wayland/tests/test-runner.c

393 lines
8.7 KiB
C
Raw Normal View History

/*
* Copyright © 2012 Intel Corporation
*
* Permission is hereby granted, free of charge, to any person obtaining
* a copy of this software and associated documentation files (the
* "Software"), to deal in the Software without restriction, including
* without limitation the rights to use, copy, modify, merge, publish,
* distribute, sublicense, and/or sell copies of the Software, and to
* permit persons to whom the Software is furnished to do so, subject to
* the following conditions:
*
* The above copyright notice and this permission notice (including the
* next paragraph) shall be included in all copies or substantial
* portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#define _GNU_SOURCE
#include <unistd.h>
#include <stdio.h>
#include <stdlib.h>
#include <sys/types.h>
#include <sys/wait.h>
#include <sys/stat.h>
2012-03-03 12:42:07 +08:00
#include <string.h>
#include <assert.h>
#include <dlfcn.h>
#include <errno.h>
#include <limits.h>
#include <sys/ptrace.h>
tests: Fix FAIL in sanity-test (*timeout*) when Yama LSM enabled This fixes a regression in the testsuite since c3653f7f, where four of the timeout tests fail with "Timeouts suppressed" messages. The timeouts are being suppressed because the testsuite is erroneously detecting that a debugger is attached. This detection mechanism (adopted from libinput) uses ptrace to test if there is a debugger parent process that can be attached. Unfortunately, this is an unreliable test: Kernel security policies exist to restrict the scope of ptrace to prevent processes from snooping on one another.[1] This security policy is set as the default on Ubuntu, and potentially other Linux distributions.[2] The Yama documentation suggests, "For software that has defined application-specific relationships between a debugging process and its inferior (crash handlers, etc), prctl(PR_SET_PTRACER, pid, ...) can be used. An inferior can declare which other process (and its descendents) are allowed to call PTRACE_ATTACH against it." This prctl call has no effect if Yama LSM is not loaded. The child needs to be synchronized to the client to prevent a race condition where the child might try to operate before the parent has finished its prctl call. This synchronization is done via pipes. This patch can be tested by running sanity-test with /proc/sys/kernel/yama/ptrace_scope set to 0 or 1; the test must pass for either value. 1: http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=2d514487faf188938a4ee4fb3464eeecfbdcf8eb 2: https://wiki.ubuntu.com/SecurityTeam/Roadmap/KernelHardening#ptrace_Protection Signed-off-by: Bryce Harrington <bryce@osg.samsung.com> Reviewed-by: Marek Chalupa <mchqwerty@gmail.com> Reviewed-by: Daniel Stone <daniels@collabora.com> Reviewed-by: Derek Foreman <derekf@osg.samsung.com> v4: Allow parent to communicate error state to child to prevent leaving child in zombie state if parent hits an error. v5: Check errno instead of rc for error. Don't waitpid on ppid. Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
2015-01-08 03:56:54 +08:00
#include <sys/prctl.h>
#ifndef PR_SET_PTRACER
# define PR_SET_PTRACER 0x59616d61
#endif
#include "test-runner.h"
/* when set to 1, check if tests are not leaking opened files.
