From: Michael S. Tsirkin <mst@redhat.com>
Sent: Monday, October 30, 2023 9:29 PM
On Mon, Oct 30, 2023 at 03:51:40PM +0000, Parav Pandit wrote:

quoted

quoted

From: Michael S. Tsirkin <mst@redhat.com>
Sent: Monday, October 30, 2023 1:53 AM

On Sun, Oct 29, 2023 at 05:59:45PM +0200, Yishai Hadas wrote:

quoted

From: Feng Liu <redacted>

Introduce support for the admin virtqueue. By negotiating
VIRTIO_F_ADMIN_VQ feature, driver detects capability and creates
one administration virtqueue. Administration virtqueue
implementation in virtio pci generic layer, enables multiple types
of upper layer drivers such as vfio, net, blk to utilize it.

Signed-off-by: Feng Liu <redacted>
Reviewed-by: Parav Pandit <redacted>
Reviewed-by: Jiri Pirko <redacted>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
---
 drivers/virtio/virtio.c                | 37 ++++++++++++++--
 drivers/virtio/virtio_pci_common.c     |  3 ++
 drivers/virtio/virtio_pci_common.h     | 15 ++++++-
 drivers/virtio/virtio_pci_modern.c     | 61 +++++++++++++++++++++++++-
 drivers/virtio/virtio_pci_modern_dev.c | 18 ++++++++
 include/linux/virtio_config.h          |  4 ++
 include/linux/virtio_pci_modern.h      |  5 +++
 7 files changed, 137 insertions(+), 6 deletions(-)

diff --git a/drivers/virtio/virtio.c b/drivers/virtio/virtio.c

index
3893dc29eb26..f4080692b351 100644

--- a/drivers/virtio/virtio.c
+++ b/drivers/virtio/virtio.c

@@ -302,9 +302,15 @@ static int virtio_dev_probe(struct device *_d)
 	if (err)
 		goto err;

+	if (dev->config->create_avq) {
+		err = dev->config->create_avq(dev);
+		if (err)
+			goto err;
+	}
+
 	err = drv->probe(dev);
 	if (err)
-		goto err;
+		goto err_probe;

 	/* If probe didn't do it, mark device DRIVER_OK ourselves. */
 	if (!(dev->config->get_status(dev) & VIRTIO_CONFIG_S_DRIVER_OK))

Hmm I am not all that happy that we are just creating avq unconditionally.
Can't we do it on demand to avoid wasting resources if no one uses it?

Virtio queues must be enabled before driver_ok as we discussed in

F_DYNAMIC bit exercise.

quoted

So creating AQ when first legacy command is invoked, would be too late.

Well we didn't release the spec with AQ so I am pretty sure there are no devices
using the feature. Do we want to already make an exception for AQ and allow
creating AQs after DRIVER_OK even without F_DYNAMIC?

No. it would abuse the init time config registers for the dynamic things like this.
For flow filters and others there is need for dynamic q creation with multiple physical address anyway.
So creating virtqueues dynamically using a generic scheme is desired with new feature bit.
_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization