diff mbox series

[v7,07/11] proc: flush task dcache entries from all procfs instances

Message ID 20200125130541.450409-8-gladkov.alexey@gmail.com (mailing list archive)
State New, archived
Headers show
Series proc: modernize proc to support multiple private instances | expand

Commit Message

Alexey Gladkov Jan. 25, 2020, 1:05 p.m. UTC
This allows to flush dcache entries of a task on multiple procfs mounts
per pid namespace.

Cc: Kees Cook <keescook@chromium.org>
Cc: Andy Lutomirski <luto@kernel.org>
Signed-off-by: Djalal Harouni <tixxdz@gmail.com>
Signed-off-by: Alexey Gladkov <gladkov.alexey@gmail.com>
---
 fs/proc/base.c                | 23 ++++++++++++++-----
 fs/proc/root.c                | 14 ++++++++++++
 include/linux/pid_namespace.h | 42 +++++++++++++++++++++++++++++++++++
 include/linux/proc_fs.h       |  2 ++
 4 files changed, 76 insertions(+), 5 deletions(-)

Comments

Linus Torvalds Jan. 25, 2020, 6:45 p.m. UTC | #1
On Sat, Jan 25, 2020 at 5:06 AM Alexey Gladkov <gladkov.alexey@gmail.com> wrote:
>
> This allows to flush dcache entries of a task on multiple procfs mounts
> per pid namespace.

From a quick read-through, this is the only one I really react negatively to.

The locking looks odd. It only seems to protect the new proc_mounts
list, but then it's a whole big rwsem, and it's taken over all of
proc_flush_task_mnt(), and the locking is exported to all over as a
result of that - including the dummy functions for "there is no proc"
case.

And proc_flush_task_mnt() itself should need no locking over any of
it, so it's all just for the silly looping over the list.

So

 (a) this looks fishy and feels wrong - I get a very strong feeling
that the locking is wrong to begin with, and could/should have been
done differently

 (b) all the locking should have been internal to /proc, and those
wrappers shouldn't exist in a common header file (and certainly not
for the non-proc case).

Yes, (a) is just a feeling, and I don't have any great suggestions.
Maybe make it an RCU list and use a spinlock for updating it?

But (b) is pretty much a non-starter in this form. Those wrappers
shouldn't be in a globally exported core header file. No way.

               Linus
Alexey Gladkov Jan. 27, 2020, 11:23 p.m. UTC | #2
On Sat, Jan 25, 2020 at 10:45:25AM -0800, Linus Torvalds wrote:
> On Sat, Jan 25, 2020 at 5:06 AM Alexey Gladkov <gladkov.alexey@gmail.com> wrote:
> >
> > This allows to flush dcache entries of a task on multiple procfs mounts
> > per pid namespace.
> 
> From a quick read-through, this is the only one I really react negatively to.
> 
> The locking looks odd. It only seems to protect the new proc_mounts
> list, but then it's a whole big rwsem, and it's taken over all of
> proc_flush_task_mnt(), and the locking is exported to all over as a
> result of that - including the dummy functions for "there is no proc"
> case.
> 
> And proc_flush_task_mnt() itself should need no locking over any of
> it, so it's all just for the silly looping over the list.

Thank you, I will rework this part.

> So
> 
>  (a) this looks fishy and feels wrong - I get a very strong feeling
> that the locking is wrong to begin with, and could/should have been
> done differently
> 
>  (b) all the locking should have been internal to /proc, and those
> wrappers shouldn't exist in a common header file (and certainly not
> for the non-proc case).
> 
> Yes, (a) is just a feeling, and I don't have any great suggestions.
> Maybe make it an RCU list and use a spinlock for updating it?

I’m thinking, is it possible to get rid of proc_flush_task at all ?
Maybe we can try to flush dcache during readdir for example.

> But (b) is pretty much a non-starter in this form. Those wrappers
> shouldn't be in a globally exported core header file. No way.
> 
>                Linus
>
diff mbox series

Patch

diff --git a/fs/proc/base.c b/fs/proc/base.c
index 4ccb280a3e79..f4f1bcb28603 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -3133,7 +3133,7 @@  static const struct inode_operations proc_tgid_base_inode_operations = {
 	.permission	= proc_pid_permission,
 };
 
-static void proc_flush_task_mnt(struct vfsmount *mnt, pid_t pid, pid_t tgid)
+static void proc_flush_task_mnt_root(struct dentry *mnt_root, pid_t pid, pid_t tgid)
 {
 	struct dentry *dentry, *leader, *dir;
 	char buf[10 + 1];
@@ -3142,7 +3142,7 @@  static void proc_flush_task_mnt(struct vfsmount *mnt, pid_t pid, pid_t tgid)
 	name.name = buf;
 	name.len = snprintf(buf, sizeof(buf), "%u", pid);
 	/* no ->d_hash() rejects on procfs */
-	dentry = d_hash_and_lookup(mnt->mnt_root, &name);
+	dentry = d_hash_and_lookup(mnt_root, &name);
 	if (dentry) {
 		d_invalidate(dentry);
 		dput(dentry);
@@ -3153,7 +3153,7 @@  static void proc_flush_task_mnt(struct vfsmount *mnt, pid_t pid, pid_t tgid)
 
 	name.name = buf;
 	name.len = snprintf(buf, sizeof(buf), "%u", tgid);
-	leader = d_hash_and_lookup(mnt->mnt_root, &name);
+	leader = d_hash_and_lookup(mnt_root, &name);
 	if (!leader)
 		goto out;
 
