| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <202207110238.dtnrv1KM-lkp@intel.com>
Date: Mon, 11 Jul 2022 02:49:38 +0800
From: kernel test robot <lkp@...el.com>
To: Jiri Olsa <jolsa@...nel.org>
Cc: kbuild-all@...ts.01.org, linux-kernel@...r.kernel.org
Subject: [jolsa-perf:bpf/tramp_27 15/33] kernel/bpf/syscall.c:2502:22: error:
implicit declaration of function 'is_tracing_multi'
tree: https://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf.git bpf/tramp_27
head: 08201713df8992b94d140b43fd4f20f1ad4aedb3
commit: ceab99fad55af7d8b76d900a15176ed8672f9191 [15/33] bpf: Add multi tracing attach types
config: x86_64-randconfig-a013 (https://download.01.org/0day-ci/archive/20220711/202207110238.dtnrv1KM-lkp@intel.com/config)
compiler: gcc-11 (Debian 11.3.0-3) 11.3.0
reproduce (this is a W=1 build):
# https://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf.git/commit/?id=ceab99fad55af7d8b76d900a15176ed8672f9191
git remote add jolsa-perf https://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf.git
git fetch --no-tags jolsa-perf bpf/tramp_27
git checkout ceab99fad55af7d8b76d900a15176ed8672f9191
# save the config file
mkdir build_dir && cp config build_dir/.config
make W=1 O=build_dir ARCH=x86_64 SHELL=/bin/bash
If you fix the issue, kindly add following tag where applicable
Reported-by: kernel test robot <lkp@...el.com>
All errors (new ones prefixed by >>):
kernel/bpf/syscall.c: In function 'bpf_prog_load':
>> kernel/bpf/syscall.c:2502:22: error: implicit declaration of function 'is_tracing_multi' [-Werror=implicit-function-declaration]
2502 | 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:15098:20: error: implicit declaration of function 'is_tracing_multi' [-Werror=implicit-function-declaration]
15098 | } else if (is_tracing_multi(prog->expected_attach_type))
| ^~~~~~~~~~~~~~~~
cc1: some warnings being treated as errors
vim +/is_tracing_multi +2502 kernel/bpf/syscall.c
2452
2453 static int bpf_prog_load(union bpf_attr *attr, bpfptr_t uattr)
2454 {
2455 enum bpf_prog_type type = attr->prog_type;
2456 struct bpf_prog *prog, *dst_prog = NULL;
2457 struct btf *attach_btf = NULL;
2458 int err;
2459 char license[128];
2460 bool multi_func;
2461 bool is_gpl;
2462
2463 if (CHECK_ATTR(BPF_PROG_LOAD))
2464 return -EINVAL;
2465
2466 if (attr->prog_flags & ~(BPF_F_STRICT_ALIGNMENT |
2467 BPF_F_ANY_ALIGNMENT |
2468 BPF_F_TEST_STATE_FREQ |
2469 BPF_F_SLEEPABLE |
2470 BPF_F_TEST_RND_HI32 |
2471 BPF_F_XDP_HAS_FRAGS))
2472 return -EINVAL;
2473
2474 if (!IS_ENABLED(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) &&
2475 (attr->prog_flags & BPF_F_ANY_ALIGNMENT) &&
2476 !bpf_capable())
2477 return -EPERM;
2478
2479 /* copy eBPF program license from user space */
2480 if (strncpy_from_bpfptr(license,
2481 make_bpfptr(attr->license, uattr.is_kernel),
2482 sizeof(license) - 1) < 0)
2483 return -EFAULT;
2484 license[sizeof(license) - 1] = 0;
2485
2486 /* eBPF programs must be GPL compatible to use GPL-ed functions */
2487 is_gpl = license_is_gpl_compatible(license);
2488
2489 if (attr->insn_cnt == 0 ||
2490 attr->insn_cnt > (bpf_capable() ? BPF_COMPLEXITY_LIMIT_INSNS : BPF_MAXINSNS))
2491 return -E2BIG;
2492 if (type != BPF_PROG_TYPE_SOCKET_FILTER &&
2493 type != BPF_PROG_TYPE_CGROUP_SKB &&
2494 !bpf_capable())
2495 return -EPERM;
2496
2497 if (is_net_admin_prog_type(type) && !capable(CAP_NET_ADMIN) && !capable(CAP_SYS_ADMIN))
2498 return -EPERM;
2499 if (is_perfmon_prog_type(type) && !perfmon_capable())
2500 return -EPERM;
2501
> 2502 multi_func = is_tracing_multi(attr->expected_attach_type);
2503
2504 /* attach_prog_fd/attach_btf_obj_fd can specify fd of either bpf_prog
2505 * or btf, we need to check which one it is
2506 */
2507 if (attr->attach_prog_fd) {
2508 dst_prog = bpf_prog_get(attr->attach_prog_fd);
2509 if (IS_ERR(dst_prog)) {
2510 dst_prog = NULL;
2511 attach_btf = btf_get_by_fd(attr->attach_btf_obj_fd);
2512 if (IS_ERR(attach_btf))
2513 return -EINVAL;
2514 if (!btf_is_kernel(attach_btf)) {
2515 /* attaching through specifying bpf_prog's BTF
2516 * objects directly might be supported eventually
2517 */
