Test the new COREDUMP_RECORDS and COREDUMP_SPARSE flags. Signed-off-by: Christian Brauner (Amutable) --- .../coredump/coredump_socket_protocol_test.c | 407 +++++++++++++++++++++ .../selftests/coredump/coredump_test_helpers.c | 236 +++++++++++- .../selftests/coredump/coredump_test_helpers.h | 8 + 3 files changed, 650 insertions(+), 1 deletion(-) diff --git a/tools/testing/selftests/coredump/coredump_socket_protocol_test.c b/tools/testing/selftests/coredump/coredump_socket_protocol_test.c index 60a357e628eb..abf6e2c4c354 100644 --- a/tools/testing/selftests/coredump/coredump_socket_protocol_test.c +++ b/tools/testing/selftests/coredump/coredump_socket_protocol_test.c @@ -1573,4 +1573,411 @@ TEST_F_TIMEOUT(coredump, socket_multiple_crashing_coredumps_epoll_workers, 500) wait_and_check_coredump_server(pid_coredump_server, _metadata, self); } +/* + * Reassemble a record stream and check that what comes out is an ELF + * core file. The records themselves are validated by recv_coredump_records(). + */ +TEST_F(coredump, socket_request_sparse_reassemble) +{ + int fd_core_file, pidfd, status; + pid_t pid, pid_coredump_server; + struct pidfd_info info = {}; + int ipc_sockets[2]; + char c; + + ASSERT_EQ(socketpair(AF_UNIX, SOCK_STREAM | SOCK_CLOEXEC, 0, ipc_sockets), 0); + ASSERT_TRUE(set_core_pattern("@@/tmp/coredump.socket")); + + pid_coredump_server = fork(); + ASSERT_GE(pid_coredump_server, 0); + if (pid_coredump_server == 0) { + int fd_server = -1, fd_coredump = -1, fd_peer_pidfd = -1; + int fd_file = -1; + int exit_code = EXIT_FAILURE; + struct coredump_req req = {}; + + close(ipc_sockets[0]); + + fd_server = create_and_listen_unix_socket("/tmp/coredump.socket"); + if (fd_server < 0) + goto out; + + if (write_nointr(ipc_sockets[1], "1", 1) < 0) + goto out; + + close(ipc_sockets[1]); + + fd_coredump = accept4(fd_server, NULL, NULL, SOCK_CLOEXEC); + if (fd_coredump < 0) + goto out; + + fd_peer_pidfd = get_peer_pidfd(fd_coredump); + if (fd_peer_pidfd < 0) + goto out; + + fd_file = creat("/tmp/coredump.file", 0644); + if (fd_file < 0) + goto out; + + if (!read_coredump_req(fd_coredump, &req)) + goto out; + + if (!check_coredump_req(&req)) + goto out; + + if (!send_coredump_ack(fd_coredump, &req, + COREDUMP_KERNEL | COREDUMP_RECORDS | + COREDUMP_SPARSE | COREDUMP_WAIT, 0)) + goto out; + + if (!read_marker(fd_coredump, COREDUMP_MARK_REQACK)) + goto out; + + if (recv_coredump_records(fd_coredump, fd_file, NULL, NULL, -1) < 0) + goto out; + + exit_code = EXIT_SUCCESS; +out: + if (fd_file >= 0) + close(fd_file); + if (fd_peer_pidfd >= 0) + close(fd_peer_pidfd); + if (fd_coredump >= 0) + close(fd_coredump); + if (fd_server >= 0) + close(fd_server); + _exit(exit_code); + } + self->pid_coredump_server = pid_coredump_server; + + EXPECT_EQ(close(ipc_sockets[1]), 0); + ASSERT_EQ(read_nointr(ipc_sockets[0], &c, 1), 1); + EXPECT_EQ(close(ipc_sockets[0]), 0); + + pid = fork(); + ASSERT_GE(pid, 0); + if (pid == 0) + crashing_child(); + + pidfd = sys_pidfd_open(pid, 0); + ASSERT_GE(pidfd, 0); + + waitpid(pid, &status, 0); + ASSERT_TRUE(WIFSIGNALED(status)); + ASSERT_TRUE(WCOREDUMP(status)); + + ASSERT_TRUE(get_pidfd_info(pidfd, &info)); + ASSERT_GT((info.mask & PIDFD_INFO_COREDUMP), 0); + ASSERT_GT((info.coredump_mask & PIDFD_COREDUMPED), 0); + + wait_and_check_coredump_server(pid_coredump_server, _metadata, self); + + /* What the records reassemble into has to be an ELF core file. */ + fd_core_file = open("/tmp/coredump.file", O_RDONLY | O_CLOEXEC); + ASSERT_GE(fd_core_file, 0); + ASSERT_TRUE(is_elf_core(fd_core_file)); + EXPECT_EQ(close(fd_core_file), 0); +} + +/* + * Crash a child with a mostly-unpopulated mapping and reassemble its + * record stream, reporting what crossed the socket and the coredump + * size the records describe. With @kill_peer the server kills the task + * once the coredump is under way so the kernel has to cut it short. + */ +static void check_record_dump(struct __test_metadata *const _metadata, + FIXTURE_DATA(coredump) *self, __u64 ack_mask, + bool kill_peer, ssize_t *received, + off_t *coredump_size) +{ + bool truncated = false; + int pidfd, status; + pid_t pid, pid_coredump_server; + struct pidfd_info info = {}; + int ipc_sockets[2]; + int pipefds[2]; + char c; + + ASSERT_EQ(socketpair(AF_UNIX, SOCK_STREAM | SOCK_CLOEXEC, 0, ipc_sockets), 0); + ASSERT_EQ(pipe(pipefds), 0); + ASSERT_TRUE(set_core_pattern("@@/tmp/coredump.socket")); + + pid_coredump_server = fork(); + ASSERT_GE(pid_coredump_server, 0); + if (pid_coredump_server == 0) { + int fd_server = -1, fd_coredump = -1, fd_peer_pidfd = -1; + int fd_file = -1; + int exit_code = EXIT_FAILURE; + struct coredump_req req = {}; + bool is_truncated = false; + off_t size = 0; + ssize_t ret; + + close(ipc_sockets[0]); + close(pipefds[0]); + + fd_server = create_and_listen_unix_socket("/tmp/coredump.socket"); + if (fd_server < 0) + goto out; + + if (write_nointr(ipc_sockets[1], "1", 1) < 0) + goto out; + + close(ipc_sockets[1]); + + fd_coredump = accept4(fd_server, NULL, NULL, SOCK_CLOEXEC); + if (fd_coredump < 0) + goto out; + + fd_peer_pidfd = get_peer_pidfd(fd_coredump); + if (fd_peer_pidfd < 0) + goto out; + + /* + * The reassembled coredump is bigger than the mapping the + * child made, so keep it on the detached tmpfs and sparse. + */ + fd_file = open_coredump_tmpfile(self->fd_tmpfs_detached); + if (fd_file < 0) + goto out; + + if (!read_coredump_req(fd_coredump, &req)) + goto out; + + if (!check_coredump_req(&req)) + goto out; + + if (!send_coredump_ack(fd_coredump, &req, ack_mask, 0)) + goto out; + + if (!read_marker(fd_coredump, COREDUMP_MARK_REQACK)) + goto out; + + ret = recv_coredump_records(fd_coredump, fd_file, &size, &is_truncated, + kill_peer ? fd_peer_pidfd : -1); + if (ret < 0) + goto out; + + if (write_nointr(pipefds[1], &ret, sizeof(ret)) != sizeof(ret)) + goto out; + if (write_nointr(pipefds[1], &size, sizeof(size)) != sizeof(size)) + goto out; + if (write_nointr(pipefds[1], &is_truncated, + sizeof(is_truncated)) != sizeof(is_truncated)) + goto out; + + exit_code = EXIT_SUCCESS; +out: + close(pipefds[1]); + if (fd_file >= 0) + close(fd_file); + if (fd_peer_pidfd >= 0) + close(fd_peer_pidfd); + if (fd_coredump >= 0) + close(fd_coredump); + if (fd_server >= 0) + close(fd_server); + _exit(exit_code); + } + self->pid_coredump_server = pid_coredump_server; + + EXPECT_EQ(close(ipc_sockets[1]), 0); + EXPECT_EQ(close(pipefds[1]), 0); + ASSERT_EQ(read_nointr(ipc_sockets[0], &c, 1), 1); + EXPECT_EQ(close(ipc_sockets[0]), 0); + + pid = fork(); + ASSERT_GE(pid, 0); + if (pid == 0) + crashing_child_sparse(SPARSE_MAPPING_SIZE); + + pidfd = sys_pidfd_open(pid, 0); + ASSERT_GE(pidfd, 