Linux Virtualization - Sep 2020 - [RFC PATCH 00/22] Enhance VHOST to enable SoC-to-SoC communication

On Thu, Jul 02, 2020 at 01:51:21PM +0530, Kishon Vijay Abraham I
wrote:
> This series enhances Linux Vhost support to enable SoC-to-SoC
> communication over MMIO. This series enables rpmsg communication between
> two SoCs using both PCIe RC<->EP and HOST1-NTB-HOST2
> 
> 1) Modify vhost to use standard Linux driver model
> 2) Add support in vring to access virtqueue over MMIO
> 3) Add vhost client driver for rpmsg
> 4) Add PCIe RC driver (uses virtio) and PCIe EP driver (uses vhost) for
>    rpmsg communication between two SoCs connected to each other
> 5) Add NTB Virtio driver and NTB Vhost driver for rpmsg communication
>    between two SoCs connected via NTB
> 6) Add configfs to configure the components
> 
> UseCase1 :
> 
>  VHOST RPMSG                     VIRTIO RPMSG
>       +                               +
>       |                               |
>       |                               |
>       |                               |
>       |                               |
> +-----v------+                 +------v-------+
> |   Linux    |                 |     Linux    |
> |  Endpoint  |                 | Root Complex |
> |            <----------------->              |
> |            |                 |              |
> |    SOC1    |                 |     SOC2     |
> +------------+                 +--------------+
> 
> UseCase 2:
> 
>      VHOST RPMSG                                      VIRTIO RPMSG
>           +                                                 +
>           |                                                 |
>           |                                                 |
>           |                                                 |
>           |                                                 |
>    +------v------+                                   +------v------+
>    |             |                                   |             |
>    |    HOST1    |                                   |    HOST2    |
>    |             |                                   |             |
>    +------^------+                                   +------^------+
>           |                                                 |
>           |                                                 |
> +---------------------------------------------------------------------+
> |  +------v------+                                   +------v------+  |
> |  |             |                                   |             |  |
> |  |     EP      |                                   |     EP      |  |
> |  | CONTROLLER1 |                                   | CONTROLLER2 |  |
> |  |             <----------------------------------->             | 
|
> |  |             |                                   |             |  |
> |  |             |                                   |             |  |
> |  |             |  SoC With Multiple EP Instances   |             |  |
> |  |             |  (Configured using NTB Function)  |             |  |
> |  +-------------+                                   +-------------+  |
> +---------------------------------------------------------------------+
> 
> Software Layering:
> 
> The high-level SW layering should look something like below. This series
> adds support only for RPMSG VHOST, however something similar should be
> done for net and scsi. With that any vhost device (PCI, NTB, Platform
> device, user) can use any of the vhost client driver.
> 
> 
>     +----------------+  +-----------+  +------------+  +----------+
>     |  RPMSG VHOST   |  | NET VHOST |  | SCSI VHOST |  |    X     |
>     +-------^--------+  +-----^-----+  +-----^------+  +----^-----+
>             |                 |              |              |
>             |                 |              |              |
>             |                 |              |              |
> +-----------v-----------------v--------------v--------------v----------+
> |                            VHOST CORE                                |
> +--------^---------------^--------------------^------------------^-----+
>          |               |                    |                  |
>          |               |                    |                  |
>          |               |                    |                  |
> +--------v-------+  +----v------+  +----------v----------+  +----v-----+
> |  PCI EPF VHOST |  | NTB VHOST |  |PLATFORM DEVICE VHOST|  |    X     |
> +----------------+  +-----------+  +---------------------+  +----------+
> 
> This was initially proposed here [1]
> 
> [1] -> https://lore.kernel.org/r/2cf00ec4-1ed6-f66e-6897-006d1a5b6390 at
ti.com

I find this very interesting. A huge patchset so will take a bit
to review, but I certainly plan to do that. Thanks!
> 
> Kishon Vijay Abraham I (22):
>   vhost: Make _feature_ bits a property of vhost device
>   vhost: Introduce standard Linux driver model in VHOST
>   vhost: Add ops for the VHOST driver to configure VHOST device
>   vringh: Add helpers to access vring in MMIO
>   vhost: Add MMIO helpers for operations on vhost virtqueue
>   vhost: Introduce configfs entry for configuring VHOST
>   virtio_pci: Use request_threaded_irq() instead of request_irq()
>   rpmsg: virtio_rpmsg_bus: Disable receive virtqueue callback when
>     reading messages
>   rpmsg: Introduce configfs entry for configuring rpmsg
>   rpmsg: virtio_rpmsg_bus: Add Address Service Notification support
>   rpmsg: virtio_rpmsg_bus: Move generic rpmsg structure to
>     rpmsg_internal.h
>   virtio: Add ops to allocate and free buffer
>   rpmsg: virtio_rpmsg_bus: Use virtio_alloc_buffer() and
>     virtio_free_buffer()
>   rpmsg: Add VHOST based remote processor messaging bus
>   samples/rpmsg: Setup delayed work to send message
>   samples/rpmsg: Wait for address to be bound to rpdev for sending
>     message
>   rpmsg.txt: Add Documentation to configure rpmsg using configfs
>   virtio_pci: Add VIRTIO driver for VHOST on Configurable PCIe Endpoint
>     device
>   PCI: endpoint: Add EP function driver to provide VHOST interface
>   NTB: Add a new NTB client driver to implement VIRTIO functionality
>   NTB: Add a new NTB client driver to implement VHOST functionality
>   NTB: Describe the ntb_virtio and ntb_vhost client in the documentation
> 
>  Documentation/driver-api/ntb.rst              |   11 +
>  Documentation/rpmsg.txt                       |   56 +
>  drivers/ntb/Kconfig                           |   18 +
>  drivers/ntb/Makefile                          |    2 +
>  drivers/ntb/ntb_vhost.c                       |  776 +++++++++++
>  drivers/ntb/ntb_virtio.c                      |  853 ++++++++++++
>  drivers/ntb/ntb_virtio.h                      |   56 +
>  drivers/pci/endpoint/functions/Kconfig        |   11 +
>  drivers/pci/endpoint/functions/Makefile       |    1 +
>  .../pci/endpoint/functions/pci-epf-vhost.c    | 1144 ++++++++++++++++
>  drivers/rpmsg/Kconfig                         |   10 +
>  drivers/rpmsg/Makefile                        |    3 +-
>  drivers/rpmsg/rpmsg_cfs.c                     |  394 ++++++
>  drivers/rpmsg/rpmsg_core.c                    |    7 +
>  drivers/rpmsg/rpmsg_internal.h                |  136 ++
>  drivers/rpmsg/vhost_rpmsg_bus.c               | 1151 +++++++++++++++++
>  drivers/rpmsg/virtio_rpmsg_bus.c              |  184 ++-
>  drivers/vhost/Kconfig                         |    1 +
>  drivers/vhost/Makefile                        |    2 +-
>  drivers/vhost/net.c                           |   10 +-
>  drivers/vhost/scsi.c                          |   24 +-
>  drivers/vhost/test.c                          |   17 +-
>  drivers/vhost/vdpa.c                          |    2 +-
>  drivers/vhost/vhost.c                         |  730 ++++++++++-
>  drivers/vhost/vhost_cfs.c                     |  341 +++++
>  drivers/vhost/vringh.c                        |  332 +++++
>  drivers/vhost/vsock.c                         |   20 +-
>  drivers/virtio/Kconfig                        |    9 +
>  drivers/virtio/Makefile                       |    1 +
>  drivers/virtio/virtio_pci_common.c            |   25 +-
>  drivers/virtio/virtio_pci_epf.c               |  670 ++++++++++
>  include/linux/mod_devicetable.h               |    6 +
>  include/linux/rpmsg.h                         |    6 +
>  {drivers/vhost => include/linux}/vhost.h      |  132 +-
>  include/linux/virtio.h                        |    3 +
>  include/linux/virtio_config.h                 |   42 +
>  include/linux/vringh.h                        |   46 +
>  samples/rpmsg/rpmsg_client_sample.c           |   32 +-
>  tools/virtio/virtio_test.c                    |    2 +-
>  39 files changed, 7083 insertions(+), 183 deletions(-)
>  create mode 100644 drivers/ntb/ntb_vhost.c
>  create mode 100644 drivers/ntb/ntb_virtio.c
>  create mode 100644 drivers/ntb/ntb_virtio.h
>  create mode 100644 drivers/pci/endpoint/functions/pci-epf-vhost.c
>  create mode 100644 drivers/rpmsg/rpmsg_cfs.c
>  create mode 100644 drivers/rpmsg/vhost_rpmsg_bus.c
>  create mode 100644 drivers/vhost/vhost_cfs.c
>  create mode 100644 drivers/virtio/virtio_pci_epf.c
>  rename {drivers/vhost => include/linux}/vhost.h (66%)
> 
> -- 
> 2.17.1
>

On 2020/7/2 ??5:51, Michael S. Tsirkin wrote:
> On Thu, Jul 02, 2020 at 01:51:21PM +0530, Kishon Vijay Abraham I wrote:
>> This series enhances Linux Vhost support to enable SoC-to-SoC
>> communication over MMIO. This series enables rpmsg communication
between
>> two SoCs using both PCIe RC<->EP and HOST1-NTB-HOST2
>>
>> 1) Modify vhost to use standard Linux driver model
>> 2) Add support in vring to access virtqueue over MMIO
>> 3) Add vhost client driver for rpmsg
>> 4) Add PCIe RC driver (uses virtio) and PCIe EP driver (uses vhost) for
>>     rpmsg communication between two SoCs connected to each other
>> 5) Add NTB Virtio driver and NTB Vhost driver for rpmsg communication
>>     between two SoCs connected via NTB
>> 6) Add configfs to configure the components
>>
>> UseCase1 :
>>
>>   VHOST RPMSG                     VIRTIO RPMSG
>>        +                               +
>>        |                               |
>>        |                               |
>>        |                               |
>>        |                               |
>> +-----v------+                 +------v-------+
>> |   Linux    |                 |     Linux    |
>> |  Endpoint  |                 | Root Complex |
>> |            <----------------->              |
>> |            |                 |              |
>> |    SOC1    |                 |     SOC2     |
>> +------------+                 +--------------+
>>
>> UseCase 2:
>>
>>       VHOST RPMSG                                      VIRTIO RPMSG
>>            +                                                 +
>>            |                                                 |
>>            |                                                 |
>>            |                                                 |
>>            |                                                 |
>>     +------v------+                                   +------v------+
>>     |             |                                   |             |
>>     |    HOST1    |                                   |    HOST2    |
>>     |             |                                   |             |
>>     +------^------+                                   +------^------+
>>            |                                                 |
>>            |                                                 |
>> +---------------------------------------------------------------------+
>> |  +------v------+                                   +------v------+  |
>> |  |             |                                   |             |  |
>> |  |     EP      |                                   |     EP      |  |
>> |  | CONTROLLER1 |                                   | CONTROLLER2 |  |
>> |  |             <----------------------------------->           
|  |
>> |  |             |                                   |             |  |
>> |  |             |                                   |             |  |
>> |  |             |  SoC With Multiple EP Instances   |             |  |
>> |  |             |  (Configured using NTB Function)  |             |  |
>> |  +-------------+                                   +-------------+  |
>> +---------------------------------------------------------------------+
>>
>> Software Layering:
>>
>> The high-level SW layering should look something like below. This
series
>> adds support only for RPMSG VHOST, however something similar should be
>> done for net and scsi. With that any vhost device (PCI, NTB, Platform
>> device, user) can use any of the vhost client driver.
>>
>>
>>      +----------------+  +-----------+  +------------+  +----------+
>>      |  RPMSG VHOST   |  | NET VHOST |  | SCSI VHOST |  |    X     |
>>      +-------^--------+  +-----^-----+  +-----^------+  +----^-----+
>>              |                 |              |              |
>>              |                 |              |              |
>>              |                 |              |              |
>>
+-----------v-----------------v--------------v--------------v----------+
>> |                            VHOST CORE                               
|
>>
+--------^---------------^--------------------^------------------^-----+
>>           |               |                    |                  |
>>           |               |                    |                  |
>>           |               |                    |                  |
>> +--------v-------+  +----v------+  +----------v----------+ 
+----v-----+
>> |  PCI EPF VHOST |  | NTB VHOST |  |PLATFORM DEVICE VHOST|  |    X    
|
>> +----------------+  +-----------+  +---------------------+ 
+----------+
>>
>> This was initially proposed here [1]
>>
>> [1] ->
https://lore.kernel.org/r/2cf00ec4-1ed6-f66e-6897-006d1a5b6390 at ti.com
>
> I find this very interesting. A huge patchset so will take a bit
> to review, but I certainly plan to do that. Thanks!

Yes, it would be better if there's a git branch for us to have a look.

Btw, I'm not sure I get the big picture, but I vaguely feel some of the 
work is duplicated with vDPA (e.g the epf transport or vhost bus).

Have you considered to implement these through vDPA?

