[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20191023232708.GA433@redhat.com>
Date: Wed, 23 Oct 2019 19:27:08 -0400
From: Andrea Arcangeli <aarcange@...hat.com>
To: Andy Lutomirski <luto@...capital.net>
Cc: Andy Lutomirski <luto@...nel.org>, Jann Horn <jannh@...gle.com>,
Daniel Colascione <dancol@...gle.com>,
Linus Torvalds <torvalds@...ux-foundation.org>,
Pavel Emelyanov <xemul@...tuozzo.com>,
Lokesh Gidra <lokeshgidra@...gle.com>,
Nick Kralevich <nnk@...gle.com>,
Nosh Minwalla <nosh@...gle.com>,
Tim Murray <timmurray@...gle.com>,
Mike Rapoport <rppt@...ux.vnet.ibm.com>,
Linux API <linux-api@...r.kernel.org>,
LKML <linux-kernel@...r.kernel.org>,
"Dr. David Alan Gilbert" <dgilbert@...hat.com>
Subject: Re: [PATCH 3/7] Add a UFFD_SECURE flag to the userfaultfd API.
On Wed, Oct 23, 2019 at 04:01:53PM -0700, Andy Lutomirski wrote:
> Delivering events through read() is just fine. The problem is when
> delivering an event does more than just returning bytes. As far as
> I’ve noticed, uffd’s read() just returns bytes as long as FORK is
> disabled.
Yes, fork is the only case where read doesn't return bytes.
Moving only the fd creation to a separate syscall would then avoid
involuntary creation of the fd.
Powered by blists - more mailing lists