From: Alexey Dobriyan <adobriyan@gmail.com>
To: akpm@linux-foundation.org
Cc: linux-kernel@vger.kernel.org
Subject: [PATCH] proc: test /proc/*/fd a bit (+ PF_KTHREAD is ABI!)
Date: Sat, 5 May 2018 03:04:14 +0300 [thread overview]
Message-ID: <20180505000414.GA15090@avx2> (raw)
* Test lookup in /proc/self/fd.
"map_files" lookup story showed that lookup is not that simple.
* Test that all those symlinks open the same file.
Check with (st_dev, st_info).
* Test that kernel threads do not have anything in their /proc/*/fd/
directory.
Now this is where things get interesting.
First, kernel threads aren't pinned by /proc/self or equivalent,
thus some "atomicity" is required.
Second, ->comm can contain whitespace and ')'.
No, they are not escaped.
Third, the only reliable way to check if process is kernel thread
appears to be field #9 in /proc/*/stat.
This field is struct task_struct::flags in decimal!
Check is done by testing PF_KTHREAD flags like we do in kernel.
PF_KTREAD value is a part of userspace ABI !!!
Other methods for determining kernel threadness are not reliable:
* RSS can be 0 if everything is swapped, even while reading
from /proc/self.
* ->total_vm CAN BE ZERO if process is finishing
munmap(NULL, whole address space);
* /proc/*/maps and similar files can be empty because unmapping
everything works. Read returning 0 can't distinguish between
kernel thread and such suicide process.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
---
tools/testing/selftests/proc/.gitignore | 3
tools/testing/selftests/proc/Makefile | 5
tools/testing/selftests/proc/fd-001-lookup.c | 168 +++++++++++++++++++++++
tools/testing/selftests/proc/fd-002-posix-eq.c | 57 ++++++++
tools/testing/selftests/proc/fd-003-kthread.c | 178 +++++++++++++++++++++++++
tools/testing/selftests/proc/proc-uptime.h | 16 --
tools/testing/selftests/proc/proc.h | 39 +++++
tools/testing/selftests/proc/read.c | 17 --
8 files changed, 451 insertions(+), 32 deletions(-)
--- a/tools/testing/selftests/proc/.gitignore
+++ b/tools/testing/selftests/proc/.gitignore
@@ -1,3 +1,6 @@
+/fd-001-lookup
+/fd-002-posix-eq
+/fd-003-kthread
/proc-loadavg-001
/proc-self-map-files-001
/proc-self-map-files-002
--- a/tools/testing/selftests/proc/Makefile
+++ b/tools/testing/selftests/proc/Makefile
@@ -1,6 +1,9 @@
-CFLAGS += -Wall -O2
+CFLAGS += -Wall -O2 -Wno-unused-function
TEST_GEN_PROGS :=
+TEST_GEN_PROGS += fd-001-lookup
+TEST_GEN_PROGS += fd-002-posix-eq
+TEST_GEN_PROGS += fd-003-kthread
TEST_GEN_PROGS += proc-loadavg-001
TEST_GEN_PROGS += proc-self-map-files-001
TEST_GEN_PROGS += proc-self-map-files-002
new file mode 100644
--- /dev/null
+++ b/tools/testing/selftests/proc/fd-001-lookup.c
@@ -0,0 +1,168 @@
+/*
+ * Copyright © 2018 Alexey Dobriyan <adobriyan@gmail.com>
+ *
+ * Permission to use, copy, modify, and distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+// Test /proc/*/fd lookup.
