Merge branch 'bpf-misc'

Daniel Borkmann says:

====================
Misc BPF improvements

This last series for this window adds various misc
improvements to BPF, one is to mark registered map and
prog types as __ro_after_init, another one for removing
cBPF stubs in eBPF JITs and moving the stub to the core
and last also improving JITs is to make generated images
visible to the kernel and kallsyms, so they can be
seen in traces. For details, please have a look at the
individual patches.

Thanks a lot!
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Documentation/sysctl/net.txt

@@ -54,6 +54,18 @@ Values :
 	1 - enable JIT hardening for unprivileged users only
 	2 - enable JIT hardening for all users
 
+bpf_jit_kallsyms
+----------------
+
+When Berkeley Packet Filter Just in Time compiler is enabled, then compiled
+images are unknown addresses to the kernel, meaning they neither show up in
+traces nor in /proc/kallsyms. This enables export of these addresses, which
+can be used for debugging/tracing. If bpf_jit_harden is enabled, this feature
+is disabled.
+Values :
+	0 - disable JIT kallsyms export (default value)
+	1 - enable JIT kallsyms export for privileged users only
+
 dev_weight
 --------------
 

arch/arm64/net/bpf_jit_comp.c

@@ -813,11 +813,6 @@ static inline void bpf_flush_icache(void *start, void *end)
 	flush_icache_range((unsigned long)start, (unsigned long)end);
 }
 
-void bpf_jit_compile(struct bpf_prog *prog)
-{
-	/* Nothing to do here. We support Internal BPF. */
-}
-
 struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 {
 	struct bpf_prog *tmp, *orig_prog = prog;
@@ -915,18 +910,3 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 					   tmp : orig_prog);
 	return prog;
 }
-
-void bpf_jit_free(struct bpf_prog *prog)
-{
-	unsigned long addr = (unsigned long)prog->bpf_func & PAGE_MASK;
-	struct bpf_binary_header *header = (void *)addr;
-
-	if (!prog->jited)
-		goto free_filter;
-
-	set_memory_rw(addr, header->pages);
-	bpf_jit_binary_free(header);
-
-free_filter:
-	bpf_prog_unlock_free(prog);
-}

arch/powerpc/net/bpf_jit_comp64.c

@@ -961,8 +961,6 @@ static int bpf_jit_build_body(struct bpf_prog *fp, u32 *image,
 	return 0;
 }
 
-void bpf_jit_compile(struct bpf_prog *fp) { }
-
 struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *fp)
 {
 	u32 proglen;
@@ -1066,6 +1064,7 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *fp)
 	return fp;
 }
 
+/* Overriding bpf_jit_free() as we don't set images read-only. */
 void bpf_jit_free(struct bpf_prog *fp)
 {
 	unsigned long addr = (unsigned long)fp->bpf_func & PAGE_MASK;

arch/s390/net/bpf_jit_comp.c

@@ -1262,14 +1262,6 @@ static int bpf_jit_prog(struct bpf_jit *jit, struct bpf_prog *fp)
 	return 0;
 }
 
-/*
- * Classic BPF function stub. BPF programs will be converted into
- * eBPF and then bpf_int_jit_compile() will be called.
- */
-void bpf_jit_compile(struct bpf_prog *fp)
-{
-}
-
 /*
  * Compile eBPF program "fp"
  */
@@ -1347,21 +1339,3 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *fp)
 					   tmp : orig_fp);
 	return fp;
 }
-
-/*
- * Free eBPF program
- */
-void bpf_jit_free(struct bpf_prog *fp)
-{
-	unsigned long addr = (unsigned long)fp->bpf_func & PAGE_MASK;
-	struct bpf_binary_header *header = (void *)addr;
-
-	if (!fp->jited)
-		goto free_filter;
-
-	set_memory_rw(addr, header->pages);
-	bpf_jit_binary_free(header);
-
-free_filter:
-	bpf_prog_unlock_free(fp);
-}

arch/x86/net/bpf_jit_comp.c

@@ -1067,13 +1067,13 @@ xadd:			if (is_imm8(insn->off))
 
 		ilen = prog - temp;
 		if (ilen > BPF_MAX_INSN_SIZE) {
-			pr_err("bpf_jit_compile fatal insn size error\n");
+			pr_err("bpf_jit: fatal insn size error\n");
 			return -EFAULT;
 		}
 
 		if (image) {
 			if (unlikely(proglen + ilen > oldproglen)) {
-				pr_err("bpf_jit_compile fatal error\n");
+				pr_err("bpf_jit: fatal error\n");
 				return -EFAULT;
 			}
 			memcpy(image + proglen, temp, ilen);
@@ -1085,10 +1085,6 @@ xadd:			if (is_imm8(insn->off))
 	return proglen;
 }
 