* It is turned on by default. It can be turned off by
* WAYLAND_TEST_NO_LEAK_CHECK environment variable. */
int fd_leak_check_enabled;
/* when this var is set to 0, every call to test_set_timeout() is
* suppressed - handy when debugging the test. Can be set by
* WAYLAND_TEST_NO_TIMEOUTS environment variable. */
static int timeouts_enabled = 1;
/* set to one if the output goes to the terminal */
static int is_atty = 0;
extern const struct test __start_test_section, __stop_test_section;
static const struct test *
find_test(const char *name)
{
const struct test *t;
for (t = &__start_test_section; t < &__stop_test_section; t++)
if (strcmp(t->name, name) == 0)
return t;
return NULL;
}
static void
usage(const char *name, int status)
{
const struct test *t;
fprintf(stderr, "Usage: %s [TEST]\n\n"
"With no arguments, run all test. Specify test case to run\n"
"only that test without forking. Available tests:\n\n",
name);
for (t = &__start_test_section; t < &__stop_test_section; t++)
fprintf(stderr, " %s\n", t->name);
fprintf(stderr, "\n");
exit(status);
}
void
test_set_timeout(unsigned int to)
{
int re;
if (!timeouts_enabled) {
fprintf(stderr, "Timeouts suppressed.\n");
return;
}
re = alarm(to);
fprintf(stderr, "Timeout was %sset", re ? "re-" : "");
if (to != 0)
fprintf(stderr, " to %d second%s from now.\n",
to, to > 1 ? "s" : "");
else
fprintf(stderr, " off.\n");
}
static void
sigalrm_handler(int signum)
{
fprintf(stderr, "Test timed out.\n");
abort();
}
void
check_fd_leaks(int supposed_fds)
{
int num_fds;
if (fd_leak_check_enabled) {
num_fds = count_open_fds();
if (supposed_fds != num_fds) {
fprintf(stderr, "fd leak detected in test. "
"Opened %d files, unclosed %d\n", num_fds,
num_fds - supposed_fds);
abort();
}
} else {
fprintf(stderr, "FD leak checks disabled\n");
}
}
static void
run_test(const struct test *t)
{
int cur_fds;
struct sigaction sa;
if (timeouts_enabled) {
sa.sa_handler = sigalrm_handler;
sa.sa_flags = 0;
sigemptyset(&sa.sa_mask);
assert(sigaction(SIGALRM, &sa, NULL) == 0);
}
//cur_alloc = get_current_alloc_num();
cur_fds = count_open_fds();
t->run();
/* turn off timeout (if any) after test completion */
if (timeouts_enabled)
alarm(0);
check_fd_leaks(cur_fds);
exit(EXIT_SUCCESS);
}
#ifndef PATH_MAX
#define PATH_MAX 256
#endif
static void
set_xdg_runtime_dir(void)
{
char xdg_runtime_dir[PATH_MAX];
const char *xrd_env;
xrd_env = getenv("XDG_RUNTIME_DIR");
/* if XDG_RUNTIME_DIR is not set in environ, fallback to /tmp */
assert((snprintf(xdg_runtime_dir, PATH_MAX, "%s/wayland-tests-XXXXXX",
xrd_env ? xrd_env : "/tmp") < PATH_MAX)
&& "test error: XDG_RUNTIME_DIR too long");
assert(mkdtemp(xdg_runtime_dir) && "test error: mkdtemp failed");
if (mkdir(xdg_runtime_dir, 0700) == -1)
if (errno != EEXIST) {
perror("Creating XDG_RUNTIME_DIR");
abort();
}
if (setenv("XDG_RUNTIME_DIR", xdg_runtime_dir, 1) == -1) {
perror("Setting XDG_RUNTIME_DIR");
abort();
}
}
static void
rmdir_xdg_runtime_dir(void)
{
const char *xrd_env = getenv("XDG_RUNTIME_DIR");
assert(xrd_env && "No XDG_RUNTIME_DIR set");
/* rmdir may fail if some test didn't do clean up */
if (rmdir(xrd_env) == -1)
perror("Cleaning XDG_RUNTIME_DIR");
}
#define RED "\033[31m"
#define GREEN "\033[32m"
static void
stderr_set_color(const char *color)
{
/* use colors only when the output is connected to
* the terminal */
if (is_atty)
fprintf(stderr, "%s", color);
}
static void
stderr_reset_color(void)
{
if (is_atty)
fprintf(stderr, "\033[0m");
}
/* this function is taken from libinput/test/litest.c
* (rev 028513a0a723e97941c39)
tests: Fix FAIL in sanity-test (*timeout*) when Yama LSM enabled This fixes a regression in the testsuite since c3653f7f, where four of the timeout tests fail with "Timeouts suppressed" messages. The timeouts are being suppressed because the testsuite is erroneously detecting that a debugger is attached. This detection mechanism (adopted from libinput) uses ptrace to test if there is a debugger parent process that can be attached. Unfortunately, this is an unreliable test: Kernel security policies exist to restrict the scope of ptrace to prevent processes from snooping on one another.[1] This security policy is set as the default on Ubuntu, and potentially other Linux distributions.