+#define _GNU_SOURCE
+#undef NDEBUG
+#include <assert.h>
+#include <dirent.h>
+#include <errno.h>
+#include <limits.h>
+#include <sched.h>
+#include <stdio.h>
+#include <unistd.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+
+#include "proc.h"
+
+/* lstat(2) has more "coverage" in case non-symlink pops up somehow. */
+static void test_lookup_pass(const char *pathname)
+{
+ struct stat st;
+ ssize_t rv;
+
+ memset(&st, 0, sizeof(struct stat));
+ rv = lstat(pathname, &st);
+ assert(rv == 0);
+ assert(S_ISLNK(st.st_mode));
+}
+
+static void test_lookup_fail(const char *pathname)
+{
+ struct stat st;
+ ssize_t rv;
+
+ rv = lstat(pathname, &st);
+ assert(rv == -1 && errno == ENOENT);
+}
+
+static void test_lookup(unsigned int fd)
+{
+ char buf[64];
+ unsigned int c;
+ unsigned int u;
+ int i;
+
+ snprintf(buf, sizeof(buf), "/proc/self/fd/%u", fd);
+ test_lookup_pass(buf);
+
+ /* leading junk */
+ for (c = 1; c <= 255; c++) {
+ if (c == '/')
+ continue;
+ snprintf(buf, sizeof(buf), "/proc/self/fd/%c%u", c, fd);
+ test_lookup_fail(buf);
+ }
+
+ /* trailing junk */
+ for (c = 1; c <= 255; c++) {
+ if (c == '/')
+ continue;
+ snprintf(buf, sizeof(buf), "/proc/self/fd/%u%c", fd, c);
+ test_lookup_fail(buf);
+ }
+
+ for (i = INT_MIN; i < INT_MIN + 1024; i++) {
+ snprintf(buf, sizeof(buf), "/proc/self/fd/%d", i);
+ test_lookup_fail(buf);
+ }
+ for (i = -1024; i < 0; i++) {
+ snprintf(buf, sizeof(buf), "/proc/self/fd/%d", i);
+ test_lookup_fail(buf);
+ }
+ for (u = INT_MAX - 1024; u <= (unsigned int)INT_MAX + 1024; u++) {
+ snprintf(buf, sizeof(buf), "/proc/self/fd/%u", u);
+ test_lookup_fail(buf);
+ }
+ for (u = UINT_MAX - 1024; u != 0; u++) {
+ snprintf(buf, sizeof(buf), "/proc/self/fd/%u", u);
+ test_lookup_fail(buf);
+ }
+
+
+}
+
+int main(void)
+{
+ struct dirent *de;
+ unsigned int fd, target_fd;
+
+ if (unshare(CLONE_FILES) == -1)
+ return 1;
+
+ /* Wipe fdtable. */
+ do {
+ DIR *d;
+
+ d = opendir("/proc/self/fd");
+ if (!d)
+ return 1;
+
+ de = xreaddir(d);
+ assert(de->d_type == DT_DIR);
+ assert(streq(de->d_name, "."));
+
+ de = xreaddir(d);
+ assert(de->d_type == DT_DIR);
+ assert(streq(de->d_name, ".."));
+next:
+ de = xreaddir(d);
+ if (de) {
+ unsigned long long fd_ull;
+ unsigned int fd;
+ char *end;
+
+ assert(de->d_type == DT_LNK);
+
+ fd_ull = xstrtoull(de->d_name, &end);
+ assert(*end == '\0');
+ assert(fd_ull == (unsigned int)fd_ull);
+
+ fd = fd_ull;
+ if (fd == dirfd(d))
+ goto next;
+ close(fd);
+ }
+
+ closedir(d);
+ } while (de);
+
+ /* Now fdtable is clean. */
+
+ fd = open("/", O_PATH|O_DIRECTORY);
+ assert(fd == 0);
+ test_lookup(fd);
+ close(fd);
+
+ /* Clean again! */
+
+ fd = open("/", O_PATH|O_DIRECTORY);
+ assert(fd == 0);
+ /* Default RLIMIT_NOFILE-1 */
+ target_fd = 1023;
+ while (target_fd > 0) {
+ if (dup2(fd, target_fd) == target_fd)
+ break;
+ target_fd /= 2;
+ }
+ assert(target_fd > 0);
+ close(fd);
+ test_lookup(target_fd);
+ close(target_fd);
+
+ return 0;
+}
new file mode 100644
--- /dev/null
+++ b/tools/testing/selftests/proc/fd-002-posix-eq.c
@@ -0,0 +1,57 @@
+/*
+ * Copyright © 2018 Alexey Dobriyan <adobriyan@gmail.com>
+ *
+ * Permission to use, copy, modify, and distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+// Test that open(/proc/*/fd/*) opens the same file.
