On Thu, 2010-08-12 at 01:37 +0100, Xu, Dongxiao wrote:
> Hi Ian,
>
> Currently logic (2) is not needed to make s/r work, since sring is set
> to NULL after resume. However I think this check could avoid error if
> later any code sets sring to NULL in other places. So I still place it
> there.
OK, that's fair enough.
Ian.
>
> Thanks,
> Dongxiao
>
> ________________________________________
> From: Ian Campbell [Ian.Campbell@xxxxxxxxxxxxx]
> Sent: Wednesday, August 11, 2010 8:01 AM
> To: Xu, Dongxiao
> Cc: Jeremy Fitzhardinge; Ian Jackson; Xen-devel@xxxxxxxxxxxxxxxxxxx
> Subject: RE: [Xen-devel] [xen-unstable test] 1959: FAIL [and 1 more messages]
>
> On Wed, 2010-08-11 at 04:19 +0100, Xu, Dongxiao wrote:
> > From 791fc8ed6e2888af3bd398f22562776c757cc4ac Mon Sep 17 00:00:00 2001
> > From: Dongxiao Xu <dongxiao.xu@xxxxxxxxx>
> > Date: Wed, 11 Aug 2010 11:06:06 +0800
> > Subject: [PATCH] Netfront: Fix save/restore after enabled smart poll
> > feature
> >
> > When s/r the guest, the shared ring will be set to NULL, and
> > in this case the polling timer interrupt should stop.
> >
> > This fix includes the two parts:
> > 1) Stop hrtimer when guest suspends.
> > 2) Add check to avoid NULL pointer dereference.
>
> Looks good to me but do you really need 2) given 1)?
>
> Ian.
>
>
> >
> > Thanks Ian J for reporting the problem.
> >
> > Signed-off-by: Dongxiao Xu <dongxiao.xu@xxxxxxxxx>
> > ---
> > drivers/net/xen-netfront.c | 13 +++++++++++++
> > 1 files changed, 13 insertions(+), 0 deletions(-)
> >
> > diff --git a/drivers/net/xen-netfront.c b/drivers/net/xen-netfront.c
> > index 328fe40..eaea132 100644
> > --- a/drivers/net/xen-netfront.c
> > +++ b/drivers/net/xen-netfront.c
> > @@ -1289,6 +1289,14 @@ static void xennet_disconnect_backend(struct
> > netfront_info *info)
> > info->rx.sring = NULL;
> > }
> >
> > +static int netfront_suspend(struct xenbus_device *dev, pm_message_t
> > state)
> > +{
> > + struct netfront_info *info = dev_get_drvdata(&dev->dev);
> > + struct hrtimer *timer = &info->smart_poll.timer;
> > + hrtimer_cancel(timer);
> > + return 0;
> > +}
> > +
> > /**
> > * We are reconnecting to the backend, due to a suspend/resume, or a
> > backend
> > * driver restart. We tear down our netif structure and recreate it,
> > but
> > @@ -1340,6 +1348,10 @@ static enum hrtimer_restart
> > smart_poll_function(struct hrtimer *timer)
> > np = netdev_priv(dev);
> >
> > spin_lock_irqsave(&np->tx_lock, flags);
> > +
> > + if (!np->rx.sring)
> > + goto end;
> > +
> > np->smart_poll.counter++;
> >
> > if (likely(netif_carrier_ok(dev))) {
> > @@ -1910,6 +1922,7 @@ static struct xenbus_driver netfront_driver = {
> > .ids = netfront_ids,
> > .probe = netfront_probe,
> > .remove = __devexit_p(xennet_remove),
> > + .suspend = netfront_suspend,
> > .resume = netfront_resume,
> > .otherend_changed = netback_changed,
> > };
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|
Home