fix a race in release_task when flushing the dentry
v2 cover fix a race in release_task when flushing the dentry
The dentries such as /proc/<pid>/ns/ have the DCACHE_OP_DELETE flag, they should be deleted when the process exits.
Suppose the following race appears:
release_task dput
-> proc_flush_task
-> dentry->d_op->d_delete(dentry)
-> __exit_signal
-> dentry->d_lockref.count-- and return. In the proc_flush_task(), if another process is using this dentry, it will not be deleted. At the same time, in dput(), d_op->d_delete() can be executed before __exit_signal(pid has not been hashed), d_delete returns false, so this dentry still cannot be deleted.
This dentry will always be cached (although its count is 0 and the DCACHE_OP_DELETE flag is set), its parent denry will also be cached too, and these dentries can only be deleted when drop_caches is manually triggered.
This will result in wasted memory. What's more troublesome is that these dentries reference pid, according to the commit f333c700c610 ("pidns: Add a limit on the number of pid namespaces"), if the pid cannot be released, it may result in the inability to create a new pid_ns.
This issue was introduced by 60347f6716aa ("pid namespaces: prepare proc_flust_task() to flush entries from multiple proc trees"), exposed by f333c700c610 ("pidns: Add a limit on the number of pid namespaces"), and then fixed by 7bc3e6e55acf ("proc: Use a list of inodes to flush from proc").
Alexey Dobriyan (1):
proc: use %u for pid printing and slightly less stack
Andreas Gruenbacher (1):
proc: Pass file mode to proc_pid_make_inode
Christian Brauner (1):
clone: add CLONE_PIDFD
Eric W. Biederman (6):
proc: Better ownership of files for non-dumpable tasks in user
namespaces
proc: Rename in proc_inode rename sysctl_inodes sibling_inodes
proc: Generalize proc_sys_prune_dcache into proc_prune_siblings_dcache
proc: Clear the pieces of proc_inode that proc_evict_inode cares about
proc: Use d_invalidate in proc_prune_siblings_dcache
proc: Use a list of inodes to flush from proc
Joel Fernandes (Google) (1):
pidfd: add polling support
fs/proc/base.c | 242 ++++++++++++++++++++------------------------- fs/proc/fd.c | 20 +--- fs/proc/inode.c | 67 ++++++++++++- fs/proc/internal.h | 22 ++--- fs/proc/namespaces.c | 3 +- fs/proc/proc_sysctl.c | 45 ++------- fs/proc/self.c | 6 +- fs/proc/thread_self.c | 5 +- include/linux/pid.h | 5 + include/linux/proc_fs.h | 4 +- include/uapi/linux/sched.h | 1 + kernel/exit.c | 5 +- kernel/fork.c | 131 +++++++++++++++++++++++- kernel/pid.c | 3 + kernel/signal.c | 11 +++ security/selinux/hooks.c | 1 + 16 files changed, 343 insertions(+), 228 deletions(-)
Signature · 1 line
1.8.3.1
PATCH v2 1/10 clone: add CLONE_PIDFD
From: Christian Brauner <[email protected]>
[ Upstream commit b3e5838252665ee4cfa76b82bdf1198dca81e5be ]
This patchset makes it possible to retrieve pid file descriptors at process creation time by introducing the new flag CLONE_PIDFD to the clone() system call. Linus originally suggested to implement this as a new flag to clone() instead of making it a separate system call. As spotted by Linus, there is exactly one bit for clone() left.
CLONE_PIDFD creates file descriptors based on the anonymous inode implementation in the kernel that will also be used to implement the new mount api. They serve as a simple opaque handle on pids. Logically, this makes it possible to interpret a pidfd differently, narrowing or widening the scope of various operations (e.g. signal sending). Thus, a pidfd cannot just refer to a tgid, but also a tid, or in theory - given appropriate flag arguments in relevant syscalls - a process group or session. A pidfd does not represent a privilege. This does not imply it cannot ever be that way but for now this is not the case.
A pidfd comes with additional information in fdinfo if the kernel supports procfs. The fdinfo file contains the pid of the process in the callers pid namespace in the same format as the procfs status file, i.e. "Pid:\t%d".
As suggested by Oleg, with CLONE_PIDFD the pidfd is returned in the parent_tidptr argument of clone. This has the advantage that we can give back the associated pid and the pidfd at the same time.
To remove worries about missing metadata access this patchset comes with a sample program that illustrates how a combination of CLONE_PIDFD, and pidfd_send_signal() can be used to gain race-free access to process metadata through /proc/<pid>. The sample program can easily be translated into a helper that would be suitable for inclusion in libc so that users don't have to worry about writing it themselves.
(clone: fix up cherry-pick conflicts for b3e583825266)
--- include/linux/pid.h | 1 + include/uapi/linux/sched.h | 1 + kernel/fork.c | 105 +++++++++++++++++++++++++++++++++++++++++++-- 3 files changed, 103 insertions(+), 4 deletions(-)
include/linux/pid.h +1 -0
@@ -73,6 +73,7 @@ struct pid_linkstruct hlist_node node;struct pid *pid; }; +extern const struct file_operations pidfd_fops;static inline struct pid *get_pid (struct pid *pid) {
include/uapi/linux/sched.h +1 -0
@@ -9,6 +9,7 @@#define CLONE_FS 0x00000200/* set if fs info shared between processes */ #define CLONE_FILES 0x00000400/* set if open files shared between processes */ #define CLONE_SIGHAND 0x00000800/* set if signal handlers and blocked signals shared */ +#define CLONE_PIDFD 0x00001000/* set if a pidfd should be placed in parent */ #define CLONE_PTRACE 0x00002000/* set if we want to let tracing continue on the child too */ #define CLONE_VFORK 0x00004000/* set if the parent wants the child to wake it up on mm_release */ #define CLONE_PARENT 0x00008000/* set if we want to have the same parent as the cloner */
kernel/fork.c +101 -4
165 lines not on this page · raw message
Signature · 1 line
1.8.3.1
PATCH v2 2/10 pidfd: add polling support
In reply to Wen YangFrom: "Joel Fernandes (Google)" <[email protected]>
[ Upstream commit b53b0b9d9a613c418057f6cb921c2f40a6f78c24 ]
This patch adds polling support to pidfd.
Android low memory killer (LMK) needs to know when a process dies once it is sent the kill signal. It does so by checking for the existence of /proc/pid which is both racy and slow. For example, if a PID is reused between when LMK sends a kill signal and checks for existence of the PID, since the wrong PID is now possibly checked for existence. Using the polling support, LMK will be able to get notified when a process exists in race-free and fast way, and allows the LMK to do other things (such as by polling on other fds) while awaiting the process being killed to die.
For notification to polling processes, we follow the same existing mechanism in the kernel used when the parent of the task group is to be notified of a child's death (do_notify_parent). This is precisely when the tasks waiting on a poll of pidfd are also awakened in this patch.
We have decided to include the waitqueue in struct pid for the following reasons:
1. The wait queue has to survive for the lifetime of the poll. Including it in task_struct would not be option in this case because the task can be reaped and destroyed before the poll returns.
2. By including the struct pid for the waitqueue means that during de_thread(), the new thread group leader automatically gets the new waitqueue/pid even though its task_struct is different.
Appropriate test cases are added in the second patch to provide coverage of all the cases the patch is handling.
