Merge branch 'bpf: Add kprobe multi link'

Jiri Olsa says:

====================

hi,
this patchset adds new link type BPF_TRACE_KPROBE_MULTI that attaches
kprobe program through fprobe API [1] instroduced by Masami.

The fprobe API allows to attach probe on multiple functions at once very
fast, because it works on top of ftrace. On the other hand this limits
the probe point to the function entry or return.

With bpftrace support I see following attach speed:

  # perf stat --null -r 5 ./src/bpftrace -e 'kprobe:x* { } i:ms:1 { exit(); } '
  Attaching 2 probes...
  Attaching 3342 functions
  ...

  1.4960 +- 0.0285 seconds time elapsed  ( +-  1.91% )

v3 changes:
  - based on latest fprobe post from Masami [2]
  - add acks
  - add extra comment to kprobe_multi_link_handler wrt entry ip setup [Masami]
  - keep swap_words_64 static and swap values directly in
    bpf_kprobe_multi_cookie_swap [Andrii]
  - rearrange locking/migrate setup in kprobe_multi_link_prog_run [Andrii]
  - move uapi fields [Andrii]
  - add bpf_program__attach_kprobe_multi_opts function [Andrii]
  - many small test changes [Andrii]
  - added tests for bpf_program__attach_kprobe_multi_opts
  - make kallsyms_lookup_name check for empty string [Andrii]

v2 changes:
  - based on latest fprobe changes [1]
  - renaming the uapi interface to kprobe multi
  - adding support for sort_r to pass user pointer for swap functions
    and using that in cookie support to keep just single functions array
  - moving new link to kernel/trace/bpf_trace.c file
  - using single fprobe callback function for entry and exit
  - using kvzalloc, libbpf_ensure_mem functions
  - adding new k[ret]probe.multi sections instead of using current kprobe
  - used glob_match from test_progs.c, added '?' matching
  - move bpf_get_func_ip verifier inline change to seprate change
  - couple of other minor fixes

Also available at:
  https://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf.git
  bpf/kprobe_multi

thanks,
jirka

[1] https://lore.kernel.org/bpf/164458044634.586276.3261555265565111183.stgit@devnote2/
[2] https://lore.kernel.org/bpf/164735281449.1084943.12438881786173547153.stgit@devnote2/
====================

Signed-off-by: Alexei Starovoitov <ast@kernel.org>

include/linux/bpf_types.h

@@ -140,3 +140,4 @@ BPF_LINK_TYPE(BPF_LINK_TYPE_XDP, xdp)
 #ifdef CONFIG_PERF_EVENTS
 BPF_LINK_TYPE(BPF_LINK_TYPE_PERF_EVENT, perf)
 #endif
+BPF_LINK_TYPE(BPF_LINK_TYPE_KPROBE_MULTI, kprobe_multi)

include/linux/sort.h

@@ -6,7 +6,7 @@
 
 void sort_r(void *base, size_t num, size_t size,
 	    cmp_r_func_t cmp_func,
-	    swap_func_t swap_func,
+	    swap_r_func_t swap_func,
 	    const void *priv);
 
 void sort(void *base, size_t num, size_t size,

include/linux/trace_events.h

@@ -15,6 +15,7 @@ struct array_buffer;
 struct tracer;
 struct dentry;
 struct bpf_prog;
+union bpf_attr;
 
 const char *trace_print_flags_seq(struct trace_seq *p, const char *delim,
 				  unsigned long flags,
@@ -738,6 +739,7 @@ void bpf_put_raw_tracepoint(struct bpf_raw_event_map *btp);
 int bpf_get_perf_event_info(const struct perf_event *event, u32 *prog_id,
 			    u32 *fd_type, const char **buf,
 			    u64 *probe_offset, u64 *probe_addr);
+int bpf_kprobe_multi_link_attach(const union bpf_attr *attr, struct bpf_prog *prog);
 #else
 static inline unsigned int trace_call_bpf(struct trace_event_call *call, void *ctx)
 {
@@ -779,6 +781,11 @@ static inline int bpf_get_perf_event_info(const struct perf_event *event,
 {
 	return -EOPNOTSUPP;
 }
+static inline int
+bpf_kprobe_multi_link_attach(const union bpf_attr *attr, struct bpf_prog *prog)
+{
+	return -EOPNOTSUPP;
+}
 #endif
 
 enum {

include/linux/types.h

@@ -226,6 +226,7 @@ struct callback_head {
 typedef void (*rcu_callback_t)(struct rcu_head *head);
 typedef void (*call_rcu_func_t)(struct rcu_head *head, rcu_callback_t func);
 
