| Message ID | 1424934286-7099-9-git-send-email-jasowang@redhat.com |
|---|---|
| State | New |
| Headers | show |
On Thu, Feb 26, 2015 at 03:04:43PM +0800, Jason Wang wrote: > Instead of depending on a macro, switch to use a bus specific queue > limit. > > Cc: Anthony Liguori <aliguori@amazon.com> > Cc: Michael S. Tsirkin <mst@redhat.com> > Signed-off-by: Jason Wang <jasowang@redhat.com> > --- > hw/virtio/virtio-pci.c | 10 +++++----- > 1 file changed, 5 insertions(+), 5 deletions(-) > > diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c > index 7fa8141..23c4649 100644 > --- a/hw/virtio/virtio-pci.c > +++ b/hw/virtio/virtio-pci.c > @@ -215,7 +215,7 @@ static void virtio_pci_start_ioeventfd(VirtIOPCIProxy *proxy) > return; > } > > - for (n = 0; n < VIRTIO_PCI_QUEUE_MAX; n++) { > + for (n = 0; n < virtio_get_queue_max(vdev); n++) { > if (!virtio_queue_get_num(vdev, n)) { > continue; > } This is done on guest IO, and I think after applying the next patch which increases the number to >500 for pci, it's too much work: VCPU is blocked meanwhile. Same applies to other places. At minimum, we'll need a faster way to locate active VQs.
On Thu, Feb 26, 2015 at 6:20 PM, Michael S. Tsirkin <mst@redhat.com> wrote: > On Thu, Feb 26, 2015 at 03:04:43PM +0800, Jason Wang wrote: >> Instead of depending on a macro, switch to use a bus specific queue >> limit. >> >> Cc: Anthony Liguori <aliguori@amazon.com> >> Cc: Michael S. Tsirkin <mst@redhat.com> >> Signed-off-by: Jason Wang <jasowang@redhat.com> >> --- >> hw/virtio/virtio-pci.c | 10 +++++----- >> 1 file changed, 5 insertions(+), 5 deletions(-) >> >> diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c >> index 7fa8141..23c4649 100644 >> --- a/hw/virtio/virtio-pci.c >> +++ b/hw/virtio/virtio-pci.c >> @@ -215,7 +215,7 @@ static void >> virtio_pci_start_ioeventfd(VirtIOPCIProxy *proxy) >> return; >> } >> >> - for (n = 0; n < VIRTIO_PCI_QUEUE_MAX; n++) { >> + for (n = 0; n < virtio_get_queue_max(vdev); n++) { >> if (!virtio_queue_get_num(vdev, n)) { >> continue; >> } > > This is done on guest IO, and I think after applying the > next patch which increases the number to >500 for pci, it's too much > work: VCPU is blocked meanwhile. Same applies to other places. > > At minimum, we'll need a faster way to locate active VQs. > Maybe just replace the continue with break in above. Since it looks like virtqueue index used by all virtio devices are contiguous?
diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c index 7fa8141..23c4649 100644 --- a/hw/virtio/virtio-pci.c +++ b/hw/virtio/virtio-pci.c @@ -215,7 +215,7 @@ static void virtio_pci_start_ioeventfd(VirtIOPCIProxy *proxy) return; } - for (n = 0; n < VIRTIO_PCI_QUEUE_MAX; n++) { + for (n = 0; n < virtio_get_queue_max(vdev); n++) { if (!virtio_queue_get_num(vdev, n)) { continue; } @@ -251,7 +251,7 @@ static void virtio_pci_stop_ioeventfd(VirtIOPCIProxy *proxy) return; } - for (n = 0; n < VIRTIO_PCI_QUEUE_MAX; n++) { + for (n = 0; n < virtio_get_queue_max(vdev); n++) { if (!virtio_queue_get_num(vdev, n)) { continue; } @@ -287,11 +287,11 @@ static void virtio_ioport_write(void *opaque, uint32_t addr, uint32_t val) virtio_queue_set_addr(vdev, vdev->queue_sel, pa); break; case VIRTIO_PCI_QUEUE_SEL: - if (val < VIRTIO_PCI_QUEUE_MAX) + if (val < virtio_get_queue_max(vdev)) vdev->queue_sel = val; break; case VIRTIO_PCI_QUEUE_NOTIFY: - if (val < VIRTIO_PCI_QUEUE_MAX) { + if (val < virtio_get_queue_max(vdev)) { virtio_queue_notify(vdev, val); } break; @@ -792,7 +792,7 @@ static int virtio_pci_set_guest_notifiers(DeviceState *d, int nvqs, bool assign) bool with_irqfd = msix_enabled(&proxy->pci_dev) && kvm_msi_via_irqfd_enabled(); - nvqs = MIN(nvqs, VIRTIO_PCI_QUEUE_MAX); + nvqs = MIN(nvqs, virtio_get_queue_max(vdev)); /* When deassigning, pass a consistent nvqs value * to avoid leaking notifiers.
Instead of depending on a macro, switch to use a bus specific queue limit. Cc: Anthony Liguori <aliguori@amazon.com> Cc: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Jason Wang <jasowang@redhat.com> --- hw/virtio/virtio-pci.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-)