summaryrefslogtreecommitdiffstats
path: root/kernel
diff options
context:
space:
mode:
authorTakashi Iwai <tiwai@suse.de>2019-02-08 08:20:32 -0500
committerTakashi Iwai <tiwai@suse.de>2019-02-08 08:20:32 -0500
commitd02cac152c97dffcb0cdd91e09b54fd6e2cca63d (patch)
tree68e4c6bd842703009f3edbf8f0e0e9326e4b2fad /kernel
parent36e4617c01153757cde9e5fcd375a75a8f8425c3 (diff)
parenta50e32694fbcdbf55875095258b9398e2eabd71f (diff)
Merge tag 'asoc-v5.1' of https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound into for-next
ASoC: Updates for v5.1 Lots and lots of new drivers so far, a highlight being the MediaTek BTCVSD which is a driver for a Bluetooth radio chip - the first such driver we've had upstream. Hopefully we will soon also see a baseband with an upstream driver! - Support for only powering up channels that are actively being used. - Quite a few improvements to simplify the generic card drivers, especially the merge of the SCU cards into the main generic drivers. - Lots of fixes for probing on Intel systems, trying to rationalize things to look more standard from a framework point of view. - New drivers for Asahi Kasei Microdevices AK4497, Cirrus Logic CS4341, Google ChromeOS embedded controllers, Ingenic JZ4725B, MediaTek BTCVSD, MT8183 and MT6358, NXP MICFIL, Rockchip RK3328, Spreadtrum DMA controllers, Qualcomm WCD9335, Xilinx S/PDIF and PCM formatters.
Diffstat (limited to 'kernel')
-rw-r--r--kernel/bpf/btf.c14
-rw-r--r--kernel/bpf/cgroup.c1
-rw-r--r--kernel/bpf/map_in_map.c17
-rw-r--r--kernel/bpf/stackmap.c12
-rw-r--r--kernel/bpf/verifier.c61
-rw-r--r--kernel/dma/swiotlb.c2
-rw-r--r--kernel/exit.c3
-rw-r--r--kernel/fork.c14
-rw-r--r--kernel/futex.c13
-rw-r--r--kernel/irq/irqdesc.c2
-rw-r--r--kernel/irq/manage.c3
-rw-r--r--kernel/locking/rwsem-xadd.c11
-rw-r--r--kernel/sched/core.c19
-rw-r--r--kernel/seccomp.c4
-rw-r--r--kernel/sys.c3
-rw-r--r--kernel/time/posix-cpu-timers.c1
-rw-r--r--kernel/trace/trace_kprobe.c12
-rw-r--r--kernel/umh.c33
18 files changed, 180 insertions, 45 deletions
diff --git a/kernel/bpf/btf.c b/kernel/bpf/btf.c
index 715f9fcf4712..befe570be5ba 100644
--- a/kernel/bpf/btf.c
+++ b/kernel/bpf/btf.c
@@ -467,7 +467,7 @@ static const struct btf_kind_operations *btf_type_ops(const struct btf_type *t)
467 return kind_ops[BTF_INFO_KIND(t->info)]; 467 return kind_ops[BTF_INFO_KIND(t->info)];
468} 468}
469 469
470bool btf_name_offset_valid(const struct btf *btf, u32 offset) 470static bool btf_name_offset_valid(const struct btf *btf, u32 offset)
471{ 471{
472 return BTF_STR_OFFSET_VALID(offset) && 472 return BTF_STR_OFFSET_VALID(offset) &&
473 offset < btf->hdr.str_len; 473 offset < btf->hdr.str_len;
@@ -1219,8 +1219,6 @@ static void btf_bitfield_seq_show(void *data, u8 bits_offset,
1219 u8 nr_copy_bits; 1219 u8 nr_copy_bits;
1220 u64 print_num; 1220 u64 print_num;
1221 1221
1222 data += BITS_ROUNDDOWN_BYTES(bits_offset);
1223 bits_offset = BITS_PER_BYTE_MASKED(bits_offset);
1224 nr_copy_bits = nr_bits + bits_offset; 1222 nr_copy_bits = nr_bits + bits_offset;
1225 nr_copy_bytes = BITS_ROUNDUP_BYTES(nr_copy_bits); 1223 nr_copy_bytes = BITS_ROUNDUP_BYTES(nr_copy_bits);
1226 1224
@@ -1255,7 +1253,9 @@ static void btf_int_bits_seq_show(const struct btf *btf,
1255 * BTF_INT_OFFSET() cannot exceed 64 bits. 1253 * BTF_INT_OFFSET() cannot exceed 64 bits.