[2] The Yama documentation suggests, "For software that has defined application-specific relationships between a debugging process and its inferior (crash handlers, etc), prctl(PR_SET_PTRACER, pid, ...) can be used. An inferior can declare which other process (and its descendents) are allowed to call PTRACE_ATTACH against it." This prctl call has no effect if Yama LSM is not loaded. The child needs to be synchronized to the client to prevent a race condition where the child might try to operate before the parent has finished its prctl call. This synchronization is done via pipes. This patch can be tested by running sanity-test with /proc/sys/kernel/yama/ptrace_scope set to 0 or 1; the test must pass for either value. 1: http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=2d514487faf188938a4ee4fb3464eeecfbdcf8eb 2: https://wiki.ubuntu.com/SecurityTeam/Roadmap/KernelHardening#ptrace_Protection Signed-off-by: Bryce Harrington <bryce@osg.samsung.com> Reviewed-by: Marek Chalupa <mchqwerty@gmail.com> Reviewed-by: Daniel Stone <daniels@collabora.com> Reviewed-by: Derek Foreman <derekf@osg.samsung.com> v4: Allow parent to communicate error state to child to prevent leaving child in zombie state if parent hits an error. v5: Check errno instead of rc for error. Don't waitpid on ppid. Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
2015-01-08 03:56:54 +08:00
*
* Returns: 1 if a debugger is confirmed present; 0 if no debugger is
* present or if it can't be determined.
*/
static int
is_debugger_attached(void)
{
int status;
int rc;
tests: Fix FAIL in sanity-test (*timeout*) when Yama LSM enabled This fixes a regression in the testsuite since c3653f7f, where four of the timeout tests fail with "Timeouts suppressed" messages. The timeouts are being suppressed because the testsuite is erroneously detecting that a debugger is attached. This detection mechanism (adopted from libinput) uses ptrace to test if there is a debugger parent process that can be attached. Unfortunately, this is an unreliable test: Kernel security policies exist to restrict the scope of ptrace to prevent processes from snooping on one another.[1] This security policy is set as the default on Ubuntu, and potentially other Linux distributions.[2] The Yama documentation suggests, "For software that has defined application-specific relationships between a debugging process and its inferior (crash handlers, etc), prctl(PR_SET_PTRACER, pid, ...) can be used. An inferior can declare which other process (and its descendents) are allowed to call PTRACE_ATTACH against it." This prctl call has no effect if Yama LSM is not loaded. The child needs to be synchronized to the client to prevent a race condition where the child might try to operate before the parent has finished its prctl call. This synchronization is done via pipes. This patch can be tested by running sanity-test with /proc/sys/kernel/yama/ptrace_scope set to 0 or 1; the test must pass for either value. 1: http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=2d514487faf188938a4ee4fb3464eeecfbdcf8eb 2: https://wiki.ubuntu.com/SecurityTeam/Roadmap/KernelHardening#ptrace_Protection Signed-off-by: Bryce Harrington <bryce@osg.samsung.com> Reviewed-by: Marek Chalupa <mchqwerty@gmail.com> Reviewed-by: Daniel Stone <daniels@collabora.com> Reviewed-by: Derek Foreman <derekf@osg.samsung.com> v4: Allow parent to communicate error state to child to prevent leaving child in zombie state if parent hits an error. v5: Check errno instead of rc for error. Don't waitpid on ppid. Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
2015-01-08 03:56:54 +08:00
pid_t pid;
int pipefd[2];