0); + + waitpid(pid, &status, 0); + ASSERT_TRUE(WIFSIGNALED(status)); + + ASSERT_EQ(read_nointr(pipefds[0], received, sizeof(*received)), + sizeof(*received)); + ASSERT_EQ(read_nointr(pipefds[0], coredump_size, sizeof(*coredump_size)), + sizeof(*coredump_size)); + ASSERT_EQ(read_nointr(pipefds[0], &truncated, sizeof(truncated)), + sizeof(truncated)); + EXPECT_EQ(close(pipefds[0]), 0); + + wait_and_check_coredump_server(pid_coredump_server, _metadata, self); + + if (kill_peer) { + /* The kernel gave up partway, so no end record closed the stream. */ + ASSERT_TRUE(truncated); + ASSERT_FALSE(WCOREDUMP(status)); + ASSERT_LT(*coredump_size, (off_t)SPARSE_MAPPING_SIZE); + return; + } + + ASSERT_FALSE(truncated); + ASSERT_TRUE(WCOREDUMP(status)); + + ASSERT_TRUE(get_pidfd_info(pidfd, &info)); + ASSERT_GT((info.mask & PIDFD_INFO_COREDUMP), 0); + ASSERT_GT((info.coredump_mask & PIDFD_COREDUMPED), 0); + + /* The mapping is in the coredump, holes included. */ + ASSERT_GT(*coredump_size, (off_t)SPARSE_MAPPING_SIZE); +} + +/* + * A mapping that has been written to is dumped whole, including the parts + * of it that were never faulted in. With COREDUMP_SPARSE the holes stay + * off the wire. + */ +TEST_F(coredump, socket_request_sparse_hole) +{ + off_t coredump_size = 0; + ssize_t received = 0; + + check_record_dump(_metadata, self, + COREDUMP_KERNEL | COREDUMP_RECORDS | + COREDUMP_SPARSE | COREDUMP_WAIT, + false, &received, &coredump_size); + + /* The holes didn't have to go over the socket. */ + ASSERT_LT(received, coredump_size / 8); +} + +/* + * COREDUMP_RECORDS alone splits the stream into records but elides + * nothing: the holes cross the socket as data records. + */ +TEST_F(coredump, socket_request_records_hole) +{ + off_t coredump_size = 0; + ssize_t received = 0; + + check_record_dump(_metadata, self, + COREDUMP_KERNEL | COREDUMP_RECORDS | COREDUMP_WAIT, + false, &received, &coredump_size); + + /* Records alone elide nothing, so everything crossed the socket. */ + ASSERT_GT(received, coredump_size); +} + +/* + * A coredump the kernel gives up on halfway still ends in an end record, + * and that record says the coredump is incomplete. COREDUMP_SPARSE is left + * out on purpose: the holes have to cross the socket so the coredump is + * far larger than the socket buffer and the kernel is still writing it + * when the kill lands. + */ +TEST_F(coredump, socket_request_records_truncated) +{ + off_t coredump_size = 0; + ssize_t received = 0; + + check_record_dump(_metadata, self, + COREDUMP_KERNEL | COREDUMP_RECORDS | COREDUMP_WAIT, + true, &received, &coredump_size); + + /* The end record crossed the socket even though the task was killed. */ + ASSERT_GT(received, 0); +} + +/* Ack @ack_mask, expect the kernel to refuse it as conflicting. */ +static void check_conflicting_ack(struct __test_metadata *const _metadata, + FIXTURE_DATA(coredump) *self, __u64 ack_mask) +{ + int pidfd, status; + pid_t pid, pid_coredump_server; + struct pidfd_info info = {}; + int ipc_sockets[2]; + char c; + + ASSERT_EQ(socketpair(AF_UNIX, SOCK_STREAM | SOCK_CLOEXEC, 0, ipc_sockets), 0); + ASSERT_TRUE(set_core_pattern("@@/tmp/coredump.socket")); + + pid_coredump_server = fork(); + ASSERT_GE(pid_coredump_server, 0); + if (pid_coredump_server == 0) { + int fd_server = -1, fd_coredump = -1, fd_peer_pidfd = -1; + int