1256 */ 1254 */
1257 total_bits_offset = bits_offset + BTF_INT_OFFSET(int_data); 1255 total_bits_offset = bits_offset + BTF_INT_OFFSET(int_data);
1258 btf_bitfield_seq_show(data, total_bits_offset, nr_bits, m); 1256 data += BITS_ROUNDDOWN_BYTES(total_bits_offset);
1257 bits_offset = BITS_PER_BYTE_MASKED(total_bits_offset);
1258 btf_bitfield_seq_show(data, bits_offset, nr_bits, m);
1259} 1259}
1260 1260
1261static void btf_int_seq_show(const struct btf *btf, const struct btf_type *t, 1261static void btf_int_seq_show(const struct btf *btf, const struct btf_type *t,
@@ -2001,12 +2001,12 @@ static void btf_struct_seq_show(const struct btf *btf, const struct btf_type *t,
2001 2001
2002 member_offset = btf_member_bit_offset(t, member); 2002 member_offset = btf_member_bit_offset(t, member);
2003 bitfield_size = btf_member_bitfield_size(t, member); 2003 bitfield_size = btf_member_bitfield_size(t, member);
2004 bytes_offset = BITS_ROUNDDOWN_BYTES(member_offset);
2005 bits8_offset = BITS_PER_BYTE_MASKED(member_offset);
2004 if (bitfield_size) { 2006 if (bitfield_size) {
2005 btf_bitfield_seq_show(data, member_offset, 2007 btf_bitfield_seq_show(data + bytes_offset, bits8_offset,
2006 bitfield_size, m); 2008 bitfield_size, m);
2007 } else { 2009 } else {
2008 bytes_offset = BITS_ROUNDDOWN_BYTES(member_offset);
2009 bits8_offset = BITS_PER_BYTE_MASKED(member_offset);
2010 ops = btf_type_ops(member_type); 2010 ops = btf_type_ops(member_type);
2011 ops->seq_show(btf, member_type, member->type, 2011 ops->seq_show(btf, member_type, member->type,
2012 data + bytes_offset, bits8_offset, m); 2012 data + bytes_offset, bits8_offset, m);
diff --git a/kernel/bpf/cgroup.c b/kernel/bpf/cgroup.c
index 9425c2fb872f..ab612fe9862f 100644
--- a/kernel/bpf/cgroup.c
+++ b/kernel/bpf/cgroup.c
@@ -718,6 +718,7 @@ cgroup_dev_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
718 case BPF_FUNC_trace_printk: 718 case BPF_FUNC_trace_printk:
719 if (capable(CAP_SYS_ADMIN)) 719 if (capable(CAP_SYS_ADMIN))
720 return bpf_get_trace_printk_proto(); 720 return bpf_get_trace_printk_proto();
721 /* fall through */
721 default: 722 default:
722 return NULL; 723 return NULL;
723 } 724 }
diff --git a/kernel/bpf/map_in_map.c b/kernel/bpf/map_in_map.c
index 99d243e1ad6e..52378d3e34b3 100644
--- a/kernel/bpf/map_in_map.c
+++ b/kernel/bpf/map_in_map.c
@@ -12,6 +12,7 @@
12struct bpf_map *bpf_map_meta_alloc(int inner_map_ufd) 12struct bpf_map *bpf_map_meta_alloc(int inner_map_ufd)
13{ 13{
14 struct bpf_map *inner_map, *inner_map_meta; 14 struct bpf_map *inner_map, *inner_map_meta;
15 u32 inner_map_meta_size;
15 struct fd f; 16 struct fd f;
16 17
17 f = fdget(inner_map_ufd); 18 f = fdget(inner_map_ufd);
@@ -36,7 +37,12 @@ struct bpf_map *bpf_map_meta_alloc(int inner_map_ufd)
36 return ERR_PTR(-EINVAL); 37 return ERR_PTR(-EINVAL);
37 } 38 }
38 39
39 inner_map_meta = kzalloc(sizeof(*inner_map_meta), GFP_USER); 40 inner_map_meta_size = sizeof(*inner_map_meta);
41 /* In some cases verifier needs to access beyond just base map. */
42 if (inner_map->ops == &array_map_ops)
43 inner_map_meta_size = sizeof(struct bpf_array);
44
45 inner_map_meta = kzalloc(inner_map_meta_size, GFP_USER);
40 if (!inner_map_meta) { 46 if (!inner_map_meta) {
41 fdput(f); 47 fdput(f);
42 return ERR_PTR(-ENOMEM); 48 return ERR_PTR(-ENOMEM);
@@ -46,9 +52,16 @@ struct bpf_map *bpf_map_meta_alloc(int inner_map_ufd)
46 inner_map_meta->key_size = inner_map->key_size; 52 inner_map_meta->key_size = inner_map->key_size;
47 inner_map_meta->value_size = inner_map->value_size; 53 inner_map_meta->value_size = inner_map->value_size;
48 inner_map_meta->map_flags = inner_map->map_flags; 54 inner_map_meta->map_flags = inner_map->map_flags;
49 inner_map_meta->ops = inner_map->ops;
50 inner_map_meta->max_entries = inner_map->max_entries; 55 inner_map_meta->max_entries = inner_map->max_entries;
51 56
