Re: [Xen-devel] [PATCH 2/5] xen/pciback: Support pci_reset_function, aka FLR or D3 support.

[Don Dutile <ddutile@xxxxxxxxxx>]

On 01/05/2012 05:24 AM, Ian Campbell wrote:
> On Thu, 2012-01-05 at 00:46 +0000, Konrad Rzeszutek Wilk wrote:
> > We use the __pci_reset_function_locked to perform the action.
> > Also on attaching ("bind") and detaching ("unbind") we save and
> > restore the configuration states. When the device is disconnected
> > from a guest we use the "pci_reset_function" to also reset the
> > device before being passed to another guest.
> Currently I thought the toolstack was supposed to do the reset (by
> writing to the reset node in sysfs) as it was (de)assigning devices
> to/from guests. That's not to say that there isn't an argument for
> preferring to do it kernels side but it would be interesting to make
> that argument explicitly.
>
> I guess the toolstack doesn't currently save/restore the configuration
> state, could it though? I guess the state is all available in sysfs. On
pci_reset_function() saves the config state before doing a fcn-level reset.

the libvirt toolstack handles the unbind/reset/bind functionality
and is used by kvm for it's device assignment.

> the other hand the kernel provides us with a very handy function which
> Just Does It...
>
> Ian.
>
> > Signed-off-by: Konrad Rzeszutek Wilk<konrad.wilk@xxxxxxxxxx>
> > ---
> >   drivers/xen/xen-pciback/pci_stub.c |   41 +++++++++++++++++++++++++++++++++--
> >   drivers/xen/xen-pciback/pciback.h  |    1 +
> >   2 files changed, 39 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/xen/xen-pciback/pci_stub.c 
> > b/drivers/xen/xen-pciback/pci_stub.c
> > index 8f06e1e..0ce0dc6 100644
> > --- a/drivers/xen/xen-pciback/pci_stub.c
> > +++ b/drivers/xen/xen-pciback/pci_stub.c
> > @@ -85,19 +85,34 @@ static struct pcistub_device *pcistub_device_alloc(struct 
> > pci_dev *dev)
> >   static void pcistub_device_release(struct kref *kref)
> >   {
> >         struct pcistub_device *psdev;
> > +       struct xen_pcibk_dev_data *dev_data;
> >
> >         psdev = container_of(kref, struct pcistub_device, kref);
> > +       dev_data = pci_get_drvdata(psdev->dev);
> >
> >         dev_dbg(&psdev->dev->dev, "pcistub_device_release\n");
> >
> >         xen_unregister_device_domain_owner(psdev->dev);
> >
> > -       /* Clean-up the device */
> > +       /* Call the reset function which does not take lock as this
> > +        * is called from "unbind" which takes a device_lock mutex.
> > +        */
> > +       __pci_reset_function_locked(psdev->dev);
> > +       if (pci_load_and_free_saved_state(psdev->dev,
> > +                                       &dev_data->pci_saved_state)) {
> > +               dev_dbg(&psdev->dev->dev, "Could not reload PCI state\n");
> > +       } else
> > +               pci_restore_state(psdev->dev);
> > +
> > +       /* Disable the device */
> >         xen_pcibk_reset_device(psdev->dev);
> > +
> > +       kfree(dev_data);
> > +       pci_set_drvdata(psdev->dev, NULL);
> > +
> > +       /* Clean-up the device */
> >         xen_pcibk_config_free_dyn_fields(psdev->dev);
> >         xen_pcibk_config_free_dev(psdev->dev);
> > -       kfree(pci_get_drvdata(psdev->dev));
> > -       pci_set_drvdata(psdev->dev, NULL);
> >
> >         pci_dev_put(psdev->dev);
> >
> > @@ -230,7 +245,17 @@ void pcistub_put_pci_dev(struct pci_dev *dev)
> >         /* Cleanup our device
> >          * (so it's ready for the next domain)
> >          */
> > +
> > +       /* This is OK - we are running from workqueue context
> > +        * and want to inhibit the user from fiddling with 'reset'
> > +        */
> > +       pci_reset_function(dev);
> > +       pci_restore_state(psdev->dev);
> > +
> > +       /* This disables the device. */
> >         xen_pcibk_reset_device(found_psdev->dev);
> > +
> > +       /* And cleanup up our emulated fields. */
> >         xen_pcibk_config_free_dyn_fields(found_psdev->dev);
> >         xen_pcibk_config_reset_dev(found_psdev->dev);
> >
> > @@ -325,6 +350,16 @@ static int __devinit pcistub_init_device(struct pci_dev 
> > *dev)
> >         if (err)
> >                 goto config_release;
> >
> > +       dev_dbg(&dev->dev, "reseting (FLR, D3, etc) the device\n");
> > +       __pci_reset_function_locked(dev);
> > +
> > +       /* We need the device active to save the state. */
> > +       dev_dbg(&dev->dev, "save state of device\n");
> > +       pci_save_state(dev);
> > +       dev_data->pci_saved_state = pci_store_saved_state(dev);
> > +       if (!dev_data->pci_saved_state)
> > +               dev_err(&dev->dev, "Could not store PCI conf saved state!\n");
> > +
> >         /* Now disable the device (this also ensures some private device
> >          * data is setup before we export)
> >          */
> > diff --git a/drivers/xen/xen-pciback/pciback.h 
> > b/drivers/xen/xen-pciback/pciback.h
> > index e9b4011..a7def01 100644
> > --- a/drivers/xen/xen-pciback/pciback.h
> > +++ b/drivers/xen/xen-pciback/pciback.h
> > @@ -41,6 +41,7 @@ struct xen_pcibk_device {
> >
> >   struct xen_pcibk_dev_data {
> >         struct list_head config_fields;
> > +       struct pci_saved_state *pci_saved_state;
> >         unsigned int permissive:1;
> >         unsigned int warned_on_write:1;
> >         unsigned int enable_intx:1;
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-pci" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel