[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200527142143.GC1194141@xz-x1>
Date: Wed, 27 May 2020 10:21:43 -0400
From: Peter Xu <peterx@...hat.com>
To: Xiaoming Ni <nixiaoming@...wei.com>
Cc: aarcange@...hat.com, akpm@...ux-foundation.org, cracauer@...s.org,
dplotnikov@...tuozzo.com, gokhale2@...l.gov, hannes@...xchg.org,
hughd@...gle.com, jglisse@...hat.com, kirill@...temov.name,
linux-kernel@...r.kernel.org, linux-mm@...ck.org,
maxime.coquelin@...hat.com, mcfadden8@...l.gov, mcgrof@...nel.org,
mgorman@...e.de, mike.kravetz@...cle.com, pbonzini@...hat.com,
rppt@...ux.vnet.ibm.com, xemul@...tuozzo.com, keescook@...omium.org
Subject: Re: [PATCH v2 1/1] userfaultfd/sysctl: add
vm.unprivileged_userfaultfd
On Wed, May 27, 2020 at 02:54:13PM +0800, Xiaoming Ni wrote:
>
> On Tue, Mar 19, 2019 at 11:07:22AM +0800, Peter Xu wrote:
> > Add a global sysctl knob "vm.unprivileged_userfaultfd" to control
> > whether userfaultfd is allowed by unprivileged users. When this is
> > set to zero, only privileged users (root user, or users with the
> > CAP_SYS_PTRACE capability) will be able to use the userfaultfd
> > syscalls.
>
> Hello
Hi, Xiaoming,
> I am a bit confused about this patch, can you help to answer it.
>
> Why the sysctl interface of fs/userfaultfd.c belongs to vm_table instead of
> fs_table ?
>
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=cefdca0a86be517bc390fc4541e3674b8e7803b0
Because I think it makes more sense to put the new key into where it suites
better, irrelevant to which directory the variable is declared. To me,
unprivileged_userfaultfd is definitely more suitable for vm rather than fs,
because userfaultfd is really about memory management rather than file system.
Thanks,
--
Peter Xu
Powered by blists - more mailing lists