Re: [RESEND RFC] translate_pid API

From: Nagarathnam Muthusamy
Date: Wed Mar 14 2018 - 17:28:53 EST




On 03/13/2018 08:29 PM, ebiederm@xxxxxxxxxxxx wrote:
nagarathnam.muthusamy@xxxxxxxxxx writes:

Resending the RFC with participants of previous discussions
in the list.
Nacked-by: "Eric W. Biederman" <ebiederm@xxxxxxxxxxxx>

Please read below to see why.

Following patch which is a variation of a solution discussed
in https://lwn.net/Articles/736330/ provides the users of
pid namespace, the functionality of pid translation between
namespaces using a namespace identifier. The topic of
pid translation has been discussed in the community few times
but there has always been a resistance to adding new solution
for this problem.
So far there has been no clear cut case that is worth the maintenance
challenge. As I recall the conversations they have always resulted
in a better way discovered.

I will outline the planned usecase of pid namespace by oracle
database and explain why any of the existing solution cannot
be used to solve their problem.

Consider a system in which several PID namespaces with multiple
nested levels exists in parallel with monitor processes managing
all the namespaces. PID translation is required for controlling
and accessing information about the processes by the monitors
and other processes down the hierarchy of namespaces. Controlling
primarily involves sending signals or using ptrace by a process in
parent namespace on any of the processes in its child namespace.
Accessing information deals with the reading /proc/<pid>/* files
of processes in child namespace. None of the processes have
root/CAP_SYS_ADMIN privileges.
So one process inside of a pid namespace one process outside of
the pid namespace and they are talking about some other process
inside the pid namespace.

Agreed that is a usecase that likely needs pid translation.

Existing methods:

*) SCM_CREDENTIALS message:
Ref: http://man7.org/linux/man-pages/man7/unix.7.html
The sender can translate its pid from its own namespace
to a pid in the target namespace by sending and receiving the
SCM_CREDENTIALS message. The drawback of this method is the
requirement of a socket communication channel to pid translation
which adds to the management overhead. This method does not enable
the sender to translate the pid of other process unless it is root
or it has CAP_SYS_ADMIN.
Or it has CAP_SYS_ADMIN (in the sending pid namespace).
That is fair.

*) /proc/<pid>/status file
Ref: http://man7.org/linux/man-pages/man5/proc.5.html
Ref: https://patchwork.kernel.org/patch/5861791/
/proc/<pid>/status file provides a way to find the pids
associated with a process in different namespaces. Pid translation
from child namespace to parent namespace from parent namespace would
require searching all the status file in the parent namespace to find
the desired pid at desired level which is inefficient.
That is fair.
Method under review:

*) pidns: introduce syscall translate_pid
Ref: https://lwn.net/Articles/736330/
This solution though not integrated yet, when used would require
monitor to have open file descriptor for every level of namespace it
has to monitor which results in explosion of number of file descriptors
the monitor has to keep open.
You know what tough nuggies.

File descriptors aren't or at least shouldn't be that expensive. And
you are already committing to tracking the pid namespaces anyway. An
int for a file descriptor is cheaper than a u64 for the userspace
process.

The size of individual file descriptor was not the issue but the
number of file descriptors to track for each level in a nested
PID namespace is the issue.

The cost of that ``cheaper'' u64 that is not in any namespace is that
you now have to go and implement a namespace of namespaces. You haven't
even attempted it. So just no. Anything that brings us to needing
a namespace of namespaces is a bad design.

I am not trying to implement a namespace of namespaces. The ID
exported through /proc/<pid>/ns/pidns_id is associated with the
pid namespace of the process. The function of this ID is very similar to
that of the inode number exported by /proc/<pid>/ns/pid file (ie) uniquely
identify a pid namespace in the system.

Inode numbers can be recycled very fast and hence I proposed a
separate ID. Instead of exporting inode number through /proc/<pid>/ns/pid
file we could directly export this new ID but I instead moved it to
a separate file so that any existing application which depends on this
number being 32-bit wont break.

Following patch uses a 64-bit ID for namespace exported by procfs
for pid translation through a new file /proc/<pid>/ns/pidns_id.
And this design detail is what brings the automatic nack.

Use file descriptros and it sounds like your use case justifies what you
are trying to do.