57 /* Misc members not needed in bpf_map_meta_equal() check. */
58 inner_map_meta->ops = inner_map->ops;
59 if (inner_map->ops == &array_map_ops) {
60 inner_map_meta->unpriv_array = inner_map->unpriv_array;
61 container_of(inner_map_meta, struct bpf_array, map)->index_mask =
62 container_of(inner_map, struct bpf_array, map)->index_mask;
63 }
64
52 fdput(f); 65 fdput(f);
53 return inner_map_meta; 66 return inner_map_meta;
54} 67}
diff --git a/kernel/bpf/stackmap.c b/kernel/bpf/stackmap.c
index 90daf285de03..d43b14535827 100644
--- a/kernel/bpf/stackmap.c
+++ b/kernel/bpf/stackmap.c
@@ -180,11 +180,14 @@ static inline int stack_map_parse_build_id(void *page_addr,
180 180
181 if (nhdr->n_type == BPF_BUILD_ID && 181 if (nhdr->n_type == BPF_BUILD_ID &&
182 nhdr->n_namesz == sizeof("GNU") && 182 nhdr->n_namesz == sizeof("GNU") &&
183 nhdr->n_descsz == BPF_BUILD_ID_SIZE) { 183 nhdr->n_descsz > 0 &&
184 nhdr->n_descsz <= BPF_BUILD_ID_SIZE) {
184 memcpy(build_id, 185 memcpy(build_id,
185 note_start + note_offs + 186 note_start + note_offs +
186 ALIGN(sizeof("GNU"), 4) + sizeof(Elf32_Nhdr), 187 ALIGN(sizeof("GNU"), 4) + sizeof(Elf32_Nhdr),
187 BPF_BUILD_ID_SIZE); 188 nhdr->n_descsz);
189 memset(build_id + nhdr->n_descsz, 0,
190 BPF_BUILD_ID_SIZE - nhdr->n_descsz);
188 return 0; 191 return 0;
189 } 192 }
190 new_offs = note_offs + sizeof(Elf32_Nhdr) + 193 new_offs = note_offs + sizeof(Elf32_Nhdr) +
@@ -260,7 +263,7 @@ static int stack_map_get_build_id(struct vm_area_struct *vma,
260 return -EFAULT; /* page not mapped */ 263 return -EFAULT; /* page not mapped */
261 264
262 ret = -EINVAL; 265 ret = -EINVAL;
263 page_addr = page_address(page); 266 page_addr = kmap_atomic(page);
264 ehdr = (Elf32_Ehdr *)page_addr; 267 ehdr = (Elf32_Ehdr *)page_addr;
265 268
266 /* compare magic x7f "ELF" */ 269 /* compare magic x7f "ELF" */
@@ -276,6 +279,7 @@ static int stack_map_get_build_id(struct vm_area_struct *vma,
276 else if (ehdr->e_ident[EI_CLASS] == ELFCLASS64) 279 else if (ehdr->e_ident[EI_CLASS] == ELFCLASS64)
277 ret = stack_map_get_build_id_64(page_addr, build_id); 280 ret = stack_map_get_build_id_64(page_addr, build_id);
278out: 281out:
282 kunmap_atomic(page_addr);
279 put_page(page); 283 put_page(page);
280 return ret; 284 return ret;
281} 285}
@@ -310,6 +314,7 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs,
310 for (i = 0; i < trace_nr; i++) { 314 for (i = 0; i < trace_nr; i++) {
311 id_offs[i].status = BPF_STACK_BUILD_ID_IP; 315 id_offs[i].status = BPF_STACK_BUILD_ID_IP;
312 id_offs[i].ip = ips[i]; 316 id_offs[i].ip = ips[i];
317 memset(id_offs[i].build_id, 0, BPF_BUILD_ID_SIZE);
313 } 318 }
314 return; 319 return;
315 } 320 }
@@ -320,6 +325,7 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs,
320 /* per entry fall back to ips */ 325 /* per entry fall back to ips */
321 id_offs[i].status = BPF_STACK_BUILD_ID_IP; 326 id_offs[i].status = BPF_STACK_BUILD_ID_IP;
322 id_offs[i].ip = ips[i]; 327 id_offs[i].ip = ips[i];
328 memset(id_offs[i].build_id, 0, BPF_BUILD_ID_SIZE);
323 continue; 329 continue;
324 } 330 }
325 id_offs[i].offset = (vma->vm_pgoff << PAGE_SHIFT) + ips[i] 331 id_offs[i].offset = (vma->vm_pgoff << PAGE_SHIFT) + ips[i]
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index f6bc62a9ee8e..56674a7c3778 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -3103,6 +3103,40 @@ static int retrieve_ptr_limit(const struct bpf_reg_state *ptr_reg,
3103 } 3103 }
3104} 3104}
3105 3105
3106static bool can_skip_alu_sanitation(const struct bpf_verifier_env *env,
3107 const struct bpf_insn *insn)
3108{
3109 return env->allow_ptr_leaks || BPF_SRC(insn->code) == BPF_K;
3110}
3111
3112static int update_alu_sanitation_state(struct bpf_insn_aux_data *aux,
3113 u32 alu_state, u32 alu_limit)
3114{
3115 /* If we arrived here from different branches with different
3116 * state or limits to sanitize, then this won't work.
