[jolsa-perf:bpf/tracing_multi_4 14/34] kernel/bpf/syscall.c:2522:22: error: implicit declaration of function 'is_tracing_multi'
From: kernel test robot
Date: Thu Jul 28 2022 - 00:39:48 EST
tree: https://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf.git bpf/tracing_multi_4
head: 1637b6b5bec11596e52cdc0a6eadfa45a15276c3
commit: bea966904e542c1a3a8d12a19f56042211d4e302 [14/34] bpf: Add multi tracing attach types
config: riscv-randconfig-r042-20220728 (https://download.01.org/0day-ci/archive/20220728/202207281228.VdBNGu2A-lkp@xxxxxxxxx/config)
compiler: riscv64-linux-gcc (GCC) 12.1.0
reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# https://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf.git/commit/?id=bea966904e542c1a3a8d12a19f56042211d4e302
git remote add jolsa-perf https://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf.git
git fetch --no-tags jolsa-perf bpf/tracing_multi_4
git checkout bea966904e542c1a3a8d12a19f56042211d4e302
# save the config file
mkdir build_dir && cp config build_dir/.config
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=riscv SHELL=/bin/bash kernel/bpf/ kernel/trace/
If you fix the issue, kindly add following tag where applicable
Reported-by: kernel test robot <lkp@xxxxxxxxx>
All errors (new ones prefixed by >>):
kernel/bpf/syscall.c: In function 'bpf_prog_load':
>> kernel/bpf/syscall.c:2522:22: error: implicit declaration of function 'is_tracing_multi' [-Werror=implicit-function-declaration]
2522 | multi_func = is_tracing_multi(attr->expected_attach_type);
| ^~~~~~~~~~~~~~~~
cc1: some warnings being treated as errors
--
kernel/bpf/verifier.c: In function 'check_attach_btf_id':
>> kernel/bpf/verifier.c:15099:20: error: implicit declaration of function 'is_tracing_multi' [-Werror=implicit-function-declaration]
15099 | } else if (is_tracing_multi(prog->expected_attach_type))
| ^~~~~~~~~~~~~~~~
cc1: some warnings being treated as errors
vim +/is_tracing_multi +2522 kernel/bpf/syscall.c
2472
2473 static int bpf_prog_load(union bpf_attr *attr, bpfptr_t uattr)
2474 {
2475 enum bpf_prog_type type = attr->prog_type;
2476 struct bpf_prog *prog, *dst_prog = NULL;
2477 struct btf *attach_btf = NULL;
2478 int err;
2479 char license[128];
2480 bool multi_func;
2481 bool is_gpl;
2482
2483 if (CHECK_ATTR(BPF_PROG_LOAD))
2484 return -EINVAL;
2485
2486 if (attr->prog_flags & ~(BPF_F_STRICT_ALIGNMENT |
2487 BPF_F_ANY_ALIGNMENT |
2488 BPF_F_TEST_STATE_FREQ |
2489 BPF_F_SLEEPABLE |
2490 BPF_F_TEST_RND_HI32 |
2491 BPF_F_XDP_HAS_FRAGS))
2492 return -EINVAL;
2493
2494 if (!IS_ENABLED(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) &&
2495 (attr->prog_flags & BPF_F_ANY_ALIGNMENT) &&
2496 !bpf_capable())
2497 return -EPERM;
2498
2499 /* copy eBPF program license from user space */
2500 if (strncpy_from_bpfptr(license,
2501 make_bpfptr(attr->license, uattr.is_kernel),
2502 sizeof(license) - 1) < 0)
2503 return -EFAULT;
2504 license[sizeof(license) - 1] = 0;
2505
2506 /* eBPF programs must be GPL compatible to use GPL-ed functions */
2507 is_gpl = license_is_gpl_compatible(license);
2508
2509 if (attr->insn_cnt == 0 ||
2510 attr->insn_cnt > (bpf_capable() ? BPF_COMPLEXITY_LIMIT_INSNS : BPF_MAXINSNS))
2511 return -E2BIG;
2512 if (type != BPF_PROG_TYPE_SOCKET_FILTER &&
2513 type != BPF_PROG_TYPE_CGROUP_SKB &&
2514 !bpf_capable())
2515 return -EPERM;
2516
2517 if (is_net_admin_prog_type(type) && !capable(CAP_NET_ADMIN) && !capable(CAP_SYS_ADMIN))
2518 return -EPERM;
2519 if (is_perfmon_prog_type(type) && !perfmon_capable())
2520 return -EPERM;
2521
> 2522 multi_func = is_tracing_multi(attr->expected_attach_type);
2523
2524 /* attach_prog_fd/attach_btf_obj_fd can specify fd of either bpf_prog
2525 * or btf, we need to check which one it is
2526 */
2527 if (attr->attach_prog_fd) {
2528 dst_prog = bpf_prog_get(attr->attach_prog_fd);
2529 if (IS_ERR(dst_prog)) {
2530 dst_prog = NULL;
2531 attach_btf = btf_get_by_fd(attr->attach_btf_obj_fd);
2532 if (IS_ERR(attach_btf))
2533 return -EINVAL;
2534 if (!btf_is_kernel(attach_btf)) {
2535 /* attaching through specifying bpf_prog's BTF
2536 * objects directly might be supported eventually
2537 */
2538 btf_put(attach_btf);
2539 return -ENOTSUPP;
2540 }
2541 }
2542 } else if (attr->attach_btf_id || multi_func) {
2543 /* fall back to vmlinux BTF, if BTF type ID is specified */
2544 attach_btf = bpf_get_btf_vmlinux();
2545 if (IS_ERR(attach_btf))
2546 return PTR_ERR(attach_btf);
2547 if (!attach_btf)
2548 return -EINVAL;
2549 btf_get(attach_btf);
2550 }
2551
2552 bpf_prog_load_fixup_attach_type(attr);
2553 if (bpf_prog_load_check_attach(type, attr->expected_attach_type,
2554 attach_btf, attr->attach_btf_id,
2555 dst_prog, multi_func)) {
2556 if (dst_prog)
2557 bpf_prog_put(dst_prog);
2558 if (attach_btf)
2559 btf_put(attach_btf);
2560 return -EINVAL;
2561 }
2562
2563 /* plain bpf_prog allocation */
2564 prog = bpf_prog_alloc(bpf_prog_size(attr->insn_cnt), GFP_USER);
2565 if (!prog) {
2566 if (dst_prog)
2567 bpf_prog_put(dst_prog);
2568 if (attach_btf)
2569 btf_put(attach_btf);
2570 return -ENOMEM;
2571 }
2572
2573 prog->expected_attach_type = attr->expected_attach_type;
2574 prog->aux->attach_btf = attach_btf;
2575 prog->aux->attach_btf_id = multi_func ? bpf_multi_func_btf_id[0] : attr->attach_btf_id;
2576 prog->aux->dst_prog = dst_prog;
2577 prog->aux->offload_requested = !!attr->prog_ifindex;
2578 prog->aux->sleepable = attr->prog_flags & BPF_F_SLEEPABLE;
2579 prog->aux->xdp_has_frags = attr->prog_flags & BPF_F_XDP_HAS_FRAGS;
2580
2581 err = security_bpf_prog_alloc(prog->aux);
2582 if (err)
2583 goto free_prog;
2584
2585 prog->aux->user = get_current_user();
2586 prog->len = attr->insn_cnt;
2587
2588 err = -EFAULT;
2589 if (copy_from_bpfptr(prog->insns,
2590 make_bpfptr(attr->insns, uattr.is_kernel),
2591 bpf_prog_insn_size(prog)) != 0)
2592 goto free_prog_sec;
2593
2594 prog->orig_prog = NULL;
2595 prog->jited = 0;
2596
2597 atomic64_set(&prog->aux->refcnt, 1);
2598 prog->gpl_compatible = is_gpl ? 1 : 0;
2599
2600 if (bpf_prog_is_dev_bound(prog->aux)) {
2601 err = bpf_prog_offload_init(prog, attr);
2602 if (err)
2603 goto free_prog_sec;
2604 }
2605
2606 /* find program type: socket_filter vs tracing_filter */
2607 err = find_prog_type(type, prog);
2608 if (err < 0)
2609 goto free_prog_sec;
2610
2611 prog->aux->load_time = ktime_get_boottime_ns();
2612 err = bpf_obj_name_cpy(prog->aux->name, attr->prog_name,
2613 sizeof(attr->prog_name));
2614 if (err < 0)
2615 goto free_prog_sec;
2616
2617 /* run eBPF verifier */
2618 err = bpf_check(&prog, attr, uattr);
2619 if (err < 0)
2620 goto free_used_maps;
2621
2622 prog = bpf_prog_select_runtime(prog, &err);
2623 if (err < 0)
2624 goto free_used_maps;
2625
2626 err = bpf_prog_alloc_id(prog);
2627 if (err)
2628 goto free_used_maps;
2629
2630 /* Upon success of bpf_prog_alloc_id(), the BPF prog is
2631 * effectively publicly exposed. However, retrieving via
2632 * bpf_prog_get_fd_by_id() will take another reference,
2633 * therefore it cannot be gone underneath us.
2634 *
2635 * Only for the time /after/ successful bpf_prog_new_fd()
2636 * and before returning to userspace, we might just hold
2637 * one reference and any parallel close on that fd could
2638 * rip everything out. Hence, below notifications must
2639 * happen before bpf_prog_new_fd().
2640 *
2641 * Also, any failure handling from this point onwards must
2642 * be using bpf_prog_put() given the program is exposed.
2643 */
2644 bpf_prog_kallsyms_add(prog);
2645 perf_event_bpf_event(prog, PERF_BPF_EVENT_PROG_LOAD, 0);
2646 bpf_audit_prog(prog, BPF_AUDIT_LOAD);
2647
2648 err = bpf_prog_new_fd(prog);
2649 if (err < 0)
2650 bpf_prog_put(prog);
2651 return err;
2652
2653 free_used_maps:
2654 /* In case we have subprogs, we need to wait for a grace
2655 * period before we can tear down JIT memory since symbols
2656 * are already exposed under kallsyms.
2657 */
2658 __bpf_prog_put_noref(prog, prog->aux->func_cnt);
2659 return err;
2660 free_prog_sec:
2661 free_uid(prog->aux->user);
2662 security_bpf_prog_free(prog->aux);
2663 free_prog:
2664 if (prog->aux->attach_btf)
2665 btf_put(prog->aux->attach_btf);
2666 bpf_prog_free(prog);
2667 return err;
2668 }
2669
--
0-DAY CI Kernel Test Service
https://01.org/lkp