Skip to contentSkip to conversation

fork: extend clone3() to support CLONE_SET_TID

PATCH In review Adrian Reber 3 participants 8 messages Aug 11, 2019 – Aug 12, 2019

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

PATCH v5 1/2 fork: extend clone3() to support CLONE_SET_TID

The main motivation to add set_tid to clone3() is CRIU.

To restore a process with the same PID/TID CRIU currently uses /proc/sys/kernel/ns_last_pid. It writes the desired (PID - 1) to ns_last_pid and then (quickly) does a clone(). This works most of the time, but it is racy. It is also slow as it requires multiple syscalls.

Extending clone3() to support set_tid makes it possible restore a process using CRIU without accessing /proc/sys/kernel/ns_last_pid and race free (as long as the desired PID/TID is available).

This clone3() extension places the same restrictions (CAP_SYS_ADMIN) on clone3() with set_tid as they are currently in place for ns_last_pid.

---
v2:
- Removed (size < sizeof(struct clone_args)) as discussed with Christian and Dmitry
- Added comment to ((set_tid != 1) && idr_get_cursor() <= 1) (Oleg)
- Use idr_alloc() instead of idr_alloc_cyclic() (Oleg)

v3:
- Return EEXIST if PID is already in use (Christian)
- Drop CLONE_SET_TID (Christian and Oleg)
- Use idr_is_empty() instead of idr_get_cursor() (Oleg)
- Handle different struct clone_args sizes (Dmitry)

v4:
- Rework struct size check with defines (Christian)
- Reduce number of set_tid checks (Oleg)
- Less parentheses and more robust code (Oleg)
- Do ns_capable() on correct user_ns (Oleg, Christian)

v5:
- make set_tid checks earlier in alloc_pid() (Christian)
- remove unnecessary comment and struct size check (Christian)

---
 include/linux/pid.h        |  2 +-
 include/linux/sched/task.h |  1 +
 include/uapi/linux/sched.h |  1 +
 kernel/fork.c              | 22 ++++++++++++++++++++--
 kernel/pid.c               | 36 +++++++++++++++++++++++++++++-------
 5 files changed, 52 insertions(+), 10 deletions(-)
include/linux/pid.h +1 -1
index 2a83e434db9d..052000db0ced 100644
@@ -116,7 +116,7 @@ extern struct pid *find_vpid(int nr);
 extern struct pid *find_get_pid(int nr);
 extern struct pid *find_ge_pid(int nr, struct pid_namespace *);
 
-extern struct pid *alloc_pid(struct pid_namespace *ns);
+extern struct pid *alloc_pid(struct pid_namespace *ns, pid_t set_tid);
 extern void free_pid(struct pid *pid);
 extern void disable_pid_allocation(struct pid_namespace *ns);
 
include/linux/sched/task.h +1 -0
index 0497091e40c1..4f2a80564332 100644
@@ -26,6 +26,7 @@ struct kernel_clone_args {
 	unsigned long stack;
 	unsigned long stack_size;
 	unsigned long tls;
+	pid_t set_tid;
 };
 
 /*
include/uapi/linux/sched.h +1 -0
index b3105ac1381a..e1ce103a2c47 100644
@@ -45,6 +45,7 @@ struct clone_args {
 	__aligned_u64 stack;
 	__aligned_u64 stack_size;
 	__aligned_u64 tls;
+	__aligned_u64 set_tid;
 };
 
 /*
kernel/fork.c +20 -2
index 2852d0e76ea3..0365041243b1 100644
@@ -117,6 +117,13 @@
  */
 #define MAX_THREADS FUTEX_TID_MASK
 
+/*
+ * Different sizes of struct clone_args
+ */
+#define CLONE3_ARGS_SIZE_V0 64
+/* V1 includes set_tid */
+#define CLONE3_ARGS_SIZE_V1 72
+
 /*
  * Protected counters by write_lock_irq(&tasklist_lock)
  */
@@ -2031,7 +2038,13 @@ static __latent_entropy struct task_struct *copy_process(
 	stackleak_task_init(p);
 
 	if (pid != &init_struct_pid) {
-		pid = alloc_pid(p->nsproxy->pid_ns_for_children);
+		if (args->set_tid && !ns_capable(
+				p->nsproxy->pid_ns_for_children->user_ns,
+				CAP_SYS_ADMIN)) {
+			retval = -EPERM;
+			goto bad_fork_cleanup_thread;
+		}
+		pid = alloc_pid(p->nsproxy->pid_ns_for_children, args->set_tid);
 		if (IS_ERR(pid)) {
 			retval = PTR_ERR(pid);
 			goto bad_fork_cleanup_thread;
@@ -2535,9 +2548,13 @@ noinline static int copy_clone_args_from_user(struct kernel_clone_args *kargs,
 	if (unlikely(size > PAGE_SIZE))
 		return -E2BIG;
 