3117 */
3118 if (aux->alu_state &&
3119 (aux->alu_state != alu_state ||
3120 aux->alu_limit != alu_limit))
3121 return -EACCES;
3122
3123 /* Corresponding fixup done in fixup_bpf_calls(). */
3124 aux->alu_state = alu_state;
3125 aux->alu_limit = alu_limit;
3126 return 0;
3127}
3128
3129static int sanitize_val_alu(struct bpf_verifier_env *env,
3130 struct bpf_insn *insn)
3131{
3132 struct bpf_insn_aux_data *aux = cur_aux(env);
3133
3134 if (can_skip_alu_sanitation(env, insn))
3135 return 0;
3136
3137 return update_alu_sanitation_state(aux, BPF_ALU_NON_POINTER, 0);
3138}
3139
3106static int sanitize_ptr_alu(struct bpf_verifier_env *env, 3140static int sanitize_ptr_alu(struct bpf_verifier_env *env,
3107 struct bpf_insn *insn, 3141 struct bpf_insn *insn,
3108 const struct bpf_reg_state *ptr_reg, 3142 const struct bpf_reg_state *ptr_reg,
@@ -3117,7 +3151,7 @@ static int sanitize_ptr_alu(struct bpf_verifier_env *env,
3117 struct bpf_reg_state tmp; 3151 struct bpf_reg_state tmp;
3118 bool ret; 3152 bool ret;
3119 3153
3120 if (env->allow_ptr_leaks || BPF_SRC(insn->code) == BPF_K) 3154 if (can_skip_alu_sanitation(env, insn))
3121 return 0; 3155 return 0;
3122 3156
3123 /* We already marked aux for masking from non-speculative 3157 /* We already marked aux for masking from non-speculative
@@ -3133,19 +3167,8 @@ static int sanitize_ptr_alu(struct bpf_verifier_env *env,
3133 3167
3134 if (retrieve_ptr_limit(ptr_reg, &alu_limit, opcode, off_is_neg)) 3168 if (retrieve_ptr_limit(ptr_reg, &alu_limit, opcode, off_is_neg))
3135 return 0; 3169 return 0;
3136 3170 if (update_alu_sanitation_state(aux, alu_state, alu_limit))
3137 /* If we arrived here from different branches with different
3138 * limits to sanitize, then this won't work.
3139 */
3140 if (aux->alu_state &&
3141 (aux->alu_state != alu_state ||
3142 aux->alu_limit != alu_limit))
3143 return -EACCES; 3171 return -EACCES;
3144
3145 /* Corresponding fixup done in fixup_bpf_calls(). */
3146 aux->alu_state = alu_state;
3147 aux->alu_limit = alu_limit;
3148
3149do_sim: 3172do_sim:
3150 /* Simulate and find potential out-of-bounds access under 3173 /* Simulate and find potential out-of-bounds access under
3151 * speculative execution from truncation as a result of 3174 * speculative execution from truncation as a result of
@@ -3418,6 +3441,8 @@ static int adjust_scalar_min_max_vals(struct bpf_verifier_env *env,
3418 s64 smin_val, smax_val; 3441 s64 smin_val, smax_val;
3419 u64 umin_val, umax_val; 3442 u64 umin_val, umax_val;
3420 u64 insn_bitness = (BPF_CLASS(insn->code) == BPF_ALU64) ? 64 : 32; 3443 u64 insn_bitness = (BPF_CLASS(insn->code) == BPF_ALU64) ? 64 : 32;
3444 u32 dst = insn->dst_reg;
3445 int ret;
3421 3446
3422 if (insn_bitness == 32) { 3447 if (insn_bitness == 32) {
3423 /* Relevant for 32-bit RSH: Information can propagate towards 3448 /* Relevant for 32-bit RSH: Information can propagate towards
@@ -3452,6 +3477,11 @@ static int adjust_scalar_min_max_vals(struct bpf_verifier_env *env,
3452 3477
3453 switch (opcode) { 3478 switch (opcode) {
3454 case BPF_ADD: 3479 case BPF_ADD:
3480 ret = sanitize_val_alu(env, insn);
3481 if (ret < 0) {
3482 verbose(env, "R%d tried to add from different pointers or scalars\n", dst);
3483 return ret;
3484 }
3455 if (signed_add_overflows(dst_reg->smin_value, smin_val) || 3485 if (signed_add_overflows(dst_reg->smin_value, smin_val) ||
3456 signed_add_overflows(dst_reg->smax_value, smax_val)) { 3486 signed_add_overflows(dst_reg->smax_value, smax_val)) {
3457 dst_reg->smin_value = S64_MIN; 3487 dst_reg->smin_value = S64_MIN;
@@ -3471,6 +3501,11 @@ static int adjust_scalar_min_max_vals(struct bpf_verifier_env *env,