+typedef void (*swap_r_func_t)(void *a, void *b, int size, const void *priv);
 typedef void (*swap_func_t)(void *a, void *b, int size);
 
 typedef int (*cmp_r_func_t)(const void *a, const void *b, const void *priv);

include/uapi/linux/bpf.h

@@ -997,6 +997,7 @@ enum bpf_attach_type {
 	BPF_SK_REUSEPORT_SELECT,
 	BPF_SK_REUSEPORT_SELECT_OR_MIGRATE,
 	BPF_PERF_EVENT,
+	BPF_TRACE_KPROBE_MULTI,
 	__MAX_BPF_ATTACH_TYPE
 };
 
@@ -1011,6 +1012,7 @@ enum bpf_link_type {
 	BPF_LINK_TYPE_NETNS = 5,
 	BPF_LINK_TYPE_XDP = 6,
 	BPF_LINK_TYPE_PERF_EVENT = 7,
+	BPF_LINK_TYPE_KPROBE_MULTI = 8,
 
 	MAX_BPF_LINK_TYPE,
 };
@@ -1118,6 +1120,11 @@ enum bpf_link_type {
  */
 #define BPF_F_XDP_HAS_FRAGS	(1U << 5)
 
+/* link_create.kprobe_multi.flags used in LINK_CREATE command for
+ * BPF_TRACE_KPROBE_MULTI attach type to create return probe.
+ */
+#define BPF_F_KPROBE_MULTI_RETURN	(1U << 0)
+
 /* When BPF ldimm64's insn[0].src_reg != 0 then this can have
  * the following extensions:
  *
@@ -1475,6 +1482,13 @@ union bpf_attr {
 				 */
 				__u64		bpf_cookie;
 			} perf_event;
+			struct {
+				__u32		flags;
+				__u32		cnt;
+				__aligned_u64	syms;
+				__aligned_u64	addrs;
+				__aligned_u64	cookies;
+			} kprobe_multi;
 		};
 	} link_create;
 

kernel/bpf/syscall.c

@@ -32,6 +32,7 @@
 #include <linux/bpf-netns.h>
 #include <linux/rcupdate_trace.h>
 #include <linux/memcontrol.h>
+#include <linux/trace_events.h>
 
 #define IS_FD_ARRAY(map) ((map)->map_type == BPF_MAP_TYPE_PERF_EVENT_ARRAY || \
 			  (map)->map_type == BPF_MAP_TYPE_CGROUP_ARRAY || \
@@ -3022,6 +3023,11 @@ static int bpf_perf_link_attach(const union bpf_attr *attr, struct bpf_prog *pro
 	fput(perf_file);
 	return err;
 }
+#else
+static int bpf_perf_link_attach(const union bpf_attr *attr, struct bpf_prog *prog)
+{
+	return -EOPNOTSUPP;
+}
 #endif /* CONFIG_PERF_EVENTS */
 
 #define BPF_RAW_TRACEPOINT_OPEN_LAST_FIELD raw_tracepoint.prog_fd
@@ -4255,7 +4261,7 @@ static int tracing_bpf_link_attach(const union bpf_attr *attr, bpfptr_t uattr,
 	return -EINVAL;
 }
 