tests: Fix FAIL in sanity-test (*timeout*) when Yama LSM enabled This fixes a regression in the testsuite since c3653f7f, where four of the timeout tests fail with "Timeouts suppressed" messages. The timeouts are being suppressed because the testsuite is erroneously detecting that a debugger is attached. This detection mechanism (adopted from libinput) uses ptrace to test if there is a debugger parent process that can be attached. Unfortunately, this is an unreliable test: Kernel security policies exist to restrict the scope of ptrace to prevent processes from snooping on one another.[1] This security policy is set as the default on Ubuntu, and potentially other Linux distributions.[2] The Yama documentation suggests, "For software that has defined application-specific relationships between a debugging process and its inferior (crash handlers, etc), prctl(PR_SET_PTRACER, pid, ...) can be used. An inferior can declare which other process (and its descendents) are allowed to call PTRACE_ATTACH against it." This prctl call has no effect if Yama LSM is not loaded. The child needs to be synchronized to the client to prevent a race condition where the child might try to operate before the parent has finished its prctl call. This synchronization is done via pipes. This patch can be tested by running sanity-test with /proc/sys/kernel/yama/ptrace_scope set to 0 or 1; the test must pass for either value. 1: http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=2d514487faf188938a4ee4fb3464eeecfbdcf8eb 2: https://wiki.ubuntu.com/SecurityTeam/Roadmap/KernelHardening#ptrace_Protection Signed-off-by: Bryce Harrington <bryce@osg.samsung.com> Reviewed-by: Marek Chalupa <mchqwerty@gmail.com> Reviewed-by: Daniel Stone <daniels@collabora.com> Reviewed-by: Derek Foreman <derekf@osg.samsung.com> v4: Allow parent to communicate error state to child to prevent leaving child in zombie state if parent hits an error. v5: Check errno instead of rc for error. Don't waitpid on ppid. Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
2015-01-08 03:56:54 +08:00
if (pipe(pipefd) == -1) {
perror("pipe");
return 0;
tests: Fix FAIL in sanity-test (*timeout*) when Yama LSM enabled This fixes a regression in the testsuite since c3653f7f, where four of the timeout tests fail with "Timeouts suppressed" messages. The timeouts are being suppressed because the testsuite is erroneously detecting that a debugger is attached. This detection mechanism (adopted from libinput) uses ptrace to test if there is a debugger parent process that can be attached. Unfortunately, this is an unreliable test: Kernel security policies exist to restrict the scope of ptrace to prevent processes from snooping on one another.[1] This security policy is set as the default on Ubuntu, and potentially other Linux distributions.[2] The Yama documentation suggests, "For software that has defined application-specific relationships between a debugging process and its inferior (crash handlers, etc), prctl(PR_SET_PTRACER, pid, ...) can be used. An inferior can declare which other process (and its descendents) are allowed to call PTRACE_ATTACH against it." This prctl call has no effect if Yama LSM is not loaded. The child needs to be synchronized to the client to prevent a race condition where the child might try to operate before the parent has finished its prctl call. This synchronization is done via pipes. This patch can be tested by running sanity-test with /proc/sys/kernel/yama/ptrace_scope set to 0 or 1; the test must pass for either value. 1: http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=2d514487faf188938a4ee4fb3464eeecfbdcf8eb 2: https://wiki.ubuntu.com/SecurityTeam/Roadmap/KernelHardening#ptrace_Protection Signed-off-by: Bryce Harrington <bryce@osg.samsung.com> Reviewed-by: Marek Chalupa <mchqwerty@gmail.com> Reviewed-by: Daniel Stone <daniels@collabora.com> Reviewed-by: Derek Foreman <derekf@osg.samsung.com> v4: Allow parent to communicate error state to child to prevent leaving child in zombie state if parent hits an error. v5: Check errno instead of rc for error. Don't waitpid on ppid. Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
2015-01-08 03:56:54 +08:00
}