Thanks

>
>> Kishon Vijay Abraham I (22):
>>    vhost: Make _feature_ bits a property of vhost device
>>    vhost: Introduce standard Linux driver model in VHOST
>>    vhost: Add ops for the VHOST driver to configure VHOST device
>>    vringh: Add helpers to access vring in MMIO
>>    vhost: Add MMIO helpers for operations on vhost virtqueue
>>    vhost: Introduce configfs entry for configuring VHOST
>>    virtio_pci: Use request_threaded_irq() instead of request_irq()
>>    rpmsg: virtio_rpmsg_bus: Disable receive virtqueue callback when
>>      reading messages
>>    rpmsg: Introduce configfs entry for configuring rpmsg
>>    rpmsg: virtio_rpmsg_bus: Add Address Service Notification support
>>    rpmsg: virtio_rpmsg_bus: Move generic rpmsg structure to
>>      rpmsg_internal.h
>>    virtio: Add ops to allocate and free buffer
>>    rpmsg: virtio_rpmsg_bus: Use virtio_alloc_buffer() and
>>      virtio_free_buffer()
>>    rpmsg: Add VHOST based remote processor messaging bus
>>    samples/rpmsg: Setup delayed work to send message
>>    samples/rpmsg: Wait for address to be bound to rpdev for sending
>>      message
>>    rpmsg.txt: Add Documentation to configure rpmsg using configfs
>>    virtio_pci: Add VIRTIO driver for VHOST on Configurable PCIe
Endpoint
>>      device
>>    PCI: endpoint: Add EP function driver to provide VHOST interface
>>    NTB: Add a new NTB client driver to implement VIRTIO functionality
>>    NTB: Add a new NTB client driver to implement VHOST functionality
>>    NTB: Describe the ntb_virtio and ntb_vhost client in the
documentation
>>
>>   Documentation/driver-api/ntb.rst              |   11 +
>>   Documentation/rpmsg.txt                       |   56 +
>>   drivers/ntb/Kconfig                           |   18 +
>>   drivers/ntb/Makefile                          |    2 +
>>   drivers/ntb/ntb_vhost.c                       |  776 +++++++++++
>>   drivers/ntb/ntb_virtio.c                      |  853 ++++++++++++
>>   drivers/ntb/ntb_virtio.h                      |   56 +
>>   drivers/pci/endpoint/functions/Kconfig        |   11 +
>>   drivers/pci/endpoint/functions/Makefile       |    1 +
>>   .../pci/endpoint/functions/pci-epf-vhost.c    | 1144 ++++++++++++++++
>>   drivers/rpmsg/Kconfig                         |   10 +
>>   drivers/rpmsg/Makefile                        |    3 +-
>>   drivers/rpmsg/rpmsg_cfs.c                     |  394 ++++++
>>   drivers/rpmsg/rpmsg_core.c                    |    7 +
>>   drivers/rpmsg/rpmsg_internal.h                |  136 ++
>>   drivers/rpmsg/vhost_rpmsg_bus.c               | 1151
+++++++++++++++++
>>   drivers/rpmsg/virtio_rpmsg_bus.c              |  184 ++-
>>   drivers/vhost/Kconfig                         |    1 +
>>   drivers/vhost/Makefile                        |    2 +-
>>   drivers/vhost/net.c                           |   10 +-
>>   drivers/vhost/scsi.c                          |   24 +-
>>   drivers/vhost/test.c                          |   17 +-
>>   drivers/vhost/vdpa.c                          |    2 +-
>>   drivers/vhost/vhost.c                         |  730 ++++++++++-
>>   drivers/vhost/vhost_cfs.c                     |  341 +++++
>>   drivers/vhost/vringh.c                        |  332 +++++
>>   drivers/vhost/vsock.c                         |   20 +-
>>   drivers/virtio/Kconfig                        |    9 +
>>   drivers/virtio/Makefile                       |    1 +
>>   drivers/virtio/virtio_pci_common.c            |   25 +-
>>   drivers/virtio/virtio_pci_epf.c               |  670 ++++++++++
>>   include/linux/mod_devicetable.h               |    6 +
>>   include/linux/rpmsg.h                         |    6 +
>>   {drivers/vhost => include/linux}/vhost.h      |  132 +-
>>   include/linux/virtio.h                        |    3 +
>>   include/linux/virtio_config.h                 |   42 +
>>   include/linux/vringh.h                        |   46 +
>>   samples/rpmsg/rpmsg_client_sample.c           |   32 +-
>>   tools/virtio/virtio_test.c                    |    2 +-
>>   39 files changed, 7083 insertions(+), 183 deletions(-)
>>   create mode 100644 drivers/ntb/ntb_vhost.c
>>   create mode 100644 drivers/ntb/ntb_virtio.c
>>   create mode 100644 drivers/ntb/ntb_virtio.h
>>   create mode 100644 drivers/pci/endpoint/functions/pci-epf-vhost.c
>>   create mode 100644 drivers/rpmsg/rpmsg_cfs.c
>>   create mode 100644 drivers/rpmsg/vhost_rpmsg_bus.c
>>   create mode 100644 drivers/vhost/vhost_cfs.c
>>   create mode 100644 drivers/virtio/virtio_pci_epf.c
>>   rename {drivers/vhost => include/linux}/vhost.h (66%)
>>
>> -- 
>> 2.17.1
>>

On 2020/7/2 ??9:35, Kishon Vijay Abraham I wrote:
> Hi Jason,
>
> On 7/2/2020 3:40 PM, Jason Wang wrote:
>> On 2020/7/2 ??5:51, Michael S. Tsirkin wrote:
>>> On Thu, Jul 02, 2020 at 01:51:21PM +0530, Kishon Vijay Abraham I
wrote:
>>>> This series enhances Linux Vhost support to enable SoC-to-SoC
>>>> communication over MMIO. This series enables rpmsg
communication between
>>>> two SoCs using both PCIe RC<->EP and HOST1-NTB-HOST2
>>>>
>>>> 1) Modify vhost to use standard Linux driver model
>>>> 2) Add support in vring to access virtqueue over MMIO
>>>> 3) Add vhost client driver for rpmsg
>>>> 4) Add PCIe RC driver (uses virtio) and PCIe EP driver (uses
vhost) for
>>>>  ??? rpmsg communication between two SoCs connected to each
other
>>>> 5) Add NTB Virtio driver and NTB Vhost driver for rpmsg
communication
>>>>  ??? between two SoCs connected via NTB
>>>> 6) Add configfs to configure the components
>>>>
>>>> UseCase1 :
>>>>
>>>>  ? VHOST RPMSG???????????????????? VIRTIO RPMSG
>>>>  ?????? +?????????????????????????????? +
>>>>  ?????? |?????????????????????????????? |
>>>>  ?????? |?????????????????????????????? |
>>>>  ?????? |?????????????????????????????? |
>>>>  ?????? |?????????????????????????????? |
>>>> +-----v------+???????????????? +------v-------+
>>>> |?? Linux??? |???????????????? |???? Linux??? |
>>>> |? Endpoint? |???????????????? | Root Complex |
>>>> |??????????? <----------------->????????????? |
>>>> |??????????? |???????????????? |????????????? |
>>>> |??? SOC1??? |???????????????? |???? SOC2???? |
>>>> +------------+???????????????? +--------------+
>>>>
>>>> UseCase 2:
>>>>
>>>>  ????? VHOST RPMSG????????????????????????????????????? VIRTIO
RPMSG
>>>>  ?????????? +???????????????????????????????????????????????? +
>>>>  ?????????? |???????????????????????????????????????????????? |
>>>>  ?????????? |???????????????????????????????????????????????? |
>>>>  ?????????? |???????????????????????????????????????????????? |
>>>>  ?????????? |???????????????????????????????????????????????? |
>>>>  ??? +------v------+??????????????????????????????????
+------v------+
>>>>  ??? |???????????? |??????????????????????????????????
|???????????? |
>>>>  ??? |??? HOST1??? |?????????????????????????????????? |???
HOST2??? |
>>>>  ??? |???????????? |??????????????????????????????????
|???????????? |
>>>>  ??? +------^------+??????????????????????????????????
+------^------+
>>>>  ?????????? |???????????????????????????????????????????????? |
>>>>  ?????????? |???????????????????????????????????????????????? |
>>>>
+---------------------------------------------------------------------+
>>>> |? +------v------+??????????????????????????????????
+------v------+? |
>>>> |? |???????????? |??????????????????????????????????
|???????????? |? |
>>>> |? |???? EP????? |?????????????????????????????????? |????
EP????? |? |
>>>> |? | CONTROLLER1 |?????????????????????????????????? |
CONTROLLER2 |? |
>>>> |? |????????????
<----------------------------------->???????????? |? |
>>>> |? |???????????? |??????????????????????????????????
|???????????? |? |
>>>> |? |???????????? |??????????????????????????????????
|???????????? |? |
>>>> |? |???????????? |? SoC With Multiple EP Instances??
|???????????? |? |
>>>> |? |???????????? |? (Configured using NTB Function)?
|???????????? |? |
>>>> |? +-------------+??????????????????????????????????
+-------------+? |
>>>>
+---------------------------------------------------------------------+
>>>>
>>>> Software Layering:
>>>>
>>>> The high-level SW layering should look something like below.
This series
>>>> adds support only for RPMSG VHOST, however something similar
should be
>>>> done for net and scsi. With that any vhost device (PCI, NTB,
Platform
>>>> device, user) can use any of the vhost client driver.
>>>>
>>>>
>>>>  ???? +----------------+? +-----------+? +------------+?
+----------+
>>>>  ???? |? RPMSG VHOST?? |? | NET VHOST |? | SCSI VHOST |? |???
X???? |
>>>>  ???? +-------^--------+? +-----^-----+? +-----^------+?
+----^-----+
>>>>  ???????????? |???????????????? |????????????? |????????????? |
>>>>  ???????????? |???????????????? |????????????? |????????????? |
>>>>  ???????????? |???????????????? |????????????? |????????????? |
>>>>
+-----------v-----------------v--------------v--------------v----------+
>>>> |??????????????????????????? VHOST
CORE??????????????????????????????? |
>>>>
+--------^---------------^--------------------^------------------^-----+
>>>>  ????????? |?????????????? |???????????????????
|????????????????? |
>>>>  ????????? |?????????????? |???????????????????
|????????????????? |
>>>>  ????????? |?????????????? |???????????????????
|????????????????? |
>>>> +--------v-------+? +----v------+? +----------v----------+?
+----v-----+
>>>> |? PCI EPF VHOST |? | NTB VHOST |? |PLATFORM DEVICE VHOST|?
|??? X???? |
>>>> +----------------+? +-----------+? +---------------------+?
+----------+
>>>>
>>>> This was initially proposed here [1]
>>>>
>>>> [1] ->
https://lore.kernel.org/r/2cf00ec4-1ed6-f66e-6897-006d1a5b6390 at ti.com
>>> I find this very interesting. A huge patchset so will take a bit
>>> to review, but I certainly plan to do that. Thanks!
>>
>> Yes, it would be better if there's a git branch for us to have a
look.
> I've pushed the branch
> https://github.com/kishon/linux-wip.git vhost_rpmsg_pci_ntb_rfc

Thanks

>> Btw, I'm not sure I get the big picture, but I vaguely feel some of
the work is
>> duplicated with vDPA (e.g the epf transport or vhost bus).
> This is about connecting two different HW systems both running Linux and
> doesn't necessarily involve virtualization.

Right, this is something similar to VOP 
(Documentation/misc-devices/mic/mic_overview.rst). The different is the 
hardware I guess and VOP use userspace application to implement the device.

>   So there is no guest or host as in
> virtualization but two entirely different systems connected via PCIe cable,
one
> acting as guest and one as host. So one system will provide virtio
> functionality reserving memory for virtqueues and the other provides vhost
> functionality providing a way to access the virtqueues in virtio memory.
One is
> source and the other is sink and there is no intermediate entity. (vhost
was
> probably intermediate entity in virtualization?)

(Not a native English speaker) but "vhost" could introduce some 
confusion for me since it was use for implementing virtio backend for 
userspace drivers. I guess "vringh" could be better.

>
>> Have you considered to implement these through vDPA?
> IIUC vDPA only provides an interface to userspace and an in-kernel rpmsg
driver
> or vhost net driver is not provided.
>
> The HW connection looks something like https://pasteboard.co/JfMVVHC.jpg
> (usecase2 above),

I see.

>   all the boards run Linux. The middle board provides NTB
> functionality and board on either side provides virtio/vhost functionality
and
> transfer data using rpmsg.

So I wonder whether it's worthwhile for a new bus. Can we use the 
existed virtio-bus/drivers? It might work as, except for the epf 
transport, we can introduce a epf "vhost" transport driver.

It will have virtqueues but only used for the communication between 
itself and uppter virtio driver. And it will have vringh queues which 
will be probe by virtio epf transport drivers. And it needs to do 
datacopy between virtqueue and vringh queues.

It works like:

virtio drivers <- virtqueue/virtio-bus -> epf vhost drivers <- vringh 
queue/epf>

The advantages is that there's no need for writing new buses and drivers.

Does this make sense?

