Merge tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next

Daniel Borkmann says:

====================
pull-request: bpf-next 2024-03-25

We've added 38 non-merge commits during the last 13 day(s) which contain
a total of 50 files changed, 867 insertions(+), 274 deletions(-).

The main changes are:

1) Add the ability to specify and retrieve BPF cookie also for raw
   tracepoint programs in order to ease migration from classic to raw
   tracepoints, from Andrii Nakryiko.

2) Allow the use of bpf_get_{ns_,}current_pid_tgid() helper for all
   program types and add additional BPF selftests, from Yonghong Song.

3) Several improvements to bpftool and its build, for example, enabling
   libbpf logs when loading pid_iter in debug mode, from Quentin Monnet.

4) Check the return code of all BPF-related set_memory_*() functions during
   load and bail out in case they fail, from Christophe Leroy.

5) Avoid a goto in regs_refine_cond_op() such that the verifier can
   be better integrated into Agni tool which doesn't support backedges
   yet, from Harishankar Vishwanathan.

6) Add a small BPF trie perf improvement by always inlining
   longest_prefix_match, from Jesper Dangaard Brouer.

7) Small BPF selftest refactor in bpf_tcp_ca.c to utilize start_server()
   helper instead of open-coding it, from Geliang Tang.

8) Improve test_tc_tunnel.sh BPF selftest to prevent client connect
   before the server bind, from Alessandro Carminati.

9) Fix BPF selftest benchmark for older glibc and use syscall(SYS_gettid)
   instead of gettid(), from Alan Maguire.

10) Implement a backward-compatible method for struct_ops types with
    additional fields which are not present in older kernels,
    from Kui-Feng Lee.

11) Add a small helper to check if an instruction is addr_space_cast
    from as(0) to as(1) and utilize it in x86-64 JIT, from Puranjay Mohan.

12) Small cleanup to remove unnecessary error check in
    bpf_struct_ops_map_update_elem, from Martin KaFai Lau.

13) Improvements to libbpf fd validity checks for BPF map/programs,
    from Mykyta Yatsenko.

* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (38 commits)
  selftests/bpf: Fix flaky test btf_map_in_map/lookup_update
  bpf: implement insn_is_cast_user() helper for JITs
  bpf: Avoid get_kernel_nofault() to fetch kprobe entry IP
  selftests/bpf: Use start_server in bpf_tcp_ca
  bpf: Sync uapi bpf.h to tools directory
  libbpf: Add new sec_def "sk_skb/verdict"
  selftests/bpf: Mark uprobe trigger functions with nocf_check attribute
  selftests/bpf: Use syscall(SYS_gettid) instead of gettid() wrapper in bench
  bpf-next: Avoid goto in regs_refine_cond_op()
  bpftool: Clean up HOST_CFLAGS, HOST_LDFLAGS for bootstrap bpftool
  selftests/bpf: scale benchmark counting by using per-CPU counters
  bpftool: Remove unnecessary source files from bootstrap version
  bpftool: Enable libbpf logs when loading pid_iter in debug mode
  selftests/bpf: add raw_tp/tp_btf BPF cookie subtests
  libbpf: add support for BPF cookie for raw_tp/tp_btf programs
  bpf: support BPF cookie in raw tracepoint (raw_tp, tp_btf) programs
  bpf: pass whole link instead of prog when triggering raw tracepoint
  bpf: flatten bpf_probe_register call chain
  selftests/bpf: Prevent client connect before server bind in test_tc_tunnel.sh
  selftests/bpf: Add a sk_msg prog bpf_get_ns_current_pid_tgid() test
  ...
====================

Link: https://lore.kernel.org/r/20240325233940.7154-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

1 files changed, 32 insertions, 20 deletions

diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
index 0a5c4efc73c3..6d0c95638e1b 100644
--- a/kernel/trace/bpf_trace.c
+++ b/kernel/trace/bpf_trace.c
@@ -1053,9 +1053,15 @@ static unsigned long get_entry_ip(unsigned long fentry_ip)
 {
 	u32 instr;
 