3471 dst_reg->var_off = tnum_add(dst_reg->var_off, src_reg.var_off); 3501 dst_reg->var_off = tnum_add(dst_reg->var_off, src_reg.var_off);
3472 break; 3502 break;
3473 case BPF_SUB: 3503 case BPF_SUB:
3504 ret = sanitize_val_alu(env, insn);
3505 if (ret < 0) {
3506 verbose(env, "R%d tried to sub from different pointers or scalars\n", dst);
3507 return ret;
3508 }
3474 if (signed_sub_overflows(dst_reg->smin_value, smax_val) || 3509 if (signed_sub_overflows(dst_reg->smin_value, smax_val) ||
3475 signed_sub_overflows(dst_reg->smax_value, smin_val)) { 3510 signed_sub_overflows(dst_reg->smax_value, smin_val)) {
3476 /* Overflow possible, we know nothing */ 3511 /* Overflow possible, we know nothing */
diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index d6361776dc5c..1fb6fd68b9c7 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -378,6 +378,8 @@ void __init swiotlb_exit(void)
378 memblock_free_late(io_tlb_start, 378 memblock_free_late(io_tlb_start,
379 PAGE_ALIGN(io_tlb_nslabs << IO_TLB_SHIFT)); 379 PAGE_ALIGN(io_tlb_nslabs << IO_TLB_SHIFT));
380 } 380 }
381 io_tlb_start = 0;
382 io_tlb_end = 0;
381 io_tlb_nslabs = 0; 383 io_tlb_nslabs = 0;
382 max_segment = 0; 384 max_segment = 0;
383} 385}
diff --git a/kernel/exit.c b/kernel/exit.c
index 2d14979577ee..3fb7be001964 100644
--- a/kernel/exit.c
+++ b/kernel/exit.c
@@ -307,7 +307,7 @@ void rcuwait_wake_up(struct rcuwait *w)
307 * MB (A) MB (B) 307 * MB (A) MB (B)
308 * [L] cond [L] tsk 308 * [L] cond [L] tsk
309 */ 309 */
310 smp_rmb(); /* (B) */ 310 smp_mb(); /* (B) */
311 311
312 /* 312 /*
313 * Avoid using task_rcu_dereference() magic as long as we are careful, 313 * Avoid using task_rcu_dereference() magic as long as we are careful,
@@ -866,6 +866,7 @@ void __noreturn do_exit(long code)
866 exit_task_namespaces(tsk); 866 exit_task_namespaces(tsk);
867 exit_task_work(tsk); 867 exit_task_work(tsk);
868 exit_thread(tsk); 868 exit_thread(tsk);
869 exit_umh(tsk);
869 870
870 /* 871 /*
871 * Flush inherited counters to the parent - before the parent 872 * Flush inherited counters to the parent - before the parent
diff --git a/kernel/fork.c b/kernel/fork.c
index a60459947f18..b69248e6f0e0 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -217,6 +217,7 @@ static unsigned long *alloc_thread_stack_node(struct task_struct *tsk, int node)
217 memset(s->addr, 0, THREAD_SIZE); 217 memset(s->addr, 0, THREAD_SIZE);
218 218
219 tsk->stack_vm_area = s; 219 tsk->stack_vm_area = s;
220 tsk->stack = s->addr;
220 return s->addr; 221 return s->addr;
221 } 222 }
222 223
@@ -1833,8 +1834,6 @@ static __latent_entropy struct task_struct *copy_process(
1833 1834
1834 posix_cpu_timers_init(p); 1835 posix_cpu_timers_init(p);
1835 1836
1836 p->start_time = ktime_get_ns();
1837 p->real_start_time = ktime_get_boot_ns();
1838 p->io_context = NULL; 1837 p->io_context = NULL;
1839 audit_set_context(p, NULL); 1838 audit_set_context(p, NULL);
1840 cgroup_fork(p); 1839 cgroup_fork(p);
@@ -2001,6 +2000,17 @@ static __latent_entropy struct task_struct *copy_process(
2001 goto bad_fork_free_pid; 2000 goto bad_fork_free_pid;
2002 2001
2003 /* 2002 /*
2003 * From this point on we must avoid any synchronous user-space
2004 * communication until we take the tasklist-lock. In particular, we do
2005 * not want user-space to be able to predict the process start-time by
2006 * stalling fork(2) after we recorded the start_time but before it is
2007 * visible to the system.
2008 */
2009
2010 p->start_time = ktime_get_ns();
2011 p->real_start_time = ktime_get_boot_ns();
2012
2013 /*
2004 * Make it visible to the rest of the system, but dont wake it up yet. 2014 * Make it visible to the rest of the system, but dont wake it up yet.