Thanks

>
> Thanks
> Kishon
>
>> Thanks
>>
>>
>>>> Kishon Vijay Abraham I (22):
>>>>  ?? vhost: Make _feature_ bits a property of vhost device
>>>>  ?? vhost: Introduce standard Linux driver model in VHOST
>>>>  ?? vhost: Add ops for the VHOST driver to configure VHOST
device
>>>>  ?? vringh: Add helpers to access vring in MMIO
>>>>  ?? vhost: Add MMIO helpers for operations on vhost virtqueue
>>>>  ?? vhost: Introduce configfs entry for configuring VHOST
>>>>  ?? virtio_pci: Use request_threaded_irq() instead of
request_irq()
>>>>  ?? rpmsg: virtio_rpmsg_bus: Disable receive virtqueue callback
when
>>>>  ???? reading messages
>>>>  ?? rpmsg: Introduce configfs entry for configuring rpmsg
>>>>  ?? rpmsg: virtio_rpmsg_bus: Add Address Service Notification
support
>>>>  ?? rpmsg: virtio_rpmsg_bus: Move generic rpmsg structure to
>>>>  ???? rpmsg_internal.h
>>>>  ?? virtio: Add ops to allocate and free buffer
>>>>  ?? rpmsg: virtio_rpmsg_bus: Use virtio_alloc_buffer() and
>>>>  ???? virtio_free_buffer()
>>>>  ?? rpmsg: Add VHOST based remote processor messaging bus
>>>>  ?? samples/rpmsg: Setup delayed work to send message
>>>>  ?? samples/rpmsg: Wait for address to be bound to rpdev for
sending
>>>>  ???? message
>>>>  ?? rpmsg.txt: Add Documentation to configure rpmsg using
configfs
>>>>  ?? virtio_pci: Add VIRTIO driver for VHOST on Configurable
PCIe Endpoint
>>>>  ???? device
>>>>  ?? PCI: endpoint: Add EP function driver to provide VHOST
interface
>>>>  ?? NTB: Add a new NTB client driver to implement VIRTIO
functionality
>>>>  ?? NTB: Add a new NTB client driver to implement VHOST
functionality
>>>>  ?? NTB: Describe the ntb_virtio and ntb_vhost client in the
documentation
>>>>
>>>>  ? Documentation/driver-api/ntb.rst????????????? |?? 11 +
>>>>  ? Documentation/rpmsg.txt?????????????????????? |?? 56 +
>>>>  ? drivers/ntb/Kconfig?????????????????????????? |?? 18 +
>>>>  ? drivers/ntb/Makefile????????????????????????? |??? 2 +
>>>>  ? drivers/ntb/ntb_vhost.c?????????????????????? |? 776
+++++++++++
>>>>  ? drivers/ntb/ntb_virtio.c????????????????????? |? 853
++++++++++++
>>>>  ? drivers/ntb/ntb_virtio.h????????????????????? |?? 56 +
>>>>  ? drivers/pci/endpoint/functions/Kconfig??????? |?? 11 +
>>>>  ? drivers/pci/endpoint/functions/Makefile?????? |??? 1 +
>>>>  ? .../pci/endpoint/functions/pci-epf-vhost.c??? | 1144
++++++++++++++++
>>>>  ? drivers/rpmsg/Kconfig???????????????????????? |?? 10 +
>>>>  ? drivers/rpmsg/Makefile??????????????????????? |??? 3 +-
>>>>  ? drivers/rpmsg/rpmsg_cfs.c???????????????????? |? 394 ++++++
>>>>  ? drivers/rpmsg/rpmsg_core.c??????????????????? |??? 7 +
>>>>  ? drivers/rpmsg/rpmsg_internal.h??????????????? |? 136 ++
>>>>  ? drivers/rpmsg/vhost_rpmsg_bus.c?????????????? | 1151
+++++++++++++++++
>>>>  ? drivers/rpmsg/virtio_rpmsg_bus.c????????????? |? 184 ++-
>>>>  ? drivers/vhost/Kconfig???????????????????????? |??? 1 +
>>>>  ? drivers/vhost/Makefile??????????????????????? |??? 2 +-
>>>>  ? drivers/vhost/net.c?????????????????????????? |?? 10 +-
>>>>  ? drivers/vhost/scsi.c????????????????????????? |?? 24 +-
>>>>  ? drivers/vhost/test.c????????????????????????? |?? 17 +-
>>>>  ? drivers/vhost/vdpa.c????????????????????????? |??? 2 +-
>>>>  ? drivers/vhost/vhost.c???????????????????????? |? 730
++++++++++-
>>>>  ? drivers/vhost/vhost_cfs.c???????????????????? |? 341 +++++
>>>>  ? drivers/vhost/vringh.c??????????????????????? |? 332 +++++
>>>>  ? drivers/vhost/vsock.c???????????????????????? |?? 20 +-
>>>>  ? drivers/virtio/Kconfig??????????????????????? |??? 9 +
>>>>  ? drivers/virtio/Makefile?????????????????????? |??? 1 +
>>>>  ? drivers/virtio/virtio_pci_common.c??????????? |?? 25 +-
>>>>  ? drivers/virtio/virtio_pci_epf.c?????????????? |? 670
++++++++++
>>>>  ? include/linux/mod_devicetable.h?????????????? |??? 6 +
>>>>  ? include/linux/rpmsg.h???????????????????????? |??? 6 +
>>>>  ? {drivers/vhost => include/linux}/vhost.h????? |? 132 +-
>>>>  ? include/linux/virtio.h??????????????????????? |??? 3 +
>>>>  ? include/linux/virtio_config.h???????????????? |?? 42 +
>>>>  ? include/linux/vringh.h??????????????????????? |?? 46 +
>>>>  ? samples/rpmsg/rpmsg_client_sample.c?????????? |?? 32 +-
>>>>  ? tools/virtio/virtio_test.c??????????????????? |??? 2 +-
>>>>  ? 39 files changed, 7083 insertions(+), 183 deletions(-)
>>>>  ? create mode 100644 drivers/ntb/ntb_vhost.c
>>>>  ? create mode 100644 drivers/ntb/ntb_virtio.c
>>>>  ? create mode 100644 drivers/ntb/ntb_virtio.h
>>>>  ? create mode 100644
drivers/pci/endpoint/functions/pci-epf-vhost.c
>>>>  ? create mode 100644 drivers/rpmsg/rpmsg_cfs.c
>>>>  ? create mode 100644 drivers/rpmsg/vhost_rpmsg_bus.c
>>>>  ? create mode 100644 drivers/vhost/vhost_cfs.c
>>>>  ? create mode 100644 drivers/virtio/virtio_pci_epf.c
>>>>  ? rename {drivers/vhost => include/linux}/vhost.h (66%)
>>>>
>>>> -- 
>>>> 2.17.1
>>>>

On 2020/7/6 ??5:32, Kishon Vijay Abraham I wrote:
> Hi Jason,
>
> On 7/3/2020 12:46 PM, Jason Wang wrote:
>> On 2020/7/2 ??9:35, Kishon Vijay Abraham I wrote:
>>> Hi Jason,
>>>
>>> On 7/2/2020 3:40 PM, Jason Wang wrote:
>>>> On 2020/7/2 ??5:51, Michael S. Tsirkin wrote:
>>>>> On Thu, Jul 02, 2020 at 01:51:21PM +0530, Kishon Vijay
Abraham I wrote:
>>>>>> This series enhances Linux Vhost support to enable
SoC-to-SoC
>>>>>> communication over MMIO. This series enables rpmsg
communication between
>>>>>> two SoCs using both PCIe RC<->EP and
HOST1-NTB-HOST2
>>>>>>
>>>>>> 1) Modify vhost to use standard Linux driver model
>>>>>> 2) Add support in vring to access virtqueue over MMIO
>>>>>> 3) Add vhost client driver for rpmsg
>>>>>> 4) Add PCIe RC driver (uses virtio) and PCIe EP driver
(uses vhost) for
>>>>>>  ???? rpmsg communication between two SoCs connected to
each other
>>>>>> 5) Add NTB Virtio driver and NTB Vhost driver for rpmsg
communication
>>>>>>  ???? between two SoCs connected via NTB
>>>>>> 6) Add configfs to configure the components
>>>>>>
>>>>>> UseCase1 :
>>>>>>
>>>>>>  ?? VHOST RPMSG???????????????????? VIRTIO RPMSG
>>>>>>  ??????? +?????????????????????????????? +
>>>>>>  ??????? |?????????????????????????????? |
>>>>>>  ??????? |?????????????????????????????? |
>>>>>>  ??????? |?????????????????????????????? |
>>>>>>  ??????? |?????????????????????????????? |
>>>>>> +-----v------+???????????????? +------v-------+
>>>>>> |?? Linux??? |???????????????? |???? Linux??? |
>>>>>> |? Endpoint? |???????????????? | Root Complex |
>>>>>> |??????????? <----------------->????????????? |
>>>>>> |??????????? |???????????????? |????????????? |
>>>>>> |??? SOC1??? |???????????????? |???? SOC2???? |
>>>>>> +------------+???????????????? +--------------+
>>>>>>
>>>>>> UseCase 2:
>>>>>>
>>>>>>  ?????? VHOST
RPMSG????????????????????????????????????? VIRTIO RPMSG
>>>>>>  ???????????
+???????????????????????????????????????????????? +
>>>>>>  ???????????
|???????????????????????????????????????????????? |
>>>>>>  ???????????
|???????????????????????????????????????????????? |
>>>>>>  ???????????
|???????????????????????????????????????????????? |
>>>>>>  ???????????
|???????????????????????????????????????????????? |
>>>>>>  ???? +------v------+??????????????????????????????????
+------v------+
>>>>>>  ???? |???????????? |??????????????????????????????????
|???????????? |
>>>>>>  ???? |??? HOST1??? |??????????????????????????????????
|??? HOST2??? |
>>>>>>  ???? |???????????? |??????????????????????????????????
|???????????? |
>>>>>>  ???? +------^------+??????????????????????????????????
+------^------+
>>>>>>  ???????????
|???????????????????????????????????????????????? |
>>>>>>  ???????????
|???????????????????????????????????????????????? |
>>>>>>
+---------------------------------------------------------------------+
>>>>>> |? +------v------+??????????????????????????????????
+------v------+? |
>>>>>> |? |???????????? |??????????????????????????????????
|???????????? |? |
>>>>>> |? |???? EP????? |??????????????????????????????????
|???? EP????? |? |
>>>>>> |? | CONTROLLER1 |?????????????????????????????????? |
CONTROLLER2 |? |
>>>>>> |? |????????????
<----------------------------------->???????????? |? |
>>>>>> |? |???????????? |??????????????????????????????????
|???????????? |? |
>>>>>> |? |???????????? |??????????????????????????????????
|???????????? |? |
>>>>>> |? |???????????? |? SoC With Multiple EP Instances??
|???????????? |? |
>>>>>> |? |???????????? |? (Configured using NTB Function)?
|???????????? |? |
>>>>>> |? +-------------+??????????????????????????????????
+-------------+? |
>>>>>>
+---------------------------------------------------------------------+
>>>>>>
>>>>>> Software Layering:
>>>>>>
>>>>>> The high-level SW layering should look something like
below. This series
>>>>>> adds support only for RPMSG VHOST, however something
similar should be
>>>>>> done for net and scsi. With that any vhost device (PCI,
NTB, Platform
>>>>>> device, user) can use any of the vhost client driver.
>>>>>>
>>>>>>
>>>>>>  ????? +----------------+? +-----------+?
+------------+? +----------+
>>>>>>  ????? |? RPMSG VHOST?? |? | NET VHOST |? | SCSI VHOST
|? |??? X???? |
>>>>>>  ????? +-------^--------+? +-----^-----+?
+-----^------+? +----^-----+
>>>>>>  ????????????? |???????????????? |?????????????
|????????????? |
>>>>>>  ????????????? |???????????????? |?????????????
|????????????? |
>>>>>>  ????????????? |???????????????? |?????????????
|????????????? |
>>>>>>
+-----------v-----------------v--------------v--------------v----------+
>>>>>> |??????????????????????????? VHOST
CORE??????????????????????????????? |
>>>>>>
+--------^---------------^--------------------^------------------^-----+
>>>>>>  ?????????? |?????????????? |???????????????????
|????????????????? |
>>>>>>  ?????????? |?????????????? |???????????????????
|????????????????? |
>>>>>>  ?????????? |?????????????? |???????????????????
|????????????????? |
>>>>>> +--------v-------+? +----v------+?
+----------v----------+? +----v-----+
>>>>>> |? PCI EPF VHOST |? | NTB VHOST |? |PLATFORM DEVICE
VHOST|? |??? X???? |
>>>>>> +----------------+? +-----------+?
+---------------------+? +----------+
>>>>>>
>>>>>> This was initially proposed here [1]
>>>>>>
>>>>>> [1] ->
https://lore.kernel.org/r/2cf00ec4-1ed6-f66e-6897-006d1a5b6390 at ti.com
>>>>> I find this very interesting. A huge patchset so will take
a bit
>>>>> to review, but I certainly plan to do that. Thanks!
>>>> Yes, it would be better if there's a git branch for us to
have a look.
>>> I've pushed the branch
>>> https://github.com/kishon/linux-wip.git vhost_rpmsg_pci_ntb_rfc
>>
>> Thanks
>>
>>
>>>> Btw, I'm not sure I get the big picture, but I vaguely feel
some of the work is
>>>> duplicated with vDPA (e.g the epf transport or vhost bus).
>>> This is about connecting two different HW systems both running
Linux and
>>> doesn't necessarily involve virtualization.
>>
>> Right, this is something similar to VOP
>> (Documentation/misc-devices/mic/mic_overview.rst). The different is the
>> hardware I guess and VOP use userspace application to implement the
device.
> I'd also like to point out, this series tries to have communication
between two
> SoCs in vendor agnostic way. Since this series solves for 2 usecases (PCIe
> RC<->EP and NTB), for the NTB case it directly plugs into NTB
framework and any
> of the HW in NTB below should be able to use a virtio-vhost communication
>
> #ls drivers/ntb/hw/
> amd  epf  idt  intel  mscc
>
> And similarly for the PCIe RC<->EP communication, this adds a generic
endpoint
> function driver and hence any SoC that supports configurable PCIe endpoint
can
> use virtio-vhost communication
>
> # ls drivers/pci/controller/dwc/*ep*
> drivers/pci/controller/dwc/pcie-designware-ep.c
> drivers/pci/controller/dwc/pcie-uniphier-ep.c
> drivers/pci/controller/dwc/pci-layerscape-ep.c

Thanks for those backgrounds.

>
>>
>>>  ? So there is no guest or host as in
>>> virtualization but two entirely different systems connected via
PCIe cable, one
>>> acting as guest and one as host. So one system will provide virtio
>>> functionality reserving memory for virtqueues and the other
provides vhost
>>> functionality providing a way to access the virtqueues in virtio
memory. One is
>>> source and the other is sink and there is no intermediate entity.
(vhost was
>>> probably intermediate entity in virtualization?)
>>
>> (Not a native English speaker) but "vhost" could introduce
some confusion for
>> me since it was use for implementing virtio backend for userspace
drivers. I
>> guess "vringh" could be better.
> Initially I had named this vringh but later decided to choose vhost instead
of
> vringh. vhost is still a virtio backend (not necessarily userspace) though
it
> now resides in an entirely different system. Whatever virtio is for a
frontend
> system, vhost can be that for a backend system. vring can be for accessing
> virtqueue and can be used either in frontend or backend.

