diff mbox series

[v2,2/2] syscalls/fanotify21: add new test checking the returned pidfd from fanotify in FAN_REPORT_PIDFD mode

Message ID 229fb61317343cf30d481d05da272657c7572f59.1635849607.git.repnop@google.com
State Changes Requested
Headers show
Series Test support for new fanotify FAN_REPORT_PIDFD feature | expand

Commit Message

Matthew Bobrowski Nov. 2, 2021, 10:57 a.m. UTC
A new test that performs verification on the values returned within
the struct fanotify_event_info_pidfd record when the notification
group has been intialized in FAN_REPORT_PIDFD mode.

Signed-off-by: Matthew Bobrowski <repnop@google.com>
Reviewed-by: Amir Goldstein <amir73il@gmail.com>
---
Changes since v1:
 - Adapted do_setup() to make use of the
   REQUIRE_FANOTIFY_INIT_FLAGS_SUPPORTED_BY_KERNEL() rather than open
   coding it.

 - For parsing procfs information, rather than reimplementing parsing
   utility functions, make use of the existing SAFE_FILE_LINES_SCANF()
   library function.

 testcases/kernel/syscalls/fanotify/.gitignore |   1 +
 .../kernel/syscalls/fanotify/fanotify21.c     | 306 ++++++++++++++++++
 2 files changed, 307 insertions(+)
 create mode 100644 testcases/kernel/syscalls/fanotify/fanotify21.c

Comments

Amir Goldstein Nov. 5, 2021, 4:21 a.m. UTC | #1
> +next_event:
> +               i += event->event_len;
> +               if (event->fd >= 0)
> +                       SAFE_CLOSE(event->fd);
> +
> +               if (info && info->pidfd >= 0)
> +                       SAFE_CLOSE(info->pidfd);
> +
> +               if (event_pidfd_fdinfo)
> +                       free(event_pidfd_fdinfo);

Need to reset event_pidfd_fdinfo = NULL
to avoid double free in do_cleanup

> +       }
> +}
> +
> +static void do_cleanup(void)
> +{
> +       if (fanotify_fd >= 0)
> +               SAFE_CLOSE(fanotify_fd);
> +
> +       if (self_pidfd_fdinfo)
> +               free(self_pidfd_fdinfo);
> +}
> +

Thanks,
Amir.
Matthew Bobrowski Nov. 9, 2021, 9:36 a.m. UTC | #2
On Fri, Nov 05, 2021 at 06:21:56AM +0200, Amir Goldstein wrote:
> > +next_event:
> > +               i += event->event_len;
> > +               if (event->fd >= 0)
> > +                       SAFE_CLOSE(event->fd);
> > +
> > +               if (info && info->pidfd >= 0)
> > +                       SAFE_CLOSE(info->pidfd);
> > +
> > +               if (event_pidfd_fdinfo)
> > +                       free(event_pidfd_fdinfo);
> 
> Need to reset event_pidfd_fdinfo = NULL
> to avoid double free in do_cleanup

But, the variable event_pidfd_fdinfo is not referenced nor is visible
outside the scope of the do_test() function and therefore no double
free would occur, right?

> > +       }
> > +}
> > +
> > +static void do_cleanup(void)
> > +{
> > +       if (fanotify_fd >= 0)
> > +               SAFE_CLOSE(fanotify_fd);
> > +
> > +       if (self_pidfd_fdinfo)
> > +               free(self_pidfd_fdinfo);

Variables self_pidfd_fdinfo do_cleanup() and event_pidfd_fdinfo
do_test() are different.

