Merge https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next

Daniel Borkmann says:

====================
pull-request: bpf-next 2022-03-04

We've added 32 non-merge commits during the last 14 day(s) which contain
a total of 59 files changed, 1038 insertions(+), 473 deletions(-).

The main changes are:

1) Optimize BPF stackmap's build_id retrieval by caching last valid build_id,
   as consecutive stack frames are likely to be in the same VMA and therefore
   have the same build id, from Hao Luo.

2) Several improvements to arm64 BPF JIT, that is, support for JITing
   the atomic[64]_fetch_add, atomic[64]_[fetch_]{and,or,xor} and lastly
   atomic[64]_{xchg|cmpxchg}. Also fix the BTF line info dump for JITed
   programs, from Hou Tao.

3) Optimize generic BPF map batch deletion by only enforcing synchronize_rcu()
   barrier once upon return to user space, from Eric Dumazet.

4) For kernel build parse DWARF and generate BTF through pahole with enabled
   multithreading, from Kui-Feng Lee.

5) BPF verifier usability improvements by making log info more concise and
   replacing inv with scalar type name, from Mykola Lysenko.

6) Two follow-up fixes for BPF prog JIT pack allocator, from Song Liu.

7) Add a new Kconfig to allow for loading kernel modules with non-matching
   BTF type info; their BTF info is then removed on load, from Connor O'Brien.

8) Remove reallocarray() usage from bpftool and switch to libbpf_reallocarray()
   in order to fix compilation errors for older glibc, from Mauricio Vásquez.

9) Fix libbpf to error on conflicting name in BTF when type declaration
   appears before the definition, from Xu Kuohai.

10) Fix issue in BPF preload for in-kernel light skeleton where loaded BPF
    program fds prevent init process from setting up fd 0-2, from Yucong Sun.

11) Fix libbpf reuse of pinned perf RB map when max_entries is auto-determined
    by libbpf, from Stijn Tintel.

12) Several cleanups for libbpf and a fix to enforce perf RB map #pages to be
    non-zero, from Yuntao Wang.

* https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (32 commits)
  bpf: Small BPF verifier log improvements
  libbpf: Add a check to ensure that page_cnt is non-zero
  bpf, x86: Set header->size properly before freeing it
  x86: Disable HAVE_ARCH_HUGE_VMALLOC on 32-bit x86
  bpf, test_run: Fix overflow in XDP frags bpf_test_finish
  selftests/bpf: Update btf_dump case for conflicting names
  libbpf: Skip forward declaration when counting duplicated type names
  bpf: Add some description about BPF_JIT_ALWAYS_ON in Kconfig
  bpf, docs: Add a missing colon in verifier.rst
  bpf: Cache the last valid build_id
  libbpf: Fix BPF_MAP_TYPE_PERF_EVENT_ARRAY auto-pinning
  bpf, selftests: Use raw_tp program for atomic test
  bpf, arm64: Support more atomic operations
  bpftool: Remove redundant slashes
  bpf: Add config to allow loading modules with BTF mismatches
  bpf, arm64: Feed byte-offset into bpf line info
  bpf, arm64: Call build_prologue() first in first JIT pass
  bpf: Fix issue with bpf preload module taking over stdout/stdin of kernel.
  bpftool: Bpf skeletons assert type sizes
  bpf: Cleanup comments
  ...
====================

Link: https://lore.kernel.org/r/20220304164313.31675-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

14 files changed, 82 insertions, 50 deletions

diff --git a/kernel/bpf/Kconfig b/kernel/bpf/Kconfig
index d24d518ddd63..c3cf0b86eeb2 100644
--- a/kernel/bpf/Kconfig
+++ b/kernel/bpf/Kconfig
@@ -58,6 +58,10 @@ config BPF_JIT_ALWAYS_ON
 	  Enables BPF JIT and removes BPF interpreter to avoid speculative
 	  execution of BPF instructions by the interpreter.
 
