The following (backported) patch removes a hook which has already been removed upstream. [1] Reason is that a permission denial can lead to soft lockups and zombies. A reproducer can be found in the initial report. [2] I hence consider this a bugfix which is allowed for stable branches.
Background: To reduce divergence of 4.9 to upstream before proposing to apply upstream commit 3dfc9b02864bt (LSM: Initialize security_hook_heads upon registration.) I'm checking which changes to the LSM hooks may be applicable to 4.9 as doing so after backporting 33dfc9b02864bt will lead to conflicts for each such commit.
[1] https://patchwork.kernel.org/project/linux-security-module/patch/1484069312-... [2] https://patchwork.kernel.org/project/selinux/patch/58736B2E.90201@huawei.com
Stephen Smalley (1): security,selinux,smack: kill security_task_wait hook
include/linux/lsm_hooks.h | 7 ------- include/linux/security.h | 6 ------ kernel/exit.c | 19 ++----------------- security/security.c | 6 ------ security/selinux/hooks.c | 6 ------ security/smack/smack_lsm.c | 20 -------------------- 6 files changed, 2 insertions(+), 62 deletions(-)
From: Stephen Smalley sds@tycho.nsa.gov
commit 3a2f5a59a695a73e0cde9a61e0feae5fa730e936 upstream.
As reported by yangshukui, a permission denial from security_task_wait() can lead to a soft lockup in zap_pid_ns_processes() since it only expects sys_wait4() to return 0 or -ECHILD. Further, security_task_wait() can in general lead to zombies; in the absence of some way to automatically reparent a child process upon a denial, the hook is not useful. Remove the security hook and its implementations in SELinux and Smack. Smack already removed its check from its hook.
Reported-by: yangshukui yangshukui@huawei.com Signed-off-by: Stephen Smalley sds@tycho.nsa.gov Acked-by: Casey Schaufler casey@schaufler-ca.com Acked-by: Oleg Nesterov oleg@redhat.com Signed-off-by: Paul Moore paul@paul-moore.com Signed-off-by: Alexander Grund theflamefire89@gmail.com --- include/linux/lsm_hooks.h | 7 ------- include/linux/security.h | 6 ------ kernel/exit.c | 19 ++----------------- security/security.c | 6 ------ security/selinux/hooks.c | 6 ------ security/smack/smack_lsm.c | 20 -------------------- 6 files changed, 2 insertions(+), 62 deletions(-)
diff --git a/include/linux/lsm_hooks.h b/include/linux/lsm_hooks.h index 53ac461f342b..491385a8a69d 100644 --- a/include/linux/lsm_hooks.h +++ b/include/linux/lsm_hooks.h @@ -666,11 +666,6 @@ * @sig contains the signal value. * @secid contains the sid of the process where the signal originated * Return 0 if permission is granted. - * @task_wait: - * Check permission before allowing a process to reap a child process @p - * and collect its status information. - * @p contains the task_struct for process. - * Return 0 if permission is granted. * @task_prctl: * Check permission before performing a process control operation on the * current process. @@ -1507,7 +1502,6 @@ union security_list_options { int (*task_movememory)(struct task_struct *p); int (*task_kill)(struct task_struct *p, struct siginfo *info, int sig, u32 secid); - int (*task_wait)(struct task_struct *p); int (*task_prctl)(int option, unsigned long arg2, unsigned long arg3, unsigned long arg4, unsigned long arg5); void (*task_to_inode)(struct task_struct *p, struct inode *inode); @@ -1768,7 +1762,6 @@ struct security_hook_heads { struct list_head task_getscheduler; struct list_head task_movememory; struct list_head task_kill; - struct list_head task_wait; struct list_head task_prctl; struct list_head task_to_inode; struct list_head ipc_permission; diff --git a/include/linux/security.h b/include/linux/security.h index 2f5d282bd3ec..472822a1e02b 100644 --- a/include/linux/security.h +++ b/include/linux/security.h @@ -332,7 +332,6 @@ int security_task_getscheduler(struct task_struct *p); int security_task_movememory(struct task_struct *p); int security_task_kill(struct task_struct *p, struct siginfo *info, int sig, u32 secid); -int security_task_wait(struct task_struct *p); int security_task_prctl(int option, unsigned long arg2, unsigned long arg3, unsigned long arg4, unsigned long arg5); void security_task_to_inode(struct task_struct *p, struct inode *inode); @@ -980,11 +979,6 @@ static inline int security_task_kill(struct task_struct *p, return 0; }