File descriptors are problematic for following reasons.
1) I need to open a couple of file descriptors for every pid translation request.
2) In case of nested PID namespaces, say a new pid namespace is created at level 20,
ÂÂÂ with unique ID, I could just record this ID in a shared memory for interested process
ÂÂÂ to use. In case of file descriptors, every level has to figure out the process ID of the
ÂÂÂ newly created namespace's init process and open a file descriptor to track it.

Thanks,
Nagarathnam.

[root@brm-x4600-01 ~]# ls -l /proc/4690/ns
total 0
lrwxrwxrwx 1 root root 0 Mar 1 15:55 cgroup -> cgroup:[4026531835]
lrwxrwxrwx 1 root root 0 Mar 1 15:55 ipc -> ipc:[4026531839]
lrwxrwxrwx 1 root root 0 Mar 1 15:55 mnt -> mnt:[4026532324]
lrwxrwxrwx 1 root root 0 Mar 1 15:55 net -> net:[4026531993]
lrwxrwxrwx 1 root root 0 Mar 1 15:55 pid -> pid:[4026532325]
lrwxrwxrwx 1 root root 0 Mar 1 15:55 pid_for_children -> pid:[4026532325]
lrwxrwxrwx 1 root root 0 Mar 1 15:55 pidns_id -> [14994344851325148137]
lrwxrwxrwx 1 root root 0 Mar 1 15:55 user -> user:[4026531837]
lrwxrwxrwx 1 root root 0 Mar 1 15:55 uts -> uts:[4026531838]

Though there is a problem of ID being recycled in longterm, managing
an ID per namespace is easier than having open file descriptors per
pid namespace. Existing namespace inode numbers recycles faster and hence
their usability as ID for this API is less.
No.

A namespace of namespaces is a much harder problem. We aren't going
there.

It sounds like you may be struggling with lifetime issues and when
to close/get rid of your pid namespace identifier. That could deserve
a conversation. Idenitifiers that are not namespaced are absolutely not
the solution.

Eric


Signed-off-by: Nagarathnam Muthusamy <nagarathnam.muthusamy@xxxxxxxxxx>
---
arch/x86/entry/syscalls/syscall_32.tbl | 1 +
arch/x86/entry/syscalls/syscall_64.tbl | 2 +
fs/nsfs.c | 9 +-
fs/proc/namespaces.c | 1 +
include/linux/ns_common.h | 1 +
include/linux/pid_namespace.h | 3 +
include/linux/proc_ns.h | 1 +
include/linux/syscalls.h | 1 +
kernel/pid_namespace.c | 190 ++++++++++++++++++++++++++++++++-
kernel/sys_ni.c | 4 +
10 files changed, 208 insertions(+), 5 deletions(-)