exit_code = EXIT_FAILURE; + struct coredump_req req = {}; + + close(ipc_sockets[0]); + + fd_server = create_and_listen_unix_socket("/tmp/coredump.socket"); + if (fd_server < 0) + goto out; + + if (write_nointr(ipc_sockets[1], "1", 1) < 0) + goto out; + + close(ipc_sockets[1]); + + fd_coredump = accept4(fd_server, NULL, NULL, SOCK_CLOEXEC); + if (fd_coredump < 0) + goto out; + + fd_peer_pidfd = get_peer_pidfd(fd_coredump); + if (fd_peer_pidfd < 0) + goto out; + + if (!read_coredump_req(fd_coredump, &req)) + goto out; + + if (!check_coredump_req(&req)) + goto out; + + if (!send_coredump_ack(fd_coredump, &req, ack_mask, 0)) + goto out; + + if (!read_marker(fd_coredump, COREDUMP_MARK_CONFLICTING)) + goto out; + + exit_code = EXIT_SUCCESS; +out: + if (fd_peer_pidfd >= 0) + close(fd_peer_pidfd); + if (fd_coredump >= 0) + close(fd_coredump); + if (fd_server >= 0) + close(fd_server); + _exit(exit_code); + } + self->pid_coredump_server = pid_coredump_server; + + EXPECT_EQ(close(ipc_sockets[1]), 0); + ASSERT_EQ(read_nointr(ipc_sockets[0], &c, 1), 1); + EXPECT_EQ(close(ipc_sockets[0]), 0); + + pid = fork(); + ASSERT_GE(pid, 0); + if (pid == 0) + crashing_child(); + + pidfd = sys_pidfd_open(pid, 0); + ASSERT_GE(pidfd, 0); + + waitpid(pid, &status, 0); + ASSERT_TRUE(WIFSIGNALED(status)); + ASSERT_FALSE(WCOREDUMP(status)); + + ASSERT_TRUE(get_pidfd_info(pidfd, &info)); + ASSERT_GT((info.mask & PIDFD_INFO_COREDUMP), 0); + ASSERT_GT((info.coredump_mask & PIDFD_COREDUMPED), 0); + + wait_and_check_coredump_server(pid_coredump_server, _metadata, self); +} + +/* COREDUMP_RECORDS applies to a coredump the kernel writes, nothing else. */ +TEST_F(coredump, socket_request_records_without_kernel) +{ + check_conflicting_ack(_metadata, self, COREDUMP_USERSPACE | COREDUMP_RECORDS); +} + +/* A zero record can't exist outside a record stream. */ +TEST_F(coredump, socket_request_sparse_without_records) +{ + check_conflicting_ack(_metadata, self, COREDUMP_KERNEL | COREDUMP_SPARSE); +} + TEST_HARNESS_MAIN diff --git a/tools/testing/selftests/coredump/coredump_test_helpers.c b/tools/testing/selftests/coredump/coredump_test_helpers.c index a5b9cde47239..5b2ffe17f7b7 100644 --- a/tools/testing/selftests/coredump/coredump_test_helpers.c +++ b/tools/testing/selftests/coredump/coredump_test_helpers.c @@ -1,9 +1,11 @@ // SPDX-License-Identifier: GPL-2.0 #include +#include #include #include #include +#include #include #include #include @@ -13,6 +15,7 @@ #include #include #include +#include #include #include #include @@ -23,6 +26,12 @@ #include "coredump_test_helpers.h" +#if __ELF_NATIVE_CLASS == 64 +#define COREDUMP_ELFCLASS ELFCLASS64 +#else +#define COREDUMP_ELFCLASS ELFCLASS32 +#endif + void *do_nothing(void *arg) { (void)arg; @@ -44,6 +53,228 @@ void crashing_child(void) i = *(volatile int *)NULL; } +void crashing_child_sparse(size_t size) +{ + char *p; + + /* + * Touch the first page only. The whole mapping is dumped because + * it has been written to, but all of it save that one page is a + * hole. + */ + p = mmap(NULL, size, PROT_READ | PROT_WRITE, + MAP_PRIVATE | MAP_ANONYMOUS | MAP_NORESERVE, -1, 0); + if (p != MAP_FAILED) + p[0] = 'x'; + + /* crash on purpose */ + *(volatile int *)NULL = 0; +} + +/* Read @len bytes off the socket, writing them at @offset if @fd_out >= 0. */ +static ssize_t recv_record_bytes(int