-void bpf_jit_compile(struct bpf_prog *prog)
-{
-}
-
 struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 {
 	struct bpf_binary_header *header = NULL;
@@ -1184,18 +1180,3 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 					   tmp : orig_prog);
 	return prog;
 }
-
-void bpf_jit_free(struct bpf_prog *fp)
-{
-	unsigned long addr = (unsigned long)fp->bpf_func & PAGE_MASK;
-	struct bpf_binary_header *header = (void *)addr;
-
-	if (!fp->jited)
-		goto free_filter;
-
-	set_memory_rw(addr, header->pages);
-	bpf_jit_binary_free(header);
-
-free_filter:
-	bpf_prog_unlock_free(fp);
-}

include/linux/bpf.h

@@ -8,10 +8,12 @@
 #define _LINUX_BPF_H 1
 
 #include <uapi/linux/bpf.h>
+
 #include <linux/workqueue.h>
 #include <linux/file.h>
 #include <linux/percpu.h>
 #include <linux/err.h>
+#include <linux/rbtree_latch.h>
 
 struct perf_event;
 struct bpf_map;
@@ -177,6 +179,8 @@ struct bpf_prog_aux {
 	atomic_t refcnt;
 	u32 used_map_cnt;
 	u32 max_ctx_offset;
+	struct latch_tree_node ksym_tnode;
+	struct list_head ksym_lnode;
 	const struct bpf_verifier_ops *ops;
 	struct bpf_map **used_maps;
 	struct bpf_prog *prog;

include/linux/filter.h

@@ -54,6 +54,12 @@ struct bpf_prog_aux;
 #define BPF_REG_AX		MAX_BPF_REG
 #define MAX_BPF_JIT_REG		(MAX_BPF_REG + 1)
 
+/* As per nm, we expose JITed images as text (code) section for
+ * kallsyms. That way, tools like perf can find it to match
+ * addresses.
+ */
+#define BPF_SYM_ELF_TYPE	't'
+
 /* BPF program can access up to 512 bytes of stack space. */
 #define MAX_BPF_STACK	512
 
@@ -555,6 +561,11 @@ static inline void bpf_prog_unlock_ro(struct bpf_prog *fp)
 {
 	set_memory_rw((unsigned long)fp, fp->pages);
 }
+
+static inline void bpf_jit_binary_unlock_ro(struct bpf_binary_header *hdr)
+{
+	set_memory_rw((unsigned long)hdr, hdr->pages);
+}
 #else
 static inline void bpf_prog_lock_ro(struct bpf_prog *fp)
 {
@@ -563,8 +574,21 @@ static inline void bpf_prog_lock_ro(struct bpf_prog *fp)
 static inline void bpf_prog_unlock_ro(struct bpf_prog *fp)
 {
 }
+
+static inline void bpf_jit_binary_unlock_ro(struct bpf_binary_header *hdr)
+{
+}
 #endif /* CONFIG_DEBUG_SET_MODULE_RONX */
 
+static inline struct bpf_binary_header *
+bpf_jit_binary_hdr(const struct bpf_prog *fp)
+{
+	unsigned long real_start = (unsigned long)fp->bpf_func;
+	unsigned long addr = real_start & PAGE_MASK;
+
+	return (void *)addr;
+}
+
 int sk_filter_trim_cap(struct sock *sk, struct sk_buff *skb, unsigned int cap);
 static inline int sk_filter(struct sock *sk, struct sk_buff *skb)
 {
@@ -607,6 +631,7 @@ void sk_filter_uncharge(struct sock *sk, struct sk_filter *fp);
 u64 __bpf_call_base(u64 r1, u64 r2, u64 r3, u64 r4, u64 r5);
 
 struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog);
+void bpf_jit_compile(struct bpf_prog *prog);
 bool bpf_helper_changes_pkt_data(void *func);
 
 struct bpf_prog *bpf_patch_insn_single(struct bpf_prog *prog, u32 off,
@@ -616,6 +641,7 @@ void bpf_warn_invalid_xdp_action(u32 act);
 #ifdef CONFIG_BPF_JIT
 extern int bpf_jit_enable;
 extern int bpf_jit_harden;
+extern int bpf_jit_kallsyms;
 
 typedef void (*bpf_jit_fill_hole_t)(void *area, unsigned int size);
 
@@ -625,7 +651,6 @@ bpf_jit_binary_alloc(unsigned int proglen, u8 **image_ptr,
 		     bpf_jit_fill_hole_t bpf_fill_ill_insns);
 void bpf_jit_binary_free(struct bpf_binary_header *hdr);
 
-void bpf_jit_compile(struct bpf_prog *fp);
 void bpf_jit_free(struct bpf_prog *fp);
 
 struct bpf_prog *bpf_jit_blind_constants(struct bpf_prog *fp);
@@ -651,6 +676,11 @@ static inline bool bpf_jit_is_ebpf(void)
 # endif
 }
 