Ok.

>>
>>>> Have you considered to implement these through vDPA?
>>> IIUC vDPA only provides an interface to userspace and an in-kernel
rpmsg driver
>>> or vhost net driver is not provided.
>>>
>>> The HW connection looks something like
https://pasteboard.co/JfMVVHC.jpg
>>> (usecase2 above),
>>
>> I see.
>>
>>
>>>  ? all the boards run Linux. The middle board provides NTB
>>> functionality and board on either side provides virtio/vhost
functionality and
>>> transfer data using rpmsg.
>>
>> So I wonder whether it's worthwhile for a new bus. Can we use the
existed
>> virtio-bus/drivers? It might work as, except for the epf transport, we
can
>> introduce a epf "vhost" transport driver.
> IMHO we'll need two buses one for frontend and other for backend
because the
> two components can then co-operate/interact with each other to provide a
> functionality. Though both will seemingly provide similar callbacks, they
are
> both provide symmetrical or complimentary funcitonality and need not be
same or
> identical.
>
> Having the same bus can also create sequencing issues.
>
> If you look at virtio_dev_probe() of virtio_bus
>
> device_features = dev->config->get_features(dev);
>
> Now if we use same bus for both front-end and back-end, both will try to
> get_features when there has been no set_features. Ideally vhost device
should
> be initialized first with the set of features it supports. Vhost and virtio
> should use "status" and "features" complimentarily and
not identically.

Yes, but there's no need for doing status/features passthrough in epf 
vhost drivers.b

>
> virtio device (or frontend) cannot be initialized before vhost device (or
> backend) gets initialized with data such as features. Similarly vhost
(backend)
> cannot access virqueues or buffers before virtio (frontend) sets
> VIRTIO_CONFIG_S_DRIVER_OK whereas that requirement is not there for virtio
as
> the physical memory for virtqueues are created by virtio (frontend).

epf vhost drivers need to implement two devices: vhost(vringh) device 
and virtio device (which is a mediated device). The vhost(vringh) device 
is doing feature negotiation with the virtio device via RC/EP or NTB. 
The virtio device is doing feature negotiation with local virtio 
drivers. If there're feature mismatch, epf vhost drivers and do 
mediation between them.

>
>> It will have virtqueues but only used for the communication between
itself and
>> uppter virtio driver. And it will have vringh queues which will be
probe by
>> virtio epf transport drivers. And it needs to do datacopy between
virtqueue and
>> vringh queues.
>>
>> It works like:
>>
>> virtio drivers <- virtqueue/virtio-bus -> epf vhost drivers <-
vringh queue/epf>
>>
>> The advantages is that there's no need for writing new buses and
drivers.
> I think this will work however there is an addtional copy between vringh
queue
> and virtqueue,

I think not? E.g in use case 1), if we stick to virtio bus, we will have:

virtio-rpmsg (EP) <- virtio ring(1) -> epf vhost driver (EP) <- virtio 
ring(2) -> virtio pci (RC) <-> virtio rpmsg (RC)

What epf vhost driver did is to read from virtio ring(1) about the 
buffer len and addr and them DMA to Linux(RC)?

> in some cases adds latency because of forwarding interrupts
> between vhost and virtio driver, vhost drivers providing features (which
means
> it has to be aware of which virtio driver will be connected).
> virtio drivers (front end) generally access the buffers from it's local
memory
> but when in backend it can access over MMIO (like PCI EPF or NTB) or
userspace.
>> Does this make sense?
> Two copies in my opinion is an issue but lets get others opinions as well.

Sure.

>
> Thanks for your suggestions!

You're welcome.

Thanks

>
> Regards
> Kishon
>
>> Thanks
>>
>>
>>> Thanks
>>> Kishon
>>>
>>>> Thanks
>>>>
>>>>
>>>>>> Kishon Vijay Abraham I (22):
>>>>>>  ??? vhost: Make _feature_ bits a property of vhost
device
>>>>>>  ??? vhost: Introduce standard Linux driver model in
VHOST
>>>>>>  ??? vhost: Add ops for the VHOST driver to configure
VHOST device
>>>>>>  ??? vringh: Add helpers to access vring in MMIO
>>>>>>  ??? vhost: Add MMIO helpers for operations on vhost
virtqueue
>>>>>>  ??? vhost: Introduce configfs entry for configuring
VHOST
>>>>>>  ??? virtio_pci: Use request_threaded_irq() instead of
request_irq()
>>>>>>  ??? rpmsg: virtio_rpmsg_bus: Disable receive virtqueue
callback when
>>>>>>  ????? reading messages
>>>>>>  ??? rpmsg: Introduce configfs entry for configuring
rpmsg
>>>>>>  ??? rpmsg: virtio_rpmsg_bus: Add Address Service
Notification support
>>>>>>  ??? rpmsg: virtio_rpmsg_bus: Move generic rpmsg
structure to
>>>>>>  ????? rpmsg_internal.h
>>>>>>  ??? virtio: Add ops to allocate and free buffer
>>>>>>  ??? rpmsg: virtio_rpmsg_bus: Use virtio_alloc_buffer()
and
>>>>>>  ????? virtio_free_buffer()
>>>>>>  ??? rpmsg: Add VHOST based remote processor messaging
bus
>>>>>>  ??? samples/rpmsg: Setup delayed work to send message
>>>>>>  ??? samples/rpmsg: Wait for address to be bound to
rpdev for sending
>>>>>>  ????? message
>>>>>>  ??? rpmsg.txt: Add Documentation to configure rpmsg
using configfs
>>>>>>  ??? virtio_pci: Add VIRTIO driver for VHOST on
Configurable PCIe Endpoint
>>>>>>  ????? device
>>>>>>  ??? PCI: endpoint: Add EP function driver to provide
VHOST interface
>>>>>>  ??? NTB: Add a new NTB client driver to implement
VIRTIO functionality
>>>>>>  ??? NTB: Add a new NTB client driver to implement
VHOST functionality
>>>>>>  ??? NTB: Describe the ntb_virtio and ntb_vhost client
in the documentation
>>>>>>
>>>>>>  ?? Documentation/driver-api/ntb.rst????????????? |??
11 +
>>>>>>  ?? Documentation/rpmsg.txt?????????????????????? |??
56 +
>>>>>>  ?? drivers/ntb/Kconfig?????????????????????????? |??
18 +
>>>>>>  ?? drivers/ntb/Makefile????????????????????????? |???
2 +
>>>>>>  ?? drivers/ntb/ntb_vhost.c?????????????????????? |?
776 +++++++++++
>>>>>>  ?? drivers/ntb/ntb_virtio.c????????????????????? |?
853 ++++++++++++
>>>>>>  ?? drivers/ntb/ntb_virtio.h????????????????????? |??
56 +
>>>>>>  ?? drivers/pci/endpoint/functions/Kconfig??????? |??
11 +
>>>>>>  ?? drivers/pci/endpoint/functions/Makefile?????? |???
1 +
>>>>>>  ?? .../pci/endpoint/functions/pci-epf-vhost.c??? |
1144 ++++++++++++++++
>>>>>>  ?? drivers/rpmsg/Kconfig???????????????????????? |??
10 +
>>>>>>  ?? drivers/rpmsg/Makefile??????????????????????? |???
3 +-
>>>>>>  ?? drivers/rpmsg/rpmsg_cfs.c???????????????????? |?
394 ++++++
>>>>>>  ?? drivers/rpmsg/rpmsg_core.c??????????????????? |???
7 +
>>>>>>  ?? drivers/rpmsg/rpmsg_internal.h??????????????? |?
136 ++
>>>>>>  ?? drivers/rpmsg/vhost_rpmsg_bus.c?????????????? |
1151 +++++++++++++++++
>>>>>>  ?? drivers/rpmsg/virtio_rpmsg_bus.c????????????? |?
184 ++-
>>>>>>  ?? drivers/vhost/Kconfig???????????????????????? |???
1 +
>>>>>>  ?? drivers/vhost/Makefile??????????????????????? |???
2 +-
>>>>>>  ?? drivers/vhost/net.c?????????????????????????? |??
10 +-
>>>>>>  ?? drivers/vhost/scsi.c????????????????????????? |??
24 +-
>>>>>>  ?? drivers/vhost/test.c????????????????????????? |??
17 +-
>>>>>>  ?? drivers/vhost/vdpa.c????????????????????????? |???
2 +-
>>>>>>  ?? drivers/vhost/vhost.c???????????????????????? |?
730 ++++++++++-
>>>>>>  ?? drivers/vhost/vhost_cfs.c???????????????????? |?
341 +++++
>>>>>>  ?? drivers/vhost/vringh.c??????????????????????? |?
332 +++++
>>>>>>  ?? drivers/vhost/vsock.c???????????????????????? |??
20 +-
>>>>>>  ?? drivers/virtio/Kconfig??????????????????????? |???
9 +
>>>>>>  ?? drivers/virtio/Makefile?????????????????????? |???
1 +
>>>>>>  ?? drivers/virtio/virtio_pci_common.c??????????? |??
25 +-
>>>>>>  ?? drivers/virtio/virtio_pci_epf.c?????????????? |?
670 ++++++++++
>>>>>>  ?? include/linux/mod_devicetable.h?????????????? |???
6 +
>>>>>>  ?? include/linux/rpmsg.h???????????????????????? |???
6 +
>>>>>>  ?? {drivers/vhost => include/linux}/vhost.h????? |?
132 +-
>>>>>>  ?? include/linux/virtio.h??????????????????????? |???
3 +
>>>>>>  ?? include/linux/virtio_config.h???????????????? |??
42 +
>>>>>>  ?? include/linux/vringh.h??????????????????????? |??
46 +
>>>>>>  ?? samples/rpmsg/rpmsg_client_sample.c?????????? |??
32 +-
>>>>>>  ?? tools/virtio/virtio_test.c??????????????????? |???
2 +-
>>>>>>  ?? 39 files changed, 7083 insertions(+), 183
deletions(-)
>>>>>>  ?? create mode 100644 drivers/ntb/ntb_vhost.c
>>>>>>  ?? create mode 100644 drivers/ntb/ntb_virtio.c
>>>>>>  ?? create mode 100644 drivers/ntb/ntb_virtio.h
>>>>>>  ?? create mode 100644
drivers/pci/endpoint/functions/pci-epf-vhost.c
>>>>>>  ?? create mode 100644 drivers/rpmsg/rpmsg_cfs.c
>>>>>>  ?? create mode 100644 drivers/rpmsg/vhost_rpmsg_bus.c
>>>>>>  ?? create mode 100644 drivers/vhost/vhost_cfs.c
>>>>>>  ?? create mode 100644 drivers/virtio/virtio_pci_epf.c
>>>>>>  ?? rename {drivers/vhost => include/linux}/vhost.h
(66%)
>>>>>>
>>>>>> -- 
>>>>>> 2.17.1
>>>>>>

