Re: [PATCH v2 4/5] vdpa: add new vdpa attribute VDPA_ATTR_DEV_F_VERSION_1

all subsystems

On Wed, Sep 15, 2021 at 11:24:53AM +0800, Wu Zongyong wrote:
On Tue, Sep 14, 2021 at 08:58:28AM -0400, Michael S. Tsirkin wrote:
On Tue, Sep 14, 2021 at 08:24:51PM +0800, Wu Zongyong wrote:
This new attribute advertises whether the vdpa device is legacy or not.
Users can pick right virtqueue size if the vdpa device is legacy which
doesn't support to change virtqueue size.

Signed-off-by: Wu Zongyong <redacted>
So if we are bothering with legacy, I think there are
several things to do when building the interface
- support transitional devices, that is allow userspace
  to tell device it's in legacy mode
- support reporting/setting supporting endian-ness
It's true if we try to implement a general drvier for legacy.
But this series is dedicated to implement a driver only for ENI. Is it
necessary to implement what you said here in this series?
To a certain degree, yes.

I am thinking about the UAPI here. The new attribute is part of that.
E.g. userspace consuming this needs to be more or less hardware agnostic
and not depend on specifics of ENI.

Otherwise if userspace assumes legacy==eni then it will break with
other hardware.
 
One way to test how generic it all is would be adding legacy support in
the simulator.


---
 drivers/vdpa/vdpa.c          | 6 ++++++
 drivers/virtio/virtio_vdpa.c | 7 ++++++-
 include/uapi/linux/vdpa.h    | 1 +
 3 files changed, 13 insertions(+), 1 deletion(-)

diff --git a/drivers/vdpa/vdpa.c b/drivers/vdpa/vdpa.c
index 1dc121a07a93..533d7f589eee 100644
--- a/drivers/vdpa/vdpa.c
+++ b/drivers/vdpa/vdpa.c
@@ -12,6 +12,7 @@
 #include <linux/slab.h>
 #include <linux/vdpa.h>
 #include <uapi/linux/vdpa.h>
+#include <uapi/linux/virtio_config.h>
 #include <net/genetlink.h>
 #include <linux/mod_devicetable.h>
 
@@ -494,6 +495,7 @@ vdpa_dev_fill(struct vdpa_device *vdev, struct sk_buff *msg, u32 portid, u32 seq
 	u16 max_vq_size;
 	u32 device_id;
 	u32 vendor_id;
+	u64 features;
 	void *hdr;
 	int err;
 
@@ -508,6 +510,7 @@ vdpa_dev_fill(struct vdpa_device *vdev, struct sk_buff *msg, u32 portid, u32 seq
 	device_id = vdev->config->get_device_id(vdev);
 	vendor_id = vdev->config->get_vendor_id(vdev);
 	max_vq_size = vdev->config->get_vq_num_max(vdev);
+	features = vdev->config->get_features(vdev);
 
 	err = -EMSGSIZE;
 	if (nla_put_string(msg, VDPA_ATTR_DEV_NAME, dev_name(&vdev->dev)))
@@ -520,6 +523,9 @@ vdpa_dev_fill(struct vdpa_device *vdev, struct sk_buff *msg, u32 portid, u32 seq
 		goto msg_err;
 	if (nla_put_u16(msg, VDPA_ATTR_DEV_MAX_VQ_SIZE, max_vq_size))
 		goto msg_err;
+	if (features & BIT_ULL(VIRTIO_F_VERSION_1) &&
_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
On Wed, Sep 15, 2021 at 3:31 PM Michael S. Tsirkin [off-list ref] wrote:
On Wed, Sep 15, 2021 at 11:24:53AM +0800, Wu Zongyong wrote:
On Tue, Sep 14, 2021 at 08:58:28AM -0400, Michael S. Tsirkin wrote:
On Tue, Sep 14, 2021 at 08:24:51PM +0800, Wu Zongyong wrote:
This new attribute advertises whether the vdpa device is legacy or not.
Users can pick right virtqueue size if the vdpa device is legacy which
doesn't support to change virtqueue size.

Signed-off-by: Wu Zongyong <redacted>
So if we are bothering with legacy, I think there are
several things to do when building the interface
- support transitional devices, that is allow userspace
  to tell device it's in legacy mode
- support reporting/setting supporting endian-ness
It's true if we try to implement a general drvier for legacy.
But this series is dedicated to implement a driver only for ENI. Is it
necessary to implement what you said here in this series?
To a certain degree, yes.

I am thinking about the UAPI here. The new attribute is part of that.
E.g. userspace consuming this needs to be more or less hardware agnostic
and not depend on specifics of ENI.

Otherwise if userspace assumes legacy==eni then it will break with
other hardware.

One way to test how generic it all is would be adding legacy support in
the simulator.
I don't get why we need to support legacy devices e.g it doesn't have
ACCESS_PLATFORM support. I think we should re-consider to mandate 1.0
devices.

https://lore.kernel.org/lkml/20210408082648.20145-1-jasowang@redhat.com/

And it will complicate all of the different layers.