(pidfd: fix up cherry-pick conflicts for b53b0b9d9a61)
--- include/linux/pid.h | 3 +++ kernel/fork.c | 26 ++++++++++++++++++++++++++ kernel/pid.c | 2 ++ kernel/signal.c | 11 +++++++++++ 4 files changed, 42 insertions(+)
include/linux/pid.h +3 -0
@@ -2,6 +2,7 @@#define _LINUX_PID_H#include <linux/rcupdate.h> +#include <linux/wait.h> enum pid_type {
@@ -62,6 +63,8 @@ struct pidunsigned int level;/* lists of tasks that use this pid */ struct hlist_head tasks[PIDTYPE_MAX ]; +/* wait queue for pidfd notifications */ +wait_queue_head_t wait_pidfd;struct rcu_head rcu;struct upid numbers[1 ]; };
kernel/fork.c +26 -0
@@ -1481,8 +1481,34 @@ static void pidfd_show_fdinfo(struct seq_file *m, struct file *f) }#endif +/* +* Poll support for process exit notification. +*/ +static unsigned int pidfd_poll (struct file *file,struct poll_table_struct *pts) +{ +struct task_struct *task; +struct pid *pid = file->private_data; +int poll_flags =0 ; + +poll_wait (file, &pid->wait_pidfd, pts); + +rcu_read_lock (); + task =pid_task (pid,PIDTYPE_PID ); +/* +* Inform pollers only when the whole thread group exits. +* If the thread group leader exits before all other threads in the +* group, then poll(2) should block, similar to the wait(2) family. +*/ +if (!task || (task->exit_state &&thread_group_empty (task))) + poll_flags =POLLIN |POLLRDNORM ; +rcu_read_unlock (); + +return poll_flags; +} +const struct file_operations pidfd_fops = { .release = pidfd_release, + .poll = pidfd_poll, #ifdef CONFIG_PROC_FS .show_fdinfo = pidfd_show_fdinfo, #endif
kernel/pid.c +2 -0
@@ -333,6 +333,8 @@ struct pid *alloc_pid(struct pid_namespace *ns)for (type =0 ; type <PIDTYPE_MAX ; ++type)INIT_HLIST_HEAD (&pid->tasks[type]); +init_waitqueue_head (&pid->wait_pidfd); + upid = pid->numbers + ns->level;spin_lock_irq (&pidmap_lock);if (!(ns->nr_hashed &PIDNS_HASH_ADDING ))
kernel/signal.c +11 -0
@@ -1632,6 +1632,14 @@ int send_sigqueue(struct sigqueue *q, struct task_struct *t, int group)return ret; } +static void do_notify_pidfd (struct task_struct *task) +{ +struct pid *pid; + + pid =task_pid (task); +wake_up_all (&pid->wait_pidfd); +} +/* * Let a parent know about the death of a child. * For a stopped/continued status change, use do_notify_parent_cldstop instead.
@@ -1655,6 +1663,9 @@ bool do_notify_parent(struct task_struct *tsk, int sig)BUG_ON (!tsk->ptrace && (tsk->group_leader != tsk || !thread_group_empty (tsk))); +/* Wake up all pidfd waiters */ +do_notify_pidfd (tsk); +if (sig !=SIGCHLD ) {/* * This is only possible if parent == real_parent.
Signature · 1 line
1.8.3.1
PATCH v2 3/10 proc: Pass file mode to proc_pid_make_inode
In reply to Wen YangFrom: Andreas Gruenbacher <[email protected]>
[ Upstream commit db978da8fa1d0819b210c137d31a339149b88875 ]
Pass the file mode of the proc inode to be created to proc_pid_make_inode. In proc_pid_make_inode, initialize inode->i_mode before calling security_task_to_inode. This allows selinux to set isec->sclass right away without introducing "half-initialized" inode security structs.
--- fs/proc/base.c | 23 +++++++++-------------- fs/proc/fd.c | 6 ++---- fs/proc/internal.h | 2 +- fs/proc/namespaces.c | 3 +-- security/selinux/hooks.c | 1 + 5 files changed, 14 insertions(+), 21 deletions(-)
fs/proc/base.c +9 -14
@@ -1676,7 +1676,8 @@ static int proc_pid_readlink(struct dentry * dentry, char __user * buffer, int b/* building an inode */ -struct inode *proc_pid_make_inode (struct super_block * sb,struct task_struct *task) +struct inode *proc_pid_make_inode (struct super_block * sb, +struct task_struct *task,umode_t mode) {struct inode * inode;struct proc_inode *ei;
@@ -1690,6 +1691,7 @@ struct inode *proc_pid_make_inode(struct super_block * sb, struct task_struct *t/* Common stuff */ ei =PROC_I (inode); + inode->i_mode = mode; inode->i_ino =get_next_ino (); inode->i_mtime = inode->i_atime = inode->i_ctime =current_time (inode); inode->i_op = &proc_def_inode_operations;
@@ -2041,7 +2043,9 @@ struct map_files_info {struct proc_inode *ei;struct inode *inode; - inode =proc_pid_make_inode (dir->i_sb, task); + inode =proc_pid_make_inode (dir->i_sb, task,S_IFLNK | + ((mode &FMODE_READ ) ?S_IRUSR :0 ) | + ((mode &FMODE_WRITE ) ?S_IWUSR :0 ));if (!inode)return -ENOENT ;
@@ -2050,12 +2054,6 @@ struct map_files_info { inode->i_op = &proc_map_files_link_inode_operations; inode->i_size =64 ; - inode->i_mode =S_IFLNK ; - -if (mode &FMODE_READ ) - inode->i_mode |=S_IRUSR ; -if (mode &FMODE_WRITE ) - inode->i_mode |=S_IWUSR ;d_set_d_op (dentry, &tid_map_files_dentry_operations);d_add (dentry, inode);
@@ -2409,12 +2407,11 @@ static int proc_pident_instantiate(struct inode *dir,struct inode *inode;struct proc_inode *ei; - inode =proc_pid_make_inode (dir->i_sb, task); + inode =proc_pid_make_inode (dir->i_sb, task, p->mode);if (!inode) goto out; ei =PROC_I (inode); - inode->i_mode = p->mode;if (S_ISDIR (inode->i_mode))set_nlink (inode,2 );/* Use getattr to fix if necessary */ if (p->iop)
@@ -3096,11 +3093,10 @@ static int proc_pid_instantiate(struct inode *dir, {struct inode *inode; - inode =proc_pid_make_inode (dir->i_sb, task); + inode =proc_pid_make_inode (dir->i_sb, task,S_IFDIR |S_IRUGO |S_IXUGO );if (!inode) goto out; - inode->i_mode =S_IFDIR |S_IRUGO |S_IXUGO ; inode->i_op = &proc_tgid_base_inode_operations; inode->i_fop = &proc_tgid_base_operations; inode->i_flags|=S_IMMUTABLE ;
@@ -3391,11 +3387,10 @@ static int proc_task_instantiate(struct inode *dir, struct dentry *dentry, struct task_struct *task, const void *ptr) { struct inode *inode; - inode =proc_pid_make_inode (dir->i_sb, task); + inode =proc_pid_make_inode (dir->i_sb, task,S_IFDIR |S_IRUGO |S_IXUGO );if (!inode) goto out; - inode->i_mode =S_IFDIR |S_IRUGO |S_IXUGO ; inode->i_op = &proc_tid_base_inode_operations; inode->i_fop = &proc_tid_base_operations; inode->i_flags|=S_IMMUTABLE ;
fs/proc/fd.c +2 -4
@@ -183,14 +183,13 @@ static int proc_fd_link(struct dentry *dentry, struct path *path)struct proc_inode *ei;struct inode *inode; - inode =proc_pid_make_inode (dir->i_sb, task); + inode =proc_pid_make_inode (dir->i_sb, task,S_IFLNK );if (!inode) goto out; ei =PROC_I (inode); ei->fd = fd; - inode->i_mode =S_IFLNK ; inode->i_op = &proc_pid_link_inode_operations; inode->i_size =64 ;
@@ -322,14 +321,13 @@ int proc_fd_permission(struct inode *inode, int mask)struct proc_inode *ei;struct inode *inode; - inode =proc_pid_make_inode (dir->i_sb, task); + inode =proc_pid_make_inode (dir->i_sb, task,S_IFREG |S_IRUSR );if (!inode) goto out; ei =PROC_I (inode); ei->fd = fd; - inode->i_mode =S_IFREG |S_IRUSR ; inode->i_fop = &proc_fdinfo_file_operations;d_set_d_op (dentry, &tid_fd_dentry_operations);
fs/proc/internal.h +1 -1
@@ -163,7 +163,7 @@ extern int proc_pid_statm(struct seq_file *, struct pid_namespace *,extern const struct dentry_operations pid_dentry_operations;extern int pid_getattr (struct vfsmount *,struct dentry *,struct kstat *);extern int proc_setattr (struct dentry *,struct iattr *); -extern struct inode *proc_pid_make_inode (struct super_block *,struct task_struct *); +extern struct inode *proc_pid_make_inode (struct super_block *,struct task_struct *,umode_t );extern int pid_revalidate (struct dentry *,unsigned int );extern int pid_delete_dentry (const struct dentry *);extern int proc_pid_readdir (struct file *,struct dir_context *);
fs/proc/namespaces.c +1 -2
@@ -92,12 +92,11 @@ static int proc_ns_instantiate(struct inode *dir,struct inode *inode;struct proc_inode *ei; - inode =proc_pid_make_inode (dir->i_sb, task); + inode =proc_pid_make_inode (dir->i_sb, task,S_IFLNK |S_IRWXUGO );if (!inode) goto out; ei =PROC_I (inode); - inode->i_mode =S_IFLNK |S_IRWXUGO ; inode->i_op = &proc_ns_link_inode_operations; ei->ns_ops = ns_ops;
security/selinux/hooks.c +1 -0
@@ -3967,6 +3967,7 @@ static void selinux_task_to_inode(struct task_struct *p,struct inode_security_struct *isec = inode->i_security;u32 sid =task_sid (p); + isec->sclass =inode_mode_to_security_class (inode->i_mode); isec->sid = sid; isec->initialized =LABEL_INITIALIZED ; }
Signature · 1 line
1.8.3.1
PATCH v2 4/10 proc: Better ownership of files for non-dumpable tasks in user namespaces
In reply to Wen YangFrom: "Eric W. Biederman" <[email protected]>
[ Upstream commit 68eb94f16227336a5773b83ecfa8290f1d6b78ce ]
Instead of making the files owned by the GLOBAL_ROOT_USER. Make non-dumpable files whose mm has always lived in a user namespace owned by the user namespace root. This allows the container root to have things work as expected in a container.