+	  When CONFIG_BPF_JIT_ALWAYS_ON is enabled, /proc/sys/net/core/bpf_jit_enable
+	  is permanently set to 1 and setting any other value than that will
+	  return failure.
+
 config BPF_JIT_DEFAULT_ON
 	def_bool ARCH_WANT_DEFAULT_BPF_JIT || BPF_JIT_ALWAYS_ON
 	depends on HAVE_EBPF_JIT && BPF_JIT
diff --git a/kernel/bpf/bpf_local_storage.c b/kernel/bpf/bpf_local_storage.c
index 71de2a89869c..092a1ac772d7 100644
--- a/kernel/bpf/bpf_local_storage.c
+++ b/kernel/bpf/bpf_local_storage.c
@@ -136,7 +136,7 @@ bool bpf_selem_unlink_storage_nolock(struct bpf_local_storage *local_storage,
 		 * will be done by the caller.
 		 *
 		 * Although the unlock will be done under
-		 * rcu_read_lock(),  it is more intutivie to
+		 * rcu_read_lock(),  it is more intuitive to
 		 * read if the freeing of the storage is done
 		 * after the raw_spin_unlock_bh(&local_storage->lock).
 		 *
diff --git a/kernel/bpf/btf.c b/kernel/bpf/btf.c
index 02d7014417a0..b472cf0c8fdb 100644
--- a/kernel/bpf/btf.c
+++ b/kernel/bpf/btf.c
@@ -1,4 +1,4 @@
-/* SPDX-License-Identifier: GPL-2.0 */
+// SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2018 Facebook */
 
 #include <uapi/linux/btf.h>
@@ -2547,7 +2547,7 @@ static int btf_ptr_resolve(struct btf_verifier_env *env,
 	 *
 	 * We now need to continue from the last-resolved-ptr to
 	 * ensure the last-resolved-ptr will not referring back to
-	 * the currenct ptr (t).
+	 * the current ptr (t).
 	 */
 	if (btf_type_is_modifier(next_type)) {
 		const struct btf_type *resolved_type;
@@ -6148,7 +6148,7 @@ int btf_type_snprintf_show(const struct btf *btf, u32 type_id, void *obj,
 
 	btf_type_show(btf, type_id, obj, (struct btf_show *)&ssnprintf);
 
-	/* If we encontered an error, return it. */
+	/* If we encountered an error, return it. */
 	if (ssnprintf.show.state.status)
 		return ssnprintf.show.state.status;
 
@@ -6398,7 +6398,8 @@ static int btf_module_notify(struct notifier_block *nb, unsigned long op,
 			pr_warn("failed to validate module [%s] BTF: %ld\n",
 				mod->name, PTR_ERR(btf));
 			kfree(btf_mod);
-			err = PTR_ERR(btf);
+			if (!IS_ENABLED(CONFIG_MODULE_ALLOW_BTF_MISMATCH))
+				err = PTR_ERR(btf);
 			goto out;
 		}
 		err = btf_alloc_id(btf);
@@ -6706,7 +6707,7 @@ static int btf_populate_kfunc_set(struct btf *btf, enum btf_kfunc_hook hook,
 				  const struct btf_kfunc_id_set *kset)
 {
 	bool vmlinux_set = !btf_is_module(btf);
-	int type, ret;
+	int type, ret = 0;
 
