diff options
Diffstat (limited to 'tools/testing/selftests/seccomp')
| -rw-r--r-- | tools/testing/selftests/seccomp/Makefile | 3 | ||||
| -rw-r--r-- | tools/testing/selftests/seccomp/config | 1 | ||||
| -rw-r--r-- | tools/testing/selftests/seccomp/seccomp_benchmark.c | 237 | ||||
| -rw-r--r-- | tools/testing/selftests/seccomp/seccomp_bpf.c | 1063 | ||||
| -rw-r--r-- | tools/testing/selftests/seccomp/settings | 2 |
5 files changed, 1215 insertions, 91 deletions
diff --git a/tools/testing/selftests/seccomp/Makefile b/tools/testing/selftests/seccomp/Makefile index 0ebfe8b0e147..584fba487037 100644 --- a/tools/testing/selftests/seccomp/Makefile +++ b/tools/testing/selftests/seccomp/Makefile @@ -1,6 +1,7 @@ # SPDX-License-Identifier: GPL-2.0 -CFLAGS += -Wl,-no-as-needed -Wall +CFLAGS += -Wl,-no-as-needed -Wall $(KHDR_INCLUDES) LDFLAGS += -lpthread +LDLIBS += -lcap TEST_GEN_PROGS := seccomp_bpf seccomp_benchmark include ../lib.mk diff --git a/tools/testing/selftests/seccomp/config b/tools/testing/selftests/seccomp/config index 64c19d8eba79..ad431a5178fb 100644 --- a/tools/testing/selftests/seccomp/config +++ b/tools/testing/selftests/seccomp/config @@ -1,3 +1,4 @@ +CONFIG_PID_NS=y CONFIG_SECCOMP=y CONFIG_SECCOMP_FILTER=y CONFIG_USER_NS=y diff --git a/tools/testing/selftests/seccomp/seccomp_benchmark.c b/tools/testing/selftests/seccomp/seccomp_benchmark.c index 91f5a89cadac..5822e25e0217 100644 --- a/tools/testing/selftests/seccomp/seccomp_benchmark.c +++ b/tools/testing/selftests/seccomp/seccomp_benchmark.c @@ -1,20 +1,26 @@ /* * Strictly speaking, this is not a test. But it can report during test - * runs so relative performace can be measured. + * runs so relative performance can be measured. */ #define _GNU_SOURCE #include <assert.h> +#include <err.h> +#include <limits.h> +#include <sched.h> +#include <stdbool.h> +#include <stddef.h> #include <stdio.h> #include <stdlib.h> #include <time.h> #include <unistd.h> #include <linux/filter.h> #include <linux/seccomp.h> +#include <sys/param.h> #include <sys/prctl.h> #include <sys/syscall.h> #include <sys/types.h> -#define ARRAY_SIZE(a) (sizeof(a) / sizeof(a[0])) +#include "../kselftest.h" unsigned long long timing(clockid_t clk_id, unsigned long long samples) { @@ -34,10 +40,10 @@ unsigned long long timing(clockid_t clk_id, unsigned long long samples) i *= 1000000000ULL; i += finish.tv_nsec - start.tv_nsec; - printf("%lu.%09lu - %lu.%09lu = %llu (%.1fs)\n", - finish.tv_sec, finish.tv_nsec, - start.tv_sec, start.tv_nsec, - i, (double)i / 1000000000.0); + ksft_print_msg("%lu.%09lu - %lu.%09lu = %llu (%.1fs)\n", + finish.tv_sec, finish.tv_nsec, + start.tv_sec, start.tv_nsec, + i, (double)i / 1000000000.0); return i; } @@ -49,7 +55,7 @@ unsigned long long calibrate(void) pid_t pid, ret; int seconds = 15; - printf("Calibrating sample size for %d seconds worth of syscalls ...\n", seconds); + ksft_print_msg("Calibrating sample size for %d seconds worth of syscalls ...\n", seconds); samples = 0; pid = getpid(); @@ -70,66 +76,231 @@ unsigned long long calibrate(void) return samples * seconds; } +bool approx(int i_one, int i_two) +{ + /* + * This continues to be a noisy test. Instead of a 1% comparison + * go with 10%. + */ + double one = i_one, one_bump = one * 0.1; + double two = i_two, two_bump = two * 0.1; + + one_bump = one + MAX(one_bump, 2.0); + two_bump = two + MAX(two_bump, 2.0); + + /* Equal to, or within 1% or 2 digits */ + if (one == two || + (one > two && one <= two_bump) || + (two > one && two <= one_bump)) + return true; + return false; +} + +bool le(int i_one, int i_two) +{ + if (i_one <= i_two) + return true; + return false; +} + +long compare(const char *name_one, const char *name_eval, const char *name_two, + unsigned long long one, bool (*eval)(int, int), unsigned long long two, + bool skip) +{ + bool good; + + if (skip) { + ksft_test_result_skip("%s %s %s\n", name_one, name_eval, + name_two); + return 0; + } + + ksft_print_msg("\t%s %s %s (%lld %s %lld): ", name_one, name_eval, name_two, + (long long)one, name_eval, (long long)two); + if (one > INT_MAX) { + ksft_print_msg("Miscalculation! Measurement went negative: %lld\n", (long long)one); + good = false; + goto out; + } + if (two > INT_MAX) { + ksft_print_msg("Miscalculation! Measurement went negative: %lld\n", (long long)two); + good = false; + goto out; + } + + good = eval(one, two); + printf("%s\n", good ? "✔️" : "❌"); + +out: + ksft_test_result(good, "%s %s %s\n", name_one, name_eval, name_two); + + return good ? 0 : 1; +} + +/* Pin to a single CPU so the benchmark won't bounce around the system. */ +void affinity(void) +{ + long cpu; + ulong ncores = sysconf(_SC_NPROCESSORS_CONF); + cpu_set_t *setp = CPU_ALLOC(ncores); + ulong setsz = CPU_ALLOC_SIZE(ncores); + + /* + * Totally unscientific way to avoid CPUs that might be busier: + * choose the highest CPU instead of the lowest. + */ + for (cpu = ncores - 1; cpu >= 0; cpu--) { + CPU_ZERO_S(setsz, setp); + CPU_SET_S(cpu, setsz, setp); + if (sched_setaffinity(getpid(), setsz, setp) == -1) + continue; + printf("Pinned to CPU %lu of %lu\n", cpu + 1, ncores); + goto out; + } + fprintf(stderr, "Could not set CPU affinity -- calibration may not work well"); + +out: + CPU_FREE(setp); +} + int main(int argc, char *argv[]) { + struct sock_filter bitmap_filter[] = { + BPF_STMT(BPF_LD|BPF_W|BPF_ABS, offsetof(struct seccomp_data, nr)), + BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_ALLOW), + }; + struct sock_fprog bitmap_prog = { + .len = (unsigned short)ARRAY_SIZE(bitmap_filter), + .filter = bitmap_filter, + }; struct sock_filter filter[] = { + BPF_STMT(BPF_LD|BPF_W|BPF_ABS, offsetof(struct seccomp_data, args[0])), BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_ALLOW), }; struct sock_fprog prog = { .len = (unsigned short)ARRAY_SIZE(filter), .filter = filter, }; - long ret; - unsigned long long samples; - unsigned long long native, filter1, filter2; - printf("Current BPF sysctl settings:\n"); - system("sysctl net.core.bpf_jit_enable"); - system("sysctl net.core.bpf_jit_harden"); + long ret, bits; + unsigned long long samples, calc; + unsigned long long native, filter1, filter2, bitmap1, bitmap2; + unsigned long long entry, per_filter1, per_filter2; + bool skip = false; + + setbuf(stdout, NULL); + + ksft_print_header(); + ksft_set_plan(7); + + ksft_print_msg("Running on:\n"); + ksft_print_msg("%s", ""); + system("uname -a"); + + ksft_print_msg("Current BPF sysctl settings:\n"); + /* Avoid using "sysctl" which may not be installed. */ + ksft_print_msg("%s", ""); + system("grep -H . /proc/sys/net/core/bpf_jit_enable"); + ksft_print_msg("%s", ""); + system("grep -H . /proc/sys/net/core/bpf_jit_harden"); + + affinity(); if (argc > 1) samples = strtoull(argv[1], NULL, 0); else samples = calibrate(); - printf("Benchmarking %llu syscalls...