2005 * Need tasklist lock for parent etc handling! 2015 * Need tasklist lock for parent etc handling!
2006 */ 2016 */
diff --git a/kernel/futex.c b/kernel/futex.c
index be3bff2315ff..fdd312da0992 100644
--- a/kernel/futex.c
+++ b/kernel/futex.c
@@ -1452,11 +1452,7 @@ static void mark_wake_futex(struct wake_q_head *wake_q, struct futex_q *q)
1452 if (WARN(q->pi_state || q->rt_waiter, "refusing to wake PI futex\n")) 1452 if (WARN(q->pi_state || q->rt_waiter, "refusing to wake PI futex\n"))
1453 return; 1453 return;
1454 1454
1455 /* 1455 get_task_struct(p);
1456 * Queue the task for later wakeup for after we've released
1457 * the hb->lock. wake_q_add() grabs reference to p.
1458 */
1459 wake_q_add(wake_q, p);
1460 __unqueue_futex(q); 1456 __unqueue_futex(q);
1461 /* 1457 /*
1462 * The waiting task can free the futex_q as soon as q->lock_ptr = NULL 1458 * The waiting task can free the futex_q as soon as q->lock_ptr = NULL
@@ -1466,6 +1462,13 @@ static void mark_wake_futex(struct wake_q_head *wake_q, struct futex_q *q)
1466 * plist_del in __unqueue_futex(). 1462 * plist_del in __unqueue_futex().
1467 */ 1463 */
1468 smp_store_release(&q->lock_ptr, NULL); 1464 smp_store_release(&q->lock_ptr, NULL);
1465
1466 /*
1467 * Queue the task for later wakeup for after we've released
1468 * the hb->lock. wake_q_add() grabs reference to p.
1469 */
1470 wake_q_add(wake_q, p);
1471 put_task_struct(p);
1469} 1472}
1470 1473
1471/* 1474/*
diff --git a/kernel/irq/irqdesc.c b/kernel/irq/irqdesc.c
index ee062b7939d3..ef8ad36cadcf 100644
--- a/kernel/irq/irqdesc.c
+++ b/kernel/irq/irqdesc.c
@@ -457,7 +457,7 @@ static int alloc_descs(unsigned int start, unsigned int cnt, int node,
457 457
458 /* Validate affinity mask(s) */ 458 /* Validate affinity mask(s) */
459 if (affinity) { 459 if (affinity) {
460 for (i = 0; i < cnt; i++, i++) { 460 for (i = 0; i < cnt; i++) {
461 if (cpumask_empty(&affinity[i].mask)) 461 if (cpumask_empty(&affinity[i].mask))
462 return -EINVAL; 462 return -EINVAL;
463 } 463 }
diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c
index a4888ce4667a..84b54a17b95d 100644
--- a/kernel/irq/manage.c
+++ b/kernel/irq/manage.c
@@ -393,6 +393,9 @@ int irq_setup_affinity(struct irq_desc *desc)
393 } 393 }
394 394
395 cpumask_and(&mask, cpu_online_mask, set); 395 cpumask_and(&mask, cpu_online_mask, set);
396 if (cpumask_empty(&mask))
397 cpumask_copy(&mask, cpu_online_mask);
398
396 if (node != NUMA_NO_NODE) { 399 if (node != NUMA_NO_NODE) {
397 const struct cpumask *nodemask = cpumask_of_node(node); 400 const struct cpumask *nodemask = cpumask_of_node(node);
398 401
diff --git a/kernel/locking/rwsem-xadd.c b/kernel/locking/rwsem-xadd.c
index 09b180063ee1..50d9af615dc4 100644
--- a/kernel/locking/rwsem-xadd.c
+++ b/kernel/locking/rwsem-xadd.c
@@ -198,15 +198,22 @@ static void __rwsem_mark_wake(struct rw_semaphore *sem,
198 woken++; 198 woken++;
199 tsk = waiter->task; 199 tsk = waiter->task;
200 200
201 wake_q_add(wake_q, tsk); 201 get_task_struct(tsk);
202 list_del(&waiter->list); 202 list_del(&waiter->list);
203 /* 203 /*
204 * Ensure that the last operation is setting the reader 204 * Ensure calling get_task_struct() before setting the reader
205 * waiter to nil such that rwsem_down_read_failed() cannot 205 * waiter to nil such that rwsem_down_read_failed() cannot
206 * race with do_exit() by always holding a reference count 206 * race with do_exit() by always holding a reference count
207 * to the task to wakeup. 207 * to the task to wakeup.
208 */ 208 */
209 smp_store_release(&waiter->task, NULL); 209 smp_store_release(&waiter->task, NULL);
210 /*
211 * Ensure issuing the wakeup (either by us or someone else)
212 * after setting the reader waiter to nil.