+static inline bool bpf_prog_ebpf_jited(const struct bpf_prog *fp)
+{
+	return fp->jited && bpf_jit_is_ebpf();
+}
+
 static inline bool bpf_jit_blinding_enabled(void)
 {
 	/* These are the prerequisites, should someone ever have the
@@ -668,15 +698,91 @@ static inline bool bpf_jit_blinding_enabled(void)
 
 	return true;
 }
-#else
-static inline void bpf_jit_compile(struct bpf_prog *fp)
+
+static inline bool bpf_jit_kallsyms_enabled(void)
+{
+	/* There are a couple of corner cases where kallsyms should
+	 * not be enabled f.e. on hardening.
+	 */
+	if (bpf_jit_harden)
+		return false;
+	if (!bpf_jit_kallsyms)
+		return false;
+	if (bpf_jit_kallsyms == 1)
+		return true;
+
+	return false;
+}
+
+const char *__bpf_address_lookup(unsigned long addr, unsigned long *size,
+				 unsigned long *off, char *sym);
+bool is_bpf_text_address(unsigned long addr);
+int bpf_get_kallsym(unsigned int symnum, unsigned long *value, char *type,
+		    char *sym);
+
+static inline const char *
+bpf_address_lookup(unsigned long addr, unsigned long *size,
+		   unsigned long *off, char **modname, char *sym)
 {
+	const char *ret = __bpf_address_lookup(addr, size, off, sym);
+
+	if (ret && modname)
+		*modname = NULL;
+	return ret;
+}
+
+void bpf_prog_kallsyms_add(struct bpf_prog *fp);
+void bpf_prog_kallsyms_del(struct bpf_prog *fp);
+
+#else /* CONFIG_BPF_JIT */
+
+static inline bool bpf_prog_ebpf_jited(const struct bpf_prog *fp)
+{
+	return false;
 }
 
 static inline void bpf_jit_free(struct bpf_prog *fp)
 {
 	bpf_prog_unlock_free(fp);
 }
+
+static inline bool bpf_jit_kallsyms_enabled(void)
+{
+	return false;
+}
+
+static inline const char *
+__bpf_address_lookup(unsigned long addr, unsigned long *size,
+		     unsigned long *off, char *sym)
+{
+	return NULL;
+}
+
+static inline bool is_bpf_text_address(unsigned long addr)
+{
+	return false;
+}
+
+static inline int bpf_get_kallsym(unsigned int symnum, unsigned long *value,
+				  char *type, char *sym)
+{
+	return -ERANGE;
+}
+
+static inline const char *
+bpf_address_lookup(unsigned long addr, unsigned long *size,
+		   unsigned long *off, char **modname, char *sym)
+{
+	return NULL;
+}
+
+static inline void bpf_prog_kallsyms_add(struct bpf_prog *fp)
+{
+}
+
+static inline void bpf_prog_kallsyms_del(struct bpf_prog *fp)
+{
+}
 #endif /* CONFIG_BPF_JIT */
 
 #define BPF_ANC		BIT(15)

kernel/bpf/arraymap.c

@@ -269,7 +269,7 @@ static const struct bpf_map_ops array_ops = {
 	.map_delete_elem = array_map_delete_elem,
 };
 
-static struct bpf_map_type_list array_type __read_mostly = {
+static struct bpf_map_type_list array_type __ro_after_init = {
 	.ops = &array_ops,
 	.type = BPF_MAP_TYPE_ARRAY,
 };
@@ -283,7 +283,7 @@ static const struct bpf_map_ops percpu_array_ops = {
 	.map_delete_elem = array_map_delete_elem,
 };
 
-static struct bpf_map_type_list percpu_array_type __read_mostly = {
+static struct bpf_map_type_list percpu_array_type __ro_after_init = {
 	.ops = &percpu_array_ops,
 	.type = BPF_MAP_TYPE_PERCPU_ARRAY,
 };
@@ -409,7 +409,7 @@ static const struct bpf_map_ops prog_array_ops = {
 	.map_fd_put_ptr = prog_fd_array_put_ptr,
 };
 
-static struct bpf_map_type_list prog_array_type __read_mostly = {
+static struct bpf_map_type_list prog_array_type __ro_after_init = {
 	.ops = &prog_array_ops,
 	.type = BPF_MAP_TYPE_PROG_ARRAY,
 };
@@ -522,7 +522,7 @@ static const struct bpf_map_ops perf_event_array_ops = {
 	.map_release = perf_event_fd_array_release,
 };
 
-static struct bpf_map_type_list perf_event_array_type __read_mostly = {
+static struct bpf_map_type_list perf_event_array_type __ro_after_init = {
 	.ops = &perf_event_array_ops,
 	.type = BPF_MAP_TYPE_PERF_EVENT_ARRAY,
 };
@@ -564,7 +564,7 @@ static const struct bpf_map_ops cgroup_array_ops = {
 	.map_fd_put_ptr = cgroup_fd_array_put_ptr,
 };
 
