Re: [withdrawn] do_notify_parent-sanitize-the-valid_signal-checks.patch removed from -mm tree

[Oleg Nesterov <oleg@redhat.com>]

On 03/09, Andrew Morton wrote:
>
> The quilt patch titled
>      Subject: do_notify_parent: sanitize the valid_signal() checks
> has been removed from the -mm tree.  Its filename was
>      do_notify_parent-sanitize-the-valid_signal-checks.patch
>
> This patch was dropped because it was withdrawn

OK, thanks!

but then please also drop kernel-fork-validate-exit_signal-in-clone-syscall.patch
from Deepanshu (cc'ed).

Because that patch:

	- is not complete in any case

	- makes no sense (at least with its current changelog) without
	  do_notify_parent-sanitize-the-valid_signal-checks.patch

If Deepanshu sends V2 (with the updated changelog), we can reconsider
both patches from Deepanshu and me.

Oleg.

> ------------------------------------------------------
> From: Oleg Nesterov <oleg@redhat.com>
> Subject: do_notify_parent: sanitize the valid_signal() checks
> Date: Sun, 22 Feb 2026 16:23:47 +0100
> 
> The "sig" argument of do_notify_parent() must always be valid and it does
> WARN_ON_ONCE(sig == -1) at the start.  The 2nd valid_signal() check before
> __send_signal_locked() must always be true or we have a bug.
> 
> This is confusing.  Change do_notify_parent() to WARN and return early if
> valid_signal(sig) is false.
> 
> Link: https://lkml.kernel.org/r/aZsfg0Y055yuAvsq@redhat.com
> Signed-off-by: Oleg Nesterov <oleg@redhat.com>
> Cc: Christian Brauner <brauner@kernel.org>
> Cc: Kees Cook <kees@kernel.org>
> Cc: Mateusz Guzik <mjguzik@gmail.com>
> Cc; Deepanshu Kartikey <kartikey406@gmail.com>
> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
> ---
> 
>  kernel/signal.c |    5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> --- a/kernel/signal.c~do_notify_parent-sanitize-the-valid_signal-checks
> +++ a/kernel/signal.c
> @@ -2173,7 +2173,8 @@ bool do_notify_parent(struct task_struct
>  	bool autoreap = false;
>  	u64 utime, stime;
>  
> -	WARN_ON_ONCE(sig == -1);
> +	if (WARN_ON_ONCE(!valid_signal(sig)))
> +		return false;
>  
>  	/* do_notify_parent_cldstop should have been called instead.  */
>  	WARN_ON_ONCE(task_is_stopped_or_traced(tsk));
> @@ -2254,7 +2255,7 @@ bool do_notify_parent(struct task_struct
>  	 * Send with __send_signal as si_pid and si_uid are in the
>  	 * parent's namespaces.
>  	 */
> -	if (valid_signal(sig) && sig)
> +	if (sig)
>  		__send_signal_locked(sig, &info, tsk->parent, PIDTYPE_TGID, false);
>  	__wake_up_parent(tsk, tsk->parent);
>  	spin_unlock_irqrestore(&psig->siglock, flags);
> _
> 
> Patches currently in -mm which might be from oleg@redhat.com are
> 
> exit-kill-unnecessary-thread_group_leader-checks-in-exit_notify-and-do_notify_parent.patch
> complete_signal-kill-always-true-core_state-signal_group_exit-check.patch
> pid-make-sub-init-creation-retryable.patch
> pid-document-the-pidns_adding-checks-in-alloc_pid-and-copy_process.patch
>