off-by: Daniel Borkmann Acked-by: John Fastabend Acked-by: Song Liu --- kernel/bpf/sockmap.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/bpf/sockmap.c b/kernel/bpf/sockmap.c index 98e621a29e8e..60ceb0e1fa56 100644 --- a/kernel/bpf/sockmap.c +++ b/kernel/bpf/sockmap.c @@ -2140,7 +2140,9 @@ static struct bpf_map *sock_hash_alloc(union bpf_attr *attr) return ERR_PTR(-EPERM); /* check sanity of attributes */ - if (attr->max_entries == 0 || attr->value_size != 4 || + if (attr->max_entries == 0 || + attr->key_size == 0 || + attr->value_size != 4 || attr->map_flags & ~SOCK_CREATE_FLAG_MASK) return ERR_PTR(-EINVAL); -- cgit v1.2.3 From eb29429d81e31b191f3b2bd19cf820279cec6463 Mon Sep 17 00:00:00 2001 From: Daniel Borkmann Date: Wed, 22 Aug 2018 18:09:17 +0200 Subject: bpf, sockmap: fix sock hash count in alloc_sock_hash_elem When we try to allocate a new sock hash entry and the allocation fails, then sock hash map fails to reduce the map element counter, meaning we keep accounting this element although it was never used. Fix it by dropping the element counter on error. Fixes: 81110384441a ("bpf: sockmap, add hash map support") Signed-off-by: Daniel Borkmann Acked-by: John Fastabend --- kernel/bpf/sockmap.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/bpf/sockmap.c b/kernel/bpf/sockmap.c index 60ceb0e1fa56..40c6ef9fc828 100644 --- a/kernel/bpf/sockmap.c +++ b/kernel/bpf/sockmap.c @@ -2269,8 +2269,10 @@ static struct htab_elem *alloc_sock_hash_elem(struct bpf_htab *htab, } l_new = kmalloc_node(htab->elem_size, GFP_ATOMIC | __GFP_NOWARN, htab->map.numa_node); - if (!l_new) + if (!l_new) { + atomic_dec(&htab->count); return ERR_PTR(-ENOMEM); + } memcpy(l_new->key, key, key_size); l_new->sk = sk; -- cgit v1.2.3 From 9b2e0388bec8ec5427403e23faff3b58dd1c3200 Mon Sep 17 00:00:00 2001 From: John Fastabend Date: Wed, 22 Aug 2018 08:37:37 -0700 Subject: bpf: sockmap: write_space events need to be passed to TCP handler When sockmap code is using the stream parser it also handles the write space events in order to handle the case where (a) verdict redirects skb to another socket and (b) the sockmap then sends the skb but due to memory constraints (or other EAGAIN errors) needs to do a retry. But the initial code missed a third case where the skb_send_sock_locked() triggers an sk_wait_event(). A typically case would be when sndbuf size is exceeded. If this happens because we do not pass the write_space event to the lower layers we never wake up the event and it will wait for sndtimeo. Which as noted in ktls fix may be rather large and look like a hang to the user. To reproduce the best test is to reduce the sndbuf size and send 1B data chunks to stress the memory handling. To fix this pass the event from the upper layer to the lower layer. Signed-off-by: John Fastabend Signed-off-by: Daniel Borkmann --- kernel/bpf/sockmap.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel') diff --git a/kernel/bpf/sockmap.c b/kernel/bpf/sockmap.c index 40c6ef9fc828..cf5195c7c331 100644 --- a/kernel/bpf/sockmap.c +++ b/kernel/bpf/sockmap.c @@ -1427,12 +1427,15 @@ out: static void smap_write_space(struct sock *sk) { struct smap_psock *psock; + void (*write_space)(struct sock *sk); rcu_read_lock(); psock = smap_psock_sk(sk); if (likely(psock && test_bit(SMAP_TX_RUNNING, &psock->state))) schedule_work(&psock->tx_work); + write_space = psock->save_write_space; rcu_read_unlock(); + write_space(sk); } static void smap_stop_sock(struct smap_psock *psock, struct sock *sk) -- cgit v1.2.3 From c0203475765f827e7b2eaf0a87222d0766e2cc4b Mon Sep 17 00:00:00 2001 From: Daniel Borkmann Date: Wed, 22 Aug 2018 23:49:37 +0200 Subject: bpf: use per htab salt for bucket hash All BPF hash and LRU maps currently have a known and global seed we feed into jhash() which is 0. This is suboptimal, thus fix it by generating a random seed upon hashtab setup time which we can later on feed