--- fs/proc/base.c | 102 ++++++++++++++++++++++++++++++----------------------- fs/proc/fd.c | 12 +------ fs/proc/internal.h | 16 ++------- 3 files changed, 61 insertions(+), 69 deletions(-)
fs/proc/base.c +58 -44
@@ -1676,12 +1676,63 @@ static int proc_pid_readlink(struct dentry * dentry, char __user * buffer, int b/* building an inode */ +void task_dump_owner (struct task_struct *task,mode_t mode, +kuid_t *ruid,kgid_t *rgid) +{ +/* Depending on the state of dumpable compute who should own a +* proc file for a task. +*/ +const struct cred *cred; +kuid_t uid; +kgid_t gid; + +/* Default to the tasks effective ownership */ +rcu_read_lock (); + cred =__task_cred (task); + uid = cred->euid; + gid = cred->egid; +rcu_read_unlock (); + +/* +* Before the /proc/pid/status file was created the only way to read +* the effective uid of a /process was to stat /proc/pid. Reading +* /proc/pid/status is slow enough that procps and other packages +* kept stating /proc/pid. To keep the rules in /proc simple I have +* made this apply to all per process world readable and executable +* directories. +*/ +if (mode != (S_IFDIR |S_IRUGO |S_IXUGO )) { +struct mm_struct *mm; +task_lock (task); + mm = task->mm; +/* Make non-dumpable tasks owned by some root */ +if (mm) { +if (get_dumpable (mm) !=SUID_DUMP_USER ) { +struct user_namespace *user_ns = mm->user_ns; + + uid =make_kuid (user_ns,0 ); +if (!uid_valid (uid)) + uid =GLOBAL_ROOT_UID ; + + gid =make_kgid (user_ns,0 ); +if (!gid_valid (gid)) + gid =GLOBAL_ROOT_GID ; + } + }else { + uid =GLOBAL_ROOT_UID ; + gid =GLOBAL_ROOT_GID ; + } +task_unlock (task); + } + *ruid = uid; + *rgid = gid; +} +struct inode *proc_pid_make_inode (struct super_block * sb,struct task_struct *task,umode_t mode) {struct inode * inode;struct proc_inode *ei; -const struct cred *cred;/* We need a new inode */
@@ -1703,13 +1754,7 @@ struct inode *proc_pid_make_inode(struct super_block * sb,if (!ei->pid) goto out_unlock; -if (task_dumpable (task)) { -rcu_read_lock (); - cred =__task_cred (task); - inode->i_uid = cred->euid; - inode->i_gid = cred->egid; -rcu_read_unlock (); - } +task_dump_owner (task,0 , &inode->i_uid, &inode->i_gid);security_task_to_inode (task, inode); out:
@@ -1724,7 +1769,6 @@ int pid_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat) {struct inode *inode =d_inode (dentry);struct task_struct *task; -const struct cred *cred;struct pid_namespace *pid = dentry->d_sb->s_fs_info;generic_fillattr (inode, stat);
@@ -1742,12 +1786,7 @@ int pid_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat)*/ return -ENOENT ; } -if ((inode->i_mode == (S_IFDIR |S_IRUGO |S_IXUGO )) || -task_dumpable (task)) { - cred =__task_cred (task); - stat->uid = cred->euid; - stat->gid = cred->egid; - } +task_dump_owner (task, inode->i_mode, &stat->uid, &stat->gid); }rcu_read_unlock ();return 0 ;
@@ -1763,18 +1802,11 @@ int pid_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat)* Rewrite the inode's ownerships here because the owning task may have * performed a setuid(), etc. * -* Before the /proc/pid/status file was created the only way to read -* the effective uid of a /process was to stat /proc/pid. Reading -* /proc/pid/status is slow enough that procps and other packages -* kept stating /proc/pid. To keep the rules in /proc simple I have -* made this apply to all per process world readable and executable -* directories. */ int pid_revalidate (struct dentry *dentry,unsigned int flags) {struct inode *inode;struct task_struct *task; -const struct cred *cred;if (flags &LOOKUP_RCU )return -ECHILD ;
@@ -1783,17 +1815,8 @@ int pid_revalidate(struct dentry *dentry, unsigned int flags) task =get_proc_task (inode);if (task) { -if ((inode->i_mode == (S_IFDIR |S_IRUGO |S_IXUGO )) || -task_dumpable (task)) { -rcu_read_lock (); - cred =__task_cred (task); - inode->i_uid = cred->euid; - inode->i_gid = cred->egid; -rcu_read_unlock (); - }else { - inode->i_uid =GLOBAL_ROOT_UID ; - inode->i_gid =GLOBAL_ROOT_GID ; - } +task_dump_owner (task, inode->i_mode, &inode->i_uid, &inode->i_gid); + inode->i_mode &= ~(S_ISUID |S_ISGID );security_task_to_inode (task, inode);put_task_struct (task);
@@ -1915,7 +1938,6 @@ static int map_files_d_revalidate(struct dentry *dentry, unsigned int flags)bool exact_vma_exists = false;struct mm_struct *mm =NULL ;struct task_struct *task; -const struct cred *cred;struct inode *inode;int status =0 ;
@@ -1940,16 +1962,8 @@ static int map_files_d_revalidate(struct dentry *dentry, unsigned int flags)mmput (mm);if (exact_vma_exists) { -if (task_dumpable (task)) { -rcu_read_lock (); - cred =__task_cred (task); - inode->i_uid = cred->euid; - inode->i_gid = cred->egid; -rcu_read_unlock (); - }else { - inode->i_uid =GLOBAL_ROOT_UID ; - inode->i_gid =GLOBAL_ROOT_GID ; - } +task_dump_owner (task,0 , &inode->i_uid, &inode->i_gid); +security_task_to_inode (task, inode); status =1 ; }
fs/proc/fd.c +1 -11
@@ -84,7 +84,6 @@ static int tid_fd_revalidate(struct dentry *dentry, unsigned int flags) {struct files_struct *files;struct task_struct *task; -const struct cred *cred;struct inode *inode;unsigned int fd;
@@ -108,16 +107,7 @@ static int tid_fd_revalidate(struct dentry *dentry, unsigned int flags)rcu_read_unlock ();put_files_struct (files); -if (task_dumpable (task)) { -rcu_read_lock (); - cred =__task_cred (task); - inode->i_uid = cred->euid; - inode->i_gid = cred->egid; -rcu_read_unlock (); - }else { - inode->i_uid =GLOBAL_ROOT_UID ; - inode->i_gid =GLOBAL_ROOT_GID ; - } +task_dump_owner (task,0 , &inode->i_uid, &inode->i_gid);if (S_ISLNK (inode->i_mode)) {unsigned i_mode =S_IFLNK ;
fs/proc/internal.h +2 -14
@@ -98,20 +98,8 @@ static inline struct task_struct *get_proc_task(struct inode *inode)return get_pid_task (proc_pid (inode),PIDTYPE_PID ); } -static inline int task_dumpable (struct task_struct *task) -{ -int dumpable =0 ; -struct mm_struct *mm; - -task_lock (task); - mm = task->mm; -if (mm) - dumpable =get_dumpable (mm); -task_unlock (task); -if (dumpable ==SUID_DUMP_USER ) -return 1 ; -return 0 ; -} +void task_dump_owner (struct task_struct *task,mode_t mode, +kuid_t *ruid,kgid_t *rgid);static inline unsigned name_to_int (const struct qstr *qstr) {
Signature · 1 line
1.8.3.1
PATCH v2 5/10 proc: use %u for pid printing and slightly less stack
In reply to Wen YangFrom: Alexey Dobriyan <[email protected]>
[ Upstream commit e3912ac37e07a13c70675cd75020694de4841c74 ]
PROC_NUMBUF is 13 which is enough for "negative int + \n + \0".