\n", samples); + ksft_print_msg("Benchmarking %llu syscalls...\n", samples); /* Native call */ native = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples; - printf("getpid native: %llu ns\n", native); + ksft_print_msg("getpid native: %llu ns\n", native); ret = prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0); assert(ret == 0); - /* One filter */ - ret = prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &prog); + /* One filter resulting in a bitmap */ + ret = prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &bitmap_prog); assert(ret == 0); - filter1 = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples; - printf("getpid RET_ALLOW 1 filter: %llu ns\n", filter1); + bitmap1 = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples; + ksft_print_msg("getpid RET_ALLOW 1 filter (bitmap): %llu ns\n", bitmap1); + + /* Second filter resulting in a bitmap */ + ret = prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &bitmap_prog); + assert(ret == 0); - if (filter1 == native) - printf("No overhead measured!? Try running again with more samples.\n"); + bitmap2 = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples; + ksft_print_msg("getpid RET_ALLOW 2 filters (bitmap): %llu ns\n", bitmap2); - /* Two filters */ + /* Third filter, can no longer be converted to bitmap */ ret = prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &prog); assert(ret == 0); + filter1 = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples; + ksft_print_msg("getpid RET_ALLOW 3 filters (full): %llu ns\n", filter1); + + /* Fourth filter, can not be converted to bitmap because of filter 3 */ + ret = prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &bitmap_prog); + assert(ret == 0); + filter2 = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples; - printf("getpid RET_ALLOW 2 filters: %llu ns\n", filter2); + ksft_print_msg("getpid RET_ALLOW 4 filters (full): %llu ns\n", filter2); + + /* Estimations */ +#define ESTIMATE(fmt, var, what) do { \ + var = (what); \ + ksft_print_msg("Estimated " fmt ": %llu ns\n", var); \ + if (var > INT_MAX) { \ + skip = true; \ + ret |= 1; \ + } \ + } while (0) + + ESTIMATE("total seccomp overhead for 1 bitmapped filter", calc, + bitmap1 - native); + ESTIMATE("total seccomp overhead for 2 bitmapped filters", calc, + bitmap2 - native); + ESTIMATE("total seccomp overhead for 3 full filters", calc, + filter1 - native); + ESTIMATE("total seccomp overhead for 4 full filters", calc, + filter2 - native); + ESTIMATE("seccomp entry overhead", entry, + bitmap1 - native - (bitmap2 - bitmap1)); + ESTIMATE("seccomp per-filter overhead (last 2 diff)", per_filter1, + filter2 - filter1); + ESTIMATE("seccomp per-filter overhead (filters / 4)", per_filter2, + (filter2 - native - entry) / 4); - /* Calculations */ - printf("Estimated total seccomp overhead for 1 filter: %llu ns\n", - filter1 - native); + ksft_print_msg("Expectations:\n"); + ret |= compare("native", "≤", "1 bitmap", native, le, bitmap1, + skip); + bits = compare("native", "≤", "1 filter", native, le, filter1, + skip); + if (bits) + skip = true; - printf("Estimated total seccomp overhead for 2 filters: %llu ns\n", - filter2 - native); + ret |= compare("per-filter (last 2 diff)", "≈", "per-filter (filters / 4)", + per_filter1, approx, per_filter2, skip); + + bits = compare("1 bitmapped", "≈", "2 bitmapped", + bitmap1 - native, approx, bitmap2 - native, skip); + if (bits) { + ksft_print_msg("Skipping constant action bitmap expectations: they appear unsupported.\n"); + skip = true; + } - printf("Estimated seccomp per-filter overhead: %llu ns\n", - filter2 - filter1); + ret |= compare("entry", "≈", "1 bitmapped", entry, approx, + bitmap1 - native, skip); + ret |= compare("entry", "≈", "2 bitmapped", entry, approx, + bitmap2 - native, skip); + ret |= compare("native + entry + (per filter * 4)", "≈", "4 filters total", + entry + (per_filter1 * 4) + native, approx, filter2, + skip); - printf("Estimated seccomp entry overhead: %llu ns\n", - filter1 - native - (filter2 - filter1)); + if (ret) + ksft_print_msg("Saw unexpected benchmark result. Try running again with more samples?\n"); - return 0; + ksft_finished(); } diff --git a/tools/testing/selftests/seccomp/seccomp_bpf.c b/tools/testing/selftests/seccomp/seccomp_bpf.c index 4a180439ee9e..61acbd45ffaa 100644 --- a/tools/testing/selftests/seccomp/seccomp_bpf.c +++ b/tools/testing/selftests/seccomp/seccomp_bpf.c @@ -46,6 +46,8 @@ #include <sys/ioctl.h> #include <linux/kcmp.h> #include <sys/resource.h> +#include <sys/capability.h> +#include <linux/perf_event.h> #include <unistd.h> #include <sys/syscall.h> @@ -59,10 +61,18 @@ #define SKIP(s, ...) XFAIL(s, ##__VA_ARGS__) #endif +#ifndef MIN +#define MIN(X, Y) ((X) < (Y) ? (X) : (Y)) +#endif + #ifndef PR_SET_PTRACER # define PR_SET_PTRACER 0x59616d61 #endif +#ifndef noinline +#define noinline __attribute__((noinline)) +#endif + #ifndef PR_SET_NO_NEW_PRIVS #define PR_SET_NO_NEW_PRIVS 38 #define PR_GET_NO_NEW_PRIVS 39 @@ -125,6 +135,8 @@ struct seccomp_data { # define __NR_seccomp 277 # elif defined(__csky__) # define __NR_seccomp 277 +# elif defined(__loongarch__) +# define __NR_seccomp 277 # elif defined(__hppa__) # define __NR_seccomp 338 # elif defined(__powerpc__) @@ -135,12 +147,20 @@ struct seccomp_data { # define __NR_seccomp 337 # elif defined(__sh__) # define __NR_seccomp 372 +# elif defined(__mc68000__) +# define __NR_seccomp 380 # else # warning "seccomp syscall number unknown for this architecture" # define __NR_seccomp 0xffff # endif #endif +#ifndef __NR_uretprobe +# if defined(__x86_64__) +# define __NR_uretprobe 335 +# endif +#endif + #ifndef SECCOMP_SET_MODE_STRICT #define SECCOMP_SET_MODE_STRICT 0 #endif @@ -235,6 +255,10 @@ struct seccomp_notif_addfd { }; #endif +#ifndef SECCOMP_ADDFD_FLAG_SEND +#define SECCOMP_ADDFD_FLAG_SEND (1UL << 1) /* Addfd and return it, atomically */ +#endif + struct seccomp_notif_addfd_small { __u64 id; char weird[4]; @@ -264,6 +288,10 @@ struct seccomp_notif_addfd_big { #define SECCOMP_FILTER_FLAG_TSYNC_ESRCH (1UL << 4) #endif +#ifndef SECCOMP_FILTER_FLAG_WAIT_KILLABLE_RECV +#define SECCOMP_FILTER_FLAG_WAIT_KILLABLE_RECV (1UL << 5) +#endif + #ifndef seccomp int seccomp(unsigned int op, unsigned int flags, void *args) { @@ -272,12 +300,12 @@ int seccomp(unsigned int op, unsigned int flags, void *args) } #endif -#if __BYTE_ORDER == __LITTLE_ENDIAN +#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__ #define syscall_arg(_n) (offsetof(struct seccomp_data, args[_n])) -#elif __BYTE_ORDER == __BIG_ENDIAN +#elif __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__ #define syscall_arg(_n) (offsetof(struct seccomp_data, args[_n]) + sizeof(__u32)) #else -#error "wut? Unknown __BYTE_ORDER?!" +#error "wut? Unknown __BYTE_ORDER__?!" #endif #define SIBLING_EXIT_UNKILLED 0xbadbeef @@ -315,7 +343,7 @@ TEST(kcmp) ret = __filecmp(getpid(), getpid(), 1, 1); EXPECT_EQ(ret, 0); if (ret != 0 && errno == ENOSYS) - SKIP(return, "Kernel does not support kcmp() (missing CONFIG_CHECKPOINT_RESTORE?)"); + SKIP(return, "Kernel does not support kcmp() (missing CONFIG_KCMP?)"); } TEST(mode_strict_support) @@ -381,6 +409,8 @@ TEST(mode_filter_without_nnp) .filter = filter, }; long ret; + cap_t cap = cap_get_proc(); + cap_flag_value_t is_cap_sys_admin = 0; ret = prctl(PR_GET_NO_NEW_PRIVS, 0, NULL, 0, 0); ASSERT_LE(0, ret) { @@ -389,8 +419,8 @@ TEST(mode_filter_without_nnp) errno = 0; ret = prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &prog, 0, 0); /* Succeeds with CAP_SYS_ADMIN, fails without */ - /* TODO(wad) check caps not euid */ - if (geteuid()) { + cap_get_flag(cap, CAP_SYS_ADMIN, CAP_EFFECTIVE, &is_cap_sys_admin); + if (!is_cap_sys_admin) { EXPECT_EQ(-1, ret); EXPECT_EQ(EACCES, errno); } else { @@ -767,7 +797,7 @@ void *kill_thread(void *data) bool die = (bool)data; if (die) { - prctl(PR_GET_SECCOMP, 0, 0, 0, 0); + syscall(__NR_getpid); return (void *)SIBLING_EXIT_FAILURE; } @@ -786,11 +816,11 @@ void kill_thread_or_group(struct __test_metadata *_metadata, { pthread_t thread; void *status; - /* Kill only when calling __NR_prctl. */ + /* Kill only when calling __NR_getpid. */ struct sock_filter filter_thread[] = { BPF_STMT(BPF_LD|BPF_W|BPF_ABS, offsetof(struct seccomp_data, nr)), - BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_prctl, 0, 1), + BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_getpid, 0, 1), BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_KILL_THREAD), BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_ALLOW), }; @@ -798,11 +828,11 @@ void kill_thread_or_group(struct __test_metadata *_metadata, .len = (unsigned short)ARRAY_SIZE(filter_thread), .filter = filter_thread, }; - int kill = kill_how == KILL_PROCESS ? SECCOMP_RET_KILL_PROCESS : 0xAAAAAAAAA; + int kill = kill_how == KILL_PROCESS ? SECCOMP_RET_KILL_PROCESS : 0xAAAAAAAA; struct sock_filter filter_process[] = { BPF_STMT(BPF_LD|BPF_W|BPF_ABS, offsetof(struct seccomp_data, nr)), - BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_prctl, 0, 1), + BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_getpid, 0, 1), BPF_STMT(BPF_RET|BPF_K, kill), BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_ALLOW), }; @@ -951,7 +981,7 @@ TEST(ERRNO_valid) ASSERT_EQ(0, ret); EXPECT_EQ(parent, syscall(__NR_getppid)); - EXPECT_EQ(-1, read(0, NULL, 0)); + EXPECT_EQ(-1, read(-1, NULL, 0)); EXPECT_EQ(E2BIG, errno); } @@ -970,7 +1000,7 @@ TEST(ERRNO_zero) EXPECT_EQ(parent, syscall(__NR_getppid)); /* "errno" of 0 is ok. */ - EXPECT_EQ(0, read(0, NULL, 0)); + EXPECT_EQ(0, read(-1, NULL, 0)); } /* @@ -991,7 +1021,7 @@ TEST(ERRNO_capped) ASSERT_EQ(0, ret); EXPECT_EQ(parent, syscall(__NR_getppid)); - EXPECT_EQ(-1, read(0, NULL, 0)); + EXPECT_EQ(-1, read(-1, NULL, 0)); EXPECT_EQ(4095, errno); } @@ -1022,7 +1052,7 @@ TEST(ERRNO_order) ASSERT_EQ(0, ret); EXPECT_EQ(parent, syscall(__NR_getppid)); - EXPECT_EQ(-1, read(0, NULL, 0)); + EXPECT_EQ(-1, read(-1, NULL, 0)); EXPECT_EQ(12, errno); } @@ -1483,7 +1513,7 @@ TEST_F(precedence, log_is_fifth_in_any_order) #define PTRACE_EVENT_SECCOMP 7 #endif -#define IS_SECCOMP_EVENT(status) ((status >> 16) == PTRACE_EVENT_SECCOMP) +#define PTRACE_EVENT_MASK(status) ((status) >> 16) bool tracer_running; void tracer_stop(int sig) { @@ -1535,12 +1565,22 @@ void start_tracer(struct __test_metadata *_metadata, int fd, pid_t tracee, if (wait(&status) != tracee) continue; - if (WIFSIGNALED(status) || WIFEXITED(status)) - /* Child is dead. Time to go. */ + + if (WIFSIGNALED(status)) { + /* Child caught a fatal signal. */ return; + } + if (WIFEXITED(status)) { + /* Child exited with code. */ + return; + } - /* Check if this is a seccomp event. */ - ASSERT_EQ(!ptrace_syscall, IS_SECCOMP_EVENT(status)); + /* Check if we got an expected event. */ + ASSERT_EQ(WIFCONTINUED(status), false); + ASSERT_EQ(WIFSTOPPED(status), true); + ASSERT_EQ(WSTOPSIG(status) & SIGTRAP, SIGTRAP) { + TH_LOG("Unexpected WSTOPSIG: %d", WSTOPSIG(status)); + } tracer_func(_metadata, tracee, status, args); @@ -1549,7 +1589,7 @@ void start_tracer(struct __test_metadata *_metadata, int fd, pid_t tracee, ASSERT_EQ(0, ret); } /* Directly report the status of our test harness results. */ - syscall(__NR_exit, _metadata->passed ? EXIT_SUCCESS : EXIT_FAILURE); + syscall(__NR_exit, _metadata->exit_code); } /* Common tracer setup/teardown functions. */ @@ -1589,14 +1629,8 @@ void teardown_trace_fixture(struct __test_metadata *_metadata, { if (tracer) { int status; - /* - * Extract the exit code from the other process and - * adopt it for ourselves in case its asserts failed. - */ ASSERT_EQ(0, kill(tracer, SIGUSR1)); ASSERT_EQ(tracer, waitpid(tracer, &status, 0)); - if (WEXITSTATUS(status)) - _metadata->passed = 0; } } @@ -1730,6 +1764,10 @@ TEST_F(TRACE_poke, getpid_runs_normally) NT_ARM_SYSTEM_CALL, &__v)); \ } while (0) # define SYSCALL_RET(_regs) (_regs).regs[0] +#elif defined(__loongarch__) +# define ARCH_REGS struct user_pt_regs +# define SYSCALL_NUM(_regs) (_regs).regs[11] +# define SYSCALL_RET(_regs) (_regs).regs[4] #elif defined(__riscv) && __riscv_xlen == 64 # define ARCH_REGS struct user_regs_struct # define SYSCALL_NUM(_regs) (_regs).a7 @@ -1753,16 +1791,25 @@ TEST_F(TRACE_poke, getpid_runs_normally) # define