diff --git a/arch/x86/entry/syscalls/syscall_32.tbl b/arch/x86/entry/syscalls/syscall_32.tbl
index 448ac21..31bf798 100644
--- a/arch/x86/entry/syscalls/syscall_32.tbl
+++ b/arch/x86/entry/syscalls/syscall_32.tbl
@@ -391,3 +391,4 @@
382 i386 pkey_free sys_pkey_free
383 i386 statx sys_statx
384 i386 arch_prctl sys_arch_prctl compat_sys_arch_prctl
+385 i386 translate_pid sys_translate_pid compat_sys_translate_pid
diff --git a/arch/x86/entry/syscalls/syscall_64.tbl b/arch/x86/entry/syscalls/syscall_64.tbl
index 5aef183..89196c3 100644
--- a/arch/x86/entry/syscalls/syscall_64.tbl
+++ b/arch/x86/entry/syscalls/syscall_64.tbl
@@ -339,6 +339,7 @@
330 common pkey_alloc sys_pkey_alloc
331 common pkey_free sys_pkey_free
332 common statx sys_statx
+333 64 translate_pid sys_translate_pid
#
# x32-specific system call numbers start at 512 to avoid cache impact
@@ -380,3 +381,4 @@
545 x32 execveat compat_sys_execveat/ptregs
546 x32 preadv2 compat_sys_preadv64v2
547 x32 pwritev2 compat_sys_pwritev64v2
+548 x32 translate_pid compat_sys_translate_pid
diff --git a/fs/nsfs.c b/fs/nsfs.c
index 36b0772..c635465 100644
--- a/fs/nsfs.c
+++ b/fs/nsfs.c
@@ -222,8 +222,13 @@ int ns_get_name(char *buf, size_t size, struct task_struct *task,
const char *name;
ns = ns_ops->get(task);
if (ns) {
- name = ns_ops->real_ns_name ? : ns_ops->name;
- res = snprintf(buf, size, "%s:[%u]", name, ns->inum);
+ if (!strcmp(ns_ops->name, "pidns_id")) {
+ res = snprintf(buf, size, "[%llu]",
+ (unsigned long long)ns->ns_id);
+ } else {
+ name = ns_ops->real_ns_name ? : ns_ops->name;
+ res = snprintf(buf, size, "%s:[%u]", name, ns->inum);
+ }
ns_ops->put(ns);
}
return res;
diff --git a/fs/proc/namespaces.c b/fs/proc/namespaces.c
index 59b17e5..ac823ce 100644
--- a/fs/proc/namespaces.c
+++ b/fs/proc/namespaces.c
@@ -24,6 +24,7 @@
#endif
#ifdef CONFIG_PID_NS
&pidns_operations,
+ &pidns_id_operations,
&pidns_for_children_operations,
#endif
#ifdef CONFIG_USER_NS
diff --git a/include/linux/ns_common.h b/include/linux/ns_common.h
index 5fbc400..6ca3d43 100644
--- a/include/linux/ns_common.h
+++ b/include/linux/ns_common.h
@@ -8,6 +8,7 @@ struct ns_common {
atomic_long_t stashed;
const struct proc_ns_operations *ops;
unsigned int inum;
+ u64 ns_id;
};
#endif
diff --git a/include/linux/pid_namespace.h b/include/linux/pid_namespace.h
index 49538b1..11d1d57 100644
--- a/include/linux/pid_namespace.h
+++ b/include/linux/pid_namespace.h
@@ -11,6 +11,7 @@
#include <linux/kref.h>
#include <linux/ns_common.h>
#include <linux/idr.h>
+#include <linux/list_bl.h>
struct fs_pin;
@@ -44,6 +45,8 @@ struct pid_namespace {
kgid_t pid_gid;
int hide_pid;
int reboot; /* group exit code if this pidns was rebooted */
+ struct hlist_bl_node node;
+ atomic_t lookups_pending;
struct ns_common ns;
} __randomize_layout;
diff --git a/include/linux/proc_ns.h b/include/linux/proc_ns.h
index d31cb62..861e38bd 100644
--- a/include/linux/proc_ns.h
+++ b/include/linux/proc_ns.h
@@ -28,6 +28,7 @@ struct proc_ns_operations {
extern const struct proc_ns_operations utsns_operations;
extern const struct proc_ns_operations ipcns_operations;
extern const struct proc_ns_operations pidns_operations;
+extern const struct proc_ns_operations pidns_id_operations;
extern const struct proc_ns_operations pidns_for_children_operations;
extern const struct proc_ns_operations userns_operations;
extern const struct proc_ns_operations mntns_operations;
diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h
index a78186d..574349a 100644
--- a/include/linux/syscalls.h
