From: Peter Zijlstra <peterz@infradead.org>
To: Mark Rutland <mark.rutland@arm.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>,
KVM list <kvm@vger.kernel.org>,
"Michael S. Tsirkin" <mst@redhat.com>,
netdev <netdev@vger.kernel.org>,
Boqun Feng <boqun.feng@gmail.com>,
dbueso@suse.de, LKML <linux-kernel@vger.kernel.org>,
virtualization@lists.linux-foundation.org,
Paul McKenney <paulmck@linux.vnet.ibm.com>,
Linus Torvalds <torvalds@linux-foundation.org>,
Dmitry Vyukov <dvyukov@google.com>
Subject: Re: [PATCH 0/3] virtio/vringh: kill off ACCESS_ONCE()
Date: Fri, 25 Nov 2016 18:42:23 +0100 [thread overview]
Message-ID: <20161125174223.GS3092@twins.programming.kicks-ass.net> (raw)
In-Reply-To: <20161125172624.GA30811@leverpostej>
On Fri, Nov 25, 2016 at 05:28:01PM +0000, Mark Rutland wrote:
> On Fri, Nov 25, 2016 at 05:49:45PM +0100, Christian Borntraeger wrote:
> > On 11/25/2016 05:17 PM, Peter Zijlstra wrote:
> > There were several cases that I found during writing the *ONCE stuff.
> > For example there are some 32bit ppc variants with 64bit PTEs. Some for
> > others (I think sparc).
>
> We have similar on 32-bit ARM w/ LPAE. LPAE implies that a naturally
> aligned 64-bit access is single-copy atomic, which is what makes that
> ok.
>
> > And the mm/ code is perfectly fine with these PTE accesses being done
> > NOT atomic.
>
> That strikes me as surprising. Is there some mutual exclusion that
> prevents writes from occuring wherever a READ_ONCE() happens to a PTE?
>
> Otherwise, how is tearing not a problem? Does it have some pattern like
> the lockref cmpxchg?
On x86 PAE we play silly games, see arch/x86/mm/gup.c:gup_get_ptr().
Those two loads really should be READ_ONCE()/LOAD_SINGLE().
WARNING: multiple messages have this Message-ID (diff)
From: Peter Zijlstra <peterz@infradead.org>
To: Mark Rutland <mark.rutland@arm.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>,
Dmitry Vyukov <dvyukov@google.com>,
Boqun Feng <boqun.feng@gmail.com>,
"Michael S. Tsirkin" <mst@redhat.com>,
LKML <linux-kernel@vger.kernel.org>,
Davidlohr Bueso <dave@stgolabs.net>,
dbueso@suse.de, jasowang@redhat.com,
KVM list <kvm@vger.kernel.org>, netdev <netdev@vger.kernel.org>,
Paul McKenney <paulmck@linux.vnet.ibm.com>,
virtualization@lists.linux-foundation.org,
Linus Torvalds <torvalds@linux-foundation.org>
Subject: Re: [PATCH 0/3] virtio/vringh: kill off ACCESS_ONCE()
Date: Fri, 25 Nov 2016 18:42:23 +0100 [thread overview]
Message-ID: <20161125174223.GS3092@twins.programming.kicks-ass.net> (raw)
In-Reply-To: <20161125172624.GA30811@leverpostej>
On Fri, Nov 25, 2016 at 05:28:01PM +0000, Mark Rutland wrote:
> On Fri, Nov 25, 2016 at 05:49:45PM +0100, Christian Borntraeger wrote:
> > On 11/25/2016 05:17 PM, Peter Zijlstra wrote:
> > There were several cases that I found during writing the *ONCE stuff.
> > For example there are some 32bit ppc variants with 64bit PTEs. Some for
> > others (I think sparc).
>
> We have similar on 32-bit ARM w/ LPAE. LPAE implies that a naturally
> aligned 64-bit access is single-copy atomic, which is what makes that
> ok.
>
> > And the mm/ code is perfectly fine with these PTE accesses being done
> > NOT atomic.
>
> That strikes me as surprising. Is there some mutual exclusion that
> prevents writes from occuring wherever a READ_ONCE() happens to a PTE?
>
> Otherwise, how is tearing not a problem? Does it have some pattern like
> the lockref cmpxchg?
On x86 PAE we play silly games, see arch/x86/mm/gup.c:gup_get_ptr().
