Re: Regression in kernel 3.5 as Dom0 regarding PCI Passthrough?!

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

> > > > And its due to a patch I added in v3.4
> > > > (cd9db80e5257682a7f7ab245a2459648b3c8d268)
> > > > - which did not work properly in v3.4, but with v3.5 got it working
> > > > (977f857ca566a1e68045fcbb7cfc9c4acb077cf0) which causes v3.5 to
> > now
> > > > work
> > > > anymore.
> > > >
> > > > Anyhow, for right now jsut revert
> > > > cd9db80e5257682a7f7ab245a2459648b3c8d268
> > > > and it should work for you.
> > > >
> Confirmed, after reverting that commit, VT-d will work fine.
> Will you fix this and push it to upstream Linux, Konrad?
> 
> > > Also, our team reported a VT-d bug 2 months ago.
> > > http://bugzilla.xen.org/bugzilla/show_bug.cgi?id=1824
> >

Can either one of you please test this patch, please:


diff --git a/drivers/xen/xen-pciback/pci_stub.c b/drivers/xen/xen-pciback/pci_stub.c
index 097e536..425bd0b 100644
--- a/drivers/xen/xen-pciback/pci_stub.c
+++ b/drivers/xen/xen-pciback/pci_stub.c
@@ -4,6 +4,8 @@
  * Ryan Wilson <hap9@epoch.ncsc.mil>
  * Chris Bookholt <hap10@epoch.ncsc.mil>
  */
+#define DEBUG 1
+
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/rwsem.h>
@@ -97,13 +99,15 @@ static void pcistub_device_release(struct kref *kref)
 	/* Call the reset function which does not take lock as this
 	 * is called from "unbind" which takes a device_lock mutex.
 	 */
+	dev_dbg(&psdev->dev->dev, "FLR locked..\n");
 	__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
+	} else {
+		dev_dbg(&psdev->dev->dev, "Reloading PCI state..\n");
 		pci_restore_state(psdev->dev);
-
+	}
 	/* Disable the device */
 	xen_pcibk_reset_device(psdev->dev);
 
@@ -353,16 +357,16 @@ 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");
-
+	else {
+		dev_dbg(&dev->dev, "reseting (FLR, D3, etc) the device\n");
+		__pci_reset_function_locked(dev);
+	}
 	/* Now disable the device (this also ensures some private device
 	 * data is setup before we export)
 	 */