-	if (unlikely(size < sizeof(struct clone_args)))
+	if (unlikely(size < CLONE3_ARGS_SIZE_V0))
 		return -EINVAL;
 
+	if (size < sizeof(struct clone_args))
+		memset((void *)&args + size, 0,
+				sizeof(struct clone_args) - size);
+
 	if (unlikely(!access_ok(uargs, size)))
 		return -EFAULT;
 
@@ -2571,6 +2588,7 @@ noinline static int copy_clone_args_from_user(struct kernel_clone_args *kargs,
 		.stack		= args.stack,
 		.stack_size	= args.stack_size,
 		.tls		= args.tls,
+		.set_tid	= args.set_tid,
 	};
 
 	return 0;
kernel/pid.c +29 -7
index 0a9f2e437217..c3ce82f8206b 100644
@@ -157,7 +157,7 @@ void free_pid(struct pid *pid)
 	call_rcu(&pid->rcu, delayed_put_pid);
 }
 
-struct pid *alloc_pid(struct pid_namespace *ns)
+struct pid *alloc_pid(struct pid_namespace *ns, int set_tid)
 {
 	struct pid *pid;
 	enum pid_type type;
@@ -166,6 +166,9 @@ struct pid *alloc_pid(struct pid_namespace *ns)
 	struct upid *upid;
 	int retval = -ENOMEM;
 
+	if (set_tid < 0 || set_tid >= pid_max)
+		return ERR_PTR(-EINVAL);
+
 	pid = kmem_cache_alloc(ns->pid_cachep, GFP_KERNEL);
 	if (!pid)
 		return ERR_PTR(retval);
@@ -186,12 +189,31 @@ struct pid *alloc_pid(struct pid_namespace *ns)
 		if (idr_get_cursor(&tmp->idr) > RESERVED_PIDS)
 			pid_min = RESERVED_PIDS;
 
-		/*
-		 * Store a null pointer so find_pid_ns does not find
-		 * a partially initialized PID (see below).
-		 */
-		nr = idr_alloc_cyclic(&tmp->idr, NULL, pid_min,
-				      pid_max, GFP_ATOMIC);
+		if (set_tid) {
+			/*
+			 * Also fail if a PID != 1 is requested
+			 * and no PID 1 exists.
+			 */
+			nr = -EINVAL;
+			if (set_tid == 1 || !idr_is_empty(&tmp->idr))
+				nr = idr_alloc(&tmp->idr, NULL, set_tid,
+					       set_tid + 1, GFP_ATOMIC);
+			/*
+			 * If ENOSPC is returned it means that the PID is
+			 * alreay in use. Return EEXIST in that case.
+			 */
+			if (nr == -ENOSPC)
+				nr = -EEXIST;
+			/* Only use set_tid for one PID namespace. */
+			set_tid = 0;
+		} else {
+			/*
+			 * Store a null pointer so find_pid_ns does not find
+			 * a partially initialized PID (see below).
+			 */
+			nr = idr_alloc_cyclic(&tmp->idr, NULL, pid_min,
+					      pid_max, GFP_ATOMIC);
+		}
 		spin_unlock_irq(&pidmap_lock);
 		idr_preload_end();
 
Signature · 1 line
2.21.0
#1 lore raw

PATCH v5 2/2 selftests: add tests for clone3()

This tests clone3() with and without set_tid to see if all desired PIDs are working as expected. The test tries to clone3() with a set_tid of -1, 1, pid_max, a PID which is already in use and an unused PID. The same tests are also running in PID namespace.

---
 tools/testing/selftests/clone3/.gitignore     |   2 +
 tools/testing/selftests/clone3/Makefile       |  11 ++
 tools/testing/selftests/clone3/clone3.c       | 141 +++++++++++++++
 .../testing/selftests/clone3/clone3_set_tid.c | 161 ++++++++++++++++++
 4 files changed, 315 insertions(+)
 create mode 100644 tools/testing/selftests/clone3/.gitignore
 create mode 100644 tools/testing/selftests/clone3/Makefile
 create mode 100644 tools/testing/selftests/clone3/clone3.c
 create mode 100644 tools/testing/selftests/clone3/clone3_set_tid.c
tools/testing/selftests/clone3/.gitignore new file +2 -0
new file mode 100644
index 000000000000..c63c64a78ddf
@@ -0,0 +1,2 @@
+clone3_set_tid
+clone3
tools/testing/selftests/clone3/Makefile new file +11 -0
new file mode 100644
index 000000000000..4efcf45b995b
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: GPL-2.0
+uname_M := $(shell uname -m 2>/dev/null || echo not)
+ARCH ?= $(shell echo $(uname_M) | sed -e s/i.86/i386/)
+
+CFLAGS += -I../../../../usr/include/
+
+ifeq ($(ARCH),x86_64)
+	TEST_GEN_PROGS := clone3 clone3_set_tid
+endif
+
+include ../lib.mk
tools/testing/selftests/clone3/clone3.c new file +141 -0

141 lines not on this page · raw message

tools/testing/selftests/clone3/clone3_set_tid.c new file +161 -0

161 lines not on this page · raw message

Signature · 1 line
2.21.0
#2 lore raw

Re: [PATCH v5 2/2] selftests: add tests for clone3()

Quoted 350 lines · Adrian Reber

This tests clone3() with and without set_tid to see if all desired PIDs are working as expected. The test tries to clone3() with a set_tid of -1, 1, pid_max, a PID which is already in use and an unused PID. The same tests are also running in PID namespace.