On 2020/7/7 ??10:45, Kishon Vijay Abraham I wrote:
> Hi Jason,
>
> On 7/7/2020 3:17 PM, Jason Wang wrote:
>> On 2020/7/6 ??5:32, Kishon Vijay Abraham I wrote:
>>> Hi Jason,
>>>
>>> On 7/3/2020 12:46 PM, Jason Wang wrote:
>>>> On 2020/7/2 ??9:35, Kishon Vijay Abraham I wrote:
>>>>> Hi Jason,
>>>>>
>>>>> On 7/2/2020 3:40 PM, Jason Wang wrote:
>>>>>> On 2020/7/2 ??5:51, Michael S. Tsirkin wrote:
>>>>>>> On Thu, Jul 02, 2020 at 01:51:21PM +0530, Kishon
Vijay Abraham I wrote:
>>>>>>>> This series enhances Linux Vhost support to
enable SoC-to-SoC
>>>>>>>> communication over MMIO. This series enables
rpmsg communication between
>>>>>>>> two SoCs using both PCIe RC<->EP and
HOST1-NTB-HOST2
>>>>>>>>
>>>>>>>> 1) Modify vhost to use standard Linux driver
model
>>>>>>>> 2) Add support in vring to access virtqueue
over MMIO
>>>>>>>> 3) Add vhost client driver for rpmsg
>>>>>>>> 4) Add PCIe RC driver (uses virtio) and PCIe EP
driver (uses vhost) for
>>>>>>>>  ????? rpmsg communication between two SoCs
connected to each other
>>>>>>>> 5) Add NTB Virtio driver and NTB Vhost driver
for rpmsg communication
>>>>>>>>  ????? between two SoCs connected via NTB
>>>>>>>> 6) Add configfs to configure the components
>>>>>>>>
>>>>>>>> UseCase1 :
>>>>>>>>
>>>>>>>>  ??? VHOST RPMSG???????????????????? VIRTIO
RPMSG
>>>>>>>>  ???????? +?????????????????????????????? +
>>>>>>>>  ???????? |?????????????????????????????? |
>>>>>>>>  ???????? |?????????????????????????????? |
>>>>>>>>  ???????? |?????????????????????????????? |
>>>>>>>>  ???????? |?????????????????????????????? |
>>>>>>>> +-----v------+???????????????? +------v-------+
>>>>>>>> |?? Linux??? |???????????????? |???? Linux??? |
>>>>>>>> |? Endpoint? |???????????????? | Root Complex |
>>>>>>>> |???????????
<----------------->????????????? |
>>>>>>>> |??????????? |???????????????? |????????????? |
>>>>>>>> |??? SOC1??? |???????????????? |???? SOC2???? |
>>>>>>>> +------------+???????????????? +--------------+
>>>>>>>>
>>>>>>>> UseCase 2:
>>>>>>>>
>>>>>>>>  ??????? VHOST
RPMSG????????????????????????????????????? VIRTIO RPMSG
>>>>>>>>  ????????????
+???????????????????????????????????????????????? +
>>>>>>>>  ????????????
|???????????????????????????????????????????????? |
>>>>>>>>  ????????????
|???????????????????????????????????????????????? |
>>>>>>>>  ????????????
|???????????????????????????????????????????????? |
>>>>>>>>  ????????????
|???????????????????????????????????????????????? |
>>>>>>>>  ?????
+------v------+?????????????????????????????????? +------v------+
>>>>>>>>  ????? |????????????
|?????????????????????????????????? |???????????? |
>>>>>>>>  ????? |??? HOST1???
|?????????????????????????????????? |??? HOST2??? |
>>>>>>>>  ????? |????????????
|?????????????????????????????????? |???????????? |
>>>>>>>>  ?????
+------^------+?????????????????????????????????? +------^------+
>>>>>>>>  ????????????
|???????????????????????????????????????????????? |
>>>>>>>>  ????????????
|???????????????????????????????????????????????? |
>>>>>>>>
+---------------------------------------------------------------------+
>>>>>>>> |?
+------v------+?????????????????????????????????? +------v------+? |
>>>>>>>> |? |????????????
|?????????????????????????????????? |???????????? |? |
>>>>>>>> |? |???? EP?????
|?????????????????????????????????? |???? EP????? |? |
>>>>>>>> |? | CONTROLLER1
|?????????????????????????????????? | CONTROLLER2 |? |
>>>>>>>> |? |????????????
<----------------------------------->???????????? |? |
>>>>>>>> |? |????????????
|?????????????????????????????????? |???????????? |? |
>>>>>>>> |? |????????????
|?????????????????????????????????? |???????????? |? |
>>>>>>>> |? |???????????? |? SoC With Multiple EP
Instances?? |???????????? |? |
>>>>>>>> |? |???????????? |? (Configured using NTB
Function)? |???????????? |? |
>>>>>>>> |?
+-------------+?????????????????????????????????? +-------------+? |
>>>>>>>>
+---------------------------------------------------------------------+
>>>>>>>>
>>>>>>>> Software Layering:
>>>>>>>>
>>>>>>>> The high-level SW layering should look
something like below. This series
>>>>>>>> adds support only for RPMSG VHOST, however
something similar should be
>>>>>>>> done for net and scsi. With that any vhost
device (PCI, NTB, Platform
>>>>>>>> device, user) can use any of the vhost client
driver.
>>>>>>>>
>>>>>>>>
>>>>>>>>  ?????? +----------------+? +-----------+?
+------------+? +----------+
>>>>>>>>  ?????? |? RPMSG VHOST?? |? | NET VHOST |? |
SCSI VHOST |? |??? X???? |
>>>>>>>>  ?????? +-------^--------+? +-----^-----+?
+-----^------+? +----^-----+
>>>>>>>>  ?????????????? |????????????????
|????????????? |????????????? |
>>>>>>>>  ?????????????? |????????????????
|????????????? |????????????? |
>>>>>>>>  ?????????????? |????????????????
|????????????? |????????????? |
>>>>>>>>
+-----------v-----------------v--------------v--------------v----------+
>>>>>>>> |??????????????????????????? VHOST
CORE??????????????????????????????? |
>>>>>>>>
+--------^---------------^--------------------^------------------^-----+
>>>>>>>>  ??????????? |??????????????
|??????????????????? |????????????????? |
>>>>>>>>  ??????????? |??????????????
|??????????????????? |????????????????? |
>>>>>>>>  ??????????? |??????????????
|??????????????????? |????????????????? |
>>>>>>>> +--------v-------+? +----v------+?
+----------v----------+? +----v-----+
>>>>>>>> |? PCI EPF VHOST |? | NTB VHOST |? |PLATFORM
DEVICE VHOST|? |??? X???? |
>>>>>>>> +----------------+? +-----------+?
+---------------------+? +----------+
>>>>>>>>
>>>>>>>> This was initially proposed here [1]
>>>>>>>>
>>>>>>>> [1] ->
>>>>>>>>
https://lore.kernel.org/r/2cf00ec4-1ed6-f66e-6897-006d1a5b6390 at ti.com
>>>>>>> I find this very interesting. A huge patchset so
will take a bit
>>>>>>> to review, but I certainly plan to do that. Thanks!
>>>>>> Yes, it would be better if there's a git branch for
us to have a look.
>>>>> I've pushed the branch
>>>>> https://github.com/kishon/linux-wip.git
vhost_rpmsg_pci_ntb_rfc
>>>> Thanks
>>>>
>>>>
>>>>>> Btw, I'm not sure I get the big picture, but I
vaguely feel some of the
>>>>>> work is
>>>>>> duplicated with vDPA (e.g the epf transport or vhost
bus).
>>>>> This is about connecting two different HW systems both
running Linux and
>>>>> doesn't necessarily involve virtualization.
>>>> Right, this is something similar to VOP
>>>> (Documentation/misc-devices/mic/mic_overview.rst). The
different is the
>>>> hardware I guess and VOP use userspace application to implement
the device.
>>> I'd also like to point out, this series tries to have
communication between two
>>> SoCs in vendor agnostic way. Since this series solves for 2
usecases (PCIe
>>> RC<->EP and NTB), for the NTB case it directly plugs into NTB
framework and any
>>> of the HW in NTB below should be able to use a virtio-vhost
communication
>>>
>>> #ls drivers/ntb/hw/
>>> amd? epf? idt? intel? mscc
>>>
>>> And similarly for the PCIe RC<->EP communication, this adds a
generic endpoint
>>> function driver and hence any SoC that supports configurable PCIe
endpoint can
>>> use virtio-vhost communication
>>>
>>> # ls drivers/pci/controller/dwc/*ep*
>>> drivers/pci/controller/dwc/pcie-designware-ep.c
>>> drivers/pci/controller/dwc/pcie-uniphier-ep.c
>>> drivers/pci/controller/dwc/pci-layerscape-ep.c
>>
>> Thanks for those backgrounds.
>>
>>
>>>>>  ?? So there is no guest or host as in
>>>>> virtualization but two entirely different systems connected
via PCIe cable,
>>>>> one
>>>>> acting as guest and one as host. So one system will provide
virtio
>>>>> functionality reserving memory for virtqueues and the other
provides vhost
>>>>> functionality providing a way to access the virtqueues in
virtio memory.
>>>>> One is
>>>>> source and the other is sink and there is no intermediate
entity. (vhost was
>>>>> probably intermediate entity in virtualization?)
>>>> (Not a native English speaker) but "vhost" could
introduce some confusion for
>>>> me since it was use for implementing virtio backend for
userspace drivers. I
>>>> guess "vringh" could be better.
>>> Initially I had named this vringh but later decided to choose vhost
instead of
>>> vringh. vhost is still a virtio backend (not necessarily userspace)
though it
>>> now resides in an entirely different system. Whatever virtio is for
a frontend
>>> system, vhost can be that for a backend system. vring can be for
accessing
>>> virtqueue and can be used either in frontend or backend.
>>
>> Ok.
>>
>>
>>>>>> Have you considered to implement these through vDPA?
>>>>> IIUC vDPA only provides an interface to userspace and an
in-kernel rpmsg
>>>>> driver
>>>>> or vhost net driver is not provided.
>>>>>
>>>>> The HW connection looks something like
https://pasteboard.co/JfMVVHC.jpg
>>>>> (usecase2 above),
>>>> I see.
>>>>
>>>>
>>>>>  ?? all the boards run Linux. The middle board provides NTB
>>>>> functionality and board on either side provides
virtio/vhost functionality and
>>>>> transfer data using rpmsg.
>>>> So I wonder whether it's worthwhile for a new bus. Can we
use the existed
>>>> virtio-bus/drivers? It might work as, except for the epf
transport, we can
>>>> introduce a epf "vhost" transport driver.
>>> IMHO we'll need two buses one for frontend and other for
backend because the
>>> two components can then co-operate/interact with each other to
provide a
>>> functionality. Though both will seemingly provide similar
callbacks, they are
>>> both provide symmetrical or complimentary funcitonality and need
not be same or
>>> identical.
>>>
>>> Having the same bus can also create sequencing issues.
>>>
>>> If you look at virtio_dev_probe() of virtio_bus
>>>
>>> device_features = dev->config->get_features(dev);
>>>
>>> Now if we use same bus for both front-end and back-end, both will
try to
>>> get_features when there has been no set_features. Ideally vhost
device should
>>> be initialized first with the set of features it supports. Vhost
and virtio
>>> should use "status" and "features"
complimentarily and not identically.
>>
>> Yes, but there's no need for doing status/features passthrough in
epf vhost
>> drivers.b
>>
>>
>>> virtio device (or frontend) cannot be initialized before vhost
device (or
>>> backend) gets initialized with data such as features. Similarly
vhost (backend)
>>> cannot access virqueues or buffers before virtio (frontend) sets
>>> VIRTIO_CONFIG_S_DRIVER_OK whereas that requirement is not there for
virtio as
>>> the physical memory for virtqueues are created by virtio
(frontend).
>>
>> epf vhost drivers need to implement two devices: vhost(vringh) device
and
>> virtio device (which is a mediated device). The vhost(vringh) device is
doing
>> feature negotiation with the virtio device via RC/EP or NTB. The virtio
device
>> is doing feature negotiation with local virtio drivers. If there're
feature
>> mismatch, epf vhost drivers and do mediation between them.
> Here epf vhost should be initialized with a set of features for it to
negotiate
> either as vhost device or virtio device no? Where should the initial
feature
> set for epf vhost come from?

I think it can work as:

1) Having an initial features (hard coded in the code) set X in epf vhost
2) Using this X for both virtio device and vhost(vringh) device
3) local virtio driver will negotiate with virtio device with feature set Y
4) remote virtio driver will negotiate with vringh device with feature set Z
5) mediate between feature Y and feature Z since both Y and Z are a 
subset of X

>>
>>>> It will have virtqueues but only used for the communication
between itself and
>>>> uppter virtio driver. And it will have vringh queues which will
be probe by
>>>> virtio epf transport drivers. And it needs to do datacopy
between virtqueue and
>>>> vringh queues.
>>>>
>>>> It works like:
>>>>
>>>> virtio drivers <- virtqueue/virtio-bus -> epf vhost
drivers <- vringh
>>>> queue/epf>
>>>>
>>>> The advantages is that there's no need for writing new
buses and drivers.
>>> I think this will work however there is an addtional copy between
vringh queue
>>> and virtqueue,
>>
>> I think not? E.g in use case 1), if we stick to virtio bus, we will
have:
>>
>> virtio-rpmsg (EP) <- virtio ring(1) -> epf vhost driver (EP)
<- virtio ring(2)
>> -> virtio pci (RC) <-> virtio rpmsg (RC)
> IIUC epf vhost driver (EP) will access virtio ring(2) using vringh?

Yes.

> And virtio
> ring(2) is created by virtio pci (RC).

Yes.

>> What epf vhost driver did is to read from virtio ring(1) about the
buffer len
>> and addr and them DMA to Linux(RC)?
> okay, I made some optimization here where vhost-rpmsg using a helper writes
a
> buffer from rpmsg's upper layer directly to remote Linux (RC) as
against here
> were it has to be first written to virtio ring (1).
>
> Thinking how this would look for NTB
> virtio-rpmsg (HOST1) <- virtio ring(1) -> NTB(HOST1) <->
NTB(HOST2)  <- virtio
> ring(2) -> virtio-rpmsg (HOST2)
>
> Here the NTB(HOST1) will access the virtio ring(2) using vringh?

Yes, I think so it needs to use vring to access virtio ring (1) as well.

>
> Do you also think this will work seamlessly with virtio_net.c,
virtio_blk.c?

Yes.

>
> I'd like to get clarity on two things in the approach you suggested,
one is
> features (since epf vhost should ideally be transparent to any virtio
driver)

We can have have an array of pre-defined features indexed by virtio 
device id in the code.

> and the other is how certain inputs to virtio device such as number of
buffers
> be determined.

We can start from hard coded the value like 256, or introduce some API 
for user to change the value.

>
> Thanks again for your suggestions!

You're welcome.

Note that I just want to check whether or not we can reuse the virtio 
bus/driver. It's something similar to what you proposed in Software 
Layering but we just replace "vhost core" with "virtio bus"
and move the
vhost core below epf/ntb/platform transport.