-static struct bpf_map_type_list cgroup_array_type __read_mostly = {
+static struct bpf_map_type_list cgroup_array_type __ro_after_init = {
 	.ops = &cgroup_array_ops,
 	.type = BPF_MAP_TYPE_CGROUP_ARRAY,
 };

kernel/bpf/core.c

@@ -28,6 +28,9 @@
 #include <linux/moduleloader.h>
 #include <linux/bpf.h>
 #include <linux/frame.h>
+#include <linux/rbtree_latch.h>
+#include <linux/kallsyms.h>
+#include <linux/rcupdate.h>
 
 #include <asm/unaligned.h>
 
@@ -95,6 +98,8 @@ struct bpf_prog *bpf_prog_alloc(unsigned int size, gfp_t gfp_extra_flags)
 	fp->aux = aux;
 	fp->aux->prog = fp;
 
+	INIT_LIST_HEAD_RCU(&fp->aux->ksym_lnode);
+
 	return fp;
 }
 EXPORT_SYMBOL_GPL(bpf_prog_alloc);
@@ -290,6 +295,206 @@ struct bpf_prog *bpf_patch_insn_single(struct bpf_prog *prog, u32 off,
 }
 
 #ifdef CONFIG_BPF_JIT
+static __always_inline void
+bpf_get_prog_addr_region(const struct bpf_prog *prog,
+			 unsigned long *symbol_start,
+			 unsigned long *symbol_end)
+{
+	const struct bpf_binary_header *hdr = bpf_jit_binary_hdr(prog);
+	unsigned long addr = (unsigned long)hdr;
+
+	WARN_ON_ONCE(!bpf_prog_ebpf_jited(prog));
+
+	*symbol_start = addr;
+	*symbol_end   = addr + hdr->pages * PAGE_SIZE;
+}
+
+static void bpf_get_prog_name(const struct bpf_prog *prog, char *sym)
+{
+	BUILD_BUG_ON(sizeof("bpf_prog_") +
+		     sizeof(prog->tag) * 2 + 1 > KSYM_NAME_LEN);
+
+	sym += snprintf(sym, KSYM_NAME_LEN, "bpf_prog_");
+	sym  = bin2hex(sym, prog->tag, sizeof(prog->tag));
+	*sym = 0;
+}
+
+static __always_inline unsigned long
+bpf_get_prog_addr_start(struct latch_tree_node *n)
+{
+	unsigned long symbol_start, symbol_end;
+	const struct bpf_prog_aux *aux;
+
+	aux = container_of(n, struct bpf_prog_aux, ksym_tnode);
+	bpf_get_prog_addr_region(aux->prog, &symbol_start, &symbol_end);
+
+	return symbol_start;
+}
+
+static __always_inline bool bpf_tree_less(struct latch_tree_node *a,
+					  struct latch_tree_node *b)
+{
+	return bpf_get_prog_addr_start(a) < bpf_get_prog_addr_start(b);
+}
+
+static __always_inline int bpf_tree_comp(void *key, struct latch_tree_node *n)
+{
+	unsigned long val = (unsigned long)key;
+	unsigned long symbol_start, symbol_end;
+	const struct bpf_prog_aux *aux;
+
+	aux = container_of(n, struct bpf_prog_aux, ksym_tnode);
+	bpf_get_prog_addr_region(aux->prog, &symbol_start, &symbol_end);
+
+	if (val < symbol_start)
+		return -1;
+	if (val >= symbol_end)
+		return  1;
+
+	return 0;
+}
+
+static const struct latch_tree_ops bpf_tree_ops = {
+	.less	= bpf_tree_less,
+	.comp	= bpf_tree_comp,
+};
+
+static DEFINE_SPINLOCK(bpf_lock);
+static LIST_HEAD(bpf_kallsyms);
+static struct latch_tree_root bpf_tree __cacheline_aligned;
+
+int bpf_jit_kallsyms __read_mostly;
+
+static void bpf_prog_ksym_node_add(struct bpf_prog_aux *aux)
+{
+	WARN_ON_ONCE(!list_empty(&aux->ksym_lnode));
+	list_add_tail_rcu(&aux->ksym_lnode, &bpf_kallsyms);
+	latch_tree_insert(&aux->ksym_tnode, &bpf_tree, &bpf_tree_ops);
+}
+
+static void bpf_prog_ksym_node_del(struct bpf_prog_aux *aux)
+{
+	if (list_empty(&aux->ksym_lnode))
+		return;
+
+	latch_tree_erase(&aux->ksym_tnode, &bpf_tree, &bpf_tree_ops);
+	list_del_rcu(&aux->ksym_lnode);
+}
+
+static bool bpf_prog_kallsyms_candidate(const struct bpf_prog *fp)
+{
+	return fp->jited && !bpf_prog_was_classic(fp);
+}
+
+static bool bpf_prog_kallsyms_verify_off(const struct bpf_prog *fp)
+{
+	return list_empty(&fp->aux->ksym_lnode) ||
+	       fp->aux->ksym_lnode.prev == LIST_POISON2;
+}
+
+void bpf_prog_kallsyms_add(struct bpf_prog *fp)
+{
+	unsigned long flags;
+
+	if (!bpf_prog_kallsyms_candidate(fp) ||
+	    !capable(CAP_SYS_ADMIN))
+		return;
+
+	spin_lock_irqsave(&bpf_lock, flags);
+	bpf_prog_ksym_node_add(fp->aux);
+	spin_unlock_irqrestore(&bpf_lock, flags);
+}
+
+void bpf_prog_kallsyms_del(struct bpf_prog *fp)
+{
+	unsigned long flags;
+
+	if (!bpf_prog_kallsyms_candidate(fp))
+		return;
+
+	spin_lock_irqsave(&bpf_lock, flags);
+	bpf_prog_ksym_node_del(fp->aux);
+	spin_unlock_irqrestore(&bpf_lock, flags);
+}
+
+static struct bpf_prog *bpf_prog_kallsyms_find(unsigned long addr)
+{
+	struct latch_tree_node *n;
+
+	if (!bpf_jit_kallsyms_enabled())
+		return NULL;
+
+	n = latch_tree_find((void *)addr, &bpf_tree, &bpf_tree_ops);
+	return n ?
+	       container_of(n, struct bpf_prog_aux, ksym_tnode)->prog :
+	       NULL;
+}
+
+const char *__bpf_address_lookup(unsigned long addr, unsigned long *size,
+				 unsigned long *off, char *sym)
+{
+	unsigned long symbol_start, symbol_end;
+	struct bpf_prog *prog;
+	char *ret = NULL;
+
+	rcu_read_lock();
+	prog = bpf_prog_kallsyms_find(addr);
+	if (prog) {
+		bpf_get_prog_addr_region(prog, &symbol_start, &symbol_end);
+		bpf_get_prog_name(prog, sym);
+
+		ret = sym;
+		if (size)
+			*size = symbol_end - symbol_start;
+		if (off)
+			*off  = addr - symbol_start;
+	}
+	rcu_read_unlock();
+
+	return ret;
+}
+
+bool is_bpf_text_address(unsigned long addr)
+{
+	bool ret;
+
+	rcu_read_lock();
+	ret = bpf_prog_kallsyms_find(addr) != NULL;
+	rcu_read_unlock();
+
+	return ret;
+}
+
+int bpf_get_kallsym(unsigned int symnum, unsigned long *value, char *type,
+		    char *sym)
+{
+	unsigned long symbol_start, symbol_end;
+	struct bpf_prog_aux *aux;
+	unsigned int it = 0;
+	int ret = -ERANGE;
+
+	if (!bpf_jit_kallsyms_enabled())
+		return ret;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(aux, &bpf_kallsyms, ksym_lnode) {
+		if (it++ != symnum)
+			continue;
+
+		bpf_get_prog_addr_region(aux->prog, &symbol_start, &symbol_end);
+		bpf_get_prog_name(aux->prog, sym);
+
+		*value = symbol_start;
+		*type  = BPF_SYM_ELF_TYPE;
+
+		ret = 0;
+		break;
+	}
+	rcu_read_unlock();
+
+	return ret;
+}
+
 struct bpf_binary_header *
 bpf_jit_binary_alloc(unsigned int proglen, u8 **image_ptr,
 		     unsigned int alignment,
@@ -326,6 +531,24 @@ void bpf_jit_binary_free(struct bpf_binary_header *hdr)
 	module_memfree(hdr);
 }
 