---

tools/testing/selftests/clone3/.gitignore     |   2 +
tools/testing/selftests/clone3/Makefile       |  11 ++
tools/testing/selftests/clone3/clone3.c       | 141 +++++++++++++++
.../testing/selftests/clone3/clone3_set_tid.c | 161 ++++++++++++++++++
4 files changed, 315 insertions(+)

create mode 100644 tools/testing/selftests/clone3/.gitignore create mode 100644 tools/testing/selftests/clone3/Makefile create mode 100644 tools/testing/selftests/clone3/clone3.c create mode 100644 tools/testing/selftests/clone3/clone3_set_tid.c

tools/testing/selftests/clone3/.gitignore new file +2 -0
new file mode 100644
index 000000000000..c63c64a78ddf
@@ -0,0 +1,2 @@
+clone3_set_tid
+clone3
tools/testing/selftests/clone3/Makefile new file +11 -0
new file mode 100644
index 000000000000..4efcf45b995b
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: GPL-2.0
+uname_M := $(shell uname -m 2>/dev/null || echo not)
+ARCH ?= $(shell echo $(uname_M) | sed -e s/i.86/i386/)
+
+CFLAGS += -I../../../../usr/include/
+
+ifeq ($(ARCH),x86_64)
+	TEST_GEN_PROGS := clone3 clone3_set_tid
+endif
+
+include ../lib.mk
tools/testing/selftests/clone3/clone3.c new file +141 -0
new file mode 100644
index 000000000000..55a6915566b8
@@ -0,0 +1,141 @@
+// SPDX-License-Identifier: GPL-2.0
+
+/* Based on Christian Brauner's clone3() example */
+
+#define _GNU_SOURCE
+#include <errno.h>
+#include <linux/types.h>
+#include <linux/sched.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/syscall.h>
+#include <sys/types.h>
+#include <sys/un.h>
+#include <sys/wait.h>
+#include <unistd.h>
+#include <sched.h>
+
+#include "../kselftest.h"
+
+static pid_t raw_clone(struct clone_args *args)
+{
+	return syscall(__NR_clone3, args, sizeof(struct clone_args));
+}
+
+static int call_clone3(int flags)
+{
+	struct clone_args args = {0};
+	pid_t ppid = -1;
+	pid_t pid = -1;
+	int status;
+
+	args.flags = flags;
+	args.exit_signal = SIGCHLD;
+
+	pid = raw_clone(&args);
+	if (pid < 0) {
+		ksft_print_msg("%s - Failed to create new process\n",
+				strerror(errno));
+		return -errno;
+	}
+
+	if (pid == 0) {
+		ksft_print_msg("I am the child, my PID is %d\n", getpid());
+		_exit(EXIT_SUCCESS);
+	}
+
+	ppid = getpid();
+	ksft_print_msg("I am the parent (%d). My child's pid is %d\n",
+			ppid, pid);
+
+	(void)wait(&status);
+	if (WEXITSTATUS(status))
+		return WEXITSTATUS(status);
+
+	return 0;
+}
+
+static int test_clone3(int flags, int expected)
+{
+	int ret;
+
+	ksft_print_msg("[%d] Trying clone3() with flags 0x%x\n",
+			getpid(), flags);
+	ret = call_clone3(flags);
+	ksft_print_msg("[%d] clone3() with flags says :%d expected %d\n",
+			getpid(), ret, expected);
+	if (ret != expected)
+		ksft_exit_fail_msg(
+			"[%d] Result (%d) is different than expected (%d)\n",
+			getpid(), ret, expected);
+	ksft_test_result_pass("[%d] Result (%d) matches expectation (%d)\n",
+			getpid(), ret, expected);
+	return 0;
+}
+int main(int argc, char *argv[])
+{
+	int ret = -1;
+	pid_t pid;
+
+	ksft_print_header();
+	ksft_set_plan(3);
+
+	/* Just a simple clone3() should return 0.*/
+	if (test_clone3(0, 0))
+		goto on_error;
+	/* Do a clone3() in a new PID NS.*/
+	if (test_clone3(CLONE_NEWPID, 0))
+		goto on_error;
+	ksft_print_msg("First unshare\n");
+	if (unshare(CLONE_NEWPID))
+		goto on_error;
+	/*
+	 * Before clone3()ing in a new PID NS with
+	 * CLONE_NEWPID a fork() is necessary.
+	 */
+	if (test_clone3(CLONE_NEWPID, -EINVAL))
+		goto on_error;
+	pid = fork();
+	if (pid < 0) {
+		ksft_print_msg("First fork() failed\n");
+		goto on_error;
+	}
+	if (pid > 0) {
+		(void)wait(NULL);
+		goto parent_out;
+	}
+	ksft_set_plan(6);
+	if (test_clone3(CLONE_NEWPID, 0))
+		goto on_error;
+	if (test_clone3(0, 0))
+		goto on_error;
+	ksft_print_msg("Second unshare\n");
+	if (unshare(CLONE_NEWPID))
+		goto on_error;
+	/*
+	 * Before clone3()ing in a new PID NS with
+	 * CLONE_NEWPID a fork() is necessary.
+	 */
+	if (test_clone3(CLONE_NEWPID, -EINVAL))
+		goto on_error;
+	pid = fork();
+	if (pid < 0) {
+		ksft_print_msg("Second fork() failed\n");
+		goto on_error;
+	}
+	if (pid > 0) {
+		(void)wait(NULL);
+		goto parent_out;
+	}
+	ksft_set_plan(8);
+	if (test_clone3(CLONE_NEWPID, 0))
+		goto on_error;
+	if (test_clone3(0, 0))