 	for (type = 0; type < ARRAY_SIZE(kset->sets); type++) {
 		if (!kset->sets[type])
diff --git a/kernel/bpf/cgroup.c b/kernel/bpf/cgroup.c
index 098632fdbc45..128028efda64 100644
--- a/kernel/bpf/cgroup.c
+++ b/kernel/bpf/cgroup.c
@@ -1031,7 +1031,7 @@ int cgroup_bpf_prog_query(const union bpf_attr *attr,
  * __cgroup_bpf_run_filter_skb() - Run a program for packet filtering
  * @sk: The socket sending or receiving traffic
  * @skb: The skb that is being sent or received
- * @type: The type of program to be exectuted
+ * @type: The type of program to be executed
  *
  * If no socket is passed, or the socket is not of type INET or INET6,
  * this function does nothing and returns 0.
@@ -1094,7 +1094,7 @@ EXPORT_SYMBOL(__cgroup_bpf_run_filter_skb);
 /**
  * __cgroup_bpf_run_filter_sk() - Run a program on a sock
  * @sk: sock structure to manipulate
- * @type: The type of program to be exectuted
+ * @type: The type of program to be executed
  *
  * socket is passed is expected to be of type INET or INET6.
  *
@@ -1119,7 +1119,7 @@ EXPORT_SYMBOL(__cgroup_bpf_run_filter_sk);
  *                                       provided by user sockaddr
  * @sk: sock struct that will use sockaddr
  * @uaddr: sockaddr struct provided by user
- * @type: The type of program to be exectuted
+ * @type: The type of program to be executed
  * @t_ctx: Pointer to attach type specific context
  * @flags: Pointer to u32 which contains higher bits of BPF program
  *         return value (OR'ed together).
@@ -1166,7 +1166,7 @@ EXPORT_SYMBOL(__cgroup_bpf_run_filter_sock_addr);
  * @sock_ops: bpf_sock_ops_kern struct to pass to program. Contains
  * sk with connection information (IP addresses, etc.) May not contain
  * cgroup info if it is a req sock.
- * @type: The type of program to be exectuted
+ * @type: The type of program to be executed
  *
  * socket passed is expected to be of type INET or INET6.
  *
diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
index ebb0193d07f0..ab630f773ec1 100644
--- a/kernel/bpf/core.c
+++ b/kernel/bpf/core.c
@@ -1112,13 +1112,16 @@ int bpf_jit_binary_pack_finalize(struct bpf_prog *prog,
  *   1) when the program is freed after;
  *   2) when the JIT engine fails (before bpf_jit_binary_pack_finalize).
  * For case 2), we need to free both the RO memory and the RW buffer.
- * Also, ro_header->size in 2) is not properly set yet, so rw_header->size
- * is used for uncharge.
+ *
+ * bpf_jit_binary_pack_free requires proper ro_header->size. However,
+ * bpf_jit_binary_pack_alloc does not set it. Therefore, ro_header->size
+ * must be set with either bpf_jit_binary_pack_finalize (normal path) or
+ * bpf_arch_text_copy (when jit fails).
  */
 void bpf_jit_binary_pack_free(struct bpf_binary_header *ro_header,
 			      struct bpf_binary_header *rw_header)
 {
-	u32 size = rw_header ? rw_header->size : ro_header->size;
+	u32 size = ro_header->size;
 
 	bpf_prog_pack_free(ro_header);
 	kvfree(rw_header);
diff --git a/kernel/bpf/hashtab.c b/kernel/bpf/hashtab.c
index d29af9988f37..65877967f414 100644
--- a/kernel/bpf/hashtab.c
+++ b/kernel/bpf/hashtab.c
@@ -1636,7 +1636,7 @@ __htab_map_lookup_and_delete_batch(struct bpf_map *map,
 		value_size = size * num_possible_cpus();
 	total = 0;
 	/* while experimenting with hash tables with sizes ranging from 10 to
-	 * 1000, it was observed that a bucket can have upto 5 entries.
+	 * 1000, it was observed that a bucket can have up to 5 entries.
 	 */
 	bucket_size = 5;
 
diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c
index 49817755b8c3..ae64110a98b5 100644
--- a/kernel/bpf/helpers.c
+++ b/kernel/bpf/helpers.c
@@ -1093,7 +1093,7 @@ struct bpf_hrtimer {
 struct bpf_timer_kern {
 	struct bpf_hrtimer *timer;
 	/* bpf_spin_lock is used here instead of spinlock_t to make
-	 * sure that it always fits into space resereved by struct bpf_timer
+	 * sure that it always fits into space reserved by struct bpf_timer
 	 * regardless of LOCKDEP and spinlock debug flags.
 	 */
 	struct bpf_spin_lock lock;
diff --git a/kernel/bpf/local_storage.c b/kernel/bpf/local_storage.c
index 23f7f9d08a62..497916060ac7 100644
--- a/kernel/bpf/local_storage.c
+++ b/kernel/bpf/local_storage.c
@@ -1,4 +1,4 @@
-//SPDX-License-Identifier: GPL-2.0
+// SPDX-License-Identifier: GPL-2.0
 #include <linux/bpf-cgroup.h>
 #include <linux/bpf.h>
 #include <linux/bpf_local_storage.h>
diff --git a/kernel/bpf/preload/bpf_preload_kern.c b/kernel/bpf/preload/bpf_preload_kern.c
index 30207c048d36..5106b5372f0c 100644
--- a/kernel/bpf/preload/bpf_preload_kern.c
+++ b/kernel/bpf/preload/bpf_preload_kern.c
@@ -54,6 +54,13 @@ static int load_skel(void)
 		err = PTR_ERR(progs_link);
 		goto out;
 	}
+	/* Avoid taking over stdin/stdout/stderr of init process. Zeroing out
+	 * makes skel_closenz() a no-op later in iterators_bpf__destroy().
+	 */
+	close_fd(skel->links.dump_bpf_map_fd);
+	skel->links.dump_bpf_map_fd = 0;
+	close_fd(skel->links.dump_bpf_prog_fd);
+	skel->links.dump_bpf_prog_fd = 0;
 	return 0;
 out:
 	free_links_and_skel();
diff --git a/kernel/bpf/reuseport_array.c b/kernel/bpf/reuseport_array.c
index 556a769b5b80..8251243022a2 100644
--- a/kernel/bpf/reuseport_array.c
+++ b/kernel/bpf/reuseport_array.c
@@ -143,7 +143,7 @@ static void reuseport_array_free(struct bpf_map *map)
 
 	/*
 	 * Once reaching here, all sk->sk_user_data is not
-	 * referenceing this "array".  "array" can be freed now.
+	 * referencing this "array". "array" can be freed now.
 	 */
 	bpf_map_area_free(array);
 }
diff --git a/kernel/bpf/stackmap.c b/kernel/bpf/stackmap.c
index 22c8ae94e4c1..38bdfcd06f55 100644
--- a/kernel/bpf/stackmap.c
+++ b/kernel/bpf/stackmap.c
@@ -132,7 +132,8 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs,
 	int i;
 	struct mmap_unlock_irq_work *work = NULL;
 	bool irq_work_busy = bpf_mmap_unlock_get_irq_work(&work);
-	struct vm_area_struct *vma;
+	struct vm_area_struct *vma, *prev_vma = NULL;
+	const char *prev_build_id;
 
 	/* If the irq_work is in use, fall back to report ips. Same
 	 * fallback is used for kernel stack (!user) on a stackmap with
@@ -150,6 +151,12 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs,
 	}
 
 	for (i = 0; i < trace_nr; i++) {
+		if (range_in_vma(prev_vma, ips[i], ips[i])) {
+			vma = prev_vma;
+			memcpy(id_offs[i].build_id, prev_build_id,
+			       BUILD_ID_SIZE_MAX);
+			goto build_id_valid;
+		}
 		vma = find_vma(current->mm, ips[i]);
 		if (!vma || build_id_parse(vma, id_offs[i].build_id, NULL)) {
 			/* per entry fall back to ips */
@@ -158,9 +165,12 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs,
 			memset(id_offs[i].build_id, 0, BUILD_ID_SIZE_MAX);
 			continue;
 		}
+build_id_valid:
 		id_offs[i].offset = (vma->vm_pgoff << PAGE_SHIFT) + ips[i]
 			- vma->vm_start;
 		id_offs[i].status = BPF_STACK_BUILD_ID_VALID;
+		prev_vma = vma;
+		prev_build_id = id_offs[i].build_id;
 	}
 	bpf_mmap_unlock_mm(work, current->mm);
 }
diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
index a72f63d5a7da..db402ebc5570 100644
--- a/kernel/bpf/syscall.c
+++ b/kernel/bpf/syscall.c
@@ -1352,7 +1352,6 @@ int generic_map_delete_batch(struct bpf_map *map,
 		err = map->ops->map_delete_elem(map, key);
 		rcu_read_unlock();
 		bpf_enable_instrumentation();
-		maybe_wait_bpf_programs(map);
 		if (err)
 			break;
 		cond_resched();
@@ -1361,6 +1360,8 @@ int generic_map_delete_batch(struct bpf_map *map,
 		err = -EFAULT;
 
