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] [day] [month] [year] [list]
Message-ID: <Yj3p6etC9Wr/gvr5@pevik>
Date:   Fri, 25 Mar 2022 17:12:25 +0100
From:   Petr Vorel <pvorel@...e.cz>
To:     Christian Brauner <brauner@...nel.org>
Cc:     linux-kernel@...r.kernel.org,
        Christian Brauner <christian@...uner.io>,
        Alexey Gladkov <legion@...nel.org>,
        "Eric W . Biederman" <ebiederm@...ssion.com>,
        "Serge E . Hallyn" <serge@...lyn.com>,
        Vasily Averin <vvs@...tuozzo.com>,
        Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: [PATCH 1/1] ns: Move MAX_PID_NS_LEVEL to ns_common.h, reuse it

Hi all,

> On Fri, Mar 25, 2022 at 12:21:27PM +0100, Petr Vorel wrote:
> > Move MAX_PID_NS_LEVEL to ns_common.h and reuse it in check in
> > user_namespace.c.

> > Signed-off-by: Petr Vorel <pvorel@...e.cz>
> > ---
> > Hi Christian, all,

> > I don't see putting MAX_PID_NS_LEVEL into ns_common.h as an elegant
> > solution but IMHO better than use a hardwired number or redefinition in
> > user_namespace.h.

> Hey Petr,

> Weird I either deleted that message by accident or didn't get it.

Interesting. BTW I didn't get any bounce suggesting mail to you didn't arrive.

> > Kind regards,
> > Petr

> >  include/linux/ns_common.h     | 3 +++
> >  include/linux/pid_namespace.h | 3 ---
> >  kernel/user_namespace.c       | 2 +-
> >  3 files changed, 4 insertions(+), 4 deletions(-)

> > diff --git a/include/linux/ns_common.h b/include/linux/ns_common.h
> > index 0f1d024bd958..173fab9dadf7 100644
> > --- a/include/linux/ns_common.h
> > +++ b/include/linux/ns_common.h
> > @@ -4,6 +4,9 @@

> >  #include <linux/refcount.h>

> > +/* MAX_PID_NS_LEVEL is needed for limiting size of 'struct pid' */
> > +#define MAX_PID_NS_LEVEL 32

> The only two namespaces where this applies to have the same hard-coded
> limit. If you want to get rid of the raw number you should just
> introduce a generic define that expresses the limit for both pidns
> and userns. Using "MAX_PID_NS_LEVEL" in the userns case is very
> confusing. So if you wanted to do this introducing something like
> #define MAX_NS_LEVEL 32
> and using it in both places makes more sense.

Thanks a lot, I overlooked they aren't related.

I wonder if there should be just one constant for both (i.e. MAX_NS_LEVEL) as
you suggest, 2 constants, i.e keep MAX_PID_NS_LEVEL and add MAX_USER_NS_LEVEL
(which happen to be both 32).

Also understand if you prefer just to keep the raw number (MAX_PID_NS_LEVEL has
more use than just a single place, but user namespaces need it just on single
place).

Kind regards,
Petr

> Christian

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