Re: [PATCH v5 2/6] xen/pciback: Don't deadlock when unbinding.

[Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>]

On Mon, Jul 28, 2014 at 02:06:59PM +0100, David Vrabel wrote:
> On 14/07/14 17:18, Konrad Rzeszutek Wilk wrote:
> > As commit 0a9fd0152929db372ff61b0d6c280fdd34ae8bdb
> > 'xen/pciback: Document the entry points for 'pcistub_put_pci_dev''
> > explained there are four entry points in this function.
> > Two of them are when the user fiddles in the SysFS to
> > unbind a device which might be in use by a guest or not.
> > 
> > Both 'unbind' states will cause a deadlock as the the PCI lock has
> > already been taken, which then pci_device_reset tries to take.
> > 
> > We can simplify this by requiring that all callers of
> > pcistub_put_pci_dev MUST hold the device lock. And then
> > we can just call the lockless version of pci_device_reset.
> > 
> > To make it even simpler we will modify xen_pcibk_release_pci_dev
> > to quality whether it should take a lock or not - as it ends
> > up calling xen_pcibk_release_pci_dev and needs to hold the lock.
> > 
> > CC: stable@vger.kernel.org
> 
> This deadlock is for a rather specific and uncommon use case (manually
> unbinding a PCI while it is passed-through). Is this critical enough to
> warrant a stable backport?

We seem to trip over it frequently when rebooting a server.

That is the VF's end up being unbinded while the guests
are being shutdown. And depending on the timing we end up in a deadlock.

> 
> > Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
> > Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
> 
> Reviewed-by: David Vrabel <david.vrabel@citrix.com>
> 
> David