+		goto on_error;
+
+parent_out:
+	ret = 0;
+on_error:
+
+	return !ret ? ksft_exit_pass() : ksft_exit_fail();
+}
tools/testing/selftests/clone3/clone3_set_tid.c new file +151 -0
new file mode 100644
index 000000000000..f5012e84dcb3
@@ -0,0 +1,161 @@
+// SPDX-License-Identifier: GPL-2.0
+
+/* Based on Christian Brauner's clone3() example */
+
+#define _GNU_SOURCE
+#include <errno.h>
+#include <linux/types.h>
+#include <linux/sched.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/syscall.h>
+#include <sys/types.h>
+#include <sys/un.h>
+#include <sys/wait.h>
+#include <unistd.h>
+#include <sched.h>
+
+#include "../kselftest.h"
+
+static pid_t raw_clone(struct clone_args *args)
+{
+	return syscall(__NR_clone3, args, sizeof(struct clone_args));
+}
+
+static int call_clone3_set_tid(int set_tid, int flags)
+{
+	struct clone_args args = {0};
+	pid_t ppid = -1;
+	pid_t pid = -1;
+	int status;
+
+	args.flags = flags;
+	args.exit_signal = SIGCHLD;
+	args.set_tid = set_tid;
+
+	pid = raw_clone(&args);
+	if (pid < 0) {
+		ksft_print_msg("%s - Failed to create new process\n",
+				strerror(errno));
+		return -errno;
+	}
+
+	if (pid == 0) {
+		ksft_print_msg("I am the child, my PID is %d (expected %d)\n",
+				getpid(), set_tid);
+		if (set_tid != getpid())
+			_exit(EXIT_FAILURE);
+		_exit(EXIT_SUCCESS);
+	}
+
+	ppid = getpid();
+	ksft_print_msg("I am the parent (%d). My child's pid is %d\n",
+			ppid, pid);
+
+	(void)wait(&status);
+	if (WEXITSTATUS(status))
+		return WEXITSTATUS(status);
+
+	return 0;
+}
+
+static int test_clone3_set_tid(int set_tid, int flags, int expected)
+{
+	int ret;
+
+	ksft_print_msg(
+		"[%d] Trying clone3() with CLONE_SET_TID to %d and 0x%x\n",
+		getpid(), set_tid, flags);
+	ret = call_clone3_set_tid(set_tid, flags);
+	ksft_print_msg(
+		"[%d] clone3() with CLONE_SET_TID %d says :%d - expected %d\n",
+		getpid(), set_tid, ret, expected);
+	if (ret != expected)
+		ksft_exit_fail_msg(
+			"[%d] Result (%d) is different than expected (%d)\n",
+			getpid(), ret, expected);
+	ksft_test_result_pass("[%d] Result (%d) matches expectation (%d)\n",
+			getpid(), ret, expected);
+	return 0;
+}
+int main(int argc, char *argv[])
+{
+	FILE *f;
+	int pid_max = 0;
+	pid_t pid;
+	pid_t ns_pid;
+	int ret = -1;
+
+	ksft_print_header();
+	ksft_set_plan(13);
+
+	f = fopen("/proc/sys/kernel/pid_max", "r");
+	if (f == NULL)
+		ksft_exit_fail_msg(
+			"%s - Could not open /proc/sys/kernel/pid_max\n",
+			strerror(errno));
+	fscanf(f, "%d", &pid_max);
+	fclose(f);
+	ksft_print_msg("/proc/sys/kernel/pid_max %d\n", pid_max);
+
+	/* First try with an invalid PID */
+	if (test_clone3_set_tid(-1, 0, -EINVAL))
+		goto on_error;
+	if (test_clone3_set_tid(-1, CLONE_NEWPID, -EINVAL))
+		goto on_error;
+	/* Then with PID 1 */
+	if (test_clone3_set_tid(1, 0, -EEXIST))
+		goto on_error;
+	/* PID 1 should not fail in a PID namespace */
+	if (test_clone3_set_tid(1, CLONE_NEWPID, 0))
+		goto on_error;
+	/* pid_max should fail everywhere */
+	if (test_clone3_set_tid(pid_max, 0, -EINVAL))
+		goto on_error;
+	if (test_clone3_set_tid(pid_max, CLONE_NEWPID, -EINVAL))
+		goto on_error;
+	/* Find the current active PID */
+	pid = fork();
+	if (pid == 0) {
+		ksft_print_msg("Child has PID %d\n", getpid());
+		sleep(1);
+		_exit(EXIT_SUCCESS);
+	}
+	/* Try to create a process with that PID should fail */
+	if (test_clone3_set_tid(pid, 0, -EEXIST))
+		goto on_error;
+	(void)wait(NULL);
+	/* After the child has finished, try again with the same PID */
+	if (test_clone3_set_tid(pid, 0, 0))
+		goto on_error;
+	/* This should fail as there is no PID 1 in that namespace */
+	if (test_clone3_set_tid(pid, CLONE_NEWPID, -EINVAL))
+		goto on_error;
+	unshare(CLONE_NEWPID);
+	if (test_clone3_set_tid(10, 0, -EINVAL))
+		goto on_error;
+	/* Let's create a PID 1 */
+	ns_pid = fork();
+	if (ns_pid == 0) {
+		ksft_print_msg("Child in PID namespace has PID %d\n", getpid());
+		sleep(1);
+		_exit(EXIT_SUCCESS);
+	}
+	/*
+	 * Now, after the unshare() it should be possible to create a process
+	 * with another ID than 1 in the PID namespace.
+	 */
+	if (test_clone3_set_tid(2, 0, 0))
+		goto on_error;
+	/* Use a different PID in this namespace. */
+	if (test_clone3_set_tid(2222, 0, 0))