into jhash() on lookup, update and deletions. Fixes: 0f8e4bd8a1fc8 ("bpf: add hashtable type of eBPF maps") Signed-off-by: Daniel Borkmann Acked-by: Alexei Starovoitov Acked-by: Song Liu Reviewed-by: Eduardo Valentin --- kernel/bpf/hashtab.c | 23 +++++++++++++---------- 1 file changed, 13 insertions(+), 10 deletions(-) (limited to 'kernel') diff --git a/kernel/bpf/hashtab.c b/kernel/bpf/hashtab.c index 04b8eda94e7d..03cc59ee9c95 100644 --- a/kernel/bpf/hashtab.c +++ b/kernel/bpf/hashtab.c @@ -15,6 +15,7 @@ #include #include #include +#include #include #include "percpu_freelist.h" #include "bpf_lru_list.h" @@ -41,6 +42,7 @@ struct bpf_htab { atomic_t count; /* number of elements in this hashtable */ u32 n_buckets; /* number of hash buckets */ u32 elem_size; /* size of each element in bytes */ + u32 hashrnd; }; /* each htab element is struct htab_elem + key + value */ @@ -371,6 +373,7 @@ static struct bpf_map *htab_map_alloc(union bpf_attr *attr) if (!htab->buckets) goto free_htab; + htab->hashrnd = get_random_int(); for (i = 0; i < htab->n_buckets; i++) { INIT_HLIST_NULLS_HEAD(&htab->buckets[i].head, i); raw_spin_lock_init(&htab->buckets[i].lock); @@ -402,9 +405,9 @@ free_htab: return ERR_PTR(err); } -static inline u32 htab_map_hash(const void *key, u32 key_len) +static inline u32 htab_map_hash(const void *key, u32 key_len, u32 hashrnd) { - return jhash(key, key_len, 0); + return jhash(key, key_len, hashrnd); } static inline struct bucket *__select_bucket(struct bpf_htab *htab, u32 hash) @@ -470,7 +473,7 @@ static void *__htab_map_lookup_elem(struct bpf_map *map, void *key) key_size = map->key_size; - hash = htab_map_hash(key, key_size); + hash = htab_map_hash(key, key_size, htab->hashrnd); head = select_bucket(htab, hash); @@ -597,7 +600,7 @@ static int htab_map_get_next_key(struct bpf_map *map, void *key, void *next_key) if (!key) goto find_first_elem; - hash = htab_map_hash(key, key_size); + hash = htab_map_hash(key, key_size, htab->hashrnd); head = select_bucket(htab, hash); @@ -824,7 +827,7 @@ static int htab_map_update_elem(struct bpf_map *map, void *key, void *value, key_size = map->key_size; - hash = htab_map_hash(key, key_size); + hash = htab_map_hash(key, key_size, htab->hashrnd); b = __select_bucket(htab, hash); head = &b->head; @@ -880,7 +883,7 @@ static int htab_lru_map_update_elem(struct bpf_map *map, void *key, void *value, key_size = map->key_size; - hash = htab_map_hash(key, key_size); + hash = htab_map_hash(key, key_size, htab->hashrnd); b = __select_bucket(htab, hash); head = &b->head; @@ -945,7 +948,7 @@ static int __htab_percpu_map_update_elem(struct bpf_map *map, void *key, key_size = map->key_size; - hash = htab_map_hash(key, key_size); + hash = htab_map_hash(key, key_size, htab->hashrnd); b = __select_bucket(htab, hash); head = &b->head; @@ -998,7 +1001,7 @@ static int __htab_lru_percpu_map_update_elem(struct bpf_map *map, void *key, key_size = map->key_size; - hash = htab_map_hash(key, key_size); + hash = htab_map_hash(key, key_size, htab->hashrnd); b = __select_bucket(htab, hash); head = &b->head; @@ -1071,7 +1074,7 @@ static int htab_map_delete_elem(struct bpf_map *map, void *key) key_size = map->key_size; - hash = htab_map_hash(key, key_size); + hash = htab_map_hash(key, key_size, htab->hashrnd); b = __select_bucket(htab, hash); head = &b->head; @@ -1103,7 +1106,7 @@ static int htab_lru_map_delete_elem(struct bpf_map *map, void *key) key_size = map->key_size; - hash = htab_map_hash(key, key_size); + hash = htab_map_hash(key, key_size, htab->hashrnd); b = __select_bucket(htab, hash); head = &b->head; -- cgit v1.2.3 From cb9d7fd51d9fbb329d182423bd7b92d0f8cb0e01 Mon Sep 17 00:00:00 2001 From: Vincent Whitchurch Date: Tue, 21 Aug 2018 17:25:07 +0200 Subject: watchdog: Mark watchdog touch functions as notrace Some architectures need to use stop_machine() to patch functions for ftrace, and the assumption is that the