2518 btf_put(attach_btf);
2519 return -ENOTSUPP;
2520 }
2521 }
2522 } else if (attr->attach_btf_id || multi_func) {
2523 /* fall back to vmlinux BTF, if BTF type ID is specified */
2524 attach_btf = bpf_get_btf_vmlinux();
2525 if (IS_ERR(attach_btf))
2526 return PTR_ERR(attach_btf);
2527 if (!attach_btf)
2528 return -EINVAL;
2529 btf_get(attach_btf);
2530 }
2531
2532 bpf_prog_load_fixup_attach_type(attr);
2533 if (bpf_prog_load_check_attach(type, attr->expected_attach_type,
2534 attach_btf, attr->attach_btf_id,
2535 dst_prog, multi_func)) {
2536 if (dst_prog)
2537 bpf_prog_put(dst_prog);
2538 if (attach_btf)
2539 btf_put(attach_btf);
2540 return -EINVAL;
2541 }
2542
2543 /* plain bpf_prog allocation */
2544 prog = bpf_prog_alloc(bpf_prog_size(attr->insn_cnt), GFP_USER);
2545 if (!prog) {
2546 if (dst_prog)
2547 bpf_prog_put(dst_prog);
2548 if (attach_btf)
2549 btf_put(attach_btf);
2550 return -ENOMEM;
2551 }
2552
2553 prog->expected_attach_type = attr->expected_attach_type;
2554 prog->aux->attach_btf = attach_btf;
2555 prog->aux->attach_btf_id = multi_func ? bpf_multi_func_btf_id[0] : attr->attach_btf_id;
2556 prog->aux->dst_prog = dst_prog;
2557 prog->aux->offload_requested = !!attr->prog_ifindex;
2558 prog->aux->sleepable = attr->prog_flags & BPF_F_SLEEPABLE;
2559 prog->aux->xdp_has_frags = attr->prog_flags & BPF_F_XDP_HAS_FRAGS;
2560
2561 err = security_bpf_prog_alloc(prog->aux);
2562 if (err)
2563 goto free_prog;
2564
2565 prog->aux->user = get_current_user();
2566 prog->len = attr->insn_cnt;
2567
2568 err = -EFAULT;
2569 if (copy_from_bpfptr(prog->insns,
2570 make_bpfptr(attr->insns, uattr.is_kernel),
2571 bpf_prog_insn_size(prog)) != 0)
2572 goto free_prog_sec;
2573
2574 prog->orig_prog = NULL;
2575 prog->jited = 0;
2576
2577 atomic64_set(&prog->aux->refcnt, 1);
2578 prog->gpl_compatible = is_gpl ? 1 : 0;
2579
2580 if (bpf_prog_is_dev_bound(prog->aux)) {
2581 err = bpf_prog_offload_init(prog, attr);
2582 if (err)
2583 goto free_prog_sec;
2584 }
2585
2586 /* find program type: socket_filter vs tracing_filter */
2587 err = find_prog_type(type, prog);
2588 if (err < 0)
2589 goto free_prog_sec;
2590
2591 prog->aux->load_time = ktime_get_boottime_ns();
2592 err = bpf_obj_name_cpy(prog->aux->name, attr->prog_name,
2593 sizeof(attr->prog_name));
2594 if (err < 0)
2595 goto free_prog_sec;
2596
2597 /* run eBPF verifier */
2598 err = bpf_check(&prog, attr, uattr);
2599 if (err < 0)
2600 goto free_used_maps;
2601
2602 prog = bpf_prog_select_runtime(prog, &err);
2603 if (err < 0)
2604 goto free_used_maps;
2605
2606 err = bpf_prog_alloc_id(prog);
2607 if (err)
2608 goto free_used_maps;
2609
2610 /* Upon success of bpf_prog_alloc_id(), the BPF prog is
2611 * effectively publicly exposed. However, retrieving via
2612 * bpf_prog_get_fd_by_id() will take another reference,
2613 * therefore it cannot be gone underneath us.
2614 *
2615 * Only for the time /after/ successful bpf_prog_new_fd()
2616 * and before returning to userspace, we might just hold
2617 * one reference and any parallel close on that fd could
2618 * rip everything out. Hence, below notifications must
2619 * happen before bpf_prog_new_fd().
2620 *
2621 * Also, any failure handling from this point onwards must
2622 * be using bpf_prog_put() given the program is exposed.
2623 */
2624 bpf_prog_kallsyms_add(prog);
2625 perf_event_bpf_event(prog, PERF_BPF_EVENT_PROG_LOAD, 0);
2626 bpf_audit_prog(prog, BPF_AUDIT_LOAD);
2627
2628 err = bpf_prog_new_fd(prog);
2629 if (err < 0)
2630 bpf_prog_put(prog);
2631 return err;
2632
2633 free_used_maps:
2634 /* In case we have subprogs, we need to wait for a grace
2635 * period before we can tear down JIT memory since symbols
2636 * are already exposed under kallsyms.
2637 */
2638 __bpf_prog_put_noref(prog, prog->aux->func_cnt);
2639 return err;
2640 free_prog_sec:
2641 free_uid(prog->aux->user);
2642 security_bpf_prog_free(prog->aux);
2643 free_prog:
2644 if (prog->aux->attach_btf)
2645 btf_put(prog->aux->attach_btf);
2646 bpf_prog_free(prog);
2647 return err;
2648 }
2649
--
0-DAY CI Kernel Test Service
https://01.org/lkp
Powered by blists - more mailing lists