tests: Fix FAIL in sanity-test (*timeout*) when Yama LSM enabled This fixes a regression in the testsuite since c3653f7f, where four of the timeout tests fail with "Timeouts suppressed" messages. The timeouts are being suppressed because the testsuite is erroneously detecting that a debugger is attached. This detection mechanism (adopted from libinput) uses ptrace to test if there is a debugger parent process that can be attached. Unfortunately, this is an unreliable test: Kernel security policies exist to restrict the scope of ptrace to prevent processes from snooping on one another.[1] This security policy is set as the default on Ubuntu, and potentially other Linux distributions.[2] The Yama documentation suggests, "For software that has defined application-specific relationships between a debugging process and its inferior (crash handlers, etc), prctl(PR_SET_PTRACER, pid, ...) can be used. An inferior can declare which other process (and its descendents) are allowed to call PTRACE_ATTACH against it." This prctl call has no effect if Yama LSM is not loaded. The child needs to be synchronized to the client to prevent a race condition where the child might try to operate before the parent has finished its prctl call. This synchronization is done via pipes. This patch can be tested by running sanity-test with /proc/sys/kernel/yama/ptrace_scope set to 0 or 1; the test must pass for either value. 1: http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=2d514487faf188938a4ee4fb3464eeecfbdcf8eb 2: https://wiki.ubuntu.com/SecurityTeam/Roadmap/KernelHardening#ptrace_Protection Signed-off-by: Bryce Harrington <bryce@osg.samsung.com> Reviewed-by: Marek Chalupa <mchqwerty@gmail.com> Reviewed-by: Daniel Stone <daniels@collabora.com> Reviewed-by: Derek Foreman <derekf@osg.samsung.com> v4: Allow parent to communicate error state to child to prevent leaving child in zombie state if parent hits an error. v5: Check errno instead of rc for error. Don't waitpid on ppid. Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
2015-01-08 03:56:54 +08:00
pid = fork();
if (pid == -1) {
perror("fork");
close(pipefd[0]);
close(pipefd[1]);
return 0;
} else if (pid == 0) {
char buf;
pid_t ppid = getppid();
tests: Fix FAIL in sanity-test (*timeout*) when Yama LSM enabled This fixes a regression in the testsuite since c3653f7f, where four of the timeout tests fail with "Timeouts suppressed" messages. The timeouts are being suppressed because the testsuite is erroneously detecting that a debugger is attached. This detection mechanism (adopted from libinput) uses ptrace to test if there is a debugger parent process that can be attached. Unfortunately, this is an unreliable test: Kernel security policies exist to restrict the scope of ptrace to prevent processes from snooping on one another.[1] This security policy is set as the default on Ubuntu, and potentially other Linux distributions.[2] The Yama documentation suggests, "For software that has defined application-specific relationships between a debugging process and its inferior (crash handlers, etc), prctl(PR_SET_PTRACER, pid, ...) can be used. An inferior can declare which other process (and its descendents) are allowed to call PTRACE_ATTACH against it." This prctl call has no effect if Yama LSM is not loaded. The child needs to be synchronized to the client to prevent a race condition where the child might try to operate before the parent has finished its prctl call. This synchronization is done via pipes. This patch can be tested by running sanity-test with /proc/sys/kernel/yama/ptrace_scope set to 0 or 1; the test must pass for either value. 1: http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=2d514487faf188938a4ee4fb3464eeecfbdcf8eb 2: https://wiki.ubuntu.com/SecurityTeam/Roadmap/KernelHardening#ptrace_Protection Signed-off-by: Bryce Harrington <bryce@osg.samsung.com> Reviewed-by: Marek Chalupa <mchqwerty@gmail.com> Reviewed-by: Daniel Stone <daniels@collabora.com> Reviewed-by: Derek Foreman <derekf@osg.samsung.com> v4: Allow parent to communicate error state to child to prevent leaving child in zombie state if parent hits an error. v5: Check errno instead of rc for error. Don't waitpid on ppid. Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
2015-01-08 03:56:54 +08:00
/* Wait until parent is ready */
close(pipefd[1]); /* Close unused write end */
read(pipefd[0], &buf, 1);
close(pipefd[0]);
if (buf == '-')
_exit(1);
if (ptrace(PTRACE_ATTACH, ppid, NULL, NULL) != 0)
_exit(1);
if (!waitpid(-1, NULL, 0))
_exit(1);
ptrace(PTRACE_CONT, NULL, NULL);
ptrace(PTRACE_DETACH, ppid, NULL, NULL);
_exit(0);
} else {
close(pipefd[0]);
/* Enable child to ptrace the parent process */
rc = prctl(PR_SET_PTRACER, pid);
if (rc != 0 && errno != EINVAL) {
/* An error prevents us from telling if a debugger is attached.
