linux-kernel - Re: [PATCH v4 13/25] virtio

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20141020140410.GA10994@redhat.com>
Date:	Mon, 20 Oct 2014 17:04:11 +0300
From:	"Michael S. Tsirkin" <mst@...hat.com>
To:	Cornelia Huck <cornelia.huck@...ibm.com>
Cc:	Thomas Graf <tgraf@...g.ch>, linux-kernel@...r.kernel.org,
	Rusty Russell <rusty@...tcorp.com.au>,
	virtualization@...ts.linux-foundation.org,
	linux-scsi@...r.kernel.org, linux-s390@...r.kernel.org,
	v9fs-developer@...ts.sourceforge.net, netdev@...r.kernel.org,
	kvm@...r.kernel.org, Amit Shah <amit.shah@...hat.com>,
	Christian Borntraeger <borntraeger@...ibm.com>,
	"David S. Miller" <davem@...emloft.net>,
	Paolo Bonzini <pbonzini@...hat.com>,
	Arnd Bergmann <arnd@...db.de>,
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Subject: Re: [PATCH v4 13/25] virtio_console: enable VQs early

On Mon, Oct 20, 2014 at 04:35:55PM +0300, Michael S. Tsirkin wrote:
> On Mon, Oct 20, 2014 at 02:42:23PM +0200, Cornelia Huck wrote:
> > On Mon, 20 Oct 2014 13:07:50 +0100
> > Thomas Graf <tgraf@...g.ch> wrote:
> > 
> > > On 10/13/14 at 10:50am, Michael S. Tsirkin wrote:
> > > > virtio spec requires drivers to set DRIVER_OK before using VQs.
> > > > This is set automatically after probe returns, virtio console violated this
> > > > rule by adding inbufs, which causes the VQ to be used directly within
> > > > probe.
> > > > 
> > > > To fix, call virtio_device_ready before using VQs.
> > > > 
> > > > Signed-off-by: Michael S. Tsirkin <mst@...hat.com>
> > > > ---
> > > >  drivers/char/virtio_console.c | 2 ++
> > > >  1 file changed, 2 insertions(+)
> > > > 
> > > > diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c
> > > > index b585b47..6ebe8f6 100644
> > > > --- a/drivers/char/virtio_console.c
> > > > +++ b/drivers/char/virtio_console.c
> > > > @@ -1449,6 +1449,8 @@ static int add_port(struct ports_device *portdev, u32 id)
> > > >  	spin_lock_init(&port->outvq_lock);
> > > >  	init_waitqueue_head(&port->waitqueue);
> > > >  
> > > > +	virtio_device_ready(portdev->vdev);
> > > > +
> > > >  	/* Fill the in_vq with buffers so the host can send us data. */
> > > >  	nr_added_bufs = fill_queue(port->in_vq, &port->inbuf_lock);
> > > >  	if (!nr_added_bufs) {
> > > 
> > > Seems like probe and add_port() now both set VIRTIO_CONFIG_S_DRIVER_OK
> > 
> > I think we need to set this in the probe function instead, otherwise we
> > fail for multiqueue (which also wants to use the control queue early).
> > 
> > Completely untested patch below; I can send this with proper s-o-b if
> > it helps.
> > 
> > diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c
> > index bfa6400..cf7a561 100644
> > --- a/drivers/char/virtio_console.c
> > +++ b/drivers/char/virtio_console.c
> > @@ -1449,8 +1449,6 @@ static int add_port(struct ports_device *portdev, u32 id)
> >  	spin_lock_init(&port->outvq_lock);
> >  	init_waitqueue_head(&port->waitqueue);
> >  
> > -	virtio_device_ready(portdev->vdev);
> > -
> >  	/* Fill the in_vq with buffers so the host can send us data. */
> >  	nr_added_bufs = fill_queue(port->in_vq, &port->inbuf_lock);
> >  	if (!nr_added_bufs) {
> > @@ -2026,6 +2024,8 @@ static int virtcons_probe(struct virtio_device *vdev)
> >  	spin_lock_init(&portdev->ports_lock);
> >  	INIT_LIST_HEAD(&portdev->ports);
> >  
> > +	virtio_device_ready(portdev->vdev);
> > +
> >  	if (multiport) {
> >  		unsigned int nr_added_bufs;
> >  
> 
> I wanted to set DRIVER_OK as late as possible, to both reduce
> the chance it can fail after DRIVER_OK and to reduce  the risk of
> introducing a regression since old qemu might only start sending
> interrupts after DRIVER_OK is set.
> 
> So I wanted everything completely initialized before DRIVER_OK.
> 
> You patch makes sense to me since nothing can fail,
> and we won't get interrupts before we add inbufs.
> 
> Reviewed-by: Michael S. Tsirkin <mst@...hat.com>
> 
> Testig will report shortly, pls send with sob line.

Sure enough, this helps:

Tested-by: Michael S. Tsirkin <mst@...hat.com>
Acked-by: Michael S. Tsirkin <mst@...hat.com>

Pls repost as a top-level patch.

> -- 
> MST
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/