Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Mon, 9 Mar 2020 15:51:00 -0400
From:   Peter Xu <peterx@redhat.com>
To:     David Hildenbrand <david@redhat.com>
Cc:     linux-mm@kvack.org, linux-kernel@vger.kernel.org,
        Andrea Arcangeli <aarcange@redhat.com>,
        Martin Cracauer <cracauer@cons.org>,
        Linus Torvalds <torvalds@linux-foundation.org>,
        Mike Rapoport <rppt@linux.vnet.ibm.com>,
        "Kirill A . Shutemov" <kirill@shutemov.name>,
        Johannes Weiner <hannes@cmpxchg.org>,
        "Dr . David Alan Gilbert" <dgilbert@redhat.com>,
        Bobby Powers <bobbypowers@gmail.com>,
        Maya Gokhale <gokhale2@llnl.gov>,
        Jerome Glisse <jglisse@redhat.com>,
        Mike Kravetz <mike.kravetz@oracle.com>,
        Matthew Wilcox <willy@infradead.org>,
        Marty McFadden <mcfadden8@llnl.gov>,
        Mel Gorman <mgorman@suse.de>, Hugh Dickins <hughd@google.com>,
        Brian Geffon <bgeffon@google.com>,
        Denis Plotnikov <dplotnikov@virtuozzo.com>,
        Pavel Emelyanov <xemul@virtuozzo.com>
Subject: Re: [PATCH RESEND v6 00/16] mm: Page fault enhancements
Message-ID: <20200309195100.GD4206@xz-x1>
References: <20200220155353.8676-1-peterx@redhat.com>
 <1eb7bdd4-348f-da87-47a1-0b022b70e918@redhat.com>
 <20200307214743.GA4206@xz-x1>
 <6d8ed084-0740-cee1-663e-a78a2faee432@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <6d8ed084-0740-cee1-663e-a78a2faee432@redhat.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Sun, Mar 08, 2020 at 01:12:34PM +0100, David Hildenbrand wrote:
> [...]
> 
> > Yes, IIUC the race can happen like this in your below test:
> > 
> >      main thread          uffd thread             disgard thread
> >      ===========          ===========             ==============
> >      access page
> >        uffd page fault
> >          wait for page
> >                           UFFDIO_ZEROCOPY
> >                             put a page P there
> >                                                   MADV_DONTNEED on P
> >                             wakeup main thread
> >          return from fault
> >        page still missing
> >        uffd page fault again
> >        (without ALLOW_RETRY)
> >        --> SIGBUS.
> 
> Exactly!
> 
> >> Can we please have a way to identify that this "feature" is available?
> >> I'd appreciate a new read-only UFFD_FEAT_ , so we can detect this from
> >> user space easily and use concurrent discards without crashing our applications.
> > 
> > I'm not sure how others think about it, but to me this still fells
> > into the bucket of "solving an existing problem" rather than a
> > feature.  Also note that this should change the behavior for the page
> > fault logic in general, rather than an uffd-only change. So I'm also
> > not sure whether UFFD_FEAT_* suites here even if we want it.
> 
> So, are we planning on backporting this to stable kernels?

I don't have a plan so far.  I'm still at the phase to only worry
about whether it can be at least merged in master.. :)

I would think it won't worth it to backport this to stables though,
considering that it could potentially change quite a bit for faulting
procedures, and after all the issues we're fixing shouldn't be common
to general users.

> 
> Imagine using this in QEMU/KVM to allow discards (e.g., balloon
> inflation) while postcopy is active . You certainly don't want random
> guest crashes. So either, we treat this as a fix (and backport) or as a
> change in behavior/feature.

I think we don't need to worry on that - QEMU will prohibit ballooning
during postcopy starting from the first day.  Feel free to see QEMU
commit 371ff5a3f04cd7 ("Inhibit ballooning during postcopy").

> 
> [...]
> 
> >>
> >> 2. What will happen if I don't place a page on a pagefault, but only do a UFFDIO_WAKE?
> >>    For now we were able to trigger a signal this way.
> > 
> > If I'm not mistaken the UFFDIO_WAKE will directly trigger the sigbus
> > even without the help of the MADV_DONTNEED race.
> 
> Yes, that's the current way of injecting a SIGBUS instead of resolving
> the pagefault. And AFAIKs, you're changing that behavior. (I am not
> aware of a user, there could be use cases, but somehow it's strange to
> get a signal when accessing memory that is mapped READ|WRITE and also
> represented like this in e.g., /proc/$PID/maps). So IMHO, only the new
> behavior makes really sense.

I agree, I'm not sure how other people think on ABI stability, but...
for my own preference I don't worry much on ABI breakage for a problem
like this.

Thanks,

-- 
Peter Xu