Merge remote-tracking branch 'remotes/awilliam/tags/vfio-pci-for-qemu-20140805.0' into staging

VFIO patches: Fix MSI-X vector expansion, remove MSI/X message caching

# gpg: Signature made Tue 05 Aug 2014 20:25:57 BST using RSA key ID 3BB08B22
# gpg: Can't check signature: public key not found

* remotes/awilliam/tags/vfio-pci-for-qemu-20140805.0:
  vfio: Don't cache MSIMessage
  vfio: Fix MSI-X vector expansion

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

1 files changed, 31 insertions, 15 deletions

diff --git a/hw/misc/vfio.c b/hw/misc/vfio.c
index 0b9eba0..ba08adb 100644
--- a/hw/misc/vfio.c
+++ b/hw/misc/vfio.c
@@ -120,11 +120,19 @@ typedef struct VFIOINTx {
 } VFIOINTx;
 
 typedef struct VFIOMSIVector {
-    EventNotifier interrupt; /* eventfd triggered on interrupt */
-    EventNotifier kvm_interrupt; /* eventfd triggered for KVM irqfd bypass */
+    /*
+     * Two interrupt paths are configured per vector.  The first, is only used
+     * for interrupts injected via QEMU.  This is typically the non-accel path,
+     * but may also be used when we want QEMU to handle masking and pending
+     * bits.  The KVM path bypasses QEMU and is therefore higher performance,
+     * but requires masking at the device.  virq is used to track the MSI route
+     * through KVM, thus kvm_interrupt is only available when virq is set to a
+     * valid (>= 0) value.
+     */
+    EventNotifier interrupt;
+    EventNotifier kvm_interrupt;
     struct VFIODevice *vdev; /* back pointer to device */
-    MSIMessage msg; /* cache the MSI message so we know when it changes */
-    int virq; /* KVM irqchip route for QEMU bypass */
+    int virq;
     bool use;
 } VFIOMSIVector;
 
@@ -681,13 +689,24 @@ static int vfio_enable_vectors(VFIODevice *vdev, bool msix)
     fds = (int32_t *)&irq_set->data;
 
     for (i = 0; i < vdev->nr_vectors; i++) {
-        if (!vdev->msi_vectors[i].use) {
-            fds[i] = -1;
-        } else if (vdev->msi_vectors[i].virq >= 0) {
-            fds[i] = event_notifier_get_fd(&vdev->msi_vectors[i].kvm_interrupt);
-        } else {
-            fds[i] = event_notifier_get_fd(&vdev->msi_vectors[i].interrupt);
+        int fd = -1;
+
+        /*
+         * MSI vs MSI-X - The guest has direct access to MSI mask and pending
+         * bits, therefore we always use the KVM signaling path when setup.
+         * MSI-X mask and pending bits are emulated, so we want to use the
+         * KVM signaling path only when configured and unmasked.
+         */
+        if (vdev->msi_vectors[i].use) {
+            if (vdev->msi_vectors[i].virq < 0 ||
+                (msix && msix_is_masked(&vdev->pdev, i))) {
+                fd = event_notifier_get_fd(&vdev->msi_vectors[i].interrupt);
+            } else {
+                fd = event_notifier_get_fd(&vdev->msi_vectors[i].kvm_interrupt);
+            }
         }
+
+        fds[i] = fd;
     }
 
     ret = ioctl(vdev->fd, VFIO_DEVICE_SET_IRQS, irq_set);
@@ -724,7 +743,6 @@ static void vfio_add_kvm_msi_virq(VFIOMSIVector *vector, MSIMessage *msg,
         return;
     }
 
-    vector->msg = *msg;
     vector->virq = virq;
 }
 
@@ -740,7 +758,6 @@ static void vfio_remove_kvm_msi_virq(VFIOMSIVector *vector)
 static void vfio_update_kvm_msi_virq(VFIOMSIVector *vector, MSIMessage msg)
 {
     kvm_irqchip_update_msi_route(kvm_state, vector->virq, msg);
-    vector->msg = msg;
 }
 
 static int vfio_msix_vector_do_use(PCIDevice *pdev, unsigned int nr,
@@ -919,6 +936,7 @@ retry:
 
     for (i = 0; i < vdev->nr_vectors; i++) {
         VFIOMSIVector *vector = &vdev->msi_vectors[i];
+        MSIMessage msg = msi_get_message(&vdev->pdev, i);
 
         vector->vdev = vdev;
         vector->virq = -1;
@@ -931,13 +949,11 @@ retry:
         qemu_set_fd_handler(event_notifier_get_fd(&vector->interrupt),
                             vfio_msi_interrupt, NULL, vector);
 
-        vector->msg = msi_get_message(&vdev->pdev, i);
-
         /*
          * Attempt to enable route through KVM irqchip,
          * default to userspace handling if unavailable.
          */
-        vfio_add_kvm_msi_virq(vector, &vector->msg, false);
+        vfio_add_kvm_msi_virq(vector, &msg, false);
     }
 
     /* Set interrupt type prior to possible interrupts */