* Instead of propagating the error, assume no debugger present.
* But note the error to the log as a clue for troubleshooting.
* Then flag the error state to the client by sending '-'.
*/
perror("prctl");
write(pipefd[1], "-", 1);
} else {
tests: Fix FAIL in sanity-test (*timeout*) when Yama LSM enabled This fixes a regression in the testsuite since c3653f7f, where four of the timeout tests fail with "Timeouts suppressed" messages. The timeouts are being suppressed because the testsuite is erroneously detecting that a debugger is attached. This detection mechanism (adopted from libinput) uses ptrace to test if there is a debugger parent process that can be attached. Unfortunately, this is an unreliable test: Kernel security policies exist to restrict the scope of ptrace to prevent processes from snooping on one another.[1] This security policy is set as the default on Ubuntu, and potentially other Linux distributions.[2] The Yama documentation suggests, "For software that has defined application-specific relationships between a debugging process and its inferior (crash handlers, etc), prctl(PR_SET_PTRACER, pid, ...) can be used. An inferior can declare which other process (and its descendents) are allowed to call PTRACE_ATTACH against it." This prctl call has no effect if Yama LSM is not loaded. The child needs to be synchronized to the client to prevent a race condition where the child might try to operate before the parent has finished its prctl call. This synchronization is done via pipes. This patch can be tested by running sanity-test with /proc/sys/kernel/yama/ptrace_scope set to 0 or 1; the test must pass for either value. 1: http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=2d514487faf188938a4ee4fb3464eeecfbdcf8eb 2: https://wiki.ubuntu.com/SecurityTeam/Roadmap/KernelHardening#ptrace_Protection Signed-off-by: Bryce Harrington <bryce@osg.samsung.com> Reviewed-by: Marek Chalupa <mchqwerty@gmail.com> Reviewed-by: Daniel Stone <daniels@collabora.com> Reviewed-by: Derek Foreman <derekf@osg.samsung.com> v4: Allow parent to communicate error state to child to prevent leaving child in zombie state if parent hits an error. v5: Check errno instead of rc for error. Don't waitpid on ppid. Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
2015-01-08 03:56:54 +08:00
/* Signal to client that parent is ready by passing '+' */
write(pipefd[1], "+", 1);
}
tests: Fix FAIL in sanity-test (*timeout*) when Yama LSM enabled This fixes a regression in the testsuite since c3653f7f, where four of the timeout tests fail with "Timeouts suppressed" messages. The timeouts are being suppressed because the testsuite is erroneously detecting that a debugger is attached. This detection mechanism (adopted from libinput) uses ptrace to test if there is a debugger parent process that can be attached. Unfortunately, this is an unreliable test: Kernel security policies exist to restrict the scope of ptrace to prevent processes from snooping on one another.[1] This security policy is set as the default on Ubuntu, and potentially other Linux distributions.