/M
Amir Goldstein Nov. 9, 2021, 10:34 a.m. UTC | #3
On Tue, Nov 9, 2021 at 11:36 AM Matthew Bobrowski <repnop@google.com> wrote:
>
> On Fri, Nov 05, 2021 at 06:21:56AM +0200, Amir Goldstein wrote:
> > > +next_event:
> > > +               i += event->event_len;
> > > +               if (event->fd >= 0)
> > > +                       SAFE_CLOSE(event->fd);
> > > +
> > > +               if (info && info->pidfd >= 0)
> > > +                       SAFE_CLOSE(info->pidfd);
> > > +
> > > +               if (event_pidfd_fdinfo)
> > > +                       free(event_pidfd_fdinfo);
> >
> > Need to reset event_pidfd_fdinfo = NULL
> > to avoid double free in do_cleanup
>
> But, the variable event_pidfd_fdinfo is not referenced nor is visible
> outside the scope of the do_test() function and therefore no double
> free would occur, right?
>
> > > +       }
> > > +}
> > > +
> > > +static void do_cleanup(void)
> > > +{
> > > +       if (fanotify_fd >= 0)
> > > +               SAFE_CLOSE(fanotify_fd);
> > > +
> > > +       if (self_pidfd_fdinfo)
> > > +               free(self_pidfd_fdinfo);
>
> Variables self_pidfd_fdinfo do_cleanup() and event_pidfd_fdinfo
> do_test() are different.
>

Right. my bad :)