However PIDs and TGIDs are never negative and newline is not a concern, so use just 10 per integer.
--- fs/proc/base.c | 16 ++++++++-------- fs/proc/fd.c | 2 +- fs/proc/self.c | 6 +++--- fs/proc/thread_self.c | 5 ++--- 4 files changed, 14 insertions(+), 15 deletions(-)
fs/proc/base.c +8 -8
@@ -3018,11 +3018,11 @@ static struct dentry *proc_tgid_base_lookup(struct inode *dir, struct dentry *destatic void proc_flush_task_mnt (struct vfsmount *mnt,pid_t pid,pid_t tgid) {struct dentry *dentry, *leader, *dir; -char buf[PROC_NUMBUF ]; +char buf[10 +1 ];struct qstr name; name.name = buf; - name.len =snprintf (buf,sizeof (buf),"%d" , pid); + name.len =snprintf (buf,sizeof (buf),"%u" , pid);/* no ->d_hash() rejects on procfs */ dentry =d_hash_and_lookup (mnt->mnt_root, &name);if (dentry) {
@@ -3034,7 +3034,7 @@ static void proc_flush_task_mnt(struct vfsmount *mnt, pid_t pid, pid_t tgid)return ; name.name = buf; - name.len =snprintf (buf,sizeof (buf),"%d" , tgid); + name.len =snprintf (buf,sizeof (buf),"%u" , tgid); leader =d_hash_and_lookup (mnt->mnt_root, &name);if (!leader) goto out;
@@ -3046,7 +3046,7 @@ static void proc_flush_task_mnt(struct vfsmount *mnt, pid_t pid, pid_t tgid) goto out_put_leader; name.name = buf; - name.len =snprintf (buf,sizeof (buf),"%d" , pid); + name.len =snprintf (buf,sizeof (buf),"%u" , pid); dentry =d_hash_and_lookup (dir, &name);if (dentry) {d_invalidate (dentry);
@@ -3226,14 +3226,14 @@ int proc_pid_readdir(struct file *file, struct dir_context *ctx)for (iter =next_tgid (ns, iter); iter.task; iter.tgid +=1 , iter =next_tgid (ns, iter)) { -char name[PROC_NUMBUF ]; +char name[10 +1 ];int len;cond_resched ();if (!has_pid_permissions (ns, iter.task,2 ))continue ; - len =snprintf (name,sizeof (name),"%d" , iter.tgid); + len =snprintf (name,sizeof (name),"%u" , iter.tgid); ctx->pos = iter.tgid +TGID_OFFSET ;if (!proc_fill_cache (file, ctx, name, len, proc_pid_instantiate, iter.task,NULL )) {
@@ -3557,10 +3557,10 @@ static int proc_task_readdir(struct file *file, struct dir_context *ctx)for (task =first_tid (proc_pid (inode), tid, ctx->pos -2 , ns); task; task =next_tid (task), ctx->pos++) { -char name[PROC_NUMBUF ]; +char name[10 +1 ];int len; tid =task_pid_nr_ns (task, ns); - len =snprintf (name,sizeof (name),"%d" , tid); + len =snprintf (name,sizeof (name),"%u" , tid);if (!proc_fill_cache (file, ctx, name, len, proc_task_instantiate, task,NULL )) {/* returning this tgid failed, save it as the first
fs/proc/fd.c +1 -1
@@ -235,7 +235,7 @@ static int proc_readfd_common(struct file *file, struct dir_context *ctx,for (fd = ctx->pos -2 ; fd <files_fdtable (files)->max_fds; fd++, ctx->pos++) { -char name[PROC_NUMBUF ]; +char name[10 +1 ];int len;if (!fcheck_files (files, fd))
fs/proc/self.c +3 -3
@@ -35,11 +35,11 @@ static const char *proc_self_get_link(struct dentry *dentry,if (!tgid)return ERR_PTR (-ENOENT ); -/* 11 for max length of signed int in decimal + NULL term */ - name =kmalloc (12 , dentry ?GFP_KERNEL :GFP_ATOMIC ); +/* max length of unsigned int in decimal + NULL term */ + name =kmalloc (10 +1 , dentry ?GFP_KERNEL :GFP_ATOMIC );if (unlikely (!name))return dentry ?ERR_PTR (-ENOMEM ) :ERR_PTR (-ECHILD ); -sprintf (name,"%d" , tgid); +sprintf (name,"%u" , tgid);set_delayed_call (done, kfree_link, name);return name; }
fs/proc/thread_self.c +2 -3
@@ -30,11 +30,10 @@ static const char *proc_thread_self_get_link(struct dentry *dentry,if (!pid)return ERR_PTR (-ENOENT ); - name =kmalloc (PROC_NUMBUF +6 +PROC_NUMBUF , - dentry ?GFP_KERNEL :GFP_ATOMIC ); + name =kmalloc (10 +6 +10 +1 , dentry ?GFP_KERNEL :GFP_ATOMIC );if (unlikely (!name))return dentry ?ERR_PTR (-ENOMEM ) :ERR_PTR (-ECHILD ); -sprintf (name,"%d/task/%d" , tgid, pid); +sprintf (name,"%u/task/%u" , tgid, pid);set_delayed_call (done, kfree_link, name);return name; }
Signature · 1 line
1.8.3.1
PATCH v2 6/10 proc: Rename in proc_inode rename sysctl_inodes sibling_inodes
In reply to Wen YangFrom: "Eric W. Biederman" <[email protected]>
[ Upstream commit 0afa5ca82212247456f9de1468b595a111fee633 ]
I about to need and use the same functionality for pid based inodes and there is no point in adding a second field when this field is already here and serving the same purporse.
Just give the field a generic name so it is clear that it is no longer sysctl specific.
Also for good measure initialize sibling_inodes when proc_inode is initialized.
--- fs/proc/inode.c | 1 + fs/proc/internal.h | 2 +- fs/proc/proc_sysctl.c | 8 ++++---- 3 files changed, 6 insertions(+), 5 deletions(-)
fs/proc/inode.c +1 -0
@@ -67,6 +67,7 @@ static struct inode *proc_alloc_inode(struct super_block *sb) ei->pde =NULL ; ei->sysctl =NULL ; ei->sysctl_entry =NULL ; +INIT_HLIST_NODE (&ei->sibling_inodes); ei->ns_ops =NULL ; inode = &ei->vfs_inode;return inode;
fs/proc/internal.h +1 -1
@@ -65,7 +65,7 @@ struct proc_inode {struct proc_dir_entry *pde;struct ctl_table_header *sysctl;struct ctl_table *sysctl_entry; -struct hlist_node sysctl_inodes; +struct hlist_node sibling_inodes;const struct proc_ns_operations *ns_ops;struct inode vfs_inode; };
fs/proc/proc_sysctl.c +4 -4
@@ -272,9 +272,9 @@ static void proc_sys_prune_dcache(struct ctl_table_header *head) node =hlist_first_rcu (&head->inodes);if (!node)break ; - ei = hlist_entry(node, struct proc_inode, sysctl_inodes); + ei = hlist_entry(node, struct proc_inode, sibling_inodes);spin_lock (&sysctl_lock); -hlist_del_init_rcu (&ei->sysctl_inodes); +hlist_del_init_rcu (&ei->sibling_inodes);spin_unlock (&sysctl_lock); inode = &ei->vfs_inode;
@@ -480,7 +480,7 @@ static struct inode *proc_sys_make_inode(struct super_block *sb, } ei->sysctl = head; ei->sysctl_entry = table; -hlist_add_head_rcu (&ei->sysctl_inodes, &head->inodes); +hlist_add_head_rcu (&ei->sibling_inodes, &head->inodes); head->count++;spin_unlock (&sysctl_lock);
@@ -511,7 +511,7 @@ static struct inode *proc_sys_make_inode(struct super_block *sb,void proc_sys_evict_inode (struct inode *inode,struct ctl_table_header *head) {spin_lock (&sysctl_lock); -hlist_del_init_rcu (&PROC_I (inode)->sysctl_inodes); +hlist_del_init_rcu (&PROC_I (inode)->sibling_inodes);if (!--head->count)kfree_rcu (head, rcu);spin_unlock (&sysctl_lock);
Signature · 1 line
1.8.3.1
PATCH v2 7/10 proc: Generalize proc_sys_prune_dcache into proc_prune_siblings_dcache
In reply to Wen YangFrom: "Eric W. Biederman" <[email protected]>
[ Upstream commit 26dbc60f385ff9cff475ea2a3bad02e80fd6fa43 ]
This prepares the way for allowing the pid part of proc to use this dcache pruning code as well.