@@ -3208,14 +3208,27 @@  void proc_flush_task(struct task_struct *task)
 	int i;
 	struct pid *pid, *tgid;
 	struct upid *upid;
+	struct pid_namespace *pid_ns;
+	struct dentry *mnt_root;
+	struct proc_fs_info *fs_info;
 
 	pid = task_pid(task);
 	tgid = task_tgid(task);
 
 	for (i = 0; i <= pid->level; i++) {
 		upid = &pid->numbers[i];
-		proc_flush_task_mnt(upid->ns->proc_mnt, upid->nr,
-					tgid->numbers[i].nr);
+
+		pid_ns = upid->ns;
+
+		pidns_proc_lock_shared(pid_ns);
+		list_for_each_entry(fs_info, &pid_ns->proc_mounts, pidns_entry) {
+			mnt_root = fs_info->m_super->s_root;
+			proc_flush_task_mnt_root(mnt_root, upid->nr, tgid->numbers[i].nr);
+		}
+		pidns_proc_unlock_shared(pid_ns);
+
+		mnt_root = pid_ns->proc_mnt->mnt_root;
+		proc_flush_task_mnt_root(mnt_root, upid->nr, tgid->numbers[i].nr);
 	}
 }
 
diff --git a/fs/proc/root.c b/fs/proc/root.c
index 5d5cba4c899b..3bb8df360cf7 100644
--- a/fs/proc/root.c
+++ b/fs/proc/root.c
@@ -112,6 +112,12 @@  static int proc_fill_super(struct super_block *s, struct fs_context *fc)
 
 	proc_apply_options(ctx->fs_info, fc, pid_ns, current_user_ns());
 
+	ctx->fs_info->m_super = s;
+
+	pidns_proc_lock(pid_ns);
+	list_add_tail(&ctx->fs_info->pidns_entry, &pid_ns->proc_mounts);
+	pidns_proc_unlock(pid_ns);
+
 	/* User space would break if executables or devices appear on proc */
 	s->s_iflags |= SB_I_USERNS_VISIBLE | SB_I_NOEXEC | SB_I_NODEV;
 	s->s_flags |= SB_NODIRATIME | SB_NOSUID | SB_NOEXEC;
@@ -215,6 +221,11 @@  static void proc_kill_sb(struct super_block *sb)
 		dput(fs_info->proc_self);
 	if (fs_info->proc_thread_self)
 		dput(fs_info->proc_thread_self);
+
+	pidns_proc_lock(fs_info->pid_ns);
+	list_del(&fs_info->pidns_entry);
+	pidns_proc_unlock(fs_info->pid_ns);
+
 	kill_anon_super(sb);
 	put_pid_ns(fs_info->pid_ns);
 	kfree(fs_info);
@@ -336,6 +347,9 @@  int pid_ns_prepare_proc(struct pid_namespace *ns)
 		ctx->fs_info->pid_ns = ns;
 	}
 
+	init_rwsem(&ns->rw_proc_mounts);
+	INIT_LIST_HEAD(&ns->proc_mounts);
+
 	mnt = fc_mount(fc);
 	put_fs_context(fc);
 	if (IS_ERR(mnt))
diff --git a/include/linux/pid_namespace.h b/include/linux/pid_namespace.h
index 66f47f1afe0d..297b39604312 100644
--- a/include/linux/pid_namespace.h
+++ b/include/linux/pid_namespace.h
@@ -26,6 +26,8 @@  struct pid_namespace {
 	struct pid_namespace *parent;
 #ifdef CONFIG_PROC_FS
 	struct vfsmount *proc_mnt; /* Internal proc mounted during each new pidns */
+	struct rw_semaphore rw_proc_mounts;
+	struct list_head proc_mounts; /* list of separated procfs mounts */
 #endif
 #ifdef CONFIG_BSD_PROCESS_ACCT
 	struct fs_pin *bacct;
@@ -90,4 +92,44 @@  extern struct pid_namespace *task_active_pid_ns(struct task_struct *tsk);
 void pidhash_init(void);
 void pid_idr_init(void);
 
+#ifdef CONFIG_PROC_FS
+static inline void pidns_proc_lock(struct pid_namespace *pid_ns)
+{
+	down_write(&pid_ns->rw_proc_mounts);
+}
+
+static inline void pidns_proc_unlock(struct pid_namespace *pid_ns)
+{
+	up_write(&pid_ns->rw_proc_mounts);
+}
+
+static inline void pidns_proc_lock_shared(struct pid_namespace *pid_ns)
+{
+	down_read(&pid_ns->rw_proc_mounts);
+}
+
+static inline void pidns_proc_unlock_shared(struct pid_namespace *pid_ns)
+{
+	up_read(&pid_ns->rw_proc_mounts);
+}
+#else /* !CONFIG_PROC_FS */
+
+static inline void pidns_proc_lock(struct pid_namespace *pid_ns)
+{
+}
+
+static inline void pidns_proc_unlock(struct pid_namespace *pid_ns)
+{
+}
+
+static inline void pidns_proc_lock_shared(struct pid_namespace *pid_ns)
+{
+}
+
+static inline void pidns_proc_unlock_shared(struct pid_namespace *pid_ns)
+{
+}
+
+#endif /* CONFIG_PROC_FS */
+
 #endif /* _LINUX_PID_NS_H */
diff --git a/include/linux/proc_fs.h b/include/linux/proc_fs.h
index 5f0b1b7e4271..f307940f8311 100644
--- a/include/linux/proc_fs.h
+++ b/include/linux/proc_fs.h
@@ -20,6 +20,8 @@  enum {
 };
 
 struct proc_fs_info {
+	struct list_head pidns_entry;    /* Node in procfs_mounts of a pidns */
+	struct super_block *m_super;
 	struct pid_namespace *pid_ns;
 	struct dentry *proc_self;        /* For /proc/self */
 	struct dentry *proc_thread_self; /* For /proc/thread-self */