-	/* Being extra safe in here in case entry ip is on the page-edge. */
-	if (get_kernel_nofault(instr, (u32 *) fentry_ip - 1))
-		return fentry_ip;
+	/* We want to be extra safe in case entry ip is on the page edge,
+	 * but otherwise we need to avoid get_kernel_nofault()'s overhead.
+	 */
+	if ((fentry_ip & ~PAGE_MASK) < ENDBR_INSN_SIZE) {
+		if (get_kernel_nofault(instr, (u32 *)(fentry_ip - ENDBR_INSN_SIZE)))
+			return fentry_ip;
+	} else {
+		instr = *(u32 *)(fentry_ip - ENDBR_INSN_SIZE);
+	}
 	if (is_endbr(instr))
 		fentry_ip -= ENDBR_INSN_SIZE;
 	return fentry_ip;
@@ -1525,8 +1531,6 @@ bpf_tracing_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		return &bpf_ktime_get_boot_ns_proto;
 	case BPF_FUNC_tail_call:
 		return &bpf_tail_call_proto;
-	case BPF_FUNC_get_current_pid_tgid:
-		return &bpf_get_current_pid_tgid_proto;
 	case BPF_FUNC_get_current_task:
 		return &bpf_get_current_task_proto;
 	case BPF_FUNC_get_current_task_btf:
@@ -1582,8 +1586,6 @@ bpf_tracing_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		return &bpf_send_signal_thread_proto;
 	case BPF_FUNC_perf_event_read_value:
 		return &bpf_perf_event_read_value_proto;
-	case BPF_FUNC_get_ns_current_pid_tgid:
-		return &bpf_get_ns_current_pid_tgid_proto;
 	case BPF_FUNC_ringbuf_output:
 		return &bpf_ringbuf_output_proto;
 	case BPF_FUNC_ringbuf_reserve:
@@ -2008,6 +2010,8 @@ raw_tp_prog_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		return &bpf_get_stackid_proto_raw_tp;
 	case BPF_FUNC_get_stack:
 		return &bpf_get_stack_proto_raw_tp;
+	case BPF_FUNC_get_attach_cookie:
+		return &bpf_get_attach_cookie_proto_tracing;
 	default:
 		return bpf_tracing_func_proto(func_id, prog);
 	}
@@ -2070,6 +2074,9 @@ tracing_prog_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 	case BPF_FUNC_get_func_arg_cnt:
 		return bpf_prog_has_trampoline(prog) ? &bpf_get_func_arg_cnt_proto : NULL;
 	case BPF_FUNC_get_attach_cookie:
+		if (prog->type == BPF_PROG_TYPE_TRACING &&
+		    prog->expected_attach_type == BPF_TRACE_RAW_TP)
+			return &bpf_get_attach_cookie_proto_tracing;
 		return bpf_prog_has_trampoline(prog) ? &bpf_get_attach_cookie_proto_tracing : NULL;
 	default:
 		fn = raw_tp_prog_func_proto(func_id, prog);
@@ -2370,16 +2377,26 @@ void bpf_put_raw_tracepoint(struct bpf_raw_event_map *btp)
 }
 
 static __always_inline
-void __bpf_trace_run(struct bpf_prog *prog, u64 *args)
+void __bpf_trace_run(struct bpf_raw_tp_link *link, u64 *args)
 {
+	struct bpf_prog *prog = link->link.prog;
+	struct bpf_run_ctx *old_run_ctx;
+	struct bpf_trace_run_ctx run_ctx;
+
 	cant_sleep();
 	if (unlikely(this_cpu_inc_return(*(prog->active)) != 1)) {
 		bpf_prog_inc_misses_counter(prog);
 		goto out;
 	}
+
+	run_ctx.bpf_cookie = link->cookie;
+	old_run_ctx = bpf_set_run_ctx(&run_ctx.run_ctx);
+
 	rcu_read_lock();
 	(void) bpf_prog_run(prog, args);
 	rcu_read_unlock();
+
+	bpf_reset_run_ctx(old_run_ctx);
 out:
 	this_cpu_dec(*(prog->active));
 }
@@ -2408,12 +2425,12 @@ out:
 #define __SEQ_0_11	0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11
 
 #define BPF_TRACE_DEFN_x(x)						\
-	void bpf_trace_run##x(struct bpf_prog *prog,			\
+	void bpf_trace_run##x(struct bpf_raw_tp_link *link,		\
 			      REPEAT(x, SARG, __DL_COM, __SEQ_0_11))	\
 	{								\
 		u64 args[x];						\
 		REPEAT(x, COPY, __DL_SEM, __SEQ_0_11);			\
-		__bpf_trace_run(prog, args);				\
+		__bpf_trace_run(link, args);				\
 	}								\
 	EXPORT_SYMBOL_GPL(bpf_trace_run##x)
 BPF_TRACE_DEFN_x(1);
@@ -2429,9 +2446,10 @@ BPF_TRACE_DEFN_x(10);
 BPF_TRACE_DEFN_x(11);
 BPF_TRACE_DEFN_x(12);
 
-static int __bpf_probe_register(struct bpf_raw_event_map *btp, struct bpf_prog *prog)
+int bpf_probe_register(struct bpf_raw_event_map *btp, struct bpf_raw_tp_link *link)
 {
 	struct tracepoint *tp = btp->tp;
+	struct bpf_prog *prog = link->link.prog;
 
 	/*
 	 * check that program doesn't access arguments beyond what's
@@ -2443,18 +2461,12 @@ static int __bpf_probe_register(struct bpf_raw_event_map *btp, struct bpf_prog *
 	if (prog->aux->max_tp_access > btp->writable_size)
 		return -EINVAL;
 
-	return tracepoint_probe_register_may_exist(tp, (void *)btp->bpf_func,
-						   prog);
-}
-
-int bpf_probe_register(struct bpf_raw_event_map *btp, struct bpf_prog *prog)
-{
-	return __bpf_probe_register(btp, prog);
+	return tracepoint_probe_register_may_exist(tp, (void *)btp->bpf_func, link);
 }
 
-int bpf_probe_unregister(struct bpf_raw_event_map *btp, struct bpf_prog *prog)
+int bpf_probe_unregister(struct bpf_raw_event_map *btp, struct bpf_raw_tp_link *link)
 {
-	return tracepoint_probe_unregister(btp->tp, (void *)btp->bpf_func, prog);
+	return tracepoint_probe_unregister(btp->tp, (void *)btp->bpf_func, link);
 }
 
 int bpf_get_perf_event_info(const struct perf_event *event, u32 *prog_id,