(proc: fix up cherry-pick conflicts for 26dbc60f385f)
--- fs/proc/inode.c | 38 ++++++++++++++++++++++++++++++++++++++ fs/proc/internal.h | 1 + fs/proc/proc_sysctl.c | 35 +---------------------------------- 3 files changed, 40 insertions(+), 34 deletions(-)
fs/proc/inode.c +38 -0
@@ -101,6 +101,44 @@ void __init proc_init_inodecache(void) init_once); } +void proc_prune_siblings_dcache (struct hlist_head *inodes,spinlock_t *lock) +{ +struct inode *inode; +struct proc_inode *ei; +struct hlist_node *node; +struct super_block *sb; + +rcu_read_lock (); +for (;;) { + node =hlist_first_rcu (inodes); +if (!node) +break ; + ei = hlist_entry(node, struct proc_inode, sibling_inodes); +spin_lock (lock); +hlist_del_init_rcu (&ei->sibling_inodes); +spin_unlock (lock); + + inode = &ei->vfs_inode; + sb = inode->i_sb; +if (!atomic_inc_not_zero (&sb->s_active)) +continue ; + inode =igrab (inode); +rcu_read_unlock (); +if (unlikely (!inode)) { +deactivate_super (sb); +rcu_read_lock (); +continue ; + } + +d_prune_aliases (inode); +iput (inode); +deactivate_super (sb); + +rcu_read_lock (); + } +rcu_read_unlock (); +} +static int proc_show_options (struct seq_file *seq,struct dentry *root) {struct super_block *sb = root->d_sb;
fs/proc/internal.h +1 -0
@@ -200,6 +200,7 @@ struct pde_opener {extern const struct inode_operations proc_pid_link_inode_operations;extern void proc_init_inodecache (void ); +void proc_prune_siblings_dcache (struct hlist_head *inodes,spinlock_t *lock);extern struct inode *proc_get_inode (struct super_block *,struct proc_dir_entry *);extern int proc_fill_super (struct super_block *,void *data,int flags);extern void proc_entry_rundown (struct proc_dir_entry *);
fs/proc/proc_sysctl.c +1 -34
@@ -262,40 +262,7 @@ static void unuse_table(struct ctl_table_header *p)static void proc_sys_prune_dcache (struct ctl_table_header *head) { -struct inode *inode; -struct proc_inode *ei; -struct hlist_node *node; -struct super_block *sb; - -rcu_read_lock (); -for (;;) { - node =hlist_first_rcu (&head->inodes); -if (!node) -break ; - ei = hlist_entry(node, struct proc_inode, sibling_inodes); -spin_lock (&sysctl_lock); -hlist_del_init_rcu (&ei->sibling_inodes); -spin_unlock (&sysctl_lock); - - inode = &ei->vfs_inode; - sb = inode->i_sb; -if (!atomic_inc_not_zero (&sb->s_active)) -continue ; - inode =igrab (inode); -rcu_read_unlock (); -if (unlikely (!inode)) { -deactivate_super (sb); -rcu_read_lock (); -continue ; - } - -d_prune_aliases (inode); -iput (inode); -deactivate_super (sb); - -rcu_read_lock (); - } -rcu_read_unlock (); +proc_prune_siblings_dcache (&head->inodes, &sysctl_lock); }/* called under sysctl_lock, will reacquire if has to wait */
Signature · 1 line
1.8.3.1
PATCH v2 8/10 proc: Clear the pieces of proc_inode that proc_evict_inode cares about
In reply to Wen YangFrom: "Eric W. Biederman" <[email protected]>
[ Upstream commit 71448011ea2a1cd36d8f5cbdab0ed716c454d565 ]
This just keeps everything tidier, and allows for using flags like SLAB_TYPESAFE_BY_RCU where slabs are not always cleared before reuse. I don't see reuse without reinitializing happening with the proc_inode but I had a false alarm while reworking flushing of proc dentries and indoes when a process dies that caused me to tidy this up.
The code is a little easier to follow and reason about this way so I figured the changes might as well be kept.
--- fs/proc/inode.c | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-)
fs/proc/inode.c +11 -5
@@ -32,21 +32,27 @@ static void proc_evict_inode(struct inode *inode) {struct proc_dir_entry *de;struct ctl_table_header *head; +struct proc_inode *ei =PROC_I (inode);truncate_inode_pages_final (&inode->i_data);clear_inode (inode);/* Stop tracking associated processes */ -put_pid (PROC_I (inode)->pid); +if (ei->pid) { +put_pid (ei->pid); + ei->pid =NULL ; + }/* Let go of any associated proc directory entry */ - de =PDE (inode); -if (de) + de = ei->pde; +if (de) {pde_put (de); + ei->pde =NULL ; + } - head =PROC_I (inode)->sysctl; + head = ei->sysctl;if (head) { -RCU_INIT_POINTER (PROC_I (inode)->sysctl,NULL ); +RCU_INIT_POINTER (ei->sysctl,NULL );proc_sys_evict_inode (inode, head); } }
Signature · 1 line
1.8.3.1
PATCH v2 9/10 proc: Use d_invalidate in proc_prune_siblings_dcache
In reply to Wen YangFrom: "Eric W. Biederman" <[email protected]>
[ Upstream commit f90f3cafe8d56d593fc509a4185da1d5800efea4 ]
The function d_prune_aliases has the problem that it will only prune aliases thare are completely unused. It will not remove aliases for the dcache or even think of removing mounts from the dcache. For that behavior d_invalidate is needed.
To use d_invalidate replace d_prune_aliases with d_find_alias followed by d_invalidate and dput.
For completeness the directory and the non-directory cases are separated because in theory (although not in currently in practice for proc) directories can only ever have a single dentry while non-directories can have hardlinks and thus multiple dentries. As part of this separation use d_find_any_alias for directories to spare d_find_alias the extra work of doing that.
Plus the differences between d_find_any_alias and d_find_alias makes it clear why the directory and non-directory code and not share code.
To make it clear these routines now invalidate dentries rename proc_prune_siblings_dache to proc_invalidate_siblings_dcache, and rename proc_sys_prune_dcache proc_sys_invalidate_dcache.
V2: Split the directory and non-directory cases. To make this
code robust to future changes in proc.