SYSCALL_RET_SET(_regs, _val) \ do { \ typeof(_val) _result = (_val); \ - /* \ - * A syscall error is signaled by CR0 SO bit \ - * and the code is stored as a positive value. \ - */ \ - if (_result < 0) { \ - SYSCALL_RET(_regs) = -result; \ - (_regs).ccr |= 0x10000000; \ + if ((_regs.trap & 0xfff0) == 0x3000) { \ + /* \ + * scv 0 system call uses -ve result \ + * for error, so no need to adjust. \ + */ \ + SYSCALL_RET(_regs) = _result; \ } else { \ - SYSCALL_RET(_regs) = result; \ - (_regs).ccr &= ~0x10000000; \ + /* \ + * A syscall error is signaled by the \ + * CR0 SO bit and the code is stored as \ + * a positive value. \ + */ \ + if (_result < 0) { \ + SYSCALL_RET(_regs) = -_result; \ + (_regs).ccr |= 0x10000000; \ + } else { \ + SYSCALL_RET(_regs) = _result; \ + (_regs).ccr &= ~0x10000000; \ + } \ } \ } while (0) # define SYSCALL_RET_SET_ON_PTRACE_EXIT @@ -1804,8 +1851,12 @@ TEST_F(TRACE_poke, getpid_runs_normally) #define SYSCALL_RET(_regs) (_regs).a[(_regs).windowbase * 4 + 2] #elif defined(__sh__) # define ARCH_REGS struct pt_regs -# define SYSCALL_NUM(_regs) (_regs).gpr[3] -# define SYSCALL_RET(_regs) (_regs).gpr[0] +# define SYSCALL_NUM(_regs) (_regs).regs[3] +# define SYSCALL_RET(_regs) (_regs).regs[0] +#elif defined(__mc68000__) +# define ARCH_REGS struct user_regs_struct +# define SYSCALL_NUM(_regs) (_regs).orig_d0 +# define SYSCALL_RET(_regs) (_regs).d0 #else # error "Do not know how to find your architecture's registers and syscalls" #endif @@ -1870,7 +1921,7 @@ const bool ptrace_entry_set_syscall_ret = * Use PTRACE_GETREGS and PTRACE_SETREGS when available. This is useful for * architectures without HAVE_ARCH_TRACEHOOK (e.g. User-mode Linux). */ -#if defined(__x86_64__) || defined(__i386__) || defined(__mips__) +#if defined(__x86_64__) || defined(__i386__) || defined(__mips__) || defined(__mc68000__) # define ARCH_GETREGS(_regs) ptrace(PTRACE_GETREGS, tracee, 0, &(_regs)) # define ARCH_SETREGS(_regs) ptrace(PTRACE_SETREGS, tracee, 0, &(_regs)) #else @@ -1948,6 +1999,11 @@ void tracer_seccomp(struct __test_metadata *_metadata, pid_t tracee, int ret; unsigned long msg; + EXPECT_EQ(PTRACE_EVENT_MASK(status), PTRACE_EVENT_SECCOMP) { + TH_LOG("Unexpected ptrace event: %d", PTRACE_EVENT_MASK(status)); + return; + } + /* Make sure we got the right message. */ ret = ptrace(PTRACE_GETEVENTMSG, tracee, NULL, &msg); EXPECT_EQ(0, ret); @@ -1998,6 +2054,11 @@ void tracer_ptrace(struct __test_metadata *_metadata, pid_t tracee, long *syscall_nr = NULL, *syscall_ret = NULL; FIXTURE_DATA(TRACE_syscall) *self = args; + EXPECT_EQ(WSTOPSIG(status) & 0x80, 0x80) { + TH_LOG("Unexpected WSTOPSIG: %d", WSTOPSIG(status)); + return; + } + /* * The traditional way to tell PTRACE_SYSCALL entry/exit * is by counting. @@ -2115,6 +2176,7 @@ FIXTURE_SETUP(TRACE_syscall) ret = prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0); ASSERT_EQ(0, ret); + /* Do not install seccomp rewrite filters, as we'll use ptrace instead. */ if (variant->use_ptrace) return; @@ -2129,6 +2191,9 @@ FIXTURE_TEARDOWN(TRACE_syscall) TEST(negative_ENOSYS) { +#if defined(__arm__) + SKIP(return, "arm32 does not support calling syscall -1"); +#endif /* * There should be no difference between an "internal" skip * and userspace asking for syscall "-1". @@ -2173,6 +2238,29 @@ TEST_F(TRACE_syscall, syscall_faked) EXPECT_SYSCALL_RETURN(45000, syscall(__NR_gettid)); } +TEST_F_SIGNAL(TRACE_syscall, kill_immediate, SIGSYS) +{ + struct sock_filter filter[] = { + BPF_STMT(BPF_LD|BPF_W|BPF_ABS, + offsetof(struct seccomp_data, nr)), + BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_mknodat, 0, 1), + BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_KILL_THREAD), + BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_ALLOW), + }; + struct sock_fprog prog = { + .len = (unsigned short)ARRAY_SIZE(filter), + .filter = filter, + }; + long ret; + + /* Install "kill on mknodat" filter. */ + ret = prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &prog, 0, 0); + ASSERT_EQ(0, ret); + + /* This should immediately die with SIGSYS, regardless of tracer. */ + EXPECT_EQ(-1, syscall(__NR_mknodat, -1, NULL, 0, 0)); +} + TEST_F(TRACE_syscall, skip_after) { struct sock_filter filter[] = { @@ -2566,7 +2654,7 @@ void *tsync_sibling(void *data) ret = prctl(PR_GET_NO_NEW_PRIVS, 0, 0, 0, 0); if (!ret) return (void *)SIBLING_EXIT_NEWPRIVS; - read(0, NULL, 0); + read(-1, NULL, 0); return (void *)SIBLING_EXIT_UNKILLED; } @@ -2994,7 +3082,8 @@ TEST(syscall_restart) timeout.tv_sec = 1; errno = 0; EXPECT_EQ(0, nanosleep(&timeout, NULL)) { - TH_LOG("Call to nanosleep() failed (errno %d)", errno); + TH_LOG("Call to nanosleep() failed (errno %d: %s)", + errno, strerror(errno)); } /* Read final sync from parent. */ @@ -3006,8 +3095,7 @@ TEST(syscall_restart) } /* Directly report the status of our test harness results. */ - syscall(__NR_exit, _metadata->passed ? EXIT_SUCCESS - : EXIT_FAILURE); + syscall(__NR_exit, _metadata->exit_code); } EXPECT_EQ(0, close(pipefd[0])); @@ -3072,12 +3160,15 @@ TEST(syscall_restart) ret = get_syscall(_metadata, child_pid); #if defined(__arm__) /* - * FIXME: * - native ARM registers do NOT expose true syscall. * - compat ARM registers on ARM64 DO expose true syscall. + * - values of utsbuf.machine include 'armv8l' or 'armb8b' + * for ARM64 running in compat mode. */ ASSERT_EQ(0, uname(&utsbuf)); - if (strncmp(utsbuf.machine, "arm", 3) == 0) { + if ((strncmp(utsbuf.machine, "arm", 3) == 0) && + (strncmp(utsbuf.machine, "armv8l", 6) != 0) && + (strncmp(utsbuf.machine, "armv8b", 6) != 0)) { EXPECT_EQ(__NR_nanosleep, ret); } else #endif @@ -3092,7 +3183,7 @@ TEST(syscall_restart) ASSERT_EQ(child_pid, waitpid(child_pid, &status, 0)); if (WIFSIGNALED(status) || WEXITSTATUS(status)) - _metadata->passed = 0; + _metadata->exit_code = KSFT_FAIL; } TEST_SIGNAL(filter_flag_log, SIGSYS) @@ -3627,7 +3718,12 @@ TEST(user_notification_sibling_pid_ns) ASSERT_GE(pid, 0); if (pid == 0) { - ASSERT_EQ(unshare(CLONE_NEWPID), 0); + ASSERT_EQ(unshare(CLONE_NEWPID), 0) { + if (errno == EPERM) + SKIP(return, "CLONE_NEWPID requires CAP_SYS_ADMIN"); + else if (errno == EINVAL) + SKIP(return, "CLONE_NEWPID is invalid (missing CONFIG_PID_NS?)"); + } pid2 = fork(); ASSERT_GE(pid2, 0); @@ -3645,6 +3741,8 @@ TEST(user_notification_sibling_pid_ns) ASSERT_EQ(unshare(CLONE_NEWPID), 0) { if (errno == EPERM) SKIP(return, "CLONE_NEWPID requires CAP_SYS_ADMIN"); + else if (errno == EINVAL) + SKIP(return, "CLONE_NEWPID is invalid (missing CONFIG_PID_NS?)"); } ASSERT_EQ(errno, 0); @@ -3685,7 +3783,10 @@ TEST(user_notification_fault_recv) struct seccomp_notif req = {}; struct seccomp_notif_resp resp = {}; - ASSERT_EQ(unshare(CLONE_NEWUSER), 0); + ASSERT_EQ(unshare(CLONE_NEWUSER), 0) { + if (errno == EINVAL) + SKIP(return, "kernel missing CLONE_NEWUSER support"); + } listener = user_notif_syscall(__NR_getppid, SECCOMP_FILTER_FLAG_NEW_LISTENER); @@ -3827,6 +3928,9 @@ TEST(user_notification_filter_empty) TH_LOG("Kernel does not support PR_SET_NO_NEW_PRIVS!"); } + if (__NR_clone3 < 0) + SKIP(return, "Test not built with clone3 support"); + pid = sys_clone3(&args, sizeof(args)); ASSERT_GE(pid, 0); @@ -3860,6 +3964,60 @@ TEST(user_notification_filter_empty) EXPECT_GT((pollfd.revents & POLLHUP) ?: 0, 0); } +TEST(user_ioctl_notification_filter_empty) +{ + pid_t pid; + long ret; + int status, p[2]; + struct __clone_args args = { + .flags = CLONE_FILES, + .exit_signal = SIGCHLD, + }; + struct seccomp_notif req = {}; + + ret = prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0); + ASSERT_EQ(0, ret) { + TH_LOG("Kernel does not support PR_SET_NO_NEW_PRIVS!"); + } + + if (__NR_clone3 < 0) + SKIP(return, "Test not built with clone3 support"); + + ASSERT_EQ(0, pipe(p)); + + pid = sys_clone3(&args, sizeof(args)); + ASSERT_GE(pid, 0); + + if (pid == 0) { + int listener; + + listener = user_notif_syscall(__NR_mknodat, SECCOMP_FILTER_FLAG_NEW_LISTENER); + if (listener < 0) + _exit(EXIT_FAILURE); + + if (dup2(listener, 200) != 200) + _exit(EXIT_FAILURE); + close(p[1]); + close(listener); + sleep(1); + + _exit(EXIT_SUCCESS); + } + if (read(p[0], &status, 1) != 0) + _exit(EXIT_SUCCESS); + close(p[0]); + /* + * The seccomp filter has become unused so we should be notified once + * the kernel gets around to cleaning up task struct. + */ + EXPECT_EQ(ioctl(200, SECCOMP_IOCTL_NOTIF_RECV, &req), -1); + EXPECT_EQ(errno, ENOENT); + + EXPECT_EQ(waitpid(pid, &status, 0), pid); + EXPECT_EQ(true, WIFEXITED(status)); + EXPECT_EQ(0, WEXITSTATUS(status)); +} + static void *do_thread(void *data) { return NULL; @@ -3881,6 +4039,9 @@ TEST(user_notification_filter_empty_threaded) TH_LOG("Kernel does not support PR_SET_NO_NEW_PRIVS!"); } + if (__NR_clone3 < 0) + SKIP(return, "Test not built with clone3 support"); + pid = sys_clone3(&args, sizeof(args)); ASSERT_GE(pid, 0); @@ -3946,11 +4107,21 @@ TEST(user_notification_filter_empty_threaded) EXPECT_GT((pollfd.revents & POLLHUP) ?: 0, 0); } + +int get_next_fd(int prev_fd) +{ + for (int i = prev_fd + 1; i < FD_SETSIZE; ++i) { + if (fcntl(i, F_GETFD) == -1) + return i; + } + _exit(EXIT_FAILURE); +} + TEST(user_notification_addfd) { pid_t pid; long ret; - int status, listener, memfd, fd; + int status, listener, memfd, fd, nextfd; struct seccomp_notif_addfd addfd = {}; struct seccomp_notif_addfd_small small = {}; struct seccomp_notif_addfd_big big = {}; @@ -3959,25 +4130,35 @@ TEST(user_notification_addfd) /* 100 ms */ struct timespec delay = { .tv_nsec = 100000000 }; + /* There may be arbitrary already-open fds at test start. */ memfd = memfd_create("test", 0); ASSERT_GE(memfd, 0); + nextfd = get_next_fd(memfd); ret = prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0); ASSERT_EQ(0, ret) { TH_LOG("Kernel does not support PR_SET_NO_NEW_PRIVS!"); } + /* fd: 4 */ /* Check that the basic notification machinery works */ listener = user_notif_syscall(__NR_getppid, SECCOMP_FILTER_FLAG_NEW_LISTENER); - ASSERT_GE(listener, 0); + ASSERT_EQ(listener, nextfd); + nextfd = get_next_fd(nextfd); pid = fork(); ASSERT_GE(pid, 0); if (pid == 0) { + /* fds will be added and this value is expected */ if (syscall(__NR_getppid) != USER_NOTIF_MAGIC) exit(1); + + /* Atomic addfd+send is received here. Check it is a valid fd */ + if (fcntl(syscall(__NR_getppid), F_GETFD) == -1) + exit(1); + exit(syscall(__NR_getppid) != USER_NOTIF_MAGIC); } @@ -4019,18 +4200,16 @@ TEST(user_notification_addfd) /* Verify we can set an arbitrary remote fd */ fd = ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD, &addfd); - /* - * The child has fds 0(stdin), 1(stdout), 2(stderr), 3(memfd), - * 4(listener), so the newly allocated fd should be 5. - */ - EXPECT_EQ(fd, 5); + EXPECT_EQ(fd, nextfd); + nextfd = get_next_fd(nextfd); EXPECT_EQ(filecmp(getpid(), pid, memfd, fd), 0); /* Verify we can set an arbitrary remote fd with large size */ memset(&big, 0x0, sizeof(big)); big.addfd = addfd; fd = ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD_BIG, &big); - EXPECT_EQ(fd, 6); + EXPECT_EQ(fd, nextfd); + nextfd = get_next_fd(nextfd); /* Verify we can set a specific remote fd */ addfd.newfd = 42; @@ -4060,6 +4239,33 @@ TEST(user_notification_addfd) ASSERT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_RECV, &req), 0); ASSERT_EQ(addfd.id, req.id); + /* Verify we can do an atomic addfd and send */ + addfd.newfd = 0; + addfd.flags = SECCOMP_ADDFD_FLAG_SEND; + fd = ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD, &addfd); + /* + * Child has earlier "low" fds and now 42, so we expect the next + * lowest available fd to be assigned here. + */ + EXPECT_EQ(fd, nextfd); + nextfd = get_next_fd(nextfd); + ASSERT_EQ(filecmp(getpid(), pid, memfd, fd), 0); + + /* + * This sets the ID of the ADD FD to the last request plus 1. The + * notification ID increments 1 per notification. + */ + addfd.id = req.id + 1; + + /* This spins until the underlying notification is generated */ + while (ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD, &addfd) != -1 && + errno != -EINPROGRESS) + nanosleep(&delay, NULL); + + memset(&req, 0, sizeof(req)); + ASSERT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_RECV, &req), 0); + ASSERT_EQ(addfd.id, req.id); + resp.id = req.id; resp.error = 0; resp.val = USER_NOTIF_MAGIC; @@ -4120,6 +4326,10 @@ TEST(user_notification_addfd_rlimit) EXPECT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD, &addfd), -1); EXPECT_EQ(errno, EMFILE); + addfd.flags = SECCOMP_ADDFD_FLAG_SEND; + EXPECT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD, &addfd), -1); + EXPECT_EQ(errno, EMFILE); + addfd.newfd = 100; addfd.flags = SECCOMP_ADDFD_FLAG_SETFD; EXPECT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_ADDFD, &addfd), -1); @@ -4139,6 +4349,747 @@ TEST(user_notification_addfd_rlimit) close(memfd); } +#ifndef SECCOMP_USER_NOTIF_FD_SYNC_WAKE_UP +#define SECCOMP_USER_NOTIF_FD_SYNC_WAKE_UP (1UL << 0) +#define SECCOMP_IOCTL_NOTIF_SET_FLAGS SECCOMP_IOW(4, __u64) +#endif + +TEST(user_notification_sync) +{ + struct seccomp_notif req = {}; + struct seccomp_notif_resp resp = {}; + int status, listener; + pid_t pid; + long ret; + + ret = prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0); + ASSERT_EQ(0, ret) { + TH_LOG("Kernel does not support PR_SET_NO_NEW_PRIVS!"); + } + + listener = user_notif_syscall(__NR_getppid, + SECCOMP_FILTER_FLAG_NEW_LISTENER); + ASSERT_GE(listener, 0); + + /* Try to set invalid flags. */ + EXPECT_SYSCALL_RETURN(-EINVAL, + ioctl(listener, SECCOMP_IOCTL_NOTIF_SET_FLAGS, 0xffffffff, 0)); + + ASSERT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_SET_FLAGS, + SECCOMP_USER_NOTIF_FD_SYNC_WAKE_UP, 0), 0); + + pid = fork(); + ASSERT_GE(pid, 0); + if (pid == 0) { + ret = syscall(__NR_getppid); + ASSERT_EQ(ret, USER_NOTIF_MAGIC) { + _exit(1); + } + _exit(0); + } + + req.pid = 0; + ASSERT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_RECV, &req), 0); + + ASSERT_EQ(req.data.nr, __NR_getppid); + + resp.id = req.id; + resp.error = 0; + resp.val = USER_NOTIF_MAGIC; + resp.flags = 0; + ASSERT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_SEND, &resp), 0); + + ASSERT_EQ(waitpid(pid, &status, 0), pid); + ASSERT_EQ(status, 0); +} + + +/* Make sure PTRACE_O_SUSPEND_SECCOMP requires CAP_SYS_ADMIN. */ +FIXTURE(O_SUSPEND_SECCOMP) { + pid_t pid; +}; + +FIXTURE_SETUP(O_SUSPEND_SECCOMP) +{ + ERRNO_FILTER(block_read, E2BIG); + cap_value_t cap_list[] = { CAP_SYS_ADMIN }; + cap_t caps; + + self->pid = 0; + + /* make sure we don't have CAP_SYS_ADMIN */ + caps = cap_get_proc(); + ASSERT_NE(NULL, caps); + ASSERT_EQ(0, cap_set_flag(caps, CAP_EFFECTIVE, 1, cap_list, CAP_CLEAR)); + ASSERT_EQ(0, cap_set_proc(caps)); + cap_free(caps); + + ASSERT_EQ(0, prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0)); + ASSERT_EQ(0, prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &prog_block_read)); + + self->pid = fork(); + ASSERT_GE(self->pid, 0); + + if (self->pid == 0) { + while (1) + pause(); + _exit(127); + } +} + +FIXTURE_TEARDOWN(O_SUSPEND_SECCOMP) +{ + if (self->pid) + kill(self->pid, SIGKILL); +} + +TEST_F(O_SUSPEND_SECCOMP, setoptions) +{ + int wstatus; + + ASSERT_EQ(0, ptrace(PTRACE_ATTACH, self->pid, NULL, 0)); + ASSERT_EQ(self->pid, wait(&wstatus)); + ASSERT_EQ(-1, ptrace(PTRACE_SETOPTIONS, self->pid, NULL, PTRACE_O_SUSPEND_SECCOMP)); + if (errno == EINVAL) + SKIP(return, "Kernel does not support PTRACE_O_SUSPEND_SECCOMP (missing CONFIG_CHECKPOINT_RESTORE?)"); + ASSERT_EQ(EPERM, errno); +} + +TEST_F(O_SUSPEND_SECCOMP, seize) +{ + int ret; + + ret = ptrace(PTRACE_SEIZE, self->pid, NULL, PTRACE_O_SUSPEND_SECCOMP); + ASSERT_EQ(-1, ret); + if (errno == EINVAL) + SKIP(return, "Kernel does not support PTRACE_O_SUSPEND_SECCOMP (missing CONFIG_CHECKPOINT_RESTORE?)"); + ASSERT_EQ(EPERM, errno); +} + +/* + * get_nth - Get the nth, space separated entry in a file. + * + * Returns the length of the read field. + * Throws error if field is zero-lengthed. + */ +static ssize_t get_nth(struct __test_metadata *_metadata, const char *path, + const unsigned int position, char **entry) +{ + char *line = NULL; + unsigned int i; + ssize_t nread; + size_t len = 0; + FILE *f; + + f = fopen(path, "r"); + ASSERT_NE(f, NULL) { + TH_LOG("Could not open %s: %s", path, strerror(errno)); + } + + for (i = 0; i < position; i++) { + nread = getdelim(&line, &len, ' ', f); + ASSERT_GE(nread, 0) { + TH_LOG("Failed to read %d entry in file %s", i, path); + } + } + fclose(f); + + ASSERT_GT(nread, 0) { + TH_LOG("Entry in file %s had zero length", path); + } + + *entry = line; + return nread - 1; +} + +/* For a given PID, get the task state (D, R, etc...) */ +static char get_proc_stat(struct __test_metadata *_metadata, pid_t pid) +{ + char proc_path[100] = {0}; + char status; + char *line; + + snprintf(proc_path, sizeof(proc_path), "/proc/%d/stat", pid); + ASSERT_EQ(get_nth(_metadata, proc_path, 3, &line), 1); + + status = *line; + free(line); + + return status; +} + +TEST(user_notification_fifo) +{ + struct seccomp_notif_resp resp = {}; + struct seccomp_notif req = {}; + int i, status, listener; + pid_t pid, pids[3]; + __u64 baseid; + long ret; + /* 100 ms */ + struct timespec delay = { .tv_nsec = 100000000 }; + + ret = prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0); + ASSERT_EQ(0, ret) { + TH_LOG("Kernel does not support PR_SET_NO_NEW_PRIVS!"); + } + + /* Setup a listener */ + listener = user_notif_syscall(__NR_getppid, + SECCOMP_FILTER_FLAG_NEW_LISTENER); + ASSERT_GE(listener, 0); + + pid = fork(); + ASSERT_GE(pid, 0); + + if (pid == 0) { + ret = syscall(__NR_getppid); + exit(ret != USER_NOTIF_MAGIC); + } + + EXPECT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_RECV, &req), 0); + baseid = req.id + 1; + + resp.id = req.id; + resp.error = 0; + resp.val = USER_NOTIF_MAGIC; + + /* check that we make sure flags == 0 */ + EXPECT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_SEND, &resp), 0); + + EXPECT_EQ(waitpid(pid, &status, 0), pid); + EXPECT_EQ(true, WIFEXITED(status)); + EXPECT_EQ(0, WEXITSTATUS(status)); + + /* Start children, and generate notifications */ + for (i = 0; i < ARRAY_SIZE(pids); i++) { + pid = fork(); + if (pid == 0) { + ret = syscall(__NR_getppid); + exit(ret != USER_NOTIF_MAGIC); + } + pids[i] = pid; + } + + /* This spins until all of the children are sleeping */ +restart_wait: + for (i = 0; i < ARRAY_SIZE(pids); i++) { + if (get_proc_stat(_metadata, pids[i]) != 'S') { + nanosleep(&delay, NULL); + goto restart_wait; + } + } + + /* Read the notifications in order (and respond) */ + for (i = 0; i < ARRAY_SIZE(pids); i++) { + memset(&req, 0, sizeof(req)); + EXPECT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_RECV, &req), 0); + EXPECT_EQ(req.id, baseid + i); + resp.id = req.id; + EXPECT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_SEND, &resp), 0); + } + + /* Make sure notifications were received */ + for (i = 0; i < ARRAY_SIZE(pids); i++) { + EXPECT_EQ(waitpid(pids[i], &status, 0), pids[i]); + EXPECT_EQ(true, WIFEXITED(status)); + EXPECT_EQ(0, WEXITSTATUS(status)); + } +} + +/* get_proc_syscall - Get the syscall