tools/testing/selftests/clone3/clone3_set_tid.c line 161

Quoted 10 lines
tools/testing/selftests/clone3/clone3_set_tid.c new file +10 -0
+		goto on_error;
+	if (test_clone3_set_tid(1, 0, -EEXIST))
+		goto on_error;
+	(void)wait(NULL);
+
+	ret = 0;
+on_error:
+
+	return !ret ? ksft_exit_pass() : ksft_exit_fail();
+}

Thanks for the tests!
Could you also add tests where we pass invalid or unknown struct sizes?:
- pass struct size that is too small aka < CLONE3_ARGS_SIZE_V0
- pass a struct size that is too large aka > PAGE_SIZE
- pass struct size that is smaller than the one the kernel supports
- pass struct size that is larger than the one the kernel supports; once with all unknown fields set to 0 and once with a field set to non-0

I think this might fit well with this patch since it changes the struct size. :)

Sorry, I didn't know you'd be sending out a new version today. :)

Christian

#3 lore raw
In reply to Adrian Reber
Quoted 193 lines · Adrian Reber

The main motivation to add set_tid to clone3() is CRIU.

To restore a process with the same PID/TID CRIU currently uses /proc/sys/kernel/ns_last_pid. It writes the desired (PID - 1) to ns_last_pid and then (quickly) does a clone(). This works most of the time, but it is racy. It is also slow as it requires multiple syscalls.

Extending clone3() to support set_tid makes it possible restore a process using CRIU without accessing /proc/sys/kernel/ns_last_pid and race free (as long as the desired PID/TID is available).

This clone3() extension places the same restrictions (CAP_SYS_ADMIN) on clone3() with set_tid as they are currently in place for ns_last_pid.

---
v2:
- Removed (size < sizeof(struct clone_args)) as discussed with Christian and Dmitry
- Added comment to ((set_tid != 1) && idr_get_cursor() <= 1) (Oleg)
- Use idr_alloc() instead of idr_alloc_cyclic() (Oleg)

v3:
- Return EEXIST if PID is already in use (Christian)
- Drop CLONE_SET_TID (Christian and Oleg)
- Use idr_is_empty() instead of idr_get_cursor() (Oleg)
- Handle different struct clone_args sizes (Dmitry)

v4:
- Rework struct size check with defines (Christian)
- Reduce number of set_tid checks (Oleg)
- Less parentheses and more robust code (Oleg)
- Do ns_capable() on correct user_ns (Oleg, Christian)

v5:
- make set_tid checks earlier in alloc_pid() (Christian)
- remove unnecessary comment and struct size check (Christian)
---

include/linux/pid.h        |  2 +-
include/linux/sched/task.h |  1 +
include/uapi/linux/sched.h |  1 +
kernel/fork.c              | 22 ++++++++++++++++++++--
kernel/pid.c               | 36 +++++++++++++++++++++++++++++-------
5 files changed, 52 insertions(+), 10 deletions(-)
include/linux/pid.h +1 -1
index 2a83e434db9d..052000db0ced 100644
@@ -116,7 +116,7 @@ extern struct pid *find_vpid(int nr);
 extern struct pid *find_get_pid(int nr);
 extern struct pid *find_ge_pid(int nr, struct pid_namespace *);
 