213 */
214 wake_q_add(wake_q, tsk);
215 /* wake_q_add() already take the task ref */
216 put_task_struct(tsk);
210 } 217 }
211 218
212 adjustment = woken * RWSEM_ACTIVE_READ_BIAS - adjustment; 219 adjustment = woken * RWSEM_ACTIVE_READ_BIAS - adjustment;
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index a674c7db2f29..d8d76a65cfdd 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -396,6 +396,18 @@ static bool set_nr_if_polling(struct task_struct *p)
396#endif 396#endif
397#endif 397#endif
398 398
399/**
400 * wake_q_add() - queue a wakeup for 'later' waking.
401 * @head: the wake_q_head to add @task to
402 * @task: the task to queue for 'later' wakeup
403 *
404 * Queue a task for later wakeup, most likely by the wake_up_q() call in the
405 * same context, _HOWEVER_ this is not guaranteed, the wakeup can come
406 * instantly.
407 *
408 * This function must be used as-if it were wake_up_process(); IOW the task
409 * must be ready to be woken at this location.
410 */
399void wake_q_add(struct wake_q_head *head, struct task_struct *task) 411void wake_q_add(struct wake_q_head *head, struct task_struct *task)
400{ 412{
401 struct wake_q_node *node = &task->wake_q; 413 struct wake_q_node *node = &task->wake_q;
@@ -405,10 +417,11 @@ void wake_q_add(struct wake_q_head *head, struct task_struct *task)
405 * its already queued (either by us or someone else) and will get the 417 * its already queued (either by us or someone else) and will get the
406 * wakeup due to that. 418 * wakeup due to that.
407 * 419 *
408 * This cmpxchg() executes a full barrier, which pairs with the full 420 * In order to ensure that a pending wakeup will observe our pending
409 * barrier executed by the wakeup in wake_up_q(). 421 * state, even in the failed case, an explicit smp_mb() must be used.
410 */ 422 */
411 if (cmpxchg(&node->next, NULL, WAKE_Q_TAIL)) 423 smp_mb__before_atomic();
424 if (cmpxchg_relaxed(&node->next, NULL, WAKE_Q_TAIL))
412 return; 425 return;
413 426
414 get_task_struct(task); 427 get_task_struct(task);
diff --git a/kernel/seccomp.c b/kernel/seccomp.c
index d7f538847b84..e815781ed751 100644
--- a/kernel/seccomp.c
+++ b/kernel/seccomp.c
@@ -976,6 +976,9 @@ static int seccomp_notify_release(struct inode *inode, struct file *file)
976 struct seccomp_filter *filter = file->private_data; 976 struct seccomp_filter *filter = file->private_data;
977 struct seccomp_knotif *knotif; 977 struct seccomp_knotif *knotif;
978 978
979 if (!filter)
980 return 0;
981
979 mutex_lock(&filter->notify_lock); 982 mutex_lock(&filter->notify_lock);
980 983
981 /* 984 /*
@@ -1300,6 +1303,7 @@ out:
1300out_put_fd: 1303out_put_fd:
1301 if (flags & SECCOMP_FILTER_FLAG_NEW_LISTENER) { 1304 if (flags & SECCOMP_FILTER_FLAG_NEW_LISTENER) {
1302 if (ret < 0) { 1305 if (ret < 0) {
1306 listener_f->private_data = NULL;
1303 fput(listener_f); 1307 fput(listener_f);
1304 put_unused_fd(listener); 1308 put_unused_fd(listener);
1305 } else { 1309 } else {
diff --git a/kernel/sys.c b/kernel/sys.c
index a48cbf1414b8..f7eb62eceb24 100644
--- a/kernel/sys.c
+++ b/kernel/sys.c
@@ -1207,7 +1207,8 @@ DECLARE_RWSEM(uts_sem);
1207/* 1207/*
1208 * Work around broken programs that cannot handle "Linux 3.0". 1208 * Work around broken programs that cannot handle "Linux 3.0".
1209 * Instead we map 3.x to 2.6.40+x, so e.g. 3.0 would be 2.6.40 1209 * Instead we map 3.x to 2.6.40+x, so e.g. 3.0 would be 2.6.40
1210 * And we map 4.x to 2.6.60+x, so 4.0 would be 2.6.60. 1210 * And we map 4.x and later versions to 2.6.60+x, so 4.0/5.0/6.0/... would be
1211 * 2.6.60.
1211 */ 1212 */
1212static int override_release(char __user *release, size_t len) 1213static int override_release(char __user *release, size_t len)
1213{ 1214{
diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c
index 8f0644af40be..80f955210861 100644
--- a/kernel/time/posix-cpu-timers.c
+++ b/kernel/time/posix-cpu-timers.c
@@ -685,6 +685,7 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int timer_flags,
685 * set up the signal and overrun bookkeeping. 685 * set up the signal and overrun bookkeeping.