+/* This symbol is only overridden by archs that have different
+ * requirements than the usual eBPF JITs, f.e. when they only
+ * implement cBPF JIT, do not set images read-only, etc.
+ */
+void __weak bpf_jit_free(struct bpf_prog *fp)
+{
+	if (fp->jited) {
+		struct bpf_binary_header *hdr = bpf_jit_binary_hdr(fp);
+
+		bpf_jit_binary_unlock_ro(hdr);
+		bpf_jit_binary_free(hdr);
+
+		WARN_ON_ONCE(!bpf_prog_kallsyms_verify_off(fp));
+	}
+
+	bpf_prog_unlock_free(fp);
+}
+
 int bpf_jit_harden __read_mostly;
 
 static int bpf_jit_blind_insn(const struct bpf_insn *from,
@@ -1154,12 +1377,22 @@ const struct bpf_func_proto bpf_tail_call_proto = {
 	.arg3_type	= ARG_ANYTHING,
 };
 
-/* For classic BPF JITs that don't implement bpf_int_jit_compile(). */
+/* Stub for JITs that only support cBPF. eBPF programs are interpreted.
+ * It is encouraged to implement bpf_int_jit_compile() instead, so that
+ * eBPF and implicitly also cBPF can get JITed!
+ */
 struct bpf_prog * __weak bpf_int_jit_compile(struct bpf_prog *prog)
 {
 	return prog;
 }
 