Thanks



---
 drivers/vdpa/vdpa.c          | 6 ++++++
 drivers/virtio/virtio_vdpa.c | 7 ++++++-
 include/uapi/linux/vdpa.h    | 1 +
 3 files changed, 13 insertions(+), 1 deletion(-)

diff --git a/drivers/vdpa/vdpa.c b/drivers/vdpa/vdpa.c
index 1dc121a07a93..533d7f589eee 100644
--- a/drivers/vdpa/vdpa.c
+++ b/drivers/vdpa/vdpa.c
@@ -12,6 +12,7 @@
 #include <linux/slab.h>
 #include <linux/vdpa.h>
 #include <uapi/linux/vdpa.h>
+#include <uapi/linux/virtio_config.h>
 #include <net/genetlink.h>
 #include <linux/mod_devicetable.h>

@@ -494,6 +495,7 @@ vdpa_dev_fill(struct vdpa_device *vdev, struct sk_buff *msg, u32 portid, u32 seq
  u16 max_vq_size;
  u32 device_id;
  u32 vendor_id;
+ u64 features;
  void *hdr;
  int err;

@@ -508,6 +510,7 @@ vdpa_dev_fill(struct vdpa_device *vdev, struct sk_buff *msg, u32 portid, u32 seq
  device_id = vdev->config->get_device_id(vdev);
  vendor_id = vdev->config->get_vendor_id(vdev);
  max_vq_size = vdev->config->get_vq_num_max(vdev);
+ features = vdev->config->get_features(vdev);

  err = -EMSGSIZE;
  if (nla_put_string(msg, VDPA_ATTR_DEV_NAME, dev_name(&vdev->dev)))
@@ -520,6 +523,9 @@ vdpa_dev_fill(struct vdpa_device *vdev, struct sk_buff *msg, u32 portid, u32 seq
          goto msg_err;
  if (nla_put_u16(msg, VDPA_ATTR_DEV_MAX_VQ_SIZE, max_vq_size))
          goto msg_err;
_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
On Wed, Sep 15, 2021 at 04:05:49PM +0800, Jason Wang wrote:
On Wed, Sep 15, 2021 at 3:31 PM Michael S. Tsirkin [off-list ref] wrote:
On Wed, Sep 15, 2021 at 11:24:53AM +0800, Wu Zongyong wrote:
On Tue, Sep 14, 2021 at 08:58:28AM -0400, Michael S. Tsirkin wrote:
On Tue, Sep 14, 2021 at 08:24:51PM +0800, Wu Zongyong wrote:
This new attribute advertises whether the vdpa device is legacy or not.
Users can pick right virtqueue size if the vdpa device is legacy which
doesn't support to change virtqueue size.

Signed-off-by: Wu Zongyong <redacted>
So if we are bothering with legacy, I think there are
several things to do when building the interface
- support transitional devices, that is allow userspace
  to tell device it's in legacy mode
- support reporting/setting supporting endian-ness
It's true if we try to implement a general drvier for legacy.
But this series is dedicated to implement a driver only for ENI. Is it
necessary to implement what you said here in this series?
To a certain degree, yes.

I am thinking about the UAPI here. The new attribute is part of that.
E.g. userspace consuming this needs to be more or less hardware agnostic
and not depend on specifics of ENI.

Otherwise if userspace assumes legacy==eni then it will break with
other hardware.

One way to test how generic it all is would be adding legacy support in
the simulator.
I don't get why we need to support legacy devices e.g it doesn't have
ACCESS_PLATFORM support. I think we should re-consider to mandate 1.0
devices.

https://lore.kernel.org/lkml/20210408082648.20145-1-jasowang@redhat.com/

And it will complicate all of the different layers.

It's not that we have to, it's just that imho, if we do
it's easier to just do it all in the kernel rather than spreading
parts of code around.



---
 drivers/vdpa/vdpa.c          | 6 ++++++
 drivers/virtio/virtio_vdpa.c | 7 ++++++-
 include/uapi/linux/vdpa.h    | 1 +
 3 files changed, 13 insertions(+), 1 deletion(-)

diff --git a/drivers/vdpa/vdpa.c b/drivers/vdpa/vdpa.c
index 1dc121a07a93..533d7f589eee 100644
--- a/drivers/vdpa/vdpa.c
+++ b/drivers/vdpa/vdpa.c
@@ -12,6 +12,7 @@
 #include <linux/slab.h>
 #include <linux/vdpa.h>
 #include <uapi/linux/vdpa.h>
+#include <uapi/linux/virtio_config.h>
 #include <net/genetlink.h>
 #include <linux/mod_devicetable.h>

@@ -494,6 +495,7 @@ vdpa_dev_fill(struct vdpa_device *vdev, struct sk_buff *msg, u32 portid, u32 seq
  u16 max_vq_size;
  u32 device_id;
  u32 vendor_id;
+ u64 features;
  void *hdr;
  int err;

@@ -508,6 +510,7 @@ vdpa_dev_fill(struct vdpa_device *vdev, struct sk_buff *msg, u32 portid, u32 seq
  device_id = vdev->config->get_device_id(vdev);
  vendor_id = vdev->config->get_vendor_id(vdev);
  max_vq_size = vdev->config->get_vq_num_max(vdev);
+ features = vdev->config->get_features(vdev);

  err = -EMSGSIZE;
  if (nla_put_string(msg, VDPA_ATTR_DEV_NAME, dev_name(&vdev->dev)))
@@ -520,6 +523,9 @@ vdpa_dev_fill(struct vdpa_device *vdev, struct sk_buff *msg, u32 portid, u32 seq
          goto msg_err;
  if (nla_put_u16(msg, VDPA_ATTR_DEV_MAX_VQ_SIZE, max_vq_size))
          goto msg_err;
_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

`h`	back out one level
`j`	next message in thread / next series-diff position
`k`	previous message in thread / previous series-diff position
`l`	drill in
`Esc`	close help / fold thread tree
`?`	toggle this help