[2] The Yama documentation suggests, "For software that has defined application-specific relationships between a debugging process and its inferior (crash handlers, etc), prctl(PR_SET_PTRACER, pid, ...) can be used. An inferior can declare which other process (and its descendents) are allowed to call PTRACE_ATTACH against it." This prctl call has no effect if Yama LSM is not loaded. The child needs to be synchronized to the client to prevent a race condition where the child might try to operate before the parent has finished its prctl call. This synchronization is done via pipes. This patch can be tested by running sanity-test with /proc/sys/kernel/yama/ptrace_scope set to 0 or 1; the test must pass for either value. 1: http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=2d514487faf188938a4ee4fb3464eeecfbdcf8eb 2: https://wiki.ubuntu.com/SecurityTeam/Roadmap/KernelHardening#ptrace_Protection Signed-off-by: Bryce Harrington <bryce@osg.samsung.com> Reviewed-by: Marek Chalupa <mchqwerty@gmail.com> Reviewed-by: Daniel Stone <daniels@collabora.com> Reviewed-by: Derek Foreman <derekf@osg.samsung.com> v4: Allow parent to communicate error state to child to prevent leaving child in zombie state if parent hits an error. v5: Check errno instead of rc for error. Don't waitpid on ppid. Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
2015-01-08 03:56:54 +08:00
close(pipefd[1]);
waitpid(pid, &status, 0);
rc = WEXITSTATUS(status);
}
return rc;
}
int main(int argc, char *argv[])
{
const struct test *t;
pid_t pid;
2012-03-03 12:42:07 +08:00
int total, pass;
siginfo_t info;
if (isatty(fileno(stderr)))
is_atty = 1;
if (is_debugger_attached()) {
fd_leak_check_enabled = 0;
timeouts_enabled = 0;
} else {
fd_leak_check_enabled = !getenv("WAYLAND_TEST_NO_LEAK_CHECK");
timeouts_enabled = !getenv("WAYLAND_TEST_NO_TIMEOUTS");
}
if (argc == 2 && strcmp(argv[1], "--help") == 0)
usage(argv[0], EXIT_SUCCESS);
if (argc == 2) {
t = find_test(argv[1]);
if (t == NULL) {
fprintf(stderr, "unknown test: \"%s\"\n", argv[1]);
usage(argv[0], EXIT_FAILURE);
}
set_xdg_runtime_dir();
/* run_test calls exit() */
assert(atexit(rmdir_xdg_runtime_dir) == 0);
run_test(t);
}
/* set our own XDG_RUNTIME_DIR */
set_xdg_runtime_dir();
pass = 0;
for (t = &__start_test_section; t < &__stop_test_section; t++) {
int success = 0;
pid = fork();
assert(pid >= 0);
if (pid == 0)
run_test(t); /* never returns */
if (waitid(P_PID, pid, &info, WEXITED)) {
stderr_set_color(RED);
fprintf(stderr, "waitid failed: %s\n",
strerror(errno));
stderr_reset_color();
abort();
}
switch (info.si_code) {
case CLD_EXITED:
if (info.si_status == EXIT_SUCCESS)
success = !t->must_fail;
else
success = t->must_fail;
stderr_set_color(success ? GREEN : RED);
fprintf(stderr, "test \"%s\":\texit status %d",
t->name, info.si_status);
break;
case CLD_KILLED:
case CLD_DUMPED:
if (t->must_fail)
success = 1;
stderr_set_color(success ? GREEN : RED);
fprintf(stderr, "test \"%s\":\tsignal %d",
t->name, info.si_status);
break;
}
if (success) {
pass++;
fprintf(stderr, ", pass.\n");
} else
fprintf(stderr, ", fail.\n");
stderr_reset_color();
/* print separator line */
fprintf(stderr, "----------------------------------------\n");
}
total = &__stop_test_section - &__start_test_section;
fprintf(stderr, "%d tests, %d pass, %d fail\n",
total, pass, total - pass);
/* cleaning */
rmdir_xdg_runtime_dir();
return pass == total ? EXIT_SUCCESS : EXIT_FAILURE;
}