+/* Stub for JITs that support eBPF. All cBPF code gets transformed into
+ * eBPF by the kernel and is later compiled by bpf_int_jit_compile().
+ */
+void __weak bpf_jit_compile(struct bpf_prog *prog)
+{
+}
+
 bool __weak bpf_helper_changes_pkt_data(void *func)
 {
 	return false;

kernel/bpf/hashtab.c

@@ -1023,7 +1023,7 @@ static const struct bpf_map_ops htab_ops = {
 	.map_delete_elem = htab_map_delete_elem,
 };
 
-static struct bpf_map_type_list htab_type __read_mostly = {
+static struct bpf_map_type_list htab_type __ro_after_init = {
 	.ops = &htab_ops,
 	.type = BPF_MAP_TYPE_HASH,
 };
@@ -1037,7 +1037,7 @@ static const struct bpf_map_ops htab_lru_ops = {
 	.map_delete_elem = htab_lru_map_delete_elem,
 };
 
-static struct bpf_map_type_list htab_lru_type __read_mostly = {
+static struct bpf_map_type_list htab_lru_type __ro_after_init = {
 	.ops = &htab_lru_ops,
 	.type = BPF_MAP_TYPE_LRU_HASH,
 };
@@ -1124,7 +1124,7 @@ static const struct bpf_map_ops htab_percpu_ops = {
 	.map_delete_elem = htab_map_delete_elem,
 };
 
-static struct bpf_map_type_list htab_percpu_type __read_mostly = {
+static struct bpf_map_type_list htab_percpu_type __ro_after_init = {
 	.ops = &htab_percpu_ops,
 	.type = BPF_MAP_TYPE_PERCPU_HASH,
 };
@@ -1138,7 +1138,7 @@ static const struct bpf_map_ops htab_lru_percpu_ops = {
 	.map_delete_elem = htab_lru_map_delete_elem,
 };
 
-static struct bpf_map_type_list htab_lru_percpu_type __read_mostly = {
+static struct bpf_map_type_list htab_lru_percpu_type __ro_after_init = {
 	.ops = &htab_lru_percpu_ops,
 	.type = BPF_MAP_TYPE_LRU_PERCPU_HASH,
 };

kernel/bpf/lpm_trie.c

@@ -508,7 +508,7 @@ static const struct bpf_map_ops trie_ops = {
 	.map_delete_elem = trie_delete_elem,
 };
 
-static struct bpf_map_type_list trie_type __read_mostly = {
+static struct bpf_map_type_list trie_type __ro_after_init = {
 	.ops = &trie_ops,
 	.type = BPF_MAP_TYPE_LPM_TRIE,
 };

kernel/bpf/stackmap.c

@@ -273,7 +273,7 @@ static const struct bpf_map_ops stack_map_ops = {
 	.map_delete_elem = stack_map_delete_elem,
 };
 
-static struct bpf_map_type_list stack_map_type __read_mostly = {
+static struct bpf_map_type_list stack_map_type __ro_after_init = {
 	.ops = &stack_map_ops,
 	.type = BPF_MAP_TYPE_STACK_TRACE,
 };

kernel/bpf/syscall.c

@@ -707,6 +707,7 @@ void bpf_prog_put(struct bpf_prog *prog)
 {
 	if (atomic_dec_and_test(&prog->aux->refcnt)) {
 		trace_bpf_prog_put_rcu(prog);
+		bpf_prog_kallsyms_del(prog);
 		call_rcu(&prog->aux->rcu, __bpf_prog_put_rcu);
 	}
 }
@@ -903,6 +904,7 @@ static int bpf_prog_load(union bpf_attr *attr)
 		/* failed to allocate fd */
 		goto free_used_maps;
 
+	bpf_prog_kallsyms_add(prog);
 	trace_bpf_prog_load(prog, err);
 	return err;
 

kernel/extable.c

@@ -20,6 +20,7 @@
 #include <linux/module.h>
 #include <linux/mutex.h>
 #include <linux/init.h>
+#include <linux/filter.h>
 
 #include <asm/sections.h>
 #include <linux/uaccess.h>
@@ -104,6 +105,8 @@ int __kernel_text_address(unsigned long addr)
 		return 1;
 	if (is_ftrace_trampoline(addr))
 		return 1;
+	if (is_bpf_text_address(addr))
+		return 1;
 	/*
 	 * There might be init symbols in saved stacktraces.
 	 * Give those symbols a chance to be printed in
@@ -123,7 +126,11 @@ int kernel_text_address(unsigned long addr)
 		return 1;
 	if (is_module_text_address(addr))
 		return 1;
-	return is_ftrace_trampoline(addr);
+	if (is_ftrace_trampoline(addr))
+		return 1;
+	if (is_bpf_text_address(addr))
+		return 1;
+	return 0;
 }
 