+++ b/include/linux/syscalls.h
@@ -901,6 +901,7 @@ asmlinkage long sys_open_by_handle_at(int mountdirfd,
struct file_handle __user *handle,
int flags);
asmlinkage long sys_setns(int fd, int nstype);
+asmlinkage long sys_translate_pid(pid_t pid, u64 source, u64 target);
asmlinkage long sys_process_vm_readv(pid_t pid,
const struct iovec __user *lvec,
unsigned long liovcnt,
diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c
index 0b53eef..ff83aa8 100644
--- a/kernel/pid_namespace.c
+++ b/kernel/pid_namespace.c
@@ -22,6 +22,12 @@
#include <linux/sched/task.h>
#include <linux/sched/signal.h>
#include <linux/idr.h>
+#include <linux/random.h>
+#include <linux/compat.h>
+
+#define PID_NS_ID_HASH_BITS 9
+
+struct hlist_bl_head *pid_ns_hash;
struct pid_cache {
int nr_ids;
@@ -34,6 +40,13 @@ struct pid_cache {
static DEFINE_MUTEX(pid_caches_mutex);
static struct kmem_cache *pid_ns_cachep;
+static inline struct hlist_bl_head *
+ pid_ns_hash_head(struct hlist_bl_head *hash,
+ uint64_t key)
+{
+ return &hash[hash_64(key, PID_NS_ID_HASH_BITS)];
+}
+
/*
* creates the kmem cache to allocate pids from.
* @nr_ids: the number of numerical ids this pid will have to carry
@@ -93,12 +106,24 @@ static void dec_pid_namespaces(struct ucounts *ucounts)
dec_ucount(ucounts, UCOUNT_PID_NAMESPACES);
}
+static inline u64 get_namespace_id(void)
+{
+ u64 id = 0;
+
+ while (!id)
+ get_random_bytes(&id, sizeof(id));
+
+ return id;
+}
+
static struct pid_namespace *create_pid_namespace(struct user_namespace *user_ns,
struct pid_namespace *parent_pid_ns)
{
- struct pid_namespace *ns;
- unsigned int level = parent_pid_ns->level + 1;
+ struct pid_namespace *ns, *pt;
struct ucounts *ucounts;
+ struct hlist_bl_head *head;
+ struct hlist_bl_node *dup_node;
+ unsigned int level = parent_pid_ns->level + 1;
int err;
err = -EINVAL;
@@ -135,7 +160,24 @@ static struct pid_namespace *create_pid_namespace(struct user_namespace *user_ns
ns->ucounts = ucounts;
ns->pid_allocated = PIDNS_ADDING;
INIT_WORK(&ns->proc_work, proc_cleanup_work);
-
+ ns->ns.ns_id = get_namespace_id();
+ while (1) {
+ head = pid_ns_hash_head(pid_ns_hash, ns->ns.ns_id);
+ hlist_bl_lock(head);
+ hlist_bl_for_each_entry(pt, dup_node, head, node) {
+ if (ns->ns.ns_id == pt->ns.ns_id) {
+ /*
+ * ID is taken. Move to next ID;
+ */
+ ns->ns.ns_id = get_namespace_id();
+ hlist_bl_unlock(head);
+ continue;
+ }
+ }
+ break;
+ }
+ hlist_bl_add_head(&ns->node, head);
+ hlist_bl_unlock(head);
return ns;
out_free_idr:
@@ -159,6 +201,30 @@ static void delayed_free_pidns(struct rcu_head *p)
static void destroy_pid_namespace(struct pid_namespace *ns)
{
+ struct pid_namespace *ph;
+ struct hlist_bl_head *head;
+ struct hlist_bl_node *dup_node;
+
+ /*
+ * Remove the namespace structure from hash table so
+ * now new lookups can start on it.
+ */
+ if (ns->ns.ns_id) {
+ head = pid_ns_hash_head(pid_ns_hash, ns->ns.ns_id);
+ hlist_bl_lock(head);
+ hlist_bl_for_each_entry(ph, dup_node, head, node) {
+ if (ns->ns.ns_id == ph->ns.ns_id) {
+ hlist_bl_del_init(&ph->node);
+ break;
+ }
+ }
+ hlist_bl_unlock(head);
+ }
+ /*
+ * Wait for pending lookups to complete.
+ */
+ while (atomic_read(&ns->lookups_pending))
+ cpu_relax();
ns_free_inum(&ns->ns);
idr_destroy(&ns->idr);
@@ -463,6 +529,17 @@ static struct user_namespace *pidns_owner(struct ns_common *ns)
.get_parent = pidns_get_parent,
};
+const struct proc_ns_operations pidns_id_operations = {
+ .name = "pidns_id",
+ .real_ns_name = "pid",
+ .type = CLONE_NEWPID,
+ .get = pidns_get,
+ .put = pidns_put,