in progress for a given pid + * + * Returns the current syscall number for a given process + * Returns -1 if not in syscall (running or blocked) + */ +static long get_proc_syscall(struct __test_metadata *_metadata, int pid) +{ + char proc_path[100] = {0}; + long ret = -1; + ssize_t nread; + char *line; + + snprintf(proc_path, sizeof(proc_path), "/proc/%d/syscall", pid); + nread = get_nth(_metadata, proc_path, 1, &line); + ASSERT_GT(nread, 0); + + if (!strncmp("running", line, MIN(7, nread))) + ret = strtol(line, NULL, 16); + + free(line); + return ret; +} + +/* Ensure non-fatal signals prior to receive are unmodified */ +TEST(user_notification_wait_killable_pre_notification) +{ + struct sigaction new_action = { + .sa_handler = signal_handler, + }; + int listener, status, sk_pair[2]; + pid_t pid; + long ret; + char c; + /* 100 ms */ + struct timespec delay = { .tv_nsec = 100000000 }; + + ASSERT_EQ(sigemptyset(&new_action.sa_mask), 0); + + ret = prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0); + ASSERT_EQ(0, ret) + { + TH_LOG("Kernel does not support PR_SET_NO_NEW_PRIVS!"); + } + + ASSERT_EQ(socketpair(PF_LOCAL, SOCK_SEQPACKET, 0, sk_pair), 0); + + listener = user_notif_syscall( + __NR_getppid, SECCOMP_FILTER_FLAG_NEW_LISTENER | + SECCOMP_FILTER_FLAG_WAIT_KILLABLE_RECV); + ASSERT_GE(listener, 0); + + /* + * Check that we can kill the process with SIGUSR1 prior to receiving + * the notification. SIGUSR1 is wired up to a custom signal handler, + * and make sure it gets called. + */ + pid = fork(); + ASSERT_GE(pid, 0); + + if (pid == 0) { + close(sk_pair[0]); + handled = sk_pair[1]; + + /* Setup the non-fatal sigaction without SA_RESTART */ + if (sigaction(SIGUSR1, &new_action, NULL)) { + perror("sigaction"); + exit(1); + } + + ret = syscall(__NR_getppid); + /* Make sure we got a return from a signal interruption */ + exit(ret != -1 || errno != EINTR); + } + + /* + * Make sure we've gotten to the seccomp user notification wait + * from getppid prior to sending any signals + */ + while (get_proc_syscall(_metadata, pid) != __NR_getppid && + get_proc_stat(_metadata, pid) != 'S') + nanosleep(&delay, NULL); + + /* Send non-fatal kill signal */ + EXPECT_EQ(kill(pid, SIGUSR1), 0); + + /* wait for process to exit (exit checks for EINTR) */ + EXPECT_EQ(waitpid(pid, &status, 0), pid); + EXPECT_EQ(true, WIFEXITED(status)); + EXPECT_EQ(0, WEXITSTATUS(status)); + + EXPECT_EQ(read(sk_pair[0], &c, 1), 1); +} + +/* Ensure non-fatal signals after receive are blocked */ +TEST(user_notification_wait_killable) +{ + struct sigaction new_action = { + .sa_handler = signal_handler, + }; + struct seccomp_notif_resp resp = {}; + struct seccomp_notif req = {}; + int listener, status, sk_pair[2]; + pid_t pid; + long ret; + char c; + /* 100 ms */ + struct timespec delay = { .tv_nsec = 100000000 }; + + ASSERT_EQ(sigemptyset(&new_action.sa_mask), 0); + + ret = prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0); + ASSERT_EQ(0, ret) + { + TH_LOG("Kernel does not support PR_SET_NO_NEW_PRIVS!"); + } + + ASSERT_EQ(socketpair(PF_LOCAL, SOCK_SEQPACKET, 0, sk_pair), 0); + + listener = user_notif_syscall( + __NR_getppid, SECCOMP_FILTER_FLAG_NEW_LISTENER | + SECCOMP_FILTER_FLAG_WAIT_KILLABLE_RECV); + ASSERT_GE(listener, 0); + + pid = fork(); + ASSERT_GE(pid, 0); + + if (pid == 0) { + close(sk_pair[0]); + handled = sk_pair[1]; + + /* Setup the sigaction without SA_RESTART */ + if (sigaction(SIGUSR1, &new_action, NULL)) { + perror("sigaction"); + exit(1); + } + + /* Make sure that the syscall is completed (no EINTR) */ + ret = syscall(__NR_getppid); + exit(ret != USER_NOTIF_MAGIC); + } + + /* + * Get the notification, to make move the notifying process into a + * non-preemptible (TASK_KILLABLE) state. + */ + EXPECT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_RECV, &req), 0); + /* Send non-fatal kill signal */ + EXPECT_EQ(kill(pid, SIGUSR1), 0); + + /* + * Make sure the task enters moves to TASK_KILLABLE by waiting for + * D (Disk Sleep) state after receiving non-fatal signal. + */ + while (get_proc_stat(_metadata, pid) != 'D') + nanosleep(&delay, NULL); + + resp.id = req.id; + resp.val = USER_NOTIF_MAGIC; + /* Make sure the notification is found and able to be replied to */ + EXPECT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_SEND, &resp), 0); + + /* + * Make sure that the signal handler does get called once we're back in + * userspace. + */ + EXPECT_EQ(read(sk_pair[0], &c, 1), 1); + /* wait for process to exit (exit checks for USER_NOTIF_MAGIC) */ + EXPECT_EQ(waitpid(pid, &status, 0), pid); + EXPECT_EQ(true, WIFEXITED(status)); + EXPECT_EQ(0, WEXITSTATUS(status)); +} + +/* Ensure fatal signals after receive are not blocked */ +TEST(user_notification_wait_killable_fatal) +{ + struct seccomp_notif req = {}; + int listener, status; + pid_t pid; + long ret; + /* 100 ms */ + struct timespec delay = { .tv_nsec = 100000000 }; + + ret = prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0); + ASSERT_EQ(0, ret) + { + TH_LOG("Kernel does not support PR_SET_NO_NEW_PRIVS!"); + } + + listener = user_notif_syscall( + __NR_getppid, SECCOMP_FILTER_FLAG_NEW_LISTENER | + SECCOMP_FILTER_FLAG_WAIT_KILLABLE_RECV); + ASSERT_GE(listener, 0); + + pid = fork(); + ASSERT_GE(pid, 0); + + if (pid == 0) { + /* This should never complete as it should get a SIGTERM */ + syscall(__NR_getppid); + exit(1); + } + + while (get_proc_stat(_metadata, pid) != 'S') + nanosleep(&delay, NULL); + + /* + * Get the notification, to make move the notifying process into a + * non-preemptible (TASK_KILLABLE) state. + */ + EXPECT_EQ(ioctl(listener, SECCOMP_IOCTL_NOTIF_RECV, &req), 0); + /* Kill the process with a fatal signal */ + EXPECT_EQ(kill(pid, SIGTERM), 0); + + /* + * Wait for the process to exit, and make sure the process terminated + * due to the SIGTERM signal. + */ + EXPECT_EQ(waitpid(pid, &status, 0), pid); + EXPECT_EQ(true, WIFSIGNALED(status)); + EXPECT_EQ(SIGTERM, WTERMSIG(status)); +} + +struct tsync_vs_thread_leader_args { + pthread_t leader; +}; + +static void *tsync_vs_dead_thread_leader_sibling(void *_args) +{ + struct sock_filter allow_filter[] = { + BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_ALLOW), + }; + struct sock_fprog allow_prog = { + .len = (unsigned short)ARRAY_SIZE(allow_filter), + .filter = allow_filter, + }; + struct tsync_vs_thread_leader_args *args = _args; + void *retval; + long ret; + + ret = pthread_join(args->leader, &retval); + if (ret) + exit(1); + if (retval != _args) + exit(2); + ret = seccomp(SECCOMP_SET_MODE_FILTER, SECCOMP_FILTER_FLAG_TSYNC, &allow_prog); + if (ret) + exit(3); + + exit(0); +} + +/* + * Ensure that a dead thread leader doesn't prevent installing new filters with + * SECCOMP_FILTER_FLAG_TSYNC from other threads. + */ +TEST(tsync_vs_dead_thread_leader) +{ + int status; + pid_t pid; + long ret; + + ret = prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0); + ASSERT_EQ(0, ret) { + TH_LOG("Kernel does not support PR_SET_NO_NEW_PRIVS!"); + } + + pid = fork(); + ASSERT_GE(pid, 0); + + if (pid == 0) { + struct sock_filter allow_filter[] = { + BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_ALLOW), + }; + struct sock_fprog allow_prog = { + .len = (unsigned short)ARRAY_SIZE(allow_filter), + .filter = allow_filter, + }; + struct tsync_vs_thread_leader_args *args; + pthread_t sibling; + + args = malloc(sizeof(*args)); + ASSERT_NE(NULL, args); + args->leader = pthread_self(); + + ret = pthread_create(&sibling, NULL, + tsync_vs_dead_thread_leader_sibling, args); + ASSERT_EQ(0, ret); + + /* Install a new filter just to the leader thread. */ + ret = seccomp(SECCOMP_SET_MODE_FILTER, 0, &allow_prog); + ASSERT_EQ(0, ret); + pthread_exit(args); + exit(1); + } + + EXPECT_EQ(pid, waitpid(pid, &status, 0)); + EXPECT_EQ(0, status); +} + +noinline int probed(void) +{ + return 1; +} + +static int parse_uint_from_file(const char *file, const char *fmt) +{ + int err = -1, ret; + FILE *f; + + f = fopen(file, "re"); + if (f) { + err = fscanf(f, fmt, &ret); + fclose(f); + } + return err == 1 ? ret : err; +} + +static int determine_uprobe_perf_type(void) +{ + const char *file = "/sys/bus/event_source/devices/uprobe/type"; + + return parse_uint_from_file(file, "%d\n"); +} + +static int determine_uprobe_retprobe_bit(void) +{ + const char *file = "/sys/bus/event_source/devices/uprobe/format/retprobe"; + + return parse_uint_from_file(file, "config:%d\n"); +} + +static ssize_t get_uprobe_offset(const void *addr) +{ + size_t start, base, end; + bool found = false; + char buf[256]; + FILE *f; + + f = fopen("/proc/self/maps", "r"); + if (!f) + return -1; + + while (fscanf(f, "%zx-%zx %s %zx %*[^\n]\n", &start, &end, buf, &base) == 4) { + if (buf[2] == 'x' && (uintptr_t)addr >= start && (uintptr_t)addr < end) { + found = true; + break; + } + } + fclose(f); + return found ? (uintptr_t)addr - start + base : -1; +} + +FIXTURE(URETPROBE) { + int fd; +}; + +FIXTURE_VARIANT(URETPROBE) { + /* + * All of the URETPROBE behaviors can be tested with either + * uretprobe attached or not + */ + bool attach; +}; + +FIXTURE_VARIANT_ADD(URETPROBE, attached) { + .attach = true, +}; + +FIXTURE_VARIANT_ADD(URETPROBE, not_attached) { + .attach = false, +}; + +FIXTURE_SETUP(URETPROBE) +{ + const size_t attr_sz = sizeof(struct perf_event_attr); + struct perf_event_attr attr; + ssize_t offset; + int type, bit; + +#ifndef __NR_uretprobe + SKIP(return, "__NR_uretprobe syscall not defined"); +#endif + + if (!variant->attach) + return; + + memset(&attr, 0, attr_sz); + + type = determine_uprobe_perf_type(); + ASSERT_GE(type, 0); + bit = determine_uprobe_retprobe_bit(); + ASSERT_GE(bit, 0); + offset = get_uprobe_offset(probed); + ASSERT_GE(offset, 0); + + attr.config |= 1 << bit; + attr.size = attr_sz; + attr.type = type; + attr.config1 = ptr_to_u64("/proc/self/exe"); + attr.config2 = offset; + + self->fd = syscall(__NR_perf_event_open, &attr, + getpid() /* pid */, -1 /* cpu */, -1 /* group_fd */, + PERF_FLAG_FD_CLOEXEC); +} + +FIXTURE_TEARDOWN(URETPROBE) +{ + /* we could call close(self->fd), but we'd need extra filter for + * that and since we are calling _exit right away.. + */ +} + +static int run_probed_with_filter(struct sock_fprog *prog) +{ + if (prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0) || + seccomp(SECCOMP_SET_MODE_FILTER, 0, prog)) { + return -1; + } + + probed(); + return 0; +} + +TEST_F(URETPROBE, uretprobe_default_allow) +{ + struct sock_filter filter[] = { + BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_ALLOW), + }; + struct sock_fprog prog = { + .len = (unsigned short)ARRAY_SIZE(filter), + .filter = filter, + }; + + ASSERT_EQ(0, run_probed_with_filter(&prog)); +} + +TEST_F(URETPROBE, uretprobe_default_block) +{ + struct sock_filter filter[] = { + BPF_STMT(BPF_LD|BPF_W|BPF_ABS, + offsetof(struct seccomp_data, nr)), + BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_exit_group, 1, 0), + BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_KILL), + BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_ALLOW), + }; + struct sock_fprog prog = { + .len = (unsigned short)ARRAY_SIZE(filter), + .filter = filter, + }; + + ASSERT_EQ(0, run_probed_with_filter(&prog)); +} + +TEST_F(URETPROBE, uretprobe_block_uretprobe_syscall) +{ + struct sock_filter filter[] = { + BPF_STMT(BPF_LD|BPF_W|BPF_ABS, + offsetof(struct seccomp_data, nr)), +#ifdef __NR_uretprobe + BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_uretprobe, 0, 1), +#endif + BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_KILL), + BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_ALLOW), + }; + struct sock_fprog prog = { + .len = (unsigned short)ARRAY_SIZE(filter), + .filter = filter, + }; + + ASSERT_EQ(0, run_probed_with_filter(&prog)); +} + +TEST_F(URETPROBE, uretprobe_default_block_with_uretprobe_syscall) +{ + struct sock_filter filter[] = { + BPF_STMT(BPF_LD|BPF_W|BPF_ABS, + offsetof(struct seccomp_data, nr)), +#ifdef __NR_uretprobe + BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_uretprobe, 2, 0), +#endif + BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, __NR_exit_group, 1, 0), + BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_KILL), + BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_ALLOW), + }; + struct sock_fprog prog = { + .len = (unsigned short)ARRAY_SIZE(filter), + .filter = filter, + }; + + ASSERT_EQ(0, run_probed_with_filter(&prog)); +} + /* * TODO: * - expand NNP testing diff --git a/tools/testing/selftests/seccomp/settings b/tools/testing/selftests/seccomp/settings index ba4d85f74cd6..a953c96aa16e 100644 --- a/tools/testing/selftests/seccomp/settings +++ b/tools/testing/selftests/seccomp/settings @@ -1 +1 @@ -timeout=90 +timeout=180 |