Thanks

>
> Regards
> Kishon
>
>>
>>> in some cases adds latency because of forwarding interrupts
>>> between vhost and virtio driver, vhost drivers providing features
(which means
>>> it has to be aware of which virtio driver will be connected).
>>> virtio drivers (front end) generally access the buffers from
it's local memory
>>> but when in backend it can access over MMIO (like PCI EPF or NTB)
or userspace.
>>>> Does this make sense?
>>> Two copies in my opinion is an issue but lets get others opinions
as well.
>>
>> Sure.
>>
>>
>>> Thanks for your suggestions!
>>
>> You're welcome.
>>
>> Thanks
>>
>>
>>> Regards
>>> Kishon
>>>
>>>> Thanks
>>>>
>>>>
>>>>> Thanks
>>>>> Kishon
>>>>>
>>>>>> Thanks
>>>>>>
>>>>>>
>>>>>>>> Kishon Vijay Abraham I (22):
>>>>>>>>  ???? vhost: Make _feature_ bits a property of
vhost device
>>>>>>>>  ???? vhost: Introduce standard Linux driver
model in VHOST
>>>>>>>>  ???? vhost: Add ops for the VHOST driver to
configure VHOST device
>>>>>>>>  ???? vringh: Add helpers to access vring in
MMIO
>>>>>>>>  ???? vhost: Add MMIO helpers for operations on
vhost virtqueue
>>>>>>>>  ???? vhost: Introduce configfs entry for
configuring VHOST
>>>>>>>>  ???? virtio_pci: Use request_threaded_irq()
instead of request_irq()
>>>>>>>>  ???? rpmsg: virtio_rpmsg_bus: Disable receive
virtqueue callback when
>>>>>>>>  ?????? reading messages
>>>>>>>>  ???? rpmsg: Introduce configfs entry for
configuring rpmsg
>>>>>>>>  ???? rpmsg: virtio_rpmsg_bus: Add Address
Service Notification support
>>>>>>>>  ???? rpmsg: virtio_rpmsg_bus: Move generic
rpmsg structure to
>>>>>>>>  ?????? rpmsg_internal.h
>>>>>>>>  ???? virtio: Add ops to allocate and free
buffer
>>>>>>>>  ???? rpmsg: virtio_rpmsg_bus: Use
virtio_alloc_buffer() and
>>>>>>>>  ?????? virtio_free_buffer()
>>>>>>>>  ???? rpmsg: Add VHOST based remote processor
messaging bus
>>>>>>>>  ???? samples/rpmsg: Setup delayed work to send
message
>>>>>>>>  ???? samples/rpmsg: Wait for address to be
bound to rpdev for sending
>>>>>>>>  ?????? message
>>>>>>>>  ???? rpmsg.txt: Add Documentation to configure
rpmsg using configfs
>>>>>>>>  ???? virtio_pci: Add VIRTIO driver for VHOST
on Configurable PCIe Endpoint
>>>>>>>>  ?????? device
>>>>>>>>  ???? PCI: endpoint: Add EP function driver to
provide VHOST interface
>>>>>>>>  ???? NTB: Add a new NTB client driver to
implement VIRTIO functionality
>>>>>>>>  ???? NTB: Add a new NTB client driver to
implement VHOST functionality
>>>>>>>>  ???? NTB: Describe the ntb_virtio and
ntb_vhost client in the documentation
>>>>>>>>
>>>>>>>>  ???
Documentation/driver-api/ntb.rst????????????? |?? 11 +
>>>>>>>>  ???
Documentation/rpmsg.txt?????????????????????? |?? 56 +
>>>>>>>>  ???
drivers/ntb/Kconfig?????????????????????????? |?? 18 +
>>>>>>>>  ???
drivers/ntb/Makefile????????????????????????? |??? 2 +
>>>>>>>>  ???
drivers/ntb/ntb_vhost.c?????????????????????? |? 776 +++++++++++
>>>>>>>>  ???
drivers/ntb/ntb_virtio.c????????????????????? |? 853 ++++++++++++
>>>>>>>>  ???
drivers/ntb/ntb_virtio.h????????????????????? |?? 56 +
>>>>>>>>  ???
drivers/pci/endpoint/functions/Kconfig??????? |?? 11 +
>>>>>>>>  ???
drivers/pci/endpoint/functions/Makefile?????? |??? 1 +
>>>>>>>>  ???
.../pci/endpoint/functions/pci-epf-vhost.c??? | 1144 ++++++++++++++++
>>>>>>>>  ???
drivers/rpmsg/Kconfig???????????????????????? |?? 10 +
>>>>>>>>  ???
drivers/rpmsg/Makefile??????????????????????? |??? 3 +-
>>>>>>>>  ???
drivers/rpmsg/rpmsg_cfs.c???????????????????? |? 394 ++++++
>>>>>>>>  ???
drivers/rpmsg/rpmsg_core.c??????????????????? |??? 7 +
>>>>>>>>  ???
drivers/rpmsg/rpmsg_internal.h??????????????? |? 136 ++
>>>>>>>>  ???
drivers/rpmsg/vhost_rpmsg_bus.c?????????????? | 1151 +++++++++++++++++
>>>>>>>>  ???
drivers/rpmsg/virtio_rpmsg_bus.c????????????? |? 184 ++-
>>>>>>>>  ???
drivers/vhost/Kconfig???????????????????????? |??? 1 +
>>>>>>>>  ???
drivers/vhost/Makefile??????????????????????? |??? 2 +-
>>>>>>>>  ???
drivers/vhost/net.c?????????????????????????? |?? 10 +-
>>>>>>>>  ???
drivers/vhost/scsi.c????????????????????????? |?? 24 +-
>>>>>>>>  ???
drivers/vhost/test.c????????????????????????? |?? 17 +-
>>>>>>>>  ???
drivers/vhost/vdpa.c????????????????????????? |??? 2 +-
>>>>>>>>  ???
drivers/vhost/vhost.c???????????????????????? |? 730 ++++++++++-
>>>>>>>>  ???
drivers/vhost/vhost_cfs.c???????????????????? |? 341 +++++
>>>>>>>>  ???
drivers/vhost/vringh.c??????????????????????? |? 332 +++++
>>>>>>>>  ???
drivers/vhost/vsock.c???????????????????????? |?? 20 +-
>>>>>>>>  ???
drivers/virtio/Kconfig??????????????????????? |??? 9 +
>>>>>>>>  ???
drivers/virtio/Makefile?????????????????????? |??? 1 +
>>>>>>>>  ???
drivers/virtio/virtio_pci_common.c??????????? |?? 25 +-
>>>>>>>>  ???
drivers/virtio/virtio_pci_epf.c?????????????? |? 670 ++++++++++
>>>>>>>>  ???
include/linux/mod_devicetable.h?????????????? |??? 6 +
>>>>>>>>  ???
include/linux/rpmsg.h???????????????????????? |??? 6 +
>>>>>>>>  ??? {drivers/vhost =>
include/linux}/vhost.h????? |? 132 +-
>>>>>>>>  ???
include/linux/virtio.h??????????????????????? |??? 3 +
>>>>>>>>  ???
include/linux/virtio_config.h???????????????? |?? 42 +
>>>>>>>>  ???
include/linux/vringh.h??????????????????????? |?? 46 +
>>>>>>>>  ???
samples/rpmsg/rpmsg_client_sample.c?????????? |?? 32 +-
>>>>>>>>  ???
tools/virtio/virtio_test.c??????????????????? |??? 2 +-
>>>>>>>>  ??? 39 files changed, 7083 insertions(+), 183
deletions(-)
>>>>>>>>  ??? create mode 100644 drivers/ntb/ntb_vhost.c
>>>>>>>>  ??? create mode 100644
drivers/ntb/ntb_virtio.c
>>>>>>>>  ??? create mode 100644
drivers/ntb/ntb_virtio.h
>>>>>>>>  ??? create mode 100644
drivers/pci/endpoint/functions/pci-epf-vhost.c
>>>>>>>>  ??? create mode 100644
drivers/rpmsg/rpmsg_cfs.c
>>>>>>>>  ??? create mode 100644
drivers/rpmsg/vhost_rpmsg_bus.c
>>>>>>>>  ??? create mode 100644
drivers/vhost/vhost_cfs.c
>>>>>>>>  ??? create mode 100644
drivers/virtio/virtio_pci_epf.c
>>>>>>>>  ??? rename {drivers/vhost =>
include/linux}/vhost.h (66%)
>>>>>>>>
>>>>>>>> -- 
>>>>>>>> 2.17.1
>>>>>>>>

On 2020/7/8 ??9:13, Kishon Vijay Abraham I wrote:
> Hi Jason,
>
> On 7/8/2020 4:52 PM, Jason Wang wrote:
>> On 2020/7/7 ??10:45, Kishon Vijay Abraham I wrote:
>>> Hi Jason,
>>>
>>> On 7/7/2020 3:17 PM, Jason Wang wrote:
>>>> On 2020/7/6 ??5:32, Kishon Vijay Abraham I wrote:
>>>>> Hi Jason,
>>>>>
>>>>> On 7/3/2020 12:46 PM, Jason Wang wrote:
>>>>>> On 2020/7/2 ??9:35, Kishon Vijay Abraham I wrote:
>>>>>>> Hi Jason,
>>>>>>>
>>>>>>> On 7/2/2020 3:40 PM, Jason Wang wrote:
>>>>>>>> On 2020/7/2 ??5:51, Michael S. Tsirkin wrote:
>>>>>>>>> On Thu, Jul 02, 2020 at 01:51:21PM +0530,
Kishon Vijay Abraham I wrote:
>>>>>>>>>> This series enhances Linux Vhost
support to enable SoC-to-SoC
>>>>>>>>>> communication over MMIO. This series
enables rpmsg communication between
>>>>>>>>>> two SoCs using both PCIe RC<->EP
and HOST1-NTB-HOST2
>>>>>>>>>>
>>>>>>>>>> 1) Modify vhost to use standard Linux
driver model
>>>>>>>>>> 2) Add support in vring to access
virtqueue over MMIO
>>>>>>>>>> 3) Add vhost client driver for rpmsg
>>>>>>>>>> 4) Add PCIe RC driver (uses virtio) and
PCIe EP driver (uses vhost) for
>>>>>>>>>>  ?????? rpmsg communication between two
SoCs connected to each other
>>>>>>>>>> 5) Add NTB Virtio driver and NTB Vhost
driver for rpmsg communication
>>>>>>>>>>  ?????? between two SoCs connected via
NTB
>>>>>>>>>> 6) Add configfs to configure the
components
>>>>>>>>>>
>>>>>>>>>> UseCase1 :
>>>>>>>>>>
>>>>>>>>>>  ???? VHOST RPMSG????????????????????
VIRTIO RPMSG
>>>>>>>>>>  ?????????
+?????????????????????????????? +
>>>>>>>>>>  ?????????
|?????????????????????????????? |
>>>>>>>>>>  ?????????
|?????????????????????????????? |
>>>>>>>>>>  ?????????
|?????????????????????????????? |
>>>>>>>>>>  ?????????
|?????????????????????????????? |
>>>>>>>>>> +-----v------+????????????????
+------v-------+
>>>>>>>>>> |?? Linux??? |???????????????? |????
Linux??? |
>>>>>>>>>> |? Endpoint? |???????????????? | Root
Complex |
>>>>>>>>>> |???????????
<----------------->????????????? |
>>>>>>>>>> |??????????? |????????????????
|????????????? |
>>>>>>>>>> |??? SOC1??? |???????????????? |????
SOC2???? |
>>>>>>>>>> +------------+????????????????
+--------------+
>>>>>>>>>>
>>>>>>>>>> UseCase 2:
>>>>>>>>>>
>>>>>>>>>>  ???????? VHOST
RPMSG????????????????????????????????????? VIRTIO RPMSG
>>>>>>>>>>  ?????????????
+???????????????????????????????????????????????? +
>>>>>>>>>>  ?????????????
|???????????????????????????????????????????????? |
>>>>>>>>>>  ?????????????
|???????????????????????????????????????????????? |
>>>>>>>>>>  ?????????????
|???????????????????????????????????????????????? |
>>>>>>>>>>  ?????????????
|???????????????????????????????????????????????? |
>>>>>>>>>>  ??????
+------v------+?????????????????????????????????? +------v------+
>>>>>>>>>>  ?????? |????????????
|?????????????????????????????????? |???????????? |
>>>>>>>>>>  ?????? |??? HOST1???
|?????????????????????????????????? |??? HOST2??? |
>>>>>>>>>>  ?????? |????????????
|?????????????????????????????????? |???????????? |
>>>>>>>>>>  ??????
+------^------+?????????????????????????????????? +------^------+
>>>>>>>>>>  ?????????????
|???????????????????????????????????????????????? |
>>>>>>>>>>  ?????????????
|???????????????????????????????????????????????? |
>>>>>>>>>>
+---------------------------------------------------------------------+
>>>>>>>>>> |?
+------v------+?????????????????????????????????? +------v------+? |
>>>>>>>>>> |? |????????????
|?????????????????????????????????? |???????????? |? |
>>>>>>>>>> |? |???? EP?????
|?????????????????????????????????? |???? EP????? |? |
>>>>>>>>>> |? | CONTROLLER1
|?????????????????????????????????? | CONTROLLER2 |? |
>>>>>>>>>> |? |????????????
<----------------------------------->???????????? |? |
>>>>>>>>>> |? |????????????
|?????????????????????????????????? |???????????? |? |
>>>>>>>>>> |? |????????????
|?????????????????????????????????? |???????????? |? |
>>>>>>>>>> |? |???????????? |? SoC With Multiple
EP Instances?? |???????????? |? |
>>>>>>>>>> |? |???????????? |? (Configured using
NTB Function)? |???????????? |? |
>>>>>>>>>> |?
+-------------+?????????????????????????????????? +-------------+? |
>>>>>>>>>>
+---------------------------------------------------------------------+
>>>>>>>>>>
>>>>>>>>>> Software Layering:
>>>>>>>>>>
>>>>>>>>>> The high-level SW layering should look
something like below. This series
>>>>>>>>>> adds support only for RPMSG VHOST,
however something similar should be
>>>>>>>>>> done for net and scsi. With that any
vhost device (PCI, NTB, Platform
>>>>>>>>>> device, user) can use any of the vhost
client driver.
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>  ??????? +----------------+?
+-----------+? +------------+? +----------+
>>>>>>>>>>  ??????? |? RPMSG VHOST?? |? | NET
VHOST |? | SCSI VHOST |? |??? X???? |
>>>>>>>>>>  ??????? +-------^--------+?
+-----^-----+? +-----^------+? +----^-----+
>>>>>>>>>>  ??????????????? |????????????????
|????????????? |????????????? |
>>>>>>>>>>  ??????????????? |????????????????
|????????????? |????????????? |
>>>>>>>>>>  ??????????????? |????????????????
|????????????? |????????????? |
>>>>>>>>>>
+-----------v-----------------v--------------v--------------v----------+
>>>>>>>>>> |??????????????????????????? VHOST
CORE??????????????????????????????? |
>>>>>>>>>>
+--------^---------------^--------------------^------------------^-----+
>>>>>>>>>>  ???????????? |??????????????
|??????????????????? |????????????????? |
>>>>>>>>>>  ???????????? |??????????????
|??????????????????? |????????????????? |
>>>>>>>>>>  ???????????? |??????????????
|??????????????????? |????????????????? |
>>>>>>>>>> +--------v-------+? +----v------+?
+----------v----------+? +----v-----+
>>>>>>>>>> |? PCI EPF VHOST |? | NTB VHOST |?
|PLATFORM DEVICE VHOST|? |??? X???? |
>>>>>>>>>> +----------------+? +-----------+?
+---------------------+? +----------+
>>>>>>>>>>
>>>>>>>>>> This was initially proposed here [1]
>>>>>>>>>>
>>>>>>>>>> [1] ->
>>>>>>>>>>
https://lore.kernel.org/r/2cf00ec4-1ed6-f66e-6897-006d1a5b6390 at ti.com
>>>>>>>>> I find this very interesting. A huge
patchset so will take a bit
>>>>>>>>> to review, but I certainly plan to do that.
Thanks!
>>>>>>>> Yes, it would be better if there's a git
branch for us to have a look.
>>>>>>> I've pushed the branch
>>>>>>> https://github.com/kishon/linux-wip.git
vhost_rpmsg_pci_ntb_rfc
>>>>>> Thanks
>>>>>>
>>>>>>
>>>>>>>> Btw, I'm not sure I get the big picture,
but I vaguely feel some of the
>>>>>>>> work is
>>>>>>>> duplicated with vDPA (e.g the epf transport or
vhost bus).
>>>>>>> This is about connecting two different HW systems
both running Linux and
>>>>>>> doesn't necessarily involve virtualization.
>>>>>> Right, this is something similar to VOP
>>>>>> (Documentation/misc-devices/mic/mic_overview.rst). The
different is the
>>>>>> hardware I guess and VOP use userspace application to
implement the device.
>>>>> I'd also like to point out, this series tries to have
communication between
>>>>> two
>>>>> SoCs in vendor agnostic way. Since this series solves for 2
usecases (PCIe
>>>>> RC<->EP and NTB), for the NTB case it directly plugs
into NTB framework and
>>>>> any
>>>>> of the HW in NTB below should be able to use a virtio-vhost
communication
>>>>>
>>>>> #ls drivers/ntb/hw/
>>>>> amd? epf? idt? intel? mscc
>>>>>
>>>>> And similarly for the PCIe RC<->EP communication,
this adds a generic endpoint
>>>>> function driver and hence any SoC that supports
configurable PCIe endpoint can
>>>>> use virtio-vhost communication
>>>>>
>>>>> # ls drivers/pci/controller/dwc/*ep*
>>>>> drivers/pci/controller/dwc/pcie-designware-ep.c
>>>>> drivers/pci/controller/dwc/pcie-uniphier-ep.c
>>>>> drivers/pci/controller/dwc/pci-layerscape-ep.c
>>>> Thanks for those backgrounds.
>>>>
>>>>
>>>>>>>  ??? So there is no guest or host as in
>>>>>>> virtualization but two entirely different systems
connected via PCIe cable,
>>>>>>> one
>>>>>>> acting as guest and one as host. So one system will
provide virtio
>>>>>>> functionality reserving memory for virtqueues and
the other provides vhost
>>>>>>> functionality providing a way to access the
virtqueues in virtio memory.
>>>>>>> One is
>>>>>>> source and the other is sink and there is no
intermediate entity. (vhost was
>>>>>>> probably intermediate entity in virtualization?)
>>>>>> (Not a native English speaker) but "vhost"
could introduce some confusion for
>>>>>> me since it was use for implementing virtio backend for
userspace drivers. I
>>>>>> guess "vringh" could be better.
>>>>> Initially I had named this vringh but later decided to
choose vhost instead of
>>>>> vringh. vhost is still a virtio backend (not necessarily
userspace) though it
>>>>> now resides in an entirely different system. Whatever
virtio is for a frontend
>>>>> system, vhost can be that for a backend system. vring can
be for accessing
>>>>> virtqueue and can be used either in frontend or backend.
>>>> Ok.
>>>>
>>>>
>>>>>>>> Have you considered to implement these through
vDPA?
>>>>>>> IIUC vDPA only provides an interface to userspace
and an in-kernel rpmsg
>>>>>>> driver
>>>>>>> or vhost net driver is not provided.
>>>>>>>
>>>>>>> The HW connection looks something like
https://pasteboard.co/JfMVVHC.jpg
>>>>>>> (usecase2 above),
>>>>>> I see.
>>>>>>
>>>>>>
>>>>>>>  ??? all the boards run Linux. The middle board
provides NTB
>>>>>>> functionality and board on either side provides
virtio/vhost
>>>>>>> functionality and
>>>>>>> transfer data using rpmsg.
>>>>>> So I wonder whether it's worthwhile for a new bus.
Can we use the existed
>>>>>> virtio-bus/drivers? It might work as, except for the
epf transport, we can
>>>>>> introduce a epf "vhost" transport driver.
>>>>> IMHO we'll need two buses one for frontend and other
for backend because the
>>>>> two components can then co-operate/interact with each other
to provide a
>>>>> functionality. Though both will seemingly provide similar
callbacks, they are
>>>>> both provide symmetrical or complimentary funcitonality and
need not be
>>>>> same or
>>>>> identical.
>>>>>
>>>>> Having the same bus can also create sequencing issues.
>>>>>
>>>>> If you look at virtio_dev_probe() of virtio_bus
>>>>>
>>>>> device_features = dev->config->get_features(dev);
>>>>>
>>>>> Now if we use same bus for both front-end and back-end,
both will try to
>>>>> get_features when there has been no set_features. Ideally
vhost device should
>>>>> be initialized first with the set of features it supports.
Vhost and virtio
>>>>> should use "status" and "features"
complimentarily and not identically.
>>>> Yes, but there's no need for doing status/features
passthrough in epf vhost
>>>> drivers.b
>>>>
>>>>
>>>>> virtio device (or frontend) cannot be initialized before
vhost device (or
>>>>> backend) gets initialized with data such as features.
Similarly vhost
>>>>> (backend)
>>>>> cannot access virqueues or buffers before virtio (frontend)
sets
>>>>> VIRTIO_CONFIG_S_DRIVER_OK whereas that requirement is not
there for virtio as
>>>>> the physical memory for virtqueues are created by virtio
(frontend).
>>>> epf vhost drivers need to implement two devices: vhost(vringh)
device and
>>>> virtio device (which is a mediated device). The vhost(vringh)
device is doing
>>>> feature negotiation with the virtio device via RC/EP or NTB.
The virtio device
>>>> is doing feature negotiation with local virtio drivers. If
there're feature
>>>> mismatch, epf vhost drivers and do mediation between them.
>>> Here epf vhost should be initialized with a set of features for it
to negotiate
>>> either as vhost device or virtio device no? Where should the
initial feature
>>> set for epf vhost come from?
>>
>> I think it can work as:
>>
>> 1) Having an initial features (hard coded in the code) set X in epf
vhost
>> 2) Using this X for both virtio device and vhost(vringh) device
>> 3) local virtio driver will negotiate with virtio device with feature
set Y
>> 4) remote virtio driver will negotiate with vringh device with feature
set Z
>> 5) mediate between feature Y and feature Z since both Y and Z are a
subset of X
>>
>>
> okay. I'm also thinking if we could have configfs for configuring this.
Anyways
> we could find different approaches of configuring this.