(proc: fix up cherry-pick conflicts for f90f3cafe8d5)
--- fs/proc/inode.c | 16 ++++++++++++++-- fs/proc/internal.h | 2 +- fs/proc/proc_sysctl.c | 8 ++++---- 3 files changed, 19 insertions(+), 7 deletions(-)
fs/proc/inode.c +14 -2
@@ -107,7 +107,7 @@ void __init proc_init_inodecache(void) init_once); } -void proc_prune_siblings_dcache (struct hlist_head *inodes,spinlock_t *lock) +void proc_invalidate_siblings_dcache (struct hlist_head *inodes,spinlock_t *lock) {struct inode *inode;struct proc_inode *ei;
@@ -136,7 +136,19 @@ void proc_prune_siblings_dcache(struct hlist_head *inodes, spinlock_t *lock)continue ; } -d_prune_aliases (inode); +if (S_ISDIR (inode->i_mode)) { +struct dentry *dir =d_find_any_alias (inode); +if (dir) { +d_invalidate (dir); +dput (dir); + } + }else { +struct dentry *dentry; +while ((dentry =d_find_alias (inode))) { +d_invalidate (dentry); +dput (dentry); + } + }iput (inode);deactivate_super (sb);
fs/proc/internal.h +1 -1
@@ -200,7 +200,7 @@ struct pde_opener {extern const struct inode_operations proc_pid_link_inode_operations;extern void proc_init_inodecache (void ); -void proc_prune_siblings_dcache (struct hlist_head *inodes,spinlock_t *lock); +void proc_invalidate_siblings_dcache (struct hlist_head *inodes,spinlock_t *lock);extern struct inode *proc_get_inode (struct super_block *,struct proc_dir_entry *);extern int proc_fill_super (struct super_block *,void *data,int flags);extern void proc_entry_rundown (struct proc_dir_entry *);
fs/proc/proc_sysctl.c +4 -4
@@ -260,9 +260,9 @@ static void unuse_table(struct ctl_table_header *p)complete (p->unregistering); } -static void proc_sys_prune_dcache (struct ctl_table_header *head) +static void proc_sys_invalidate_dcache (struct ctl_table_header *head) { -proc_prune_siblings_dcache (&head->inodes, &sysctl_lock); +proc_invalidate_siblings_dcache (&head->inodes, &sysctl_lock); }/* called under sysctl_lock, will reacquire if has to wait */
@@ -284,10 +284,10 @@ static void start_unregistering(struct ctl_table_header *p)spin_unlock (&sysctl_lock); }/* -* Prune dentries for unregistered sysctls: namespaced sysctls +* Invalidate dentries for unregistered sysctls: namespaced sysctls * can have duplicate names and contaminate dcache very badly. */ -proc_sys_prune_dcache (p); +proc_sys_invalidate_dcache (p);/* * do not remove from the list until nobody holds it; walking the * list in do_sysctl() relies on that.
Signature · 1 line
1.8.3.1
PATCH v2 10/10 proc: Use a list of inodes to flush from proc
In reply to Wen YangFrom: "Eric W. Biederman" <[email protected]>
[ Upstream commit 7bc3e6e55acf065500a24621f3b313e7e5998acf ]
Rework the flushing of proc to use a list of directory inodes that need to be flushed.
The list is kept on struct pid not on struct task_struct, as there is a fixed connection between proc inodes and pids but at least for the case of de_thread the pid of a task_struct changes.
This removes the dependency on proc_mnt which allows for different mounts of proc having different mount options even in the same pid namespace and this allows for the removal of proc_mnt which will trivially the first mount of proc to honor it's mount options.
This flushing remains an optimization. The functions pid_delete_dentry and pid_revalidate ensure that ordinary dcache management will not attempt to use dentries past the point their respective task has died. When unused the shrinker will eventually be able to remove these dentries.
There is a case in de_thread where proc_flush_pid can be called early for a given pid. Which winds up being safe (if suboptimal) as this is just an optiimization.
Only pid directories are put on the list as the other per pid files are children of those directories and d_invalidate on the directory will get them as well.
So that the pid can be used during flushing it's reference count is taken in release_task and dropped in proc_flush_pid. Further the call of proc_flush_pid is moved after the tasklist_lock is released in release_task so that it is certain that the pid has already been unhashed when flushing it taking place. This removes a small race where a dentry could recreated.
As struct pid is supposed to be small and I need a per pid lock I reuse the only lock that currently exists in struct pid the the wait_pidfd.lock.
The net result is that this adds all of this functionality with just a little extra list management overhead and a single extra pointer in struct pid.
v2: Initialize pid->inodes. I somehow failed to get that
initialization into the initial version of the patch. A boot failure was reported by "kernel test robot <[email protected]>", and failure to initialize that pid->inodes matches all of the reported symptoms.
namespaces")
entries from multiple proc trees")
Cc: <[email protected]> # 4.9.x: b3e583825266: clone: add CLONE_PIDFD
Cc: <[email protected]> # 4.9.x: b53b0b9d9a61: pidfd: add polling support
Cc: <[email protected]> # 4.9.x: db978da8fa1d: proc: Pass file mode to proc_pid_make_inode
Cc: <[email protected]> # 4.9.x: 68eb94f16227: proc: Better ownership of files for non-dumpable tasks in user namespaces
Cc: <[email protected]> # 4.9.x: e3912ac37e07: proc: use %u for pid printing and slightly less stack
Cc: <[email protected]> # 4.9.x: 0afa5ca82212: proc: Rename in proc_inode rename sysctl_inodes sibling_inodes
Cc: <[email protected]> # 4.9.x: 26dbc60f385f: proc: Generalize proc_sys_prune_dcache into proc_prune_siblings_dcache
Cc: <[email protected]> # 4.9.x: 71448011ea2a: proc: Clear the pieces of proc_inode that proc_evict_inode cares about
Cc: <[email protected]> # 4.9.x: f90f3cafe8d5: Use d_invalidate in proc_prune_siblings_dcache
Cc: <[email protected]> # 4.9.x
(proc: fix up cherry-pick conflicts for 7bc3e6e55acf)
--- fs/proc/base.c | 111 ++++++++++++++++-------------------------------- fs/proc/inode.c | 2 +- fs/proc/internal.h | 1 + include/linux/pid.h | 1 + include/linux/proc_fs.h | 4 +- kernel/exit.c | 5 ++- kernel/pid.c | 1 + 7 files changed, 45 insertions(+), 80 deletions(-)
fs/proc/base.c +36 -75
@@ -1728,11 +1728,25 @@ void task_dump_owner(struct task_struct *task, mode_t mode, *rgid = gid; } +void proc_pid_evict_inode (struct proc_inode *ei) +{ +struct pid *pid = ei->pid; + +if (S_ISDIR (ei->vfs_inode.i_mode)) { +spin_lock (&pid->wait_pidfd.lock); +hlist_del_init_rcu (&ei->sibling_inodes); +spin_unlock (&pid->wait_pidfd.lock); + } + +put_pid (pid); +} +struct inode *proc_pid_make_inode (struct super_block * sb,struct task_struct *task,umode_t mode) {struct inode * inode;struct proc_inode *ei; +struct pid *pid;/* We need a new inode */
@@ -1750,10 +1764,18 @@ struct inode *proc_pid_make_inode(struct super_block * sb,/* * grab the reference to task. */ - ei->pid =get_task_pid (task,PIDTYPE_PID ); -if (!ei->pid) + pid =get_task_pid (task,PIDTYPE_PID ); +if (!pid) goto out_unlock; +/* Let the pid remember us for quick removal */ + ei->pid = pid; +if (S_ISDIR (mode)) { +spin_lock (&pid->wait_pidfd.lock); +hlist_add_head_rcu (&ei->sibling_inodes, &pid->inodes); +spin_unlock (&pid->wait_pidfd.lock); + } +task_dump_owner (task,0 , &inode->i_uid, &inode->i_gid);security_task_to_inode (task, inode);