686 */ 686 */
687 timer->it.cpu.incr = timespec64_to_ns(&new->it_interval); 687 timer->it.cpu.incr = timespec64_to_ns(&new->it_interval);
688 timer->it_interval = ns_to_ktime(timer->it.cpu.incr);
688 689
689 /* 690 /*
690 * This acts as a modification timestamp for the timer, 691 * This acts as a modification timestamp for the timer,
diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c
index 5c19b8c41c7e..d5fb09ebba8b 100644
--- a/kernel/trace/trace_kprobe.c
+++ b/kernel/trace/trace_kprobe.c
@@ -607,11 +607,17 @@ static int trace_kprobe_create(int argc, const char *argv[])
607 char buf[MAX_EVENT_NAME_LEN]; 607 char buf[MAX_EVENT_NAME_LEN];
608 unsigned int flags = TPARG_FL_KERNEL; 608 unsigned int flags = TPARG_FL_KERNEL;
609 609
610 /* argc must be >= 1 */ 610 switch (argv[0][0]) {
611 if (argv[0][0] == 'r') { 611 case 'r':
612 is_return = true; 612 is_return = true;
613 flags |= TPARG_FL_RETURN; 613 flags |= TPARG_FL_RETURN;
614 } else if (argv[0][0] != 'p' || argc < 2) 614 break;
615 case 'p':
616 break;
617 default:
618 return -ECANCELED;
619 }
620 if (argc < 2)
615 return -ECANCELED; 621 return -ECANCELED;
616 622
617 event = strchr(&argv[0][1], ':'); 623 event = strchr(&argv[0][1], ':');
diff --git a/kernel/umh.c b/kernel/umh.c
index 0baa672e023c..d937cbad903a 100644
--- a/kernel/umh.c
+++ b/kernel/umh.c
@@ -37,6 +37,8 @@ static kernel_cap_t usermodehelper_bset = CAP_FULL_SET;
37static kernel_cap_t usermodehelper_inheritable = CAP_FULL_SET; 37static kernel_cap_t usermodehelper_inheritable = CAP_FULL_SET;
38static DEFINE_SPINLOCK(umh_sysctl_lock); 38static DEFINE_SPINLOCK(umh_sysctl_lock);
39static DECLARE_RWSEM(umhelper_sem); 39static DECLARE_RWSEM(umhelper_sem);
40static LIST_HEAD(umh_list);
41static DEFINE_MUTEX(umh_list_lock);
40 42
41static void call_usermodehelper_freeinfo(struct subprocess_info *info) 43static void call_usermodehelper_freeinfo(struct subprocess_info *info)
42{ 44{
@@ -100,10 +102,12 @@ static int call_usermodehelper_exec_async(void *data)
100 commit_creds(new); 102 commit_creds(new);
101 103
102 sub_info->pid = task_pid_nr(current); 104 sub_info->pid = task_pid_nr(current);
103 if (sub_info->file) 105 if (sub_info->file) {
104 retval = do_execve_file(sub_info->file, 106 retval = do_execve_file(sub_info->file,
105 sub_info->argv, sub_info->envp); 107 sub_info->argv, sub_info->envp);
106 else 108 if (!retval)
109 current->flags |= PF_UMH;
110 } else
107 retval = do_execve(getname_kernel(sub_info->path), 111 retval = do_execve(getname_kernel(sub_info->path),
108 (const char __user *const __user *)sub_info->argv, 112 (const char __user *const __user *)sub_info->argv,
109 (const char __user *const __user *)sub_info->envp); 113 (const char __user *const __user *)sub_info->envp);
@@ -517,6 +521,11 @@ int fork_usermode_blob(void *data, size_t len, struct umh_info *info)
517 goto out; 521 goto out;
518 522
519 err = call_usermodehelper_exec(sub_info, UMH_WAIT_EXEC); 523 err = call_usermodehelper_exec(sub_info, UMH_WAIT_EXEC);
524 if (!err) {
525 mutex_lock(&umh_list_lock);
526 list_add(&info->list, &umh_list);
527 mutex_unlock(&umh_list_lock);
528 }
520out: 529out:
521 fput(file); 530 fput(file);
522 return err; 531 return err;
@@ -679,6 +688,26 @@ static int proc_cap_handler(struct ctl_table *table, int write,
679 return 0; 688 return 0;
680} 689}
681 690
691void __exit_umh(struct task_struct *tsk)
692{
693 struct umh_info *info;
694 pid_t pid = tsk->pid;
695
696 mutex_lock(&umh_list_lock);
697 list_for_each_entry(info, &umh_list, list) {
698 if (info->pid == pid) {
699 list_del(&info->list);
700 mutex_unlock(&umh_list_lock);
701 goto out;
702 }
703 }
704 mutex_unlock(&umh_list_lock);
705 return;
706out:
707 if (info->cleanup)
708 info->cleanup(info);
709}
710
682struct ctl_table usermodehelper_table[] = { 711struct ctl_table usermodehelper_table[] = {
683 { 712 {
684 .procname = "bset", 713 .procname = "bset",