Yes, and I think some management API is needed even in the design of 
your "Software Layering". In that figure, rpmsg vhost need some
pre-set
or hard-coded features.

>>>>>> It will have virtqueues but only used for the
communication between itself
>>>>>> and
>>>>>> uppter virtio driver. And it will have vringh queues
which will be probe by
>>>>>> virtio epf transport drivers. And it needs to do
datacopy between
>>>>>> virtqueue and
>>>>>> vringh queues.
>>>>>>
>>>>>> It works like:
>>>>>>
>>>>>> virtio drivers <- virtqueue/virtio-bus -> epf
vhost drivers <- vringh
>>>>>> queue/epf>
>>>>>>
>>>>>> The advantages is that there's no need for writing
new buses and drivers.
>>>>> I think this will work however there is an addtional copy
between vringh queue
>>>>> and virtqueue,
>>>> I think not? E.g in use case 1), if we stick to virtio bus, we
will have:
>>>>
>>>> virtio-rpmsg (EP) <- virtio ring(1) -> epf vhost driver
(EP) <- virtio ring(2)
>>>> -> virtio pci (RC) <-> virtio rpmsg (RC)
>>> IIUC epf vhost driver (EP) will access virtio ring(2) using vringh?
>>
>> Yes.
>>
>>
>>> And virtio
>>> ring(2) is created by virtio pci (RC).
>>
>> Yes.
>>
>>
>>>> What epf vhost driver did is to read from virtio ring(1) about
the buffer len
>>>> and addr and them DMA to Linux(RC)?
>>> okay, I made some optimization here where vhost-rpmsg using a
helper writes a
>>> buffer from rpmsg's upper layer directly to remote Linux (RC)
as against here
>>> were it has to be first written to virtio ring (1).
>>>
>>> Thinking how this would look for NTB
>>> virtio-rpmsg (HOST1) <- virtio ring(1) -> NTB(HOST1)
<-> NTB(HOST2)? <- virtio
>>> ring(2) -> virtio-rpmsg (HOST2)
>>>
>>> Here the NTB(HOST1) will access the virtio ring(2) using vringh?
>>
>> Yes, I think so it needs to use vring to access virtio ring (1) as
well.
> NTB(HOST1) and virtio ring(1) will be in the same system. So it doesn't
have to
> use vring. virtio ring(1) is by the virtio device the NTB(HOST1) creates.

Right.

>>
>>> Do you also think this will work seamlessly with virtio_net.c,
virtio_blk.c?
>>
>> Yes.
> okay, I haven't looked at this but the backend of virtio_blk should
access an
> actual storage device no?

Good point, for non-peer device like storage. There's probably no need 
for it to be registered on the virtio bus and it might be better to 
behave as you proposed.

Just to make sure I understand the design, how is VHOST SCSI expected to 
work in your proposal, does it have a device for file as a backend?

>>
>>> I'd like to get clarity on two things in the approach you
suggested, one is
>>> features (since epf vhost should ideally be transparent to any
virtio driver)
>>
>> We can have have an array of pre-defined features indexed by virtio
device id
>> in the code.
>>
>>
>>> and the other is how certain inputs to virtio device such as number
of buffers
>>> be determined.
>>
>> We can start from hard coded the value like 256, or introduce some API
for user
>> to change the value.
>>
>>
>>> Thanks again for your suggestions!
>>
>> You're welcome.
>>
>> Note that I just want to check whether or not we can reuse the virtio
>> bus/driver. It's something similar to what you proposed in Software
Layering
>> but we just replace "vhost core" with "virtio bus"
and move the vhost core
>> below epf/ntb/platform transport.
> Got it. My initial design was based on my understanding of your comments
[1].

Yes, but that's just for a networking device. If we want something more 
generic, it may require more thought (bus etc).

>
> I'll try to create something based on your proposed design here.

Sure, but for coding, we'd better wait for other's opinion here.