Thanks,
Amir.
diff mbox series

Patch

diff --git a/testcases/kernel/syscalls/fanotify/.gitignore b/testcases/kernel/syscalls/fanotify/.gitignore
index c99e6fff7..35e73b91e 100644
--- a/testcases/kernel/syscalls/fanotify/.gitignore
+++ b/testcases/kernel/syscalls/fanotify/.gitignore
@@ -18,4 +18,5 @@ 
 /fanotify18
 /fanotify19
 /fanotify20
+/fanotify21
 /fanotify_child
diff --git a/testcases/kernel/syscalls/fanotify/fanotify21.c b/testcases/kernel/syscalls/fanotify/fanotify21.c
new file mode 100644
index 000000000..e842dad17
--- /dev/null
+++ b/testcases/kernel/syscalls/fanotify/fanotify21.c
@@ -0,0 +1,306 @@ 
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2021 Google. All Rights Reserved.
+ *
+ * Started by Matthew Bobrowski <repnop@google.com>
+ */
+
+/*\
+ * [Description]
+ *
+ * A test which verifies whether the returned struct
+ * fanotify_event_info_pidfd in FAN_REPORT_PIDFD mode contains the
+ * expected set of information.
+ */
+
+#define _GNU_SOURCE
+#include <stdio.h>
+#include <ctype.h>
+#include <stdlib.h>
+#include <string.h>
+#include "tst_test.h"
+#include "tst_safe_stdio.h"
+#include "lapi/pidfd_open.h"
+
+#ifdef HAVE_SYS_FANOTIFY_H
+#include "fanotify.h"
+
+#define BUF_SZ		4096
+#define MOUNT_PATH	"fs_mnt"
+#define TEST_FILE	MOUNT_PATH "/testfile"
+
+struct pidfd_fdinfo_t {
+	int pos;
+	int flags;
+	int mnt_id;
+	int pid;
+	int ns_pid;
+};
+
+struct test_case_t {
+	char *name;
+	int fork;
+	int want_pidfd_err;
+} test_cases[] = {
+	{
+		"return a valid pidfd for event created by self",
+		0,
+		0,
+	},
+	{
+		"return invalid pidfd for event created by terminated child",
+		1,
+		FAN_NOPIDFD,
+	},
+};
+
+static int fanotify_fd;
+static char event_buf[BUF_SZ];
+static struct pidfd_fdinfo_t *self_pidfd_fdinfo = NULL;
+
+static struct pidfd_fdinfo_t *read_pidfd_fdinfo(int pidfd)
+{
+	char *fdinfo_path;
+	struct pidfd_fdinfo_t *pidfd_fdinfo;
+
+	pidfd_fdinfo = SAFE_MALLOC(sizeof(struct pidfd_fdinfo_t));
+
+	SAFE_ASPRINTF(&fdinfo_path, "/proc/self/fdinfo/%d", pidfd);
+	SAFE_FILE_LINES_SCANF(fdinfo_path, "pos: %d", &pidfd_fdinfo->pos);
+	SAFE_FILE_LINES_SCANF(fdinfo_path, "flags: %d", &pidfd_fdinfo->flags);
+	SAFE_FILE_LINES_SCANF(fdinfo_path, "mnt_id: %d", &pidfd_fdinfo->mnt_id);
+	SAFE_FILE_LINES_SCANF(fdinfo_path, "Pid: %d", &pidfd_fdinfo->pid);
+	SAFE_FILE_LINES_SCANF(fdinfo_path, "NSpid: %d", &pidfd_fdinfo->ns_pid);
+
+	free(fdinfo_path);
+
+	return pidfd_fdinfo;
+}
+
+static void generate_event(void)
+{
+	int fd;
+
+	/* Generate a single FAN_OPEN event on the watched object. */
+	fd = SAFE_OPEN(TEST_FILE, O_RDONLY);
+	SAFE_CLOSE(fd);
+}
+
+static void do_fork(void)
+{
+	int status;
+	pid_t child;
+
+	child = SAFE_FORK();
+	if (child == 0) {
+		SAFE_CLOSE(fanotify_fd);
+		generate_event();
+		exit(EXIT_SUCCESS);
+	}
+
+	SAFE_WAITPID(child, &status, 0);
+	if (WIFEXITED(status) && WEXITSTATUS(status) != 0)
+		tst_brk(TBROK,
+			"child process terminated incorrectly");
+}
+
+static void do_setup(void)
+{
+	int pidfd;
+
+	SAFE_TOUCH(TEST_FILE, 0666, NULL);
+
+	/*
+	 * An explicit check for FAN_REPORT_PIDFD is performed early
+	 * on in the test initialization as it's a prerequisite for
+	 * all test cases.
+	 */
+	REQUIRE_FANOTIFY_INIT_FLAGS_SUPPORTED_BY_KERNEL(FAN_REPORT_PIDFD);
+
+	fanotify_fd = SAFE_FANOTIFY_INIT(FAN_REPORT_PIDFD, O_RDONLY);
+	SAFE_FANOTIFY_MARK(fanotify_fd, FAN_MARK_ADD, FAN_OPEN, AT_FDCWD,
+			   TEST_FILE);
+
+	pidfd = pidfd_open(getpid(), 0);
+	if (pidfd < 0) {
+		tst_brk(TBROK | TERRNO,
+			"pidfd=%d, pidfd_open(%d, 0) failed",
+			pidfd, getpid());
+	}
+
+	self_pidfd_fdinfo = read_pidfd_fdinfo(pidfd);
+	if (self_pidfd_fdinfo == NULL) {
+		tst_brk(TBROK,
+			"pidfd=%d, failed to read pidfd fdinfo",
+			pidfd);
+	}
+}
+
+static void do_test(unsigned int num)
+{
+	int i = 0, len;
+	struct test_case_t *tc = &test_cases[num];
+
+	tst_res(TINFO, "Test #%d: %s", num, tc->name);
+
+	/*
+	 * Generate the event in either self or a child process. Event
+	 * generation in a child process is done so that the FAN_NOPIDFD case
+	 * can be verified.
+	 */
+	if (tc->fork)
+		do_fork();
+	else
+		generate_event();
+
+	/*
+	 * Read all of the queued events into the provided event