stopped CPUs do not make function calls to traceable functions when they are in the stopped state. Commit ce4f06dcbb5d ("stop_machine: Touch_nmi_watchdog() after MULTI_STOP_PREPARE") added calls to the watchdog touch functions from the stopped CPUs and those functions lack notrace annotations. This leads to crashes when enabling/disabling ftrace on ARM kernels built with the Thumb-2 instruction set. Fix it by adding the necessary notrace annotations. Fixes: ce4f06dcbb5d ("stop_machine: Touch_nmi_watchdog() after MULTI_STOP_PREPARE") Signed-off-by: Vincent Whitchurch Signed-off-by: Thomas Gleixner Cc: Peter Zijlstra Cc: oleg@redhat.com Cc: tj@kernel.org Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20180821152507.18313-1-vincent.whitchurch@axis.com --- kernel/watchdog.c | 4 ++-- kernel/watchdog_hld.c | 2 +- kernel/workqueue.c | 2 +- 3 files changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel') diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 5470dce212c0..977918d5d350 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -261,7 +261,7 @@ static void __touch_watchdog(void) * entering idle state. This should only be used for scheduler events. * Use touch_softlockup_watchdog() for everything else. */ -void touch_softlockup_watchdog_sched(void) +notrace void touch_softlockup_watchdog_sched(void) { /* * Preemption can be enabled. It doesn't matter which CPU's timestamp @@ -270,7 +270,7 @@ void touch_softlockup_watchdog_sched(void) raw_cpu_write(watchdog_touch_ts, 0); } -void touch_softlockup_watchdog(void) +notrace void touch_softlockup_watchdog(void) { touch_softlockup_watchdog_sched(); wq_watchdog_touch(raw_smp_processor_id()); diff --git a/kernel/watchdog_hld.c b/kernel/watchdog_hld.c index 1f7020d65d0a..71381168dede 100644 --- a/kernel/watchdog_hld.c +++ b/kernel/watchdog_hld.c @@ -29,7 +29,7 @@ static struct cpumask dead_events_mask; static unsigned long hardlockup_allcpu_dumped; static atomic_t watchdog_cpus = ATOMIC_INIT(0); -void arch_touch_nmi_watchdog(void) +notrace void arch_touch_nmi_watchdog(void) { /* * Using __raw here because some code paths have diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 60e80198c3df..0280deac392e 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -5574,7 +5574,7 @@ static void wq_watchdog_timer_fn(struct timer_list *unused) mod_timer(&wq_watchdog_timer, jiffies + thresh); } -void wq_watchdog_touch(int cpu) +notrace void wq_watchdog_touch(int cpu) { if (cpu >= 0) per_cpu(wq_watchdog_touched_cpu, cpu) = jiffies; -- cgit v1.2.3 From 13ba17bee18e321b073b49a88dcab10881f757da Mon Sep 17 00:00:00 2001 From: Mukesh Ojha Date: Fri, 24 Aug 2018 18:03:53 +0530 Subject: notifier: Remove notifier header file wherever not used The conversion of the hotplug notifiers to a state machine left the notifier.h includes around in some places. Remove them. Signed-off-by: Mukesh Ojha Signed-off-by: Thomas Gleixner Link: https://lkml.kernel.org/r/1535114033-4605-1-git-send-email-mojha@codeaurora.org --- fs/buffer.c | 1 - kernel/printk/printk.c | 1 - lib/percpu_counter.c | 1 - mm/page-writeback.c | 1 - mm/page_alloc.c | 1 - mm/slub.c | 1 - net/core/dev.c | 1 - 7 files changed, 7 deletions(-) (limited to 'kernel') diff --git a/fs/buffer.c b/fs/buffer.c index 4cc679d5bf58..6f1ae3ac9789 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -39,7 +39,6 @@ #include #include #include -#include #include #include #include diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c index 924e37fb1620..fd6f8ed28e01 100644 --- a/kernel/printk/printk.c +++ b/kernel/printk/printk.c @@ -38,7 +38,6 @@ #include #include #include -#include #include #include #include diff --git a/lib/percpu_counter.c b/lib/percpu_counter.c index c72577e472f2..a66595ba5543 100644 --- a/lib/percpu_counter.c +++ b/lib/percpu_counter.c @@ -4,7 +4,6 @@ */ #include -#include #include #include #include diff --git a/mm/page-writeback.c b/mm/page-writeback.c