-extern struct pid *alloc_pid(struct pid_namespace *ns);
+extern struct pid *alloc_pid(struct pid_namespace *ns, pid_t set_tid);
 extern void free_pid(struct pid *pid);
 extern void disable_pid_allocation(struct pid_namespace *ns);
 
include/linux/sched/task.h +1 -0
index 0497091e40c1..4f2a80564332 100644
@@ -26,6 +26,7 @@ struct kernel_clone_args {
 	unsigned long stack;
 	unsigned long stack_size;
 	unsigned long tls;
+	pid_t set_tid;
 };
 
 /*
include/uapi/linux/sched.h +1 -0
index b3105ac1381a..e1ce103a2c47 100644
@@ -45,6 +45,7 @@ struct clone_args {
 	__aligned_u64 stack;
 	__aligned_u64 stack_size;
 	__aligned_u64 tls;
+	__aligned_u64 set_tid;
 };
 
 /*
kernel/fork.c +14 -1
index 2852d0e76ea3..0365041243b1 100644
@@ -117,6 +117,13 @@
 */
 #define MAX_THREADS FUTEX_TID_MASK
 
+/*
+ * Different sizes of struct clone_args
+ */
+#define CLONE3_ARGS_SIZE_V0 64
+/* V1 includes set_tid */
+#define CLONE3_ARGS_SIZE_V1 72
+
 /*
 * Protected counters by write_lock_irq(&tasklist_lock)
 */
@@ -2031,7 +2038,13 @@ static __latent_entropy struct task_struct
 *copy_process(
 	stackleak_task_init(p);
 
 	if (pid != &init_struct_pid) {
-		pid = alloc_pid(p->nsproxy->pid_ns_for_children);
+		if (args->set_tid && !ns_capable(
+				p->nsproxy->pid_ns_for_children->user_ns,
+				CAP_SYS_ADMIN)) {
+			retval = -EPERM;
+			goto bad_fork_cleanup_thread;
+		}
+		pid = alloc_pid(p->nsproxy->pid_ns_for_children, args->set_tid);
 		if (IS_ERR(pid)) {
 			retval = PTR_ERR(pid);
			goto bad_fork_cleanup_thread;
kernel/fork.c +6 -1
@@ -2535,9 +2548,13 @@ noinline static int
 copy_clone_args_from_user(struct kernel_clone_args *kargs,
 	if (unlikely(size > PAGE_SIZE))
 		return -E2BIG;
 
-	if (unlikely(size < sizeof(struct clone_args)))
+	if (unlikely(size < CLONE3_ARGS_SIZE_V0))
 		return -EINVAL;
 
+	if (size < sizeof(struct clone_args))
+		memset((void *)&args + size, 0,
+				sizeof(struct clone_args) - size);
+
 	if (unlikely(!access_ok(uargs, size)))
 		return -EFAULT;
@@ -2571,6 +2588,7 @@ noinline static int
 copy_clone_args_from_user(struct kernel_clone_args *kargs,
 		.stack		= args.stack,
 		.stack_size	= args.stack_size,
 		.tls		= args.tls,
+		.set_tid	= args.set_tid,
 	};
 
	return 0;
kernel/pid.c +22 -7
index 0a9f2e437217..c3ce82f8206b 100644
@@ -157,7 +157,7 @@ void free_pid(struct pid *pid)
 	call_rcu(&pid->rcu, delayed_put_pid);
 }
 
-struct pid *alloc_pid(struct pid_namespace *ns)
+struct pid *alloc_pid(struct pid_namespace *ns, int set_tid)
 {
 	struct pid *pid;
 	enum pid_type type;
@@ -166,6 +166,9 @@ struct pid *alloc_pid(struct pid_namespace *ns)
 	struct upid *upid;
 	int retval = -ENOMEM;
 
+	if (set_tid < 0 || set_tid >= pid_max)
+		return ERR_PTR(-EINVAL);
+
 	pid = kmem_cache_alloc(ns->pid_cachep, GFP_KERNEL);
 	if (!pid)
 		return ERR_PTR(retval);
@@ -186,12 +189,31 @@ struct pid *alloc_pid(struct pid_namespace *ns)
 		if (idr_get_cursor(&tmp->idr) > RESERVED_PIDS)
 			pid_min = RESERVED_PIDS;
 