+ .install = pidns_install,
+ .owner = pidns_owner,
+ .get_parent = pidns_get_parent,
+};
+
const struct proc_ns_operations pidns_for_children_operations = {
.name = "pid_for_children",
.real_ns_name = "pid",
@@ -474,9 +551,116 @@ static struct user_namespace *pidns_owner(struct ns_common *ns)
.get_parent = pidns_get_parent,
};
+/*
+ * translate_pid - convert pid in source pid-ns into target pid-ns.
+ * @pid: pid for translation
+ * @source: pid-ns id
+ * @target: pid-ns id
+ *
+ * Return pid in @target pid-ns, zero if task have no pid there,
+ * or -ESRCH of task with @pid is not found in @source pid-ns.
+ */
+SYSCALL_DEFINE3(translate_pid, pid_t, pid, u64, source,
+ u64, target)
+{
+ struct pid_namespace *source_ns = NULL, *target_ns = NULL;
+ struct pid *struct_pid;
+ struct pid_namespace *ph;
+ struct hlist_bl_head *shead = NULL;
+ struct hlist_bl_head *thead = NULL;
+ struct hlist_bl_node *dup_node;
+ pid_t result;
+
+ if (!source) {
+ source_ns = &init_pid_ns;
+ } else {
+ shead = pid_ns_hash_head(pid_ns_hash, source);
+ hlist_bl_lock(shead);
+ hlist_bl_for_each_entry(ph, dup_node, shead, node) {
+ if (source == ph->ns.ns_id) {
+ source_ns = ph;
+ break;
+ }
+ }
+ if (!source_ns) {
+ hlist_bl_unlock(shead);
+ return -EINVAL;
+ }
+ }
+ if (!ptrace_may_access(source_ns->child_reaper,
+ PTRACE_MODE_READ_FSCREDS)) {
+ if (shead)
+ hlist_bl_unlock(shead);
+ return -EPERM;
+ }
+
+ atomic_inc(&source_ns->lookups_pending);
+ if (shead)
+ hlist_bl_unlock(shead);
+
+ if (!target) {
+ target_ns = &init_pid_ns;
+ } else {
+ thead = pid_ns_hash_head(pid_ns_hash, target);
+ hlist_bl_lock(thead);
+ hlist_bl_for_each_entry(ph, dup_node, thead, node) {
+ if (target == ph->ns.ns_id) {
+ target_ns = ph;
+ break;
+ }
+ }
+ if (!target_ns) {
+ atomic_dec(&source_ns->lookups_pending);
+ hlist_bl_unlock(thead);
+ return -EINVAL;
+ }
+ }
+ if (!ptrace_may_access(target_ns->child_reaper,
+ PTRACE_MODE_READ_FSCREDS)) {
+ atomic_dec(&source_ns->lookups_pending);
+ if (thead)
+ hlist_bl_unlock(thead);
+ return -EPERM;
+ }
+ atomic_inc(&target_ns->lookups_pending);
+ if (thead)
+ hlist_bl_unlock(thead);
+
+ rcu_read_lock();
+ struct_pid = find_pid_ns(pid, source_ns);
+ result = struct_pid ? pid_nr_ns(struct_pid, target_ns) : -ESRCH;
+ rcu_read_unlock();
+ atomic_dec(&source_ns->lookups_pending);
+ atomic_dec(&target_ns->lookups_pending);
+ return result;
+}
+
+#ifdef CONFIG_COMPAT
+COMPAT_SYSCALL_DEFINE5(translate_pid, pid_t, pid, u32, s0, u32, s1,
+ u32, t0, u32, t1)
+{
+#ifdef __BIG_ENDIAN
+ return sys_translate_pid(pid, ((u64)s0 << 32) | s1,
+ ((u64)t0 << 32) | t1);
+#else
+ return sys_translate_pid(pid, ((u64)s1 << 32) | s0,
+ ((u64)t1 << 32) | t0);
+#endif
+}
+#endif
+
static __init int pid_namespaces_init(void)
{
+ unsigned long bucket_count;
+ int i;
+
+ bucket_count = (1UL << PID_NS_ID_HASH_BITS);
pid_ns_cachep = KMEM_CACHE(pid_namespace, SLAB_PANIC);
+ pid_ns_hash = kmalloc_array(bucket_count, sizeof(struct hlist_bl_head),
+ GFP_KERNEL);
+
+ for (i = 0; i < bucket_count; i++)
+ INIT_HLIST_BL_HEAD(&pid_ns_hash[i]);
#ifdef CONFIG_CHECKPOINT_RESTORE
register_sysctl_paths(kern_path, pid_ns_ctl_table);
diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c
index b518976..467255f 100644
--- a/kernel/sys_ni.c
+++ b/kernel/sys_ni.c
@@ -259,3 +259,7 @@ asmlinkage long sys_ni_syscall(void)
cond_syscall(sys_pkey_mprotect);
cond_syscall(sys_pkey_alloc);
cond_syscall(sys_pkey_free);
+
+/* pid translation */
+cond_syscall(sys_translate_pid);
+cond_syscall(compat_sys_translate_pid);