 /*

kernel/kallsyms.c

@@ -23,6 +23,7 @@
 #include <linux/mm.h>
 #include <linux/ctype.h>
 #include <linux/slab.h>
+#include <linux/filter.h>
 #include <linux/compiler.h>
 
 #include <asm/sections.h>
@@ -300,10 +301,11 @@ int kallsyms_lookup_size_offset(unsigned long addr, unsigned long *symbolsize,
 				unsigned long *offset)
 {
 	char namebuf[KSYM_NAME_LEN];
+
 	if (is_ksym_addr(addr))
 		return !!get_symbol_pos(addr, symbolsize, offset);
-
-	return !!module_address_lookup(addr, symbolsize, offset, NULL, namebuf);
+	return !!module_address_lookup(addr, symbolsize, offset, NULL, namebuf) ||
+	       !!__bpf_address_lookup(addr, symbolsize, offset, namebuf);
 }
 
 /*
@@ -318,6 +320,8 @@ const char *kallsyms_lookup(unsigned long addr,
 			    unsigned long *offset,
 			    char **modname, char *namebuf)
 {
+	const char *ret;
+
 	namebuf[KSYM_NAME_LEN - 1] = 0;
 	namebuf[0] = 0;
 
@@ -333,9 +337,13 @@ const char *kallsyms_lookup(unsigned long addr,
 		return namebuf;
 	}
 
-	/* See if it's in a module. */
-	return module_address_lookup(addr, symbolsize, offset, modname,
-				     namebuf);
+	/* See if it's in a module or a BPF JITed image. */
+	ret = module_address_lookup(addr, symbolsize, offset,
+				    modname, namebuf);
+	if (!ret)
+		ret = bpf_address_lookup(addr, symbolsize,
+					 offset, modname, namebuf);
+	return ret;
 }
 
 int lookup_symbol_name(unsigned long addr, char *symname)
@@ -471,6 +479,7 @@ EXPORT_SYMBOL(__print_symbol);
 /* To avoid using get_symbol_offset for every symbol, we carry prefix along. */
 struct kallsym_iter {
 	loff_t pos;
+	loff_t pos_mod_end;
 	unsigned long value;
 	unsigned int nameoff; /* If iterating in core kernel symbols. */
 	char type;
@@ -481,13 +490,27 @@ struct kallsym_iter {
 
 static int get_ksymbol_mod(struct kallsym_iter *iter)
 {
-	if (module_get_kallsym(iter->pos - kallsyms_num_syms, &iter->value,
-				&iter->type, iter->name, iter->module_name,
-				&iter->exported) < 0)
+	int ret = module_get_kallsym(iter->pos - kallsyms_num_syms,
+				     &iter->value, &iter->type,
+				     iter->name, iter->module_name,
+				     &iter->exported);
+	if (ret < 0) {
+		iter->pos_mod_end = iter->pos;
 		return 0;
+	}
+
 	return 1;
 }
 
+static int get_ksymbol_bpf(struct kallsym_iter *iter)
+{
+	iter->module_name[0] = '\0';
+	iter->exported = 0;
+	return bpf_get_kallsym(iter->pos - iter->pos_mod_end,
+			       &iter->value, &iter->type,
+			       iter->name) < 0 ? 0 : 1;
+}
+
 /* Returns space to next name. */
 static unsigned long get_ksymbol_core(struct kallsym_iter *iter)
 {
@@ -508,16 +531,30 @@ static void reset_iter(struct kallsym_iter *iter, loff_t new_pos)
 	iter->name[0] = '\0';
 	iter->nameoff = get_symbol_offset(new_pos);
 	iter->pos = new_pos;
+	if (new_pos == 0)
+		iter->pos_mod_end = 0;
+}
+
+static int update_iter_mod(struct kallsym_iter *iter, loff_t pos)
+{
+	iter->pos = pos;
+
+	if (iter->pos_mod_end > 0 &&
+	    iter->pos_mod_end < iter->pos)
+		return get_ksymbol_bpf(iter);
+
+	if (!get_ksymbol_mod(iter))
+		return get_ksymbol_bpf(iter);
+
+	return 1;
 }
 
 /* Returns false if pos at or past end of file. */
 static int update_iter(struct kallsym_iter *iter, loff_t pos)
 {
 	/* Module symbols can be accessed randomly. */
-	if (pos >= kallsyms_num_syms) {
-		iter->pos = pos;
-		return get_ksymbol_mod(iter);
-	}
+	if (pos >= kallsyms_num_syms)
+		return update_iter_mod(iter, pos);
 
 	/* If we're not on the desired position, reset to new position. */
 	if (pos != iter->pos)

kernel/trace/bpf_trace.c

@@ -506,7 +506,7 @@ static const struct bpf_verifier_ops kprobe_prog_ops = {
 	.is_valid_access = kprobe_prog_is_valid_access,
 };
 