-#define BPF_LINK_CREATE_LAST_FIELD link_create.iter_info_len
+#define BPF_LINK_CREATE_LAST_FIELD link_create.kprobe_multi.cookies
 static int link_create(union bpf_attr *attr, bpfptr_t uattr)
 {
 	enum bpf_prog_type ptype;
@@ -4279,14 +4285,21 @@ static int link_create(union bpf_attr *attr, bpfptr_t uattr)
 		ret = tracing_bpf_link_attach(attr, uattr, prog);
 		goto out;
 	case BPF_PROG_TYPE_PERF_EVENT:
-	case BPF_PROG_TYPE_KPROBE:
 	case BPF_PROG_TYPE_TRACEPOINT:
 		if (attr->link_create.attach_type != BPF_PERF_EVENT) {
 			ret = -EINVAL;
 			goto out;
 		}
 		ptype = prog->type;
 		break;
+	case BPF_PROG_TYPE_KPROBE:
+		if (attr->link_create.attach_type != BPF_PERF_EVENT &&
+		    attr->link_create.attach_type != BPF_TRACE_KPROBE_MULTI) {
+			ret = -EINVAL;
+			goto out;
+		}
+		ptype = prog->type;
+		break;
 	default:
 		ptype = attach_type_to_prog_type(attr->link_create.attach_type);
 		if (ptype == BPF_PROG_TYPE_UNSPEC || ptype != prog->type) {
@@ -4318,13 +4331,16 @@ static int link_create(union bpf_attr *attr, bpfptr_t uattr)
 		ret = bpf_xdp_link_attach(attr, prog);
 		break;
 #endif
-#ifdef CONFIG_PERF_EVENTS
 	case BPF_PROG_TYPE_PERF_EVENT:
 	case BPF_PROG_TYPE_TRACEPOINT:
-	case BPF_PROG_TYPE_KPROBE:
 		ret = bpf_perf_link_attach(attr, prog);
 		break;
-#endif
+	case BPF_PROG_TYPE_KPROBE:
+		if (attr->link_create.attach_type == BPF_PERF_EVENT)
+			ret = bpf_perf_link_attach(attr, prog);
+		else
+			ret = bpf_kprobe_multi_link_attach(attr, prog);
+		break;
 	default:
 		ret = -EINVAL;
 	}

kernel/bpf/verifier.c

@@ -13678,7 +13678,7 @@ static int do_misc_fixups(struct bpf_verifier_env *env)
 			continue;
 		}
 
-		/* Implement bpf_get_func_ip inline. */
+		/* Implement tracing bpf_get_func_ip inline. */
 		if (prog_type == BPF_PROG_TYPE_TRACING &&
 		    insn->imm == BPF_FUNC_get_func_ip) {
 			/* Load IP address from ctx - 16 */
@@ -13693,6 +13693,25 @@ static int do_misc_fixups(struct bpf_verifier_env *env)
 			continue;
 		}
 
+#ifdef CONFIG_X86
+		/* Implement kprobe_multi bpf_get_func_ip inline. */
+		if (prog_type == BPF_PROG_TYPE_KPROBE &&
+		    eatype == BPF_TRACE_KPROBE_MULTI &&
+		    insn->imm == BPF_FUNC_get_func_ip) {
+			/* Load IP address from ctx (struct pt_regs) ip */
+			insn_buf[0] = BPF_LDX_MEM(BPF_DW, BPF_REG_0, BPF_REG_1,
+						  offsetof(struct pt_regs, ip));
+
+			new_prog = bpf_patch_insn_data(env, i + delta, insn_buf, 1);
+			if (!new_prog)
+				return -ENOMEM;
+
+			env->prog = prog = new_prog;
+			insn      = new_prog->insnsi + i + delta;
+			continue;
+		}
+#endif
+
 patch_call_imm:
 		fn = env->ops->get_func_proto(insn->imm, env->prog);
 		/* all functions that have prototype and verifier allowed

kernel/kallsyms.c

@@ -212,6 +212,10 @@ unsigned long kallsyms_lookup_name(const char *name)
 	unsigned long i;
 	unsigned int off;
 
+	/* Skip the search for empty string. */
+	if (!*name)
+		return 0;
+
 	for (i = 0, off = 0; i < kallsyms_num_syms; i++) {
 		off = kallsyms_expand_symbol(off, namebuf, ARRAY_SIZE(namebuf));