@@ -3015,90 +3037,29 @@ static struct dentry *proc_tgid_base_lookup(struct inode *dir, struct dentry *de .permission = proc_pid_permission, }; -static void proc_flush_task_mnt (struct vfsmount *mnt,pid_t pid,pid_t tgid) -{ -struct dentry *dentry, *leader, *dir; -char buf[10 +1 ]; -struct qstr name; - - 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); -if (dentry) { -d_invalidate (dentry); -dput (dentry); - } - -if (pid == tgid) -return ; - - name.name = buf; - name.len =snprintf (buf,sizeof (buf),"%u" , tgid); - leader =d_hash_and_lookup (mnt->mnt_root, &name); -if (!leader) - goto out; - - name.name ="task" ; - name.len =strlen (name.name); - dir =d_hash_and_lookup (leader, &name); -if (!dir) - goto out_put_leader; - - name.name = buf; - name.len =snprintf (buf,sizeof (buf),"%u" , pid); - dentry =d_hash_and_lookup (dir, &name); -if (dentry) { -d_invalidate (dentry); -dput (dentry); - } - -dput (dir); -out_put_leader: -dput (leader); -out: -return ; -} -/** -* proc_flush_task - Remove dcache entries for @task from the /proc dcache. -* @task: task that should be flushed. +* proc_flush_pid - Remove dcache entries for @pid from the /proc dcache. +* @pid: pid that should be flushed. * -* When flushing dentries from proc, one needs to flush them from global -* proc (proc_mnt) and from all the namespaces' procs this task was seen -* in. This call is supposed to do all of this job. -* -* Looks in the dcache for -* /proc/@pid -* /proc/@tgid/task/@pid -* if either directory is present flushes it and all of it'ts children -* from the dcache. +* This function walks a list of inodes (that belong to any proc +* filesystem) that are attached to the pid and flushes them from +* the dentry cache. * * It is safe and reasonable to cache /proc entries for a task until * that task exits. After that they just clog up the dcache with * useless entries, possibly causing useful dcache entries to be -* flushed instead. This routine is proved to flush those useless -* dcache entries at process exit time. +* flushed instead. This routine is provided to flush those useless +* dcache entries when a process is reaped. * * NOTE: This routine is just an optimization so it does not guarantee -* that no dcache entries will exist at process exit time it -* just makes it very unlikely that any will persist. +* that no dcache entries will exist after a process is reaped +* it just makes it very unlikely that any will persist. */ -void proc_flush_task (struct task_struct *task) +void proc_flush_pid (struct pid *pid) { -int i; -struct pid *pid, *tgid; -struct upid *upid; - - 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); - } +proc_invalidate_siblings_dcache (&pid->inodes, &pid->wait_pidfd.lock); +put_pid (pid); } static int proc_pid_instantiate(struct inode *dir,
fs/proc/inode.c +1 -1
@@ -39,7 +39,7 @@ static void proc_evict_inode(struct inode *inode)/* Stop tracking associated processes */ if (ei->pid) { -put_pid (ei->pid); +proc_pid_evict_inode (ei); ei->pid =NULL ; }
fs/proc/internal.h +1 -0
@@ -151,6 +151,7 @@ extern int proc_pid_statm(struct seq_file *, struct pid_namespace *,extern const struct dentry_operations pid_dentry_operations;extern int pid_getattr (struct vfsmount *,struct dentry *,struct kstat *);extern int proc_setattr (struct dentry *,struct iattr *); +extern void proc_pid_evict_inode (struct proc_inode *);extern struct inode *proc_pid_make_inode (struct super_block *,struct task_struct *,umode_t );extern int pid_revalidate (struct dentry *,unsigned int );extern int pid_delete_dentry (const struct dentry *);
include/linux/pid.h +1 -0
@@ -63,6 +63,7 @@ struct pidunsigned int level;/* lists of tasks that use this pid */ struct hlist_head tasks[PIDTYPE_MAX ]; +struct hlist_head inodes;/* wait queue for pidfd notifications */ wait_queue_head_t wait_pidfd;struct rcu_head rcu;
include/linux/proc_fs.h +2 -2
@@ -12,7 +12,7 @@#ifdef CONFIG_PROC_FS extern void proc_root_init (void ); -extern void proc_flush_task (struct task_struct *); +extern void proc_flush_pid (struct pid *);extern struct proc_dir_entry *proc_symlink (const char *,struct proc_dir_entry *,const char *);
@@ -48,7 +48,7 @@ static inline void proc_root_init(void) { } -static inline void proc_flush_task (struct task_struct *task) +static inline void proc_flush_pid (struct pid *pid) { }
kernel/exit.c +3 -2
@@ -168,6 +168,7 @@ static void delayed_put_task_struct(struct rcu_head *rhp)void release_task (struct task_struct *p) {struct task_struct *leader; +struct pid *thread_pid;int zap_leader; repeat:/* don't need to get the RCU readlock here - the process is dead and
@@ -176,10 +177,9 @@ void release_task(struct task_struct *p)atomic_dec (&__task_cred (p)->user->processes);rcu_read_unlock (); -proc_flush_task (p); -write_lock_irq (&tasklist_lock);ptrace_release_task (p); + thread_pid =get_pid (p->pids[PIDTYPE_PID ].pid);__exit_signal (p);/*
@@ -202,6 +202,7 @@ void release_task(struct task_struct *p) }write_unlock_irq (&tasklist_lock); +proc_flush_pid (thread_pid);release_thread (p);call_rcu (&p->rcu, delayed_put_task_struct);
kernel/pid.c +1 -0
@@ -334,6 +334,7 @@ struct pid *alloc_pid(struct pid_namespace *ns)INIT_HLIST_HEAD (&pid->tasks[type]);init_waitqueue_head (&pid->wait_pidfd); +INIT_HLIST_HEAD (&pid->inodes); upid = pid->numbers + ns->level;spin_lock_irq (&pidmap_lock);
Signature · 1 line
1.8.3.1
Greg Kroah-Hartman
maintainerQuoted 25 lines · Wen Yang
The dentries such as /proc/<pid>/ns/ have the
DCACHE_OP_DELETEflag, they should be deleted when the process exits.Suppose the following race appears:
release_taskdput
->proc_flush_task-> dentry->d_op->d_delete(dentry) -> __exit_signal -> dentry->d_lockref.count-- and return.In the
proc_flush_task(), if another process is using this dentry, it will not be deleted. At the same time, indput(),d_op->d_delete()can be executed before__exit_signal(pid has not been hashed),d_deletereturns false, so this dentry still cannot be deleted.This dentry will always be cached (although its count is 0 and the
DCACHE_OP_DELETEflag is set), its parent denry will also be cached too, and these dentries can only be deleted whendrop_cachesis manually triggered.This will result in wasted memory. What's more troublesome is that these dentries reference pid, according to the commit
f333c700c610("pidns: Add a limit on the number of pid namespaces"), if the pid cannot be released, it may result in the inability to create a newpid_ns.
Quoted 4 lines
This issue was introduced by
60347f6716aa("pid namespaces: prepareproc_flust_task()to flush entries from multiple proc trees"), exposed byf333c700c610("pidns: Add a limit on the number of pid namespaces"), and then fixed by7bc3e6e55acf("proc: Use a list of inodes to flush from proc").
Why are you just submitting a series for 4.9 and 4.19, what about 4.14? We can't have users move to a newer kernel and then experience old bugs, right?
But the larger question is why are you backporting a whole new feature here? Why is CLONE_PIDFD needed? That feels really wrong...
thanks,
greg k-h
Quoted 26 lines · Greg Kroah-Hartman
Quoted 25 lines · Wen Yang
The dentries such as /proc/<pid>/ns/ have the
DCACHE_OP_DELETEflag, they should be deleted when the process exits.Suppose the following race appears:
release_taskdput
->proc_flush_task-> dentry->d_op->d_delete(dentry) -> __exit_signal -> dentry->d_lockref.count-- and return.In the
proc_flush_task(), if another process is using this dentry, it will not be deleted. At the same time, indput(),d_op->d_delete()can be executed before__exit_signal(pid has not been hashed),d_deletereturns false, so this dentry still cannot be deleted.This dentry will always be cached (although its count is 0 and the
DCACHE_OP_DELETEflag is set), its parent denry will also be cached too, and these dentries can only be deleted whendrop_cachesis manually triggered.This will result in wasted memory. What's more troublesome is that these dentries reference pid, according to the commit
f333c700c610("pidns: Add a limit on the number of pid namespaces"), if the pid cannot be released, it may result in the inability to create a newpid_ns.
Quoted 9 lines
Quoted 4 lines
This issue was introduced by
60347f6716aa("pid namespaces: prepareproc_flust_task()to flush entries from multiple proc trees"), exposed byf333c700c610("pidns: Add a limit on the number of pid namespaces"), and then fixed by7bc3e6e55acf("proc: Use a list of inodes to flush from proc").Why are you just submitting a series for 4.9 and 4.19, what about 4.14? We can't have users move to a newer kernel and then experience old bugs, right?
Okay, the patches corresponding to 4.14 will be ready later.
Quoted 3 lines
But the larger question is why are you backporting a whole new feature here? Why is
CLONE_PIDFDneeded? That feels really wrong...
The reason for backporting CLONE_PIDFD is because 7bc3e6e55acf ("proc: Use a list of inodes to flush from proc") relies on wait_pidfd.lock. There are indeed many associated modifications here. We are also testing it. Please check the code more.
Thanks.