-		/*
-		 * Store a null pointer so find_pid_ns does not find
-		 * a partially initialized PID (see below).
-		 */
-		nr = idr_alloc_cyclic(&tmp->idr, NULL, pid_min,
-				      pid_max, GFP_ATOMIC);
+		if (set_tid) {
+			/*
+			 * Also fail if a PID != 1 is requested
+			 * and no PID 1 exists.
+			 */
+			nr = -EINVAL;
+			if (set_tid == 1 || !idr_is_empty(&tmp->idr))
+				nr = idr_alloc(&tmp->idr, NULL, set_tid,
+					       set_tid + 1, GFP_ATOMIC);
+			/*
+			 * If ENOSPC is returned it means that the PID is
+			 * alreay in use. Return EEXIST in that case.
+			 */
+			if (nr == -ENOSPC)
+				nr = -EEXIST;
+			/* Only use set_tid for one PID namespace. */
+			set_tid = 0;
+		} else {

kernel/pid.c struct pid *alloc_pid(struct pid_namespace *ns) · line 219

Quoted 10 lines
kernel/pid.c +7 -0
+			/*
+			 * Store a null pointer so find_pid_ns does not find
+			 * a partially initialized PID (see below).
+			 */
+			nr = idr_alloc_cyclic(&tmp->idr, NULL, pid_min,
+					      pid_max, GFP_ATOMIC);
+		}
 		spin_unlock_irq(&pidmap_lock);
 		idr_preload_end();
 

Ah, one more nit, since I reckon you'll resend anyway because of the additional tests:
The subject of the patch is not accurate anymore.
Since we don't add a new flag but rather only a new argument referencing CLONE_SET_TID seems misleading. :)

Christian

#4 lore raw
In reply to Adrian Reber
Quoted 7 lines · Adrian Reber
 include/linux/pid.h        |  2 +-
 include/linux/sched/task.h |  1 +
 include/uapi/linux/sched.h |  1 +
 kernel/fork.c              | 22 ++++++++++++++++++++--
 kernel/pid.c               | 36 +++++++++++++++++++++++++++++-------
 5 files changed, 52 insertions(+), 10 deletions(-)

Looks good to me...

A couple of nits below, but I won't insist, feel free to ignore.

kernel/fork.c line 123

Quoted 4 lines
kernel/fork.c +4 -0
+/*
+ * Different sizes of struct clone_args
+ */
+#define CLONE3_ARGS_SIZE_V0 64

I don't really understand why do we want the "size < CLONE3_ARGS_SIZE_V0" check in copy_clone_args_from_user(), but I won't argue.

kernel/fork.c line 125

Quoted 2 lines
kernel/fork.c +2 -0
+/* V1 includes set_tid */
+#define CLONE3_ARGS_SIZE_V1 72

unused?

kernel/fork.c static __latent_entropy struct task_struct *copy_process( · line 2047

Quoted 12 lines
kernel/fork.c +7 -1
@@ -2031,7 +2038,13 @@ static __latent_entropy struct task_struct *copy_process(
 	stackleak_task_init(p);
 
 	if (pid != &init_struct_pid) {
-		pid = alloc_pid(p->nsproxy->pid_ns_for_children);
+		if (args->set_tid && !ns_capable(
+				p->nsproxy->pid_ns_for_children->user_ns,
+				CAP_SYS_ADMIN)) {
+			retval = -EPERM;
+			goto bad_fork_cleanup_thread;
+		}
+		pid = alloc_pid(p->nsproxy->pid_ns_for_children, args->set_tid);

copy_process() is already huge and complex, why not move this check into alloc_pid() ? Again, note that is accepts the same ->pid_ns_for_children.

Quoted 17 lines
kernel/pid.c +3 -3
@@ -166,6 +166,9 @@ struct pid *alloc_pid(struct pid_namespace *ns)
 	struct upid *upid;
 	int retval = -ENOMEM;
 
+	if (set_tid < 0 || set_tid >= pid_max)
+		return ERR_PTR(-EINVAL);
+
 	pid = kmem_cache_alloc(ns->pid_cachep, GFP_KERNEL);
 	if (!pid)
 		return ERR_PTR(retval);
@@ -186,12 +189,31 @@ struct pid *alloc_pid(struct pid_namespace *ns)
 		if (idr_get_cursor(&tmp->idr) > RESERVED_PIDS)
 			pid_min = RESERVED_PIDS;
 
-		/*
-		 * Store a null pointer so find_pid_ns does not find
-		 * a partially initialized PID (see below).

kernel/pid.c struct pid *alloc_pid(struct pid_namespace *ns) · line 198

Quoted 10 lines
kernel/pid.c +7 -3
-		 */
-		nr = idr_alloc_cyclic(&tmp->idr, NULL, pid_min,
-				      pid_max, GFP_ATOMIC);
+		if (set_tid) {
+			/*
+			 * Also fail if a PID != 1 is requested
+			 * and no PID 1 exists.
+			 */
+			nr = -EINVAL;
+			if (set_tid == 1 || !idr_is_empty(&tmp->idr))

On the second thought, I think we should check ns->child_reaper != NULL rather than !idr_is_empty(), this looks more robust and clean.

