linux-kernel - Re: [PATCH] clone.2: Mention that CLONE

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 21 Nov 2019 10:53:50 +0100
From:   "Michael Kerrisk (man-pages)" <mtk.manpages@...il.com>
To:     Christian Brauner <christian@...uner.io>
Cc:     mtk.manpages@...il.com, adrian@...as.de, akpm@...ux-foundation.org,
        arnd@...db.de, avagin@...il.com, christian.brauner@...ntu.com,
        dhowells@...hat.com, fweimer@...hat.com, jannh@...gle.com,
        keescook@...omium.org, linux-api@...r.kernel.org,
        linux-kernel@...r.kernel.org, linux-man@...r.kernel.org,
        mingo@...e.hu, oleg@...hat.com, xemul@...tuozzo.com
Subject: Re: [PATCH] clone.2: Mention that CLONE_PARENT is off-limits for
 inits

Hello Christian,

On 11/20/19 11:45 AM, Christian Brauner wrote:
> From: Christian Brauner <christian.brauner@...ntu.com>
> 
> The CLONE_PARENT flag cannot but used by init processes. Let's mention
> this in the manpages to prevent suprises.
> 
> Signed-off-by: Christian Brauner <christian.brauner@...ntu.com>
> ---
>  man2/clone.2 | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/man2/clone.2 b/man2/clone.2
> index f0f29d6f1..aa98ab79b 100644
> --- a/man2/clone.2
> +++ b/man2/clone.2
> @@ -646,6 +646,13 @@ if
>  .B CLONE_PARENT
>  is set, then the parent of the calling process, rather than the
>  calling process itself, will be signaled.
> +.IP
> +The kernel will not allow global init and init processes in pid
> +namespaces to use the
> +.B CLONE_PARENT
> +flag. This is done to prevent the creation of multi-rooted process
> +trees. It also avoids unreapable zombies in the initial pid
> +namespace.
>  .TP
>  .BR CLONE_PARENT_SETTID " (since Linux 2.5.49)"
>  Store the child thread ID at the location pointed to by

Thank. I applied, and then tweaked the text a little,
and noted the associated EINVAL error. In the end, the
change is as below.

Cheers,

Michael

index 60e746151..382f6b791 100644
--- a/man2/clone.2
+++ b/man2/clone.2
@@ -648,6 +648,14 @@ if
 .B CLONE_PARENT
 is set, then the parent of the calling process, rather than the
 calling process itself, will be signaled.
+.IP
+The
+.B CLONE_PARENT
+flag can't be used in clone calls by the
+global init process (PID 1 in the initial PID namespace)
+and init processes in other PID namespaces.
+This restriction prevents the creation of multi-rooted process trees
+as well as the creation of unreapable zombies in the initial PID namespace.
 .TP
 .BR CLONE_PARENT_SETTID " (since Linux 2.5.49)"
 Store the child thread ID at the location pointed to by
@@ -1273,6 +1281,11 @@ were specified in the
 .IR flags
 mask.
 .TP
+.BR EINVAL " (since Linux 2.6.32)"
+.\" commit 123be07b0b399670a7cc3d82fef0cb4f93ef885c
+.BR CLONE_PARENT
+was specified, and the caller is an init process.
+.TP
 .B EINVAL
 Returned by the glibc
 .BR clone ()



-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/