Those two loads really should be READ_ONCE()/LOAD_SINGLE().
next prev parent reply other threads:[~2016-11-25 17:42 UTC|newest]
Thread overview: 68+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-11-24 10:25 [PATCH 0/3] virtio/vringh: kill off ACCESS_ONCE() Mark Rutland
2016-11-24 10:25 ` [PATCH 1/3] tools/virtio: fix READ_ONCE() Mark Rutland
2016-11-24 11:34 ` Cornelia Huck
2016-11-24 11:34 ` Cornelia Huck
2016-11-25 2:38 ` Jason Wang
2016-11-25 2:38 ` Jason Wang
2016-11-24 10:25 ` Mark Rutland
2016-11-24 10:25 ` [PATCH 2/3] vringh: kill off ACCESS_ONCE() Mark Rutland
2016-11-24 10:25 ` Mark Rutland
2016-11-24 11:10 ` Christian Borntraeger
2016-11-24 11:10 ` Christian Borntraeger
2016-11-24 11:35 ` Cornelia Huck
2016-11-24 11:35 ` Cornelia Huck
2016-11-25 2:40 ` Jason Wang
2016-11-25 2:40 ` Jason Wang
2016-11-24 10:25 ` [PATCH 3/3] tools/virtio: use {READ,WRITE}_ONCE() in uaccess.h Mark Rutland
2016-11-24 11:37 ` Cornelia Huck
2016-11-24 11:37 ` Cornelia Huck
2016-11-25 2:40 ` Jason Wang
2016-11-25 2:40 ` Jason Wang
2016-11-24 10:25 ` Mark Rutland
2016-11-24 20:36 ` [PATCH 0/3] virtio/vringh: kill off ACCESS_ONCE() Michael S. Tsirkin
2016-11-24 20:36 ` Michael S. Tsirkin
2016-11-25 11:22 ` Mark Rutland
2016-11-25 11:22 ` Mark Rutland
2016-11-25 11:33 ` Christian Borntraeger
2016-11-25 11:33 ` Christian Borntraeger
2016-11-25 12:23 ` Mark Rutland
2016-11-25 12:23 ` Mark Rutland
2016-11-25 12:40 ` Peter Zijlstra
2016-11-25 12:40 ` Peter Zijlstra
2016-11-25 12:44 ` Peter Zijlstra
2016-11-25 12:44 ` Peter Zijlstra
2016-11-25 14:56 ` Boqun Feng
2016-11-25 14:56 ` Boqun Feng
2016-11-25 15:21 ` Dmitry Vyukov via Virtualization
2016-11-25 15:21 ` Dmitry Vyukov
2016-11-25 16:10 ` Mark Rutland
2016-11-25 16:17 ` Peter Zijlstra
2016-11-25 16:17 ` Peter Zijlstra
2016-11-25 16:32 ` Mark Rutland
2016-11-25 16:32 ` Mark Rutland
2016-11-25 16:49 ` Christian Borntraeger
2016-11-25 16:49 ` Christian Borntraeger
2016-11-25 17:28 ` Mark Rutland
2016-11-25 17:28 ` Mark Rutland
2016-11-25 17:42 ` Peter Zijlstra [this message]
2016-11-25 17:42 ` Peter Zijlstra
2016-11-25 18:46 ` Christian Borntraeger
2016-11-25 18:46 ` Christian Borntraeger
2016-11-25 21:08 ` Michael S. Tsirkin
2016-11-25 21:08 ` Michael S. Tsirkin
2016-11-25 21:45 ` Christian Borntraeger
2016-11-25 21:45 ` Christian Borntraeger
2016-11-25 17:28 ` Dmitry Vyukov
2016-11-25 17:43 ` Mark Rutland
2016-11-25 17:43 ` Mark Rutland
2016-11-25 17:52 ` Linus Torvalds
2016-11-25 17:52 ` Linus Torvalds
2016-11-25 18:07 ` Mark Rutland
2016-11-25 18:07 ` Mark Rutland
2016-11-25 18:47 ` Linus Torvalds
2016-11-25 18:47 ` Linus Torvalds
2016-11-25 17:28 ` Dmitry Vyukov via Virtualization
2016-11-25 16:10 ` Mark Rutland
2016-11-25 14:35 ` Mark Rutland
2016-11-25 14:35 ` Mark Rutland
-- strict thread matches above, loose matches on Subject: below --
2016-11-24 10:25 Mark Rutland
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20161125174223.GS3092@twins.programming.kicks-ass.net \
--to=peterz@infradead.org \
--cc=boqun.feng@gmail.com \
--cc=dave@stgolabs.net \
--cc=dbueso@suse.de \
--cc=dvyukov@google.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mark.rutland@arm.com \
--cc=mst@redhat.com \
--cc=netdev@vger.kernel.org \
--cc=paulmck@linux.vnet.ibm.com \
--cc=torvalds@linux-foundation.org \
--cc=virtualization@lists.linux-foundation.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.