And this way alloc_pid() can do everything lockless at the start,

	if (set_tid) {
		if (set_tid < 0 || set_tid >= pid_max)
			return ERR_PTR(-EINVAL);
		if (set_tid != 1 && !ns->child_reaper)
			return ERR_PTR(-EINVAL);
		if (!ns_capable(ns->user_ns, CAP_SYS_ADMIN))
			return ERR_PTR(-EPERM);
	}

and just for record... this is off-topic and I need to recheck, but today "ns->pid_allocated = 0" in free_pid() doesn't look right to me... This logic predates unshare(CLONE_PIDNS).

Oleg.

#5 lore raw
Quoted 13 lines · Oleg Nesterov
Quoted 7 lines · Adrian Reber
 include/linux/pid.h        |  2 +-
 include/linux/sched/task.h |  1 +
 include/uapi/linux/sched.h |  1 +
 kernel/fork.c              | 22 ++++++++++++++++++++--
 kernel/pid.c               | 36 +++++++++++++++++++++++++++++-------
 5 files changed, 52 insertions(+), 10 deletions(-)

Looks good to me...

A couple of nits below, but I won't insist, feel free to ignore.

Quoted 7 lines

kernel/fork.c line 123

Quoted 4 lines
kernel/fork.c +4 -0
+/*
+ * Different sizes of struct clone_args
+ */
+#define CLONE3_ARGS_SIZE_V0 64

I don't really understand why do we want the "size < CLONE3_ARGS_SIZE_V0" check in copy_clone_args_from_user(), but I won't argue.

To make sure a user can't give us a garbage sized struct that is smaller than the initial version of the struct. Hm, maybe you did make a suggestion how to detect this case that I missed in one of the previous reviews or why it's not needed?

#6 lore raw
Quoted 22 lines · Christian Brauner
Quoted 13 lines · Oleg Nesterov
Quoted 7 lines · Adrian Reber
 include/linux/pid.h        |  2 +-
 include/linux/sched/task.h |  1 +
 include/uapi/linux/sched.h |  1 +
 kernel/fork.c              | 22 ++++++++++++++++++++--
 kernel/pid.c               | 36 +++++++++++++++++++++++++++++-------
 5 files changed, 52 insertions(+), 10 deletions(-)

Looks good to me...

A couple of nits below, but I won't insist, feel free to ignore.

Quoted 7 lines

kernel/fork.c line 123

Quoted 4 lines
kernel/fork.c +4 -0
+/*
+ * Different sizes of struct clone_args
+ */
+#define CLONE3_ARGS_SIZE_V0 64

I don't really understand why do we want the "size < CLONE3_ARGS_SIZE_V0" check in copy_clone_args_from_user(), but I won't argue.

Quoted 2 lines

To make sure a user can't give us a garbage sized struct that is smaller than the initial version of the struct.

But why do we want to detect this case?

And why CLONE3_ARGS_SIZE_V0 is special?

Oleg.

#7 lore raw
Quoted 23 lines · Oleg Nesterov
Quoted 22 lines · Christian Brauner
Quoted 20 lines · Oleg Nesterov
Quoted 7 lines · Adrian Reber
 include/linux/pid.h        |  2 +-
 include/linux/sched/task.h |  1 +
 include/uapi/linux/sched.h |  1 +
 kernel/fork.c              | 22 ++++++++++++++++++++--
 kernel/pid.c               | 36 +++++++++++++++++++++++++++++-------
 5 files changed, 52 insertions(+), 10 deletions(-)

Looks good to me...

A couple of nits below, but I won't insist, feel free to ignore.

kernel/fork.c line 123

Quoted 4 lines
kernel/fork.c +4 -0
+/*
+ * Different sizes of struct clone_args
+ */
+#define CLONE3_ARGS_SIZE_V0 64

I don't really understand why do we want the "size < CLONE3_ARGS_SIZE_V0" check in copy_clone_args_from_user(), but I won't argue.

Quoted 4 lines
Quoted 2 lines

To make sure a user can't give us a garbage sized struct that is smaller than the initial version of the struct.

But why do we want to detect this case?

I have to admit I don't understand that question. Because it's a garbage sized struct that we can't do anything with. Why shouldn't we detect this case?

Quoted 2 lines

And why CLONE3_ARGS_SIZE_V0 is special?

Hm? Because that is the first-well known struct size. In fact this pattern is also used for perf and sched:

	if (!access_ok(uattr, PERF_ATTR_SIZE_VER0))
		return -EFAULT;