 	kvfree(key);
+
+	maybe_wait_bpf_programs(map);
 	return err;
 }
 
@@ -2565,7 +2566,7 @@ static int bpf_link_alloc_id(struct bpf_link *link)
  * pre-allocated resources are to be freed with bpf_cleanup() call. All the
  * transient state is passed around in struct bpf_link_primer.
  * This is preferred way to create and initialize bpf_link, especially when
- * there are complicated and expensive operations inbetween creating bpf_link
+ * there are complicated and expensive operations in between creating bpf_link
  * itself and attaching it to BPF hook. By using bpf_link_prime() and
  * bpf_link_settle() kernel code using bpf_link doesn't have to perform
  * expensive (and potentially failing) roll back operations in a rare case
diff --git a/kernel/bpf/trampoline.c b/kernel/bpf/trampoline.c
index 7224691df2ec..0b41fa993825 100644
--- a/kernel/bpf/trampoline.c
+++ b/kernel/bpf/trampoline.c
@@ -45,7 +45,7 @@ void *bpf_jit_alloc_exec_page(void)
 
 	set_vm_flush_reset_perms(image);
 	/* Keep image as writeable. The alternative is to keep flipping ro/rw
-	 * everytime new program is attached or detached.
+	 * every time new program is attached or detached.
 	 */
 	set_memory_x((long)image, 1);
 	return image;
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index d7473fee247c..a57db4b2803c 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -539,7 +539,7 @@ static const char *reg_type_str(struct bpf_verifier_env *env,
 	char postfix[16] = {0}, prefix[32] = {0};
 	static const char * const str[] = {
 		[NOT_INIT]		= "?",
-		[SCALAR_VALUE]		= "inv",
+		[SCALAR_VALUE]		= "scalar",
 		[PTR_TO_CTX]		= "ctx",
 		[CONST_PTR_TO_MAP]	= "map_ptr",
 		[PTR_TO_MAP_VALUE]	= "map_value",
@@ -685,74 +685,80 @@ static void print_verifier_state(struct bpf_verifier_env *env,
 			continue;
 		verbose(env, " R%d", i);
 		print_liveness(env, reg->live);
-		verbose(env, "=%s", reg_type_str(env, t));
+		verbose(env, "=");
 		if (t == SCALAR_VALUE && reg->precise)
 			verbose(env, "P");
 		if ((t == SCALAR_VALUE || t == PTR_TO_STACK) &&
 		    tnum_is_const(reg->var_off)) {
 			/* reg->off should be 0 for SCALAR_VALUE */
+			verbose(env, "%s", t == SCALAR_VALUE ? "" : reg_type_str(env, t));
 			verbose(env, "%lld", reg->var_off.value + reg->off);
 		} else {
+			const char *sep = "";
+
+			verbose(env, "%s", reg_type_str(env, t));
 			if (base_type(t) == PTR_TO_BTF_ID ||
 			    base_type(t) == PTR_TO_PERCPU_BTF_ID)
 				verbose(env, "%s", kernel_type_name(reg->btf, reg->btf_id));
-			verbose(env, "(id=%d", reg->id);
-			if (reg_type_may_be_refcounted_or_null(t))
-				verbose(env, ",ref_obj_id=%d", reg->ref_obj_id);
+			verbose(env, "(");
+/*
+ * _a stands for append, was shortened to avoid multiline statements below.
+ * This macro is used to output a comma separated list of attributes.
+ */
+#define verbose_a(fmt, ...) ({ verbose(env, "%s" fmt, sep, __VA_ARGS__); sep = ","; })
+
+			if (reg->id)
+				verbose_a("id=%d", reg->id);
+			if (reg_type_may_be_refcounted_or_null(t) && reg->ref_obj_id)
+				verbose_a("ref_obj_id=%d", reg->ref_obj_id);
 			if (t != SCALAR_VALUE)
-				verbose(env, ",off=%d", reg->off);
+				verbose_a("off=%d", reg->off);
 			if (type_is_pkt_pointer(t))
-				verbose(env, ",r=%d", reg->range);
+				verbose_a("r=%d", reg->range);
 			else if (base_type(t) == CONST_PTR_TO_MAP ||
 				 base_type(t) == PTR_TO_MAP_KEY ||
 				 base_type(t) == PTR_TO_MAP_VALUE)
-				verbose(env, ",ks=%d,vs=%d",
-					reg->map_ptr->key_size,
-					reg->map_ptr->value_size);
+				verbose_a("ks=%d,vs=%d",
+					  reg->map_ptr->key_size,
+					  reg->map_ptr->value_size);
 			if (tnum_is_const(reg->var_off)) {
 				/* Typically an immediate SCALAR_VALUE, but
 				 * could be a pointer whose offset is too big
 				 * for reg->off
 				 */
-				verbose(env, ",imm=%llx", reg->var_off.value);
+				verbose_a("imm=%llx", reg->var_off.value);
 			} else {
 				if (reg->smin_value != reg->umin_value &&
 				    reg->smin_value != S64_MIN)
-					verbose(env, ",smin_value=%lld",
-						(long long)reg->smin_value);
+					verbose_a("smin=%lld", (long long)reg->smin_value);
 				if (reg->smax_value != reg->umax_value &&
 				    reg->smax_value != S64_MAX)
-					verbose(env, ",smax_value=%lld",
-						(long long)reg->smax_value);
+					verbose_a("smax=%lld", (long long)reg->smax_value);
 				if (reg->umin_value != 0)
-					verbose(env, ",umin_value=%llu",
-						(unsigned long long)reg->umin_value);
+					verbose_a("umin=%llu", (unsigned long long)reg->umin_value);
 				if (reg->umax_value != U64_MAX)
-					verbose(env, ",umax_value=%llu",
-						(unsigned long long)reg->umax_value);
+					verbose_a("umax=%llu", (unsigned long long)reg->umax_value);
 				if (!tnum_is_unknown(reg->var_off)) {
 					char tn_buf[48];
 