fd_coredump, __u64 len, int fd_out, + off_t offset) +{ + ssize_t received = 0; + + while (len) { + char buffer[PAGE_SIZE]; + size_t chunk = len < sizeof(buffer) ? len : sizeof(buffer); + ssize_t ret; + + ret = recv(fd_coredump, buffer, chunk, MSG_WAITALL); + if (ret <= 0) { + fprintf(stderr, "%s: short read %zd: %m\n", + __func__, ret); + return -1; + } + + if (fd_out >= 0 && + pwrite(fd_out, buffer, ret, offset + received) != ret) { + fprintf(stderr, "%s: pwrite failed: %m\n", __func__); + return -1; + } + + received += ret; + len -= ret; + } + + return received; +} + +/* + * Reassemble a record stream. If @fd_peer_pidfd is valid the task behind + * it is killed once a data record has arrived, so the kernel has to cut + * the coredump short with the stream already under way. + */ +ssize_t recv_coredump_records(int fd_coredump, int fd_core_file, + off_t *coredump_size, bool *truncated, + int fd_peer_pidfd) +{ + ssize_t received = 0; + off_t size = 0; + bool is_truncated = false; + bool ended = false; + char trailing; + + while (!ended) { + struct coredump_record_header record = {}; + size_t known_size; + ssize_t ret; + + /* Peek the header size the way read_coredump_req() does. */ + ret = recv(fd_coredump, &record, sizeof(record.size), + MSG_PEEK | MSG_WAITALL); + if (ret == 0) { + /* Nothing closed the stream, so the coredump was cut short. */ + if (truncated) { + is_truncated = true; + break; + } + fprintf(stderr, "%s: stream ended without an end record\n", + __func__); + return -1; + } + if (ret != sizeof(record.size)) { + fprintf(stderr, "%s: short record peek %zd: %m\n", + __func__, ret); + return -1; + } + + if (record.size < COREDUMP_RECORD_HEADER_SIZE_VER0) { + fprintf(stderr, "%s: header size %u below minimum %u\n", + __func__, record.size, + COREDUMP_RECORD_HEADER_SIZE_VER0); + return -1; + } + + /* Consume as much of the header as we know about. */ + known_size = record.size < sizeof(record) ? record.size : sizeof(record); + ret = recv(fd_coredump, &record, known_size, MSG_WAITALL); + if (ret != (ssize_t)known_size) { + fprintf(stderr, "%s: short record read %zd: %m\n", + __func__, ret); + return -1; + } + received += ret; + + /* + * A flag changes what the record means, so refuse one we + * don't know rather than guess. + */ + if (record.flags) { + fprintf(stderr, "%s: unknown header flags 0x%llx\n", + __func__, (unsigned long long)record.flags); + return -1; + } + + /* Discard any part of the header we have no use for. */ + ret = recv_record_bytes(fd_coredump, record.size - known_size, -1, 0); + if (ret < 0) + return -1; + received += ret; + + /* Records are sent in order and they don't leave gaps. */ + if (record.offset != (__u64)size) { + fprintf(stderr, "%s: record at %llu, expected %llu\n", + __func__, (unsigned long long)record.offset, + (unsigned long long)size); + return -1; + } + + switch (record.type) { + case COREDUMP_RECORD_ZERO: + /* A hole. It comes with no data and needs none. */ + break; + case COREDUMP_RECORD_DATA: + ret = recv_record_bytes(fd_coredump, record.len, + fd_core_file, size); + if (ret < 0) + return -1; + received += ret; + if (fd_peer_pidfd >= 0) { + if (sys_pidfd_send_signal(fd_peer_pidfd, SIGKILL, + NULL, 0)) { + fprintf(stderr, "%s: kill failed: %m\n", + __func__); + return -1; + } + fd_peer_pidfd = -1; + } + break; + case