+#undef NDEBUG
+#include <assert.h>
+#include <stdio.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+#include <unistd.h>
+
+int main(void)
+{
+ int fd0, fd1, fd2;
+ struct stat st0, st1, st2;
+ char buf[64];
+ int rv;
+
+ fd0 = open("/", O_DIRECTORY|O_RDONLY);
+ assert(fd0 >= 0);
+
+ snprintf(buf, sizeof(buf), "/proc/self/fd/%u", fd0);
+ fd1 = open(buf, O_RDONLY);
+ assert(fd1 >= 0);
+
+ snprintf(buf, sizeof(buf), "/proc/thread-self/fd/%u", fd0);
+ fd2 = open(buf, O_RDONLY);
+ assert(fd2 >= 0);
+
+ rv = fstat(fd0, &st0);
+ assert(rv == 0);
+ rv = fstat(fd1, &st1);
+ assert(rv == 0);
+ rv = fstat(fd2, &st2);
+ assert(rv == 0);
+
+ assert(st0.st_dev == st1.st_dev);
+ assert(st0.st_ino == st1.st_ino);
+
+ assert(st0.st_dev == st2.st_dev);
+ assert(st0.st_ino == st2.st_ino);
+
+ return 0;
+}
new file mode 100644
--- /dev/null
+++ b/tools/testing/selftests/proc/fd-003-kthread.c
@@ -0,0 +1,178 @@
+/*
+ * Copyright © 2018 Alexey Dobriyan <adobriyan@gmail.com>
+ *
+ * Permission to use, copy, modify, and distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+// Test that /proc/$KERNEL_THREAD/fd/ is empty.
+#define _GNU_SOURCE
+#undef NDEBUG
+#include <sys/syscall.h>
+#include <assert.h>
+#include <dirent.h>
+#include <limits.h>
+#include <stdio.h>
+#include <string.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+#include <unistd.h>
+
+#include "proc.h"
+
+#define PF_KHTREAD 0x00200000
+
+/*
+ * Test for kernel threadness atomically with openat().
+ *
+ * Return /proc/$PID/fd descriptor if process is kernel thread.
+ * Return -1 if a process is userspace process.
+ */
+static int kernel_thread_fd(unsigned int pid)
+{
+ unsigned int flags = 0;
+ char buf[4096];
+ int dir_fd, fd;
+ ssize_t rv;
+
+ snprintf(buf, sizeof(buf), "/proc/%u", pid);
+ dir_fd = open(buf, O_RDONLY|O_DIRECTORY);
+ if (dir_fd == -1)
+ return -1;
+
+ /*
+ * Believe it or not, struct task_struct::flags is directly exposed
+ * to userspace!
+ */
+ fd = openat(dir_fd, "stat", O_RDONLY);
+ if (fd == -1) {
+ close(dir_fd);
+ return -1;
+ }
+ rv = read(fd, buf, sizeof(buf));
+ close(fd);
+ if (0 < rv && rv <= sizeof(buf)) {
+ unsigned long long flags_ull;
+ char *p, *end;
+ int i;
+
+ assert(buf[rv - 1] == '\n');
+ buf[rv - 1] = '\0';
+
+ /* Search backwards: ->comm can contain whitespace and ')'. */
+ for (i = 0; i < 43; i++) {
+ p = strrchr(buf, ' ');
+ assert(p);
+ *p = '\0';
+ }
+
+ p = strrchr(buf, ' ');
+ assert(p);
+
+ flags_ull = xstrtoull(p + 1, &end);
+ assert(*end == '\0');
+ assert(flags_ull == (unsigned int)flags_ull);
+
+ flags = flags_ull;
+ }
+
+ fd = -1;
+ if (flags & PF_KHTREAD) {
+ fd = openat(dir_fd, "fd", O_RDONLY|O_DIRECTORY);
+ }
+ close(dir_fd);
+ return fd;
+}
+
+static void test_readdir(int fd)
+{
+ DIR *d;
+ struct dirent *de;
+
+ d = fdopendir(fd);
+ assert(d);
+
+ de = xreaddir(d);
+ assert(streq(de->d_name, "."));
+ assert(de->d_type == DT_DIR);
+
+ de = xreaddir(d);
+ assert(streq(de->d_name, ".."));
+ assert(de->d_type == DT_DIR);
+
+ de = xreaddir(d);
+ assert(!de);
+}
+
+static inline int sys_statx(int dirfd, const char *pathname, int flags,
+ unsigned int mask, void *stx)
+{
+ return syscall(SYS_statx, dirfd, pathname, flags, mask, stx);
+}
+
+static void test_lookup_fail(int fd, const char *pathname)
+{
+ char stx[256] __attribute__((aligned(8)));
+ int rv;
+
+ rv = sys_statx(fd, pathname, AT_SYMLINK_NOFOLLOW, 0, (void *)stx);
+ assert(rv == -1 && errno == ENOENT);
+}
+
+static void test_lookup(int fd)
+{
+ char buf[64];
+ unsigned int u;
+ int i;
+
+ for (i = INT_MIN; i < INT_MIN + 1024; i++) {
+ snprintf(buf, sizeof(buf), "%d", i);
+ test_lookup_fail(fd, buf);
+ }
+ for (i = -1024; i < 1024; i++) {
+ snprintf(buf, sizeof(buf), "%d", i);
+ test_lookup_fail(fd, buf);
+ }
+ for (u = INT_MAX - 1024; u < (unsigned int)INT_MAX + 1024; u++) {
+ snprintf(buf, sizeof(buf), "%u", u);
+ test_lookup_fail(fd, buf);
+ }
+ for (u = UINT_MAX - 1024; u != 0; u++) {
+ snprintf(buf, sizeof(buf), "%u", u);
+ test_lookup_fail(fd, buf);
+ }
+}
+
+int main(void)
+{
+ unsigned int pid;
+ int fd;
+
+ /*
+ * In theory this will loop indefinitely if kernel threads are exiled
+ * from /proc.