Signature · 2 lines
Best wishes, Wen
Greg Kroah-Hartman
maintainerQuoted 34 lines · Wen Yang
Quoted 31 lines · Greg Kroah-Hartman
Quoted 29 lines · Wen Yang
The dentries such as /proc/<pid>/ns/ have the
DCACHE_OP_DELETEflag, they should be deleted when the process exits.Suppose the following race appears:
release_taskdput
->proc_flush_task-> dentry->d_op->d_delete(dentry) -> __exit_signal -> dentry->d_lockref.count-- and return.In the
proc_flush_task(), if another process is using this dentry, it will not be deleted. At the same time, indput(),d_op->d_delete()can be executed before__exit_signal(pid has not been hashed),d_deletereturns false, so this dentry still cannot be deleted.This dentry will always be cached (although its count is 0 and the
DCACHE_OP_DELETEflag is set), its parent denry will also be cached too, and these dentries can only be deleted whendrop_cachesis manually triggered.This will result in wasted memory. What's more troublesome is that these dentries reference pid, according to the commit
f333c700c610("pidns: Add a limit on the number of pid namespaces"), if the pid cannot be released, it may result in the inability to create a newpid_ns.This issue was introduced by
60347f6716aa("pid namespaces: prepareproc_flust_task()to flush entries from multiple proc trees"), exposed byf333c700c610("pidns: Add a limit on the number of pid namespaces"), and then fixed by7bc3e6e55acf("proc: Use a list of inodes to flush from proc").
Quoted 5 lines
Quoted 4 lines
Why are you just submitting a series for 4.9 and 4.19, what about 4.14? We can't have users move to a newer kernel and then experience old bugs, right?
Okay, the patches corresponding to 4.14 will be ready later.
Note for some reason you didn't cc: the stable list for these patches :(
Quoted 8 lines
Quoted 3 lines
But the larger question is why are you backporting a whole new feature here? Why is
CLONE_PIDFDneeded? That feels really wrong...The reason for backporting
CLONE_PIDFDis because7bc3e6e55acf("proc: Use a list of inodes to flush from proc") relies on wait_pidfd.lock. There are indeed many associated modifications here. We are also testing it. Please check the code more.
Is the only "issue" here wasted memory? Will it eventually be freed anyway even if you do not echo to the proc file to flush caches?
You mention the inability to create a new pid for a specific namespace, is that really a problem? Shouldn't the code handle such issues normally? What breaks without these changes?
I think at this point, it might just time for you to move to a newer kernel release, as adding a whole new userspace feature for this feels really really odd.
What is preventing you from doing that today? What holds you to older kernels that will not allow you to move forward?
thanks,
greg k-h
Quoted 59 lines · Greg Kroah-Hartman
Quoted 34 lines · Wen Yang
Quoted 31 lines · Greg Kroah-Hartman
Quoted 29 lines · Wen Yang
The dentries such as /proc/<pid>/ns/ have the
DCACHE_OP_DELETEflag, they should be deleted when the process exits.Suppose the following race appears:
release_taskdput
->proc_flush_task-> dentry->d_op->d_delete(dentry) -> __exit_signal -> dentry->d_lockref.count-- and return.In the
proc_flush_task(), if another process is using this dentry, it will not be deleted. At the same time, indput(),d_op->d_delete()can be executed before__exit_signal(pid has not been hashed),d_deletereturns false, so this dentry still cannot be deleted.This dentry will always be cached (although its count is 0 and the
DCACHE_OP_DELETEflag is set), its parent denry will also be cached too, and these dentries can only be deleted whendrop_cachesis manually triggered.This will result in wasted memory. What's more troublesome is that these dentries reference pid, according to the commit
f333c700c610("pidns: Add a limit on the number of pid namespaces"), if the pid cannot be released, it may result in the inability to create a newpid_ns.This issue was introduced by
60347f6716aa("pid namespaces: prepareproc_flust_task()to flush entries from multiple proc trees"), exposed byf333c700c610("pidns: Add a limit on the number of pid namespaces"), and then fixed by7bc3e6e55acf("proc: Use a list of inodes to flush from proc").Quoted 5 lines
Quoted 4 lines
Why are you just submitting a series for 4.9 and 4.19, what about 4.14? We can't have users move to a newer kernel and then experience old bugs, right?
Okay, the patches corresponding to 4.14 will be ready later.
Note for some reason you didn't cc: the stable list for these patches :(
Quoted 8 lines
Quoted 3 lines
But the larger question is why are you backporting a whole new feature here? Why is
CLONE_PIDFDneeded? That feels really wrong...The reason for backporting
CLONE_PIDFDis because7bc3e6e55acf("proc: Use a list of inodes to flush from proc") relies on wait_pidfd.lock. There are indeed many associated modifications here. We are also testing it. Please check the code more.Is the only "issue" here wasted memory? Will it eventually be freed anyway even if you do not echo to the proc file to flush caches?
You mention the inability to create a new pid for a specific namespace, is that really a problem? Shouldn't the code handle such issues normally? What breaks without these changes?
Quoted 7 lines
I think at this point, it might just time for you to move to a newer kernel release, as adding a whole new userspace feature for this feels really really odd.
What is preventing you from doing that today? What holds you to older kernels that will not allow you to move forward?
We have encountered this problem in the cloud server environment. Users will frequently create and delete containers, and the corresponding pid_ns will accumulate, eventually making it impossible to create a new container.
https://bugzilla.kernel.org/show_bug.cgi?id=208613
The kernels (4.9/4.19) used on a large scale in our current production environment (almost tens of thousands of machines) may need to be fixed.
Thanks.
Signature · 2 lines
Best wishes, Wen
Greg Kroah-Hartman
maintainerQuoted 70 lines · Wen Yang
Quoted 59 lines · Greg Kroah-Hartman
Quoted 34 lines · Wen Yang
Quoted 31 lines · Greg Kroah-Hartman
Quoted 29 lines · Wen Yang
The dentries such as /proc/<pid>/ns/ have the
DCACHE_OP_DELETEflag, they should be deleted when the process exits.Suppose the following race appears:
release_taskdput
->proc_flush_task-> dentry->d_op->d_delete(dentry) -> __exit_signal -> dentry->d_lockref.count-- and return.In the
proc_flush_task(), if another process is using this dentry, it will not be deleted. At the same time, indput(),d_op->d_delete()can be executed before__exit_signal(pid has not been hashed),d_deletereturns false, so this dentry still cannot be deleted.This dentry will always be cached (although its count is 0 and the
DCACHE_OP_DELETEflag is set), its parent denry will also be cached too, and these dentries can only be deleted whendrop_cachesis manually triggered.This will result in wasted memory. What's more troublesome is that these dentries reference pid, according to the commit
f333c700c610("pidns: Add a limit on the number of pid namespaces"), if the pid cannot be released, it may result in the inability to create a newpid_ns.This issue was introduced by
60347f6716aa("pid namespaces: prepareproc_flust_task()to flush entries from multiple proc trees"), exposed byf333c700c610("pidns: Add a limit on the number of pid namespaces"), and then fixed by7bc3e6e55acf("proc: Use a list of inodes to flush from proc").Quoted 5 lines
Quoted 4 lines
Why are you just submitting a series for 4.9 and 4.19, what about 4.14? We can't have users move to a newer kernel and then experience old bugs, right?
Okay, the patches corresponding to 4.14 will be ready later.
Note for some reason you didn't cc: the stable list for these patches :(
Quoted 8 lines
Quoted 3 lines
But the larger question is why are you backporting a whole new feature here? Why is
CLONE_PIDFDneeded? That feels really wrong...The reason for backporting
CLONE_PIDFDis because7bc3e6e55acf("proc: Use a list of inodes to flush from proc") relies on wait_pidfd.lock. There are indeed many associated modifications here. We are also testing it. Please check the code more.Is the only "issue" here wasted memory? Will it eventually be freed anyway even if you do not echo to the proc file to flush caches?
You mention the inability to create a new pid for a specific namespace, is that really a problem? Shouldn't the code handle such issues normally? What breaks without these changes?
Quoted 7 lines
I think at this point, it might just time for you to move to a newer kernel release, as adding a whole new userspace feature for this feels really really odd.
What is preventing you from doing that today? What holds you to older kernels that will not allow you to move forward?
Quoted 8 lines
We have encountered this problem in the cloud server environment. Users will frequently create and delete containers, and the corresponding
pid_nswill accumulate, eventually making it impossible to create a new container.https://bugzilla.kernel.org/show_bug.cgi?id=208613
The kernels (4.9/4.19) used on a large scale in our current production environment (almost tens of thousands of machines) may need to be fixed.
What prevents you from moving them to 5.4 or better yet, 5.10? You will have to do it soon anyway, I'm sure you have been testing those kernels to validate that all works well with them on a subset of your environment, so for those systems that have this problem, why can't you update the base kernel?
thanks,
greg k-h