COREDUMP_RECORD_END: + /* The coredump ends here and nothing follows it. */ + if (record.len) { + fprintf(stderr, "%s: end record covers %llu bytes\n", + __func__, + (unsigned long long)record.len); + return -1; + } + ended = true; + break; + default: + fprintf(stderr, "%s: unknown record type %u\n", + __func__, record.type); + return -1; + } + + size += record.len; + } + + /* The end record is the last thing on the wire. */ + if (recv(fd_coredump, &trailing, sizeof(trailing), MSG_DONTWAIT) > 0) { + fprintf(stderr, "%s: data after the end record\n", __func__); + return -1; + } + + if (truncated) + *truncated = is_truncated; + + /* + * Nothing is written for a hole, so grow the file to the size the + * records describe in case the coredump ended in one. + */ + if (ftruncate(fd_core_file, size) < 0) { + fprintf(stderr, "%s: ftruncate to %llu failed: %m\n", + __func__, (unsigned long long)size); + return -1; + } + + if (coredump_size) + *coredump_size = size; + + fprintf(stderr, "Received %zd bytes for a %s coredump of %llu bytes\n", + received, is_truncated ? "truncated" : "complete", + (unsigned long long)size); + return received; +} + +/* The ELF header of a native core file. */ +static bool is_core_ehdr(const ElfW(Ehdr) *ehdr) +{ + return !memcmp(ehdr->e_ident, ELFMAG, SELFMAG) && + ehdr->e_ident[EI_CLASS] == COREDUMP_ELFCLASS && + ehdr->e_type == ET_CORE; +} + +/* Whatever the server ends up with has to be an ELF core file. */ +bool is_elf_core(int fd) +{ + ElfW(Ehdr) ehdr; + + if (pread(fd, &ehdr, sizeof(ehdr), 0) != sizeof(ehdr)) { + fprintf(stderr, "%s: short read: %m\n", __func__); + return false; + } + + if (!is_core_ehdr(&ehdr)) { + fprintf(stderr, "%s: not an ELF core file\n", __func__); + return false; + } + + return true; +} + int create_detached_tmpfs(void) { int fd_context, fd_tmpfs; @@ -86,6 +317,7 @@ int create_and_listen_unix_socket(const char *path) return fd; out: + fprintf(stderr, "%s: %s: %m\n", __func__, path); if (fd >= 0) close(fd); return -1; @@ -264,8 +496,10 @@ bool send_coredump_ack(int fd, const struct coredump_req *req, large_ack.ack.mask = mask; large_ack.ack.size = size_ack; ret = send(fd, &large_ack, size_ack, MSG_NOSIGNAL); - if (ret != size_ack) + if (ret != size_ack) { + fprintf(stderr, "%s: short send %zd: %m\n", __func__, ret); return false; + } fprintf(stderr, "Sent coredump ack with size %zu and mask 0x%llx\n", size_ack, (unsigned long long)mask); diff --git a/tools/testing/selftests/coredump/coredump_test_helpers.h b/tools/testing/selftests/coredump/coredump_test_helpers.h index 45904bd177b8..fe0a88a71b05 100644 --- a/tools/testing/selftests/coredump/coredump_test_helpers.h +++ b/tools/testing/selftests/coredump/coredump_test_helpers.h @@ -15,9 +15,17 @@ #define NUM_THREAD_SPAWN 128 +/* Size of the mostly unpopulated mapping the sparse coredump test maps. */ +#define SPARSE_MAPPING_SIZE (256 * 1024 * 1024) + /* Shared helper function declarations */ void *do_nothing(void *arg); void crashing_child(void); +void crashing_child_sparse(size_t size); +ssize_t recv_coredump_records(int fd_coredump, int fd_core_file, + off_t *coredump_size, bool *truncated, + int fd_peer_pidfd); +bool is_elf_core(int fd); int create_detached_tmpfs(void); int create_and_listen_unix_socket(const char *path); bool set_core_pattern(const char *pattern); -- 2.53.0