-static struct bpf_prog_type_list kprobe_tl = {
+static struct bpf_prog_type_list kprobe_tl __ro_after_init = {
 	.ops	= &kprobe_prog_ops,
 	.type	= BPF_PROG_TYPE_KPROBE,
 };
@@ -589,7 +589,7 @@ static const struct bpf_verifier_ops tracepoint_prog_ops = {
 	.is_valid_access = tp_prog_is_valid_access,
 };
 
-static struct bpf_prog_type_list tracepoint_tl = {
+static struct bpf_prog_type_list tracepoint_tl __ro_after_init = {
 	.ops	= &tracepoint_prog_ops,
 	.type	= BPF_PROG_TYPE_TRACEPOINT,
 };
@@ -648,7 +648,7 @@ static const struct bpf_verifier_ops perf_event_prog_ops = {
 	.convert_ctx_access	= pe_prog_convert_ctx_access,
 };
 
-static struct bpf_prog_type_list perf_event_tl = {
+static struct bpf_prog_type_list perf_event_tl __ro_after_init = {
 	.ops	= &perf_event_prog_ops,
 	.type	= BPF_PROG_TYPE_PERF_EVENT,
 };

net/Kconfig

@@ -297,7 +297,8 @@ config BPF_JIT
 
 	  Note, admin should enable this feature changing:
 	  /proc/sys/net/core/bpf_jit_enable
-	  /proc/sys/net/core/bpf_jit_harden (optional)
+	  /proc/sys/net/core/bpf_jit_harden   (optional)
+	  /proc/sys/net/core/bpf_jit_kallsyms (optional)
 
 config NET_FLOW_LIMIT
 	bool

net/core/filter.c

@@ -3296,47 +3296,47 @@ static const struct bpf_verifier_ops cg_sock_ops = {
 	.convert_ctx_access	= sock_filter_convert_ctx_access,
 };
 
-static struct bpf_prog_type_list sk_filter_type __read_mostly = {
+static struct bpf_prog_type_list sk_filter_type __ro_after_init = {
 	.ops	= &sk_filter_ops,
 	.type	= BPF_PROG_TYPE_SOCKET_FILTER,
 };
 
-static struct bpf_prog_type_list sched_cls_type __read_mostly = {
+static struct bpf_prog_type_list sched_cls_type __ro_after_init = {
 	.ops	= &tc_cls_act_ops,
 	.type	= BPF_PROG_TYPE_SCHED_CLS,
 };
 
-static struct bpf_prog_type_list sched_act_type __read_mostly = {
+static struct bpf_prog_type_list sched_act_type __ro_after_init = {
 	.ops	= &tc_cls_act_ops,
 	.type	= BPF_PROG_TYPE_SCHED_ACT,
 };
 
-static struct bpf_prog_type_list xdp_type __read_mostly = {
+static struct bpf_prog_type_list xdp_type __ro_after_init = {
 	.ops	= &xdp_ops,
 	.type	= BPF_PROG_TYPE_XDP,
 };
 
-static struct bpf_prog_type_list cg_skb_type __read_mostly = {
+static struct bpf_prog_type_list cg_skb_type __ro_after_init = {
 	.ops	= &cg_skb_ops,
 	.type	= BPF_PROG_TYPE_CGROUP_SKB,
 };
 
-static struct bpf_prog_type_list lwt_in_type __read_mostly = {
+static struct bpf_prog_type_list lwt_in_type __ro_after_init = {
 	.ops	= &lwt_inout_ops,
 	.type	= BPF_PROG_TYPE_LWT_IN,
 };
 
-static struct bpf_prog_type_list lwt_out_type __read_mostly = {
+static struct bpf_prog_type_list lwt_out_type __ro_after_init = {
 	.ops	= &lwt_inout_ops,
 	.type	= BPF_PROG_TYPE_LWT_OUT,
 };
 
-static struct bpf_prog_type_list lwt_xmit_type __read_mostly = {
+static struct bpf_prog_type_list lwt_xmit_type __ro_after_init = {
 	.ops	= &lwt_xmit_ops,
 	.type	= BPF_PROG_TYPE_LWT_XMIT,
 };
 
-static struct bpf_prog_type_list cg_sock_type __read_mostly = {
+static struct bpf_prog_type_list cg_sock_type __ro_after_init = {
 	.ops	= &cg_sock_ops,
 	.type	= BPF_PROG_TYPE_CGROUP_SOCK
 };

net/core/sysctl_net_core.c

@@ -334,6 +334,13 @@ static struct ctl_table net_core_table[] = {
 		.mode		= 0600,
 		.proc_handler	= proc_dointvec,
 	},
+	{
+		.procname	= "bpf_jit_kallsyms",
+		.data		= &bpf_jit_kallsyms,
+		.maxlen		= sizeof(int),
+		.mode		= 0600,
+		.proc_handler	= proc_dointvec,
+	},
 # endif
 #endif
 	{