index 6551d3b0dc30..84ae9bf5858a 100644 --- a/mm/page-writeback.c +++ b/mm/page-writeback.c @@ -27,7 +27,6 @@ #include #include #include -#include #include #include #include diff --git a/mm/page_alloc.c b/mm/page_alloc.c index e75865d58ba7..05e983f42316 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -32,7 +32,6 @@ #include #include #include -#include #include #include #include diff --git a/mm/slub.c b/mm/slub.c index ce2b9e5cea77..8da34a8af53d 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -19,7 +19,6 @@ #include #include "slab.h" #include -#include #include #include #include diff --git a/net/core/dev.c b/net/core/dev.c index 325fc5088370..82114e1111e6 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -93,7 +93,6 @@ #include #include #include -#include #include #include #include -- cgit v1.2.3 From 6fb86d97207880c1286cd4cb3a7e6a598afbc727 Mon Sep 17 00:00:00 2001 From: Mukesh Ojha Date: Tue, 28 Aug 2018 12:24:54 +0530 Subject: cpu/hotplug: Remove skip_onerr field from cpuhp_step structure When notifiers were there, `skip_onerr` was used to avoid calling particular step startup/teardown callbacks in the CPU up/down rollback path, which made the hotplug asymmetric. As notifiers are gone now after the full state machine conversion, the `skip_onerr` field is no longer required. Remove it from the structure and its usage. Signed-off-by: Mukesh Ojha Signed-off-by: Thomas Gleixner Acked-by: Peter Zijlstra (Intel) Link: https://lkml.kernel.org/r/1535439294-31426-1-git-send-email-mojha@codeaurora.org --- kernel/cpu.c | 26 ++++---------------------- 1 file changed, 4 insertions(+), 22 deletions(-) (limited to 'kernel') diff --git a/kernel/cpu.c b/kernel/cpu.c index ed44d7d34c2d..aa7fe85ad62e 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -102,8 +102,6 @@ static inline void cpuhp_lock_release(bool bringup) { } * @name: Name of the step * @startup: Startup function of the step * @teardown: Teardown function of the step - * @skip_onerr: Do not invoke the functions on error rollback - * Will go away once the notifiers are gone * @cant_stop: Bringup/teardown can't be stopped at this step */ struct cpuhp_step { @@ -119,7 +117,6 @@ struct cpuhp_step { struct hlist_node *node); } teardown; struct hlist_head list; - bool skip_onerr; bool cant_stop; bool multi_instance; }; @@ -550,12 +547,8 @@ static int bringup_cpu(unsigned int cpu) static void undo_cpu_up(unsigned int cpu, struct cpuhp_cpu_state *st) { - for (st->state--; st->state > st->target; st->state--) { - struct cpuhp_step *step = cpuhp_get_step(st->state); - - if (!step->skip_onerr) - cpuhp_invoke_callback(cpu, st->state, false, NULL, NULL); - } + for (st->state--; st->state > st->target; st->state--) + cpuhp_invoke_callback(cpu, st->state, false, NULL, NULL); } static int cpuhp_up_callbacks(unsigned int cpu, struct cpuhp_cpu_state *st, @@ -644,12 +637,6 @@ static void cpuhp_thread_fun(unsigned int cpu) WARN_ON_ONCE(!cpuhp_is_ap_state(state)); - if (st->rollback) { - struct cpuhp_step *step = cpuhp_get_step(state); - if (step->skip_onerr) - goto next; - } - if (cpuhp_is_atomic_state(state)) { local_irq_disable(); st->result = cpuhp_invoke_callback(cpu, state, bringup, st->node, &st->last); @@ -673,7 +660,6 @@ static void cpuhp_thread_fun(unsigned int cpu) st->should_run = false; } -next: cpuhp_lock_release(bringup); if (!st->should_run) @@ -916,12 +902,8 @@ void cpuhp_report_idle_dead(void) static void undo_cpu_down(unsigned int cpu, struct cpuhp_cpu_state *st) { - for (st->state++; st->state < st->target; st->state++) { - struct cpuhp_step *step = cpuhp_get_step(st->state); - - if (!step->skip_onerr) - cpuhp_invoke_callback(cpu, st->state, true, NULL, NULL); - } + for (st->state++; st->state < st->target; st->state++) + cpuhp_invoke_callback(cpu, st->state, true, NULL, NULL); } static int cpuhp_down_callbacks(unsigned int cpu, struct cpuhp_cpu_state *st, -- cgit v1.2.3