-static inline int security_task_wait(struct task_struct *p) -{ - return 0; -} - static inline int security_task_prctl(int option, unsigned long arg2, unsigned long arg3, unsigned long arg4, diff --git a/kernel/exit.c b/kernel/exit.c index 8716f0780fe3..e0db254a405b 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -14,7 +14,6 @@ #include <linux/tty.h> #include <linux/iocontext.h> #include <linux/key.h> -#include <linux/security.h> #include <linux/cpu.h> #include <linux/acct.h> #include <linux/tsacct_kern.h> @@ -1342,7 +1341,7 @@ static int wait_task_continued(struct wait_opts *wo, struct task_struct *p) * Returns nonzero for a final return, when we have unlocked tasklist_lock. * Returns zero if the search for a child should continue; * then ->notask_error is 0 if @p is an eligible child, - * or another error from security_task_wait(), or still -ECHILD. + * or still -ECHILD. */ static int wait_consider_task(struct wait_opts *wo, int ptrace, struct task_struct *p) @@ -1362,20 +1361,6 @@ static int wait_consider_task(struct wait_opts *wo, int ptrace, if (!ret) return ret;
- ret = security_task_wait(p); - if (unlikely(ret < 0)) { - /* - * If we have not yet seen any eligible child, - * then let this error code replace -ECHILD. - * A permission error will give the user a clue - * to look for security policy problems, rather - * than for mysterious wait bugs. - */ - if (wo->notask_error) - wo->notask_error = ret; - return 0; - } - if (unlikely(exit_state == EXIT_TRACE)) { /* * ptrace == 0 means we are the natural parent. In this case @@ -1468,7 +1453,7 @@ static int wait_consider_task(struct wait_opts *wo, int ptrace, * Returns nonzero for a final return, when we have unlocked tasklist_lock. * Returns zero if the search for a child should continue; then * ->notask_error is 0 if there were any eligible children, - * or another error from security_task_wait(), or still -ECHILD. + * or still -ECHILD. */ static int do_wait_thread(struct wait_opts *wo, struct task_struct *tsk) { diff --git a/security/security.c b/security/security.c index 9a13d72a6446..5171c3cd1d30 100644 --- a/security/security.c +++ b/security/security.c @@ -1032,11 +1032,6 @@ int security_task_kill(struct task_struct *p, struct siginfo *info, return call_int_hook(task_kill, 0, p, info, sig, secid); }
-int security_task_wait(struct task_struct *p) -{ - return call_int_hook(task_wait, 0, p); -} - int security_task_prctl(int option, unsigned long arg2, unsigned long arg3, unsigned long arg4, unsigned long arg5) { @@ -1776,7 +1771,6 @@ struct security_hook_heads security_hook_heads = { .task_movememory = LIST_HEAD_INIT(security_hook_heads.task_movememory), .task_kill = LIST_HEAD_INIT(security_hook_heads.task_kill), - .task_wait = LIST_HEAD_INIT(security_hook_heads.task_wait), .task_prctl = LIST_HEAD_INIT(security_hook_heads.task_prctl), .task_to_inode = LIST_HEAD_INIT(security_hook_heads.task_to_inode), diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index eb9e2b4e81d9..eb503eccbacc 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -3951,11 +3951,6 @@ static int selinux_task_kill(struct task_struct *p, struct siginfo *info, return rc; }
-static int selinux_task_wait(struct task_struct *p) -{ - return task_has_perm(p, current, PROCESS__SIGCHLD); -} - static void selinux_task_to_inode(struct task_struct *p, struct inode *inode) { @@ -6220,7 +6215,6 @@ static struct security_hook_list selinux_hooks[] = { LSM_HOOK_INIT(task_getscheduler, selinux_task_getscheduler), LSM_HOOK_INIT(task_movememory, selinux_task_movememory), LSM_HOOK_INIT(task_kill, selinux_task_kill), - LSM_HOOK_INIT(task_wait, selinux_task_wait), LSM_HOOK_INIT(task_to_inode, selinux_task_to_inode),