+ *
+ * Start with kthreadd.
+ */
+ pid = 2;
+ while ((fd = kernel_thread_fd(pid)) == -1 && pid < 1024) {
+ pid++;
+ }
+ /* EACCES if run as non-root. */
+ if (pid >= 1024)
+ return 1;
+
+ test_readdir(fd);
+ test_lookup(fd);
+
+ return 0;
+}
--- a/tools/testing/selftests/proc/proc-uptime.h
+++ b/tools/testing/selftests/proc/proc-uptime.h
@@ -20,21 +20,7 @@
#include <stdlib.h>
#include <unistd.h>
-static unsigned long long xstrtoull(const char *p, char **end)
-{
- if (*p == '0') {
- *end = (char *)p + 1;
- return 0;
- } else if ('1' <= *p && *p <= '9') {
- unsigned long long val;
-
- errno = 0;
- val = strtoull(p, end, 10);
- assert(errno == 0);
- return val;
- } else
- assert(0);
-}
+#include "proc.h"
static void proc_uptime(int fd, uint64_t *uptime, uint64_t *idle)
{
new file mode 100644
--- /dev/null
+++ b/tools/testing/selftests/proc/proc.h
@@ -0,0 +1,39 @@
+#pragma once
+#undef NDEBUG
+#include <assert.h>
+#include <dirent.h>
+#include <errno.h>
+#include <stdbool.h>
+#include <stdlib.h>
+#include <string.h>
+
+static inline bool streq(const char *s1, const char *s2)
+{
+ return strcmp(s1, s2) == 0;
+}
+
+static unsigned long long xstrtoull(const char *p, char **end)
+{
+ if (*p == '0') {
+ *end = (char *)p + 1;
+ return 0;
+ } else if ('1' <= *p && *p <= '9') {
+ unsigned long long val;
+
+ errno = 0;
+ val = strtoull(p, end, 10);
+ assert(errno == 0);
+ return val;
+ } else
+ assert(0);
+}
+
+static struct dirent *xreaddir(DIR *d)
+{
+ struct dirent *de;
+
+ errno = 0;
+ de = readdir(d);
+ assert(de || errno == 0);
+ return de;
+}
--- a/tools/testing/selftests/proc/read.c
+++ b/tools/testing/selftests/proc/read.c
@@ -31,22 +31,7 @@
#include <fcntl.h>
#include <unistd.h>
-static inline bool streq(const char *s1, const char *s2)
-{
- return strcmp(s1, s2) == 0;
-}
-
-static struct dirent *xreaddir(DIR *d)
-{
- struct dirent *de;
-
- errno = 0;
- de = readdir(d);
- if (!de && errno != 0) {
- exit(1);
- }
- return de;
-}
+#include "proc.h"
static void f_reg(DIR *d, const char *filename)
{
next reply other threads:[~2018-05-05 0:04 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-05-05 0:04 Alexey Dobriyan [this message]
2018-05-08 22:39 ` [PATCH] proc: test /proc/*/fd a bit (+ PF_KTHREAD is ABI!) Andrew Morton
2018-05-09 16:34 ` Alexey Dobriyan
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180505000414.GA15090@avx2 \
--to=adobriyan@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=linux-kernel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.