RE: [PATCH 1/1] Drivers: hv: vmbus: Don't leak channel ids

[KY Srinivasan]

> -----Original Message-----
> From: Dexuan Cui
> Sent: Tuesday, March 14, 2017 5:49 AM
> To: KY Srinivasan <kys@xxxxxxxxxxxxx>; gregkh@xxxxxxxxxxxxxxxxxxx; linux-
> kernel@xxxxxxxxxxxxxxx; devel@xxxxxxxxxxxxxxxxxxxxxx; olaf@xxxxxxxxx;
> apw@xxxxxxxxxxxxx; vkuznets@xxxxxxxxxx; jasowang@xxxxxxxxxx;
> leann.ogasawara@xxxxxxxxxxxxx; marcelo.cerri@xxxxxxxxxxxxx
> Cc: stable@xxxxxxxxxxxxxxx
> Subject: RE: [PATCH 1/1] Drivers: hv: vmbus: Don't leak channel ids
> 
> > From: linux-kernel-owner@xxxxxxxxxxxxxxx [mailto:linux-kernel-
> > owner@xxxxxxxxxxxxxxx] On Behalf Of kys@xxxxxxxxxxxxxxxxxxxxxx
> >
> > diff --git a/drivers/hv/channel_mgmt.c b/drivers/hv/channel_mgmt.c
> > index e1a3ae4..0a85246 100644
> > --- a/drivers/hv/channel_mgmt.c
> > +++ b/drivers/hv/channel_mgmt.c
> > @@ -802,6 +802,7 @@ static void vmbus_onoffer(struct
> > vmbus_channel_message_header *hdr)
> >         /* Allocate the channel object and save this offer. */
> >         newchannel = alloc_channel();
> >         if (!newchannel) {
> > +               vmbus_release_relid(offer->child_relid);
> >                 pr_err("Unable to allocate channel object\n");
> >                 return;
> >         }
> 
> The patch seems good.
> 
> BTW,  vmbus_onoffer -> alloc_channel is only called in the workqueue
> context, so maybe we should change the atomic kzalloc in alloc_channel
> to GFP_KERNEL?

No; this will introduce potential reordering of execution. I am working on rescind handling
where, we need to ensure temporal ordering of events until the channel is created.

K. Y
> 
> Thanks,
> -- Dexuan