LSM_HOOK_INIT(ipc_permission, selinux_ipc_permission), diff --git a/security/smack/smack_lsm.c b/security/smack/smack_lsm.c index 84ed47195cdd..f01b69ead47e 100644 --- a/security/smack/smack_lsm.c +++ b/security/smack/smack_lsm.c @@ -2276,25 +2276,6 @@ static int smack_task_kill(struct task_struct *p, struct siginfo *info, return rc; }
-/** - * smack_task_wait - Smack access check for waiting - * @p: task to wait for - * - * Returns 0 - */ -static int smack_task_wait(struct task_struct *p) -{ - /* - * Allow the operation to succeed. - * Zombies are bad. - * In userless environments (e.g. phones) programs - * get marked with SMACK64EXEC and even if the parent - * and child shouldn't be talking the parent still - * may expect to know when the child exits. - */ - return 0; -} - /** * smack_task_to_inode - copy task smack into the inode blob * @p: task to copy from @@ -4686,7 +4667,6 @@ static struct security_hook_list smack_hooks[] = { LSM_HOOK_INIT(task_getscheduler, smack_task_getscheduler), LSM_HOOK_INIT(task_movememory, smack_task_movememory), LSM_HOOK_INIT(task_kill, smack_task_kill), - LSM_HOOK_INIT(task_wait, smack_task_wait), LSM_HOOK_INIT(task_to_inode, smack_task_to_inode),
LSM_HOOK_INIT(ipc_permission, smack_ipc_permission),
Hi Greg,
after the previous discussion about what kind of patches are acceptable for stable and your hints on how to send them to the ML in https://lore.kernel.org/all/YsrfDfe3urGkepvJ@kroah.com/ I'd like to know if this patch meets the requirements and if it can be considered.
I do have a few more similar ones which I think meet the stable requirements and finally the init-cleanup patch (upstream 3dfc9b02864bt "LSM: Initialize security_hook_heads upon registration.") which I'd like to backport to 4.9. But first I want to know whether I now got the formal requirements right before sending further patches.
Thanks, Alex
On 11.07.22 11:56, Alexander Grund wrote:
From: Stephen Smalley sds@tycho.nsa.gov
commit 3a2f5a59a695a73e0cde9a61e0feae5fa730e936 upstream.
As reported by yangshukui, a permission denial from security_task_wait() can lead to a soft lockup in zap_pid_ns_processes() since it only expects sys_wait4() to return 0 or -ECHILD. Further, security_task_wait() can in general lead to zombies; in the absence of some way to automatically reparent a child process upon a denial, the hook is not useful. Remove the security hook and its implementations in SELinux and Smack. Smack already removed its check from its hook.
<snip>
A: http://en.wikipedia.org/wiki/Top_post Q: Were do I find info about this thing called top-posting? A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail?
A: No. Q: Should I include quotations after my reply?
http://daringfireball.net/2007/07/on_top
On Sat, Jul 23, 2022 at 10:37:10AM +0200, Alexander Grund wrote:
Hi Greg,
after the previous discussion about what kind of patches are acceptable for stable and your hints on how to send them to the ML in https://lore.kernel.org/all/YsrfDfe3urGkepvJ@kroah.com/ I'd like to know if this patch meets the requirements and if it can be considered.
Yes, it is now queued up. It was delayed due to the RETBLEED mess.
I do have a few more similar ones which I think meet the stable requirements and finally the init-cleanup patch (upstream 3dfc9b02864bt "LSM: Initialize security_hook_heads upon registration.") which I'd like to backport to 4.9. But first I want to know whether I now got the formal requirements right before sending further patches.
Let's see what the other patches look like, I can't guarantee anything without reviewing them first.
thanks,
greg k-h
linux-stable-mirror@lists.linaro.org
-
Alexander Grund -
Greg KH