 					tnum_strn(tn_buf, sizeof(tn_buf), reg->var_off);
-					verbose(env, ",var_off=%s", tn_buf);
+					verbose_a("var_off=%s", tn_buf);
 				}
 				if (reg->s32_min_value != reg->smin_value &&
 				    reg->s32_min_value != S32_MIN)
-					verbose(env, ",s32_min_value=%d",
-						(int)(reg->s32_min_value));
+					verbose_a("s32_min=%d", (int)(reg->s32_min_value));
 				if (reg->s32_max_value != reg->smax_value &&
 				    reg->s32_max_value != S32_MAX)
-					verbose(env, ",s32_max_value=%d",
-						(int)(reg->s32_max_value));
+					verbose_a("s32_max=%d", (int)(reg->s32_max_value));
 				if (reg->u32_min_value != reg->umin_value &&
 				    reg->u32_min_value != U32_MIN)
-					verbose(env, ",u32_min_value=%d",
-						(int)(reg->u32_min_value));
+					verbose_a("u32_min=%d", (int)(reg->u32_min_value));
 				if (reg->u32_max_value != reg->umax_value &&
 				    reg->u32_max_value != U32_MAX)
-					verbose(env, ",u32_max_value=%d",
-						(int)(reg->u32_max_value));
+					verbose_a("u32_max=%d", (int)(reg->u32_max_value));
 			}
+#undef verbose_a
+
 			verbose(env, ")");
 		}
 	}
@@ -777,7 +783,7 @@ static void print_verifier_state(struct bpf_verifier_env *env,
 		if (is_spilled_reg(&state->stack[i])) {
 			reg = &state->stack[i].spilled_ptr;
 			t = reg->type;
-			verbose(env, "=%s", reg_type_str(env, t));
+			verbose(env, "=%s", t == SCALAR_VALUE ? "" : reg_type_str(env, t));
 			if (t == SCALAR_VALUE && reg->precise)
 				verbose(env, "P");
 			if (t == SCALAR_VALUE && tnum_is_const(reg->var_off))