[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200325190429.73k52amlfjer7epa@comp-core-i7-2640m-0182e6>
Date: Wed, 25 Mar 2020 20:04:29 +0100
From: Alexey Gladkov <gladkov.alexey@...il.com>
To: Alexey Dobriyan <adobriyan@...il.com>
Cc: Linus Torvalds <torvalds@...ux-foundation.org>,
LKML <linux-kernel@...r.kernel.org>,
Kernel Hardening <kernel-hardening@...ts.openwall.com>,
Linux API <linux-api@...r.kernel.org>,
Linux FS Devel <linux-fsdevel@...r.kernel.org>,
Linux Security Module <linux-security-module@...r.kernel.org>,
Akinobu Mita <akinobu.mita@...il.com>,
Alexander Viro <viro@...iv.linux.org.uk>,
Andrew Morton <akpm@...ux-foundation.org>,
Andy Lutomirski <luto@...nel.org>,
Daniel Micay <danielmicay@...il.com>,
Djalal Harouni <tixxdz@...il.com>,
"Dmitry V . Levin" <ldv@...linux.org>,
"Eric W . Biederman" <ebiederm@...ssion.com>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Ingo Molnar <mingo@...nel.org>,
"J . Bruce Fields" <bfields@...ldses.org>,
Jeff Layton <jlayton@...chiereds.net>,
Jonathan Corbet <corbet@....net>,
Kees Cook <keescook@...omium.org>,
Oleg Nesterov <oleg@...hat.com>
Subject: Re: [PATCH RESEND v9 3/8] proc: move hide_pid, pid_gid from
pid_namespace to proc_fs_info
On Wed, Mar 25, 2020 at 09:00:15PM +0300, Alexey Dobriyan wrote:
> On Tue, Mar 24, 2020 at 02:21:59PM -0700, Linus Torvalds wrote:
> > On Tue, Mar 24, 2020 at 1:46 PM Alexey Gladkov <gladkov.alexey@...il.com> wrote:
> > >
> > > +/* definitions for hide_pid field */
> > > +enum {
> > > + HIDEPID_OFF = 0,
> > > + HIDEPID_NO_ACCESS = 1,
> > > + HIDEPID_INVISIBLE = 2,
> > > +};
> >
> > Should this enum be named...
> >
> > > struct proc_fs_info {
> > > struct pid_namespace *pid_ns;
> > > struct dentry *proc_self; /* For /proc/self */
> > > struct dentry *proc_thread_self; /* For /proc/thread-self */
> > > + kgid_t pid_gid;
> > > + int hide_pid;
> > > };
> >
> > .. and then used here instead of "int"?
> >
> > Same goes for 'struct proc_fs_context' too, for that matter?
> >
> > And maybe in the function declarations and definitions too? In things
> > like 'has_pid_permissions()' (the series adds some other cases later,
> > like hidepid2str() etc)
> >
> > Yeah, enums and ints are kind of interchangeable in C, but even if it
> > wouldn't give us any more typechecking (except perhaps with sparse if
> > you mark it so), it would be documenting the use.
> >
> > Or am I missing something?
> >
> > Anyway, I continue to think the series looks fine, bnut would love to
> > see it in -next and perhaps comments from Al and Alexey Dobriyan..
>
> Patches are OK, except the part where "pid" is named "pidfs" and
> the suffix doesn't convey any information.
I will fix this in the final version.
> mount -t proc -o subset=pid,sysctl,misc
I have not yet figured out how to implement this. I mean subset=meminfo,misc.
--
Rgrds, legion
Powered by blists - more mailing lists