+	 * buffer.
+	 */
+	len = SAFE_READ(0, fanotify_fd, event_buf, sizeof(event_buf));
+	while(i < len) {
+		struct fanotify_event_metadata *event;
+		struct fanotify_event_info_pidfd *info;
+		struct pidfd_fdinfo_t *event_pidfd_fdinfo = NULL;
+
+		event = (struct fanotify_event_metadata *)&event_buf[i];
+		info = (struct fanotify_event_info_pidfd *)(event + 1);
+
+		/*
+		 * Checks ensuring that pidfd information record object header
+		 * fields are set correctly.
+		 */
+		if (info->hdr.info_type != FAN_EVENT_INFO_TYPE_PIDFD) {
+			tst_res(TFAIL,
+				"unexpected info_type received in info "
+				"header (expected: %d, got: %d",
+				FAN_EVENT_INFO_TYPE_PIDFD,
+				info->hdr.info_type);
+			info = NULL;
+			goto next_event;
+		} else if (info->hdr.len !=
+			   sizeof(struct fanotify_event_info_pidfd)) {
+			tst_res(TFAIL,
+				"unexpected info object length "
+				"(expected: %lu, got: %d",
+				sizeof(struct fanotify_event_info_pidfd),
+				info->hdr.len);
+			info = NULL;
+			goto next_event;
+		}
+
+		/*
+		 * Check if pidfd information object reported any errors during
+		 * creation and whether they're expected.
+		 */
+		if (info->pidfd < 0 && !tc->want_pidfd_err) {
+			tst_res(TFAIL,
+				"pidfd creation failed for pid: %u with pidfd error value "
+				"set to: %d",
+				(unsigned)event->pid,
+				info->pidfd);
+			goto next_event;
+		} else if (tc->want_pidfd_err &&
+			   info->pidfd != tc->want_pidfd_err) {
+			tst_res(TFAIL,
+				"pidfd set to an unexpected error: %d for pid: %u",
+				info->pidfd,
+				(unsigned)event->pid);
+			goto next_event;
+		} else if (tc->want_pidfd_err &&
+			   info->pidfd == tc->want_pidfd_err) {
+			tst_res(TPASS,
+				"pid: %u terminated before pidfd was created, "
+				"pidfd set to the value of: %d, as expected",
+				(unsigned)event->pid,
+				FAN_NOPIDFD);
+			goto next_event;
+		}
+
+		/*
+		 * No pidfd errors occurred, continue with verifying pidfd
+		 * fdinfo validity.
+		 */
+		event_pidfd_fdinfo = read_pidfd_fdinfo(info->pidfd);
+		if (event_pidfd_fdinfo == NULL) {
+			tst_brk(TBROK,
+				"reading fdinfo for pidfd: %d "
+				"describing pid: %u failed",
+				info->pidfd,
+				(unsigned)event->pid);
+			goto next_event;
+		} else if (event_pidfd_fdinfo->pid != event->pid) {
+			tst_res(TFAIL,
+				"pidfd provided for incorrect pid "
+				"(expected pidfd for pid: %u, got pidfd for "
+				"pid: %u)",
+				(unsigned)event->pid,
+				(unsigned)event_pidfd_fdinfo->pid);
+			goto next_event;
+		} else if (memcmp(event_pidfd_fdinfo, self_pidfd_fdinfo,
+				  sizeof(struct pidfd_fdinfo_t))) {
+			tst_res(TFAIL,
+				"pidfd fdinfo values for self and event differ "
+				"(expected pos: %d, flags: %x, mnt_id: %d, "
+				"pid: %d, ns_pid: %d, got pos: %d, "
+				"flags: %x, mnt_id: %d, pid: %d, ns_pid: %d",
+				self_pidfd_fdinfo->pos,
+				self_pidfd_fdinfo->flags,
+				self_pidfd_fdinfo->mnt_id,
+				self_pidfd_fdinfo->pid,
+				self_pidfd_fdinfo->ns_pid,
+				event_pidfd_fdinfo->pos,
+				event_pidfd_fdinfo->flags,
+				event_pidfd_fdinfo->mnt_id,
+				event_pidfd_fdinfo->pid,
+				event_pidfd_fdinfo->ns_pid);
+			goto next_event;
+		} else {
+			tst_res(TPASS,
+				"got an event with a valid pidfd info record, "
+				"mask: %lld, pid: %u, fd: %d, "
+				"pidfd: %d, info_type: %d, info_len: %d",
+				(unsigned long long)event->mask,
+				(unsigned)event->pid,
+				event->fd,
+				info->pidfd,
+				info->hdr.info_type,
+				info->hdr.len);
+		}
+
+next_event:
+		i += event->event_len;
+		if (event->fd >= 0)
+			SAFE_CLOSE(event->fd);
+
+		if (info && info->pidfd >= 0)
+			SAFE_CLOSE(info->pidfd);
+
+		if (event_pidfd_fdinfo)
+			free(event_pidfd_fdinfo);
+	}
+}
+
+static void do_cleanup(void)
+{
+	if (fanotify_fd >= 0)
+		SAFE_CLOSE(fanotify_fd);
+
+	if (self_pidfd_fdinfo)
+		free(self_pidfd_fdinfo);
+}
+
+static struct tst_test test = {
+	.setup = do_setup,
+	.test = do_test,
+	.tcnt = ARRAY_SIZE(test_cases),
+	.cleanup = do_cleanup,
+	.all_filesystems = 1,
+	.needs_root = 1,
+	.mntpoint = MOUNT_PATH,
+	.forks_child = 1,
+};
+
+#else
+	TST_TEST_TCONF("system doesn't have required fanotify support");
+#endif /* HAVE_SYS_FANOTIFY_H */