Thanks

>
> Regards
> Kishon
>
> [1] ->
> https://lore.kernel.org/linux-pci/59982499-0fc1-2e39-9ff9-993fb4dd7dcc at
redhat.com/
>> Thanks
>>
>>
>>> Regards
>>> Kishon
>>>
>>>>> in some cases adds latency because of forwarding interrupts
>>>>> between vhost and virtio driver, vhost drivers providing
features (which means
>>>>> it has to be aware of which virtio driver will be
connected).
>>>>> virtio drivers (front end) generally access the buffers
from it's local memory
>>>>> but when in backend it can access over MMIO (like PCI EPF
or NTB) or
>>>>> userspace.
>>>>>> Does this make sense?
>>>>> Two copies in my opinion is an issue but lets get others
opinions as well.
>>>> Sure.
>>>>
>>>>
>>>>> Thanks for your suggestions!
>>>> You're welcome.
>>>>
>>>> Thanks
>>>>
>>>>
>>>>> Regards
>>>>> Kishon
>>>>>
>>>>>> Thanks
>>>>>>
>>>>>>
>>>>>>> Thanks
>>>>>>> Kishon
>>>>>>>
>>>>>>>> Thanks
>>>>>>>>
>>>>>>>>
>>>>>>>>>> Kishon Vijay Abraham I (22):
>>>>>>>>>>  ????? vhost: Make _feature_ bits a
property of vhost device
>>>>>>>>>>  ????? vhost: Introduce standard Linux
driver model in VHOST
>>>>>>>>>>  ????? vhost: Add ops for the VHOST
driver to configure VHOST device
>>>>>>>>>>  ????? vringh: Add helpers to access
vring in MMIO
>>>>>>>>>>  ????? vhost: Add MMIO helpers for
operations on vhost virtqueue
>>>>>>>>>>  ????? vhost: Introduce configfs entry
for configuring VHOST
>>>>>>>>>>  ????? virtio_pci: Use
request_threaded_irq() instead of request_irq()
>>>>>>>>>>  ????? rpmsg: virtio_rpmsg_bus: Disable
receive virtqueue callback when
>>>>>>>>>>  ??????? reading messages
>>>>>>>>>>  ????? rpmsg: Introduce configfs entry
for configuring rpmsg
>>>>>>>>>>  ????? rpmsg: virtio_rpmsg_bus: Add
Address Service Notification support
>>>>>>>>>>  ????? rpmsg: virtio_rpmsg_bus: Move
generic rpmsg structure to
>>>>>>>>>>  ??????? rpmsg_internal.h
>>>>>>>>>>  ????? virtio: Add ops to allocate and
free buffer
>>>>>>>>>>  ????? rpmsg: virtio_rpmsg_bus: Use
virtio_alloc_buffer() and
>>>>>>>>>>  ??????? virtio_free_buffer()
>>>>>>>>>>  ????? rpmsg: Add VHOST based remote
processor messaging bus
>>>>>>>>>>  ????? samples/rpmsg: Setup delayed
work to send message
>>>>>>>>>>  ????? samples/rpmsg: Wait for address
to be bound to rpdev for sending
>>>>>>>>>>  ??????? message
>>>>>>>>>>  ????? rpmsg.txt: Add Documentation to
configure rpmsg using configfs
>>>>>>>>>>  ????? virtio_pci: Add VIRTIO driver
for VHOST on Configurable PCIe
>>>>>>>>>> Endpoint
>>>>>>>>>>  ??????? device
>>>>>>>>>>  ????? PCI: endpoint: Add EP function
driver to provide VHOST interface
>>>>>>>>>>  ????? NTB: Add a new NTB client driver
to implement VIRTIO functionality
>>>>>>>>>>  ????? NTB: Add a new NTB client driver
to implement VHOST functionality
>>>>>>>>>>  ????? NTB: Describe the ntb_virtio and
ntb_vhost client in the
>>>>>>>>>> documentation
>>>>>>>>>>
>>>>>>>>>>  ????
Documentation/driver-api/ntb.rst????????????? |?? 11 +
>>>>>>>>>>  ????
Documentation/rpmsg.txt?????????????????????? |?? 56 +
>>>>>>>>>>  ????
drivers/ntb/Kconfig?????????????????????????? |?? 18 +
>>>>>>>>>>  ????
drivers/ntb/Makefile????????????????????????? |??? 2 +
>>>>>>>>>>  ????
drivers/ntb/ntb_vhost.c?????????????????????? |? 776 +++++++++++
>>>>>>>>>>  ????
drivers/ntb/ntb_virtio.c????????????????????? |? 853 ++++++++++++
>>>>>>>>>>  ????
drivers/ntb/ntb_virtio.h????????????????????? |?? 56 +
>>>>>>>>>>  ????
drivers/pci/endpoint/functions/Kconfig??????? |?? 11 +
>>>>>>>>>>  ????
drivers/pci/endpoint/functions/Makefile?????? |??? 1 +
>>>>>>>>>>  ????
.../pci/endpoint/functions/pci-epf-vhost.c??? | 1144
>>>>>>>>>> ++++++++++++++++
>>>>>>>>>>  ????
drivers/rpmsg/Kconfig???????????????????????? |?? 10 +
>>>>>>>>>>  ????
drivers/rpmsg/Makefile??????????????????????? |??? 3 +-
>>>>>>>>>>  ????
drivers/rpmsg/rpmsg_cfs.c???????????????????? |? 394 ++++++
>>>>>>>>>>  ????
drivers/rpmsg/rpmsg_core.c??????????????????? |??? 7 +
>>>>>>>>>>  ????
drivers/rpmsg/rpmsg_internal.h??????????????? |? 136 ++
>>>>>>>>>>  ????
drivers/rpmsg/vhost_rpmsg_bus.c?????????????? | 1151
>>>>>>>>>> +++++++++++++++++
>>>>>>>>>>  ????
drivers/rpmsg/virtio_rpmsg_bus.c????????????? |? 184 ++-
>>>>>>>>>>  ????
drivers/vhost/Kconfig???????????????????????? |??? 1 +
>>>>>>>>>>  ????
drivers/vhost/Makefile??????????????????????? |??? 2 +-
>>>>>>>>>>  ????
drivers/vhost/net.c?????????????????????????? |?? 10 +-
>>>>>>>>>>  ????
drivers/vhost/scsi.c????????????????????????? |?? 24 +-
>>>>>>>>>>  ????
drivers/vhost/test.c????????????????????????? |?? 17 +-
>>>>>>>>>>  ????
drivers/vhost/vdpa.c????????????????????????? |??? 2 +-
>>>>>>>>>>  ????
drivers/vhost/vhost.c???????????????????????? |? 730 ++++++++++-
>>>>>>>>>>  ????
drivers/vhost/vhost_cfs.c???????????????????? |? 341 +++++
>>>>>>>>>>  ????
drivers/vhost/vringh.c??????????????????????? |? 332 +++++
>>>>>>>>>>  ????
drivers/vhost/vsock.c???????????????????????? |?? 20 +-
>>>>>>>>>>  ????
drivers/virtio/Kconfig??????????????????????? |??? 9 +
>>>>>>>>>>  ????
drivers/virtio/Makefile?????????????????????? |??? 1 +
>>>>>>>>>>  ????
drivers/virtio/virtio_pci_common.c??????????? |?? 25 +-
>>>>>>>>>>  ????
drivers/virtio/virtio_pci_epf.c?????????????? |? 670 ++++++++++
>>>>>>>>>>  ????
include/linux/mod_devicetable.h?????????????? |??? 6 +
>>>>>>>>>>  ????
include/linux/rpmsg.h???????????????????????? |??? 6 +
>>>>>>>>>>  ???? {drivers/vhost =>
include/linux}/vhost.h????? |? 132 +-
>>>>>>>>>>  ????
include/linux/virtio.h??????????????????????? |??? 3 +
>>>>>>>>>>  ????
include/linux/virtio_config.h???????????????? |?? 42 +
>>>>>>>>>>  ????
include/linux/vringh.h??????????????????????? |?? 46 +
>>>>>>>>>>  ????
samples/rpmsg/rpmsg_client_sample.c?????????? |?? 32 +-
>>>>>>>>>>  ????
tools/virtio/virtio_test.c??????????????????? |??? 2 +-
>>>>>>>>>>  ???? 39 files changed, 7083
insertions(+), 183 deletions(-)
>>>>>>>>>>  ???? create mode 100644
drivers/ntb/ntb_vhost.c
>>>>>>>>>>  ???? create mode 100644
drivers/ntb/ntb_virtio.c
>>>>>>>>>>  ???? create mode 100644
drivers/ntb/ntb_virtio.h
>>>>>>>>>>  ???? create mode 100644
drivers/pci/endpoint/functions/pci-epf-vhost.c
>>>>>>>>>>  ???? create mode 100644
drivers/rpmsg/rpmsg_cfs.c
>>>>>>>>>>  ???? create mode 100644
drivers/rpmsg/vhost_rpmsg_bus.c
>>>>>>>>>>  ???? create mode 100644
drivers/vhost/vhost_cfs.c
>>>>>>>>>>  ???? create mode 100644
drivers/virtio/virtio_pci_epf.c
>>>>>>>>>>  ???? rename {drivers/vhost =>
include/linux}/vhost.h (66%)
>>>>>>>>>>
>>>>>>>>>> -- 
>>>>>>>>>> 2.17.1
>>>>>>>>>>

On Thu, 9 Jul 2020 14:26:53 +0800
Jason Wang <jasowang at redhat.com> wrote:

[Let me note right at the beginning that I first noted this while
listening to Kishon's talk at LPC on Wednesday. I might be very
confused about the background here, so let me apologize beforehand for
any confusion I might spread.]
> On 2020/7/8 ??9:13, Kishon Vijay Abraham I wrote:
> > Hi Jason,
> >
> > On 7/8/2020 4:52 PM, Jason Wang wrote:  
> >> On 2020/7/7 ??10:45, Kishon Vijay Abraham I wrote:  
> >>> Hi Jason,
> >>>
> >>> On 7/7/2020 3:17 PM, Jason Wang wrote:  
> >>>> On 2020/7/6 ??5:32, Kishon Vijay Abraham I wrote:  
> >>>>> Hi Jason,
> >>>>>
> >>>>> On 7/3/2020 12:46 PM, Jason Wang wrote:  
> >>>>>> On 2020/7/2 ??9:35, Kishon Vijay Abraham I wrote:
> >>>>>>> Hi Jason,
> >>>>>>>
> >>>>>>> On 7/2/2020 3:40 PM, Jason Wang wrote:  
> >>>>>>>> On 2020/7/2 ??5:51, Michael S. Tsirkin
wrote:
> >>>>>>>>> On Thu, Jul 02, 2020 at 01:51:21PM
+0530, Kishon Vijay Abraham I wrote:
> >>>>>>>>>> This series enhances Linux Vhost
support to enable SoC-to-SoC
> >>>>>>>>>> communication over MMIO. This
series enables rpmsg communication between
> >>>>>>>>>> two SoCs using both PCIe
RC<->EP and HOST1-NTB-HOST2
> >>>>>>>>>>
> >>>>>>>>>> 1) Modify vhost to use standard
Linux driver model
> >>>>>>>>>> 2) Add support in vring to access
virtqueue over MMIO
> >>>>>>>>>> 3) Add vhost client driver for
rpmsg
> >>>>>>>>>> 4) Add PCIe RC driver (uses
virtio) and PCIe EP driver (uses vhost) for
> >>>>>>>>>>  ?????? rpmsg communication
between two SoCs connected to each other
> >>>>>>>>>> 5) Add NTB Virtio driver and NTB
Vhost driver for rpmsg communication
> >>>>>>>>>>  ?????? between two SoCs connected
via NTB
> >>>>>>>>>> 6) Add configfs to configure the
components
> >>>>>>>>>>
> >>>>>>>>>> UseCase1 :
> >>>>>>>>>>
> >>>>>>>>>>  ???? VHOST
RPMSG???????????????????? VIRTIO RPMSG
> >>>>>>>>>>  ?????????
+?????????????????????????????? +
> >>>>>>>>>>  ?????????
|?????????????????????????????? |
> >>>>>>>>>>  ?????????
|?????????????????????????????? |
> >>>>>>>>>>  ?????????
|?????????????????????????????? |
> >>>>>>>>>>  ?????????
|?????????????????????????????? |
> >>>>>>>>>> +-----v------+????????????????
+------v-------+
> >>>>>>>>>> |?? Linux??? |????????????????
|???? Linux??? |
> >>>>>>>>>> |? Endpoint? |???????????????? |
Root Complex |
> >>>>>>>>>> |???????????
<----------------->????????????? |
> >>>>>>>>>> |??????????? |????????????????
|????????????? |
> >>>>>>>>>> |??? SOC1??? |????????????????
|???? SOC2???? |
> >>>>>>>>>> +------------+????????????????
+--------------+
> >>>>>>>>>>
> >>>>>>>>>> UseCase 2:
> >>>>>>>>>>
> >>>>>>>>>>  ???????? VHOST
RPMSG????????????????????????????????????? VIRTIO RPMSG
> >>>>>>>>>>  ?????????????
+???????????????????????????????????????????????? +
> >>>>>>>>>>  ?????????????
|???????????????????????????????????????????????? |
> >>>>>>>>>>  ?????????????
|???????????????????????????????????????????????? |
> >>>>>>>>>>  ?????????????
|???????????????????????????????????????????????? |
> >>>>>>>>>>  ?????????????
|???????????????????????????????????????????????? |
> >>>>>>>>>>  ??????
+------v------+?????????????????????????????????? +------v------+
> >>>>>>>>>>  ?????? |????????????
|?????????????????????????????????? |???????????? |
> >>>>>>>>>>  ?????? |??? HOST1???
|?????????????????????????????????? |??? HOST2??? |
> >>>>>>>>>>  ?????? |????????????
|?????????????????????????????????? |???????????? |
> >>>>>>>>>>  ??????
+------^------+?????????????????????????????????? +------^------+
> >>>>>>>>>>  ?????????????
|???????????????????????????????????????????????? |
> >>>>>>>>>>  ?????????????
|???????????????????????????????????????????????? |
> >>>>>>>>>>
+---------------------------------------------------------------------+
> >>>>>>>>>> |?
+------v------+?????????????????????????????????? +------v------+? |
> >>>>>>>>>> |? |????????????
|?????????????????????????????????? |???????????? |? |
> >>>>>>>>>> |? |???? EP?????
|?????????????????????????????????? |???? EP????? |? |
> >>>>>>>>>> |? | CONTROLLER1
|?????????????????????????????????? | CONTROLLER2 |? |
> >>>>>>>>>> |? |????????????
<----------------------------------->???????????? |? |
> >>>>>>>>>> |? |????????????
|?????????????????????????????????? |???????????? |? |
> >>>>>>>>>> |? |????????????
|?????????????????????????????????? |???????????? |? |
> >>>>>>>>>> |? |???????????? |? SoC With
Multiple EP Instances?? |???????????? |? |
> >>>>>>>>>> |? |???????????? |? (Configured
using NTB Function)? |???????????? |? |
> >>>>>>>>>> |?
+-------------+?????????????????????????????????? +-------------+? |
> >>>>>>>>>>
+---------------------------------------------------------------------+
First of all, to clarify the terminology:
Is "vhost rpmsg" acting as what the virtio standard calls the
'device',
and "virtio rpmsg" as the 'driver'? Or is the
"vhost" part mostly just
virtqueues + the exiting vhost interfaces?
> >>>>>>>>>>
> >>>>>>>>>> Software Layering:
> >>>>>>>>>>
> >>>>>>>>>> The high-level SW layering should
look something like below. This series
> >>>>>>>>>> adds support only for RPMSG VHOST,
however something similar should be
> >>>>>>>>>> done for net and scsi. With that
any vhost device (PCI, NTB, Platform
> >>>>>>>>>> device, user) can use any of the
vhost client driver.
> >>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>>  ??????? +----------------+?
+-----------+? +------------+? +----------+
> >>>>>>>>>>  ??????? |? RPMSG VHOST?? |? | NET
VHOST |? | SCSI VHOST |? |??? X???? |
> >>>>>>>>>>  ??????? +-------^--------+?
+-----^-----+? +-----^------+? +----^-----+
> >>>>>>>>>>  ??????????????? |????????????????
|????????????? |????????????? |
> >>>>>>>>>>  ??????????????? |????????????????
|????????????? |????????????? |
> >>>>>>>>>>  ??????????????? |????????????????
|????????????? |????????????? |
> >>>>>>>>>>
+-----------v-----------------v--------------v--------------v----------+
> >>>>>>>>>> |??????????????????????????? VHOST
CORE??????????????????????????????? |
> >>>>>>>>>>
+--------^---------------^--------------------^------------------^-----+
> >>>>>>>>>>  ???????????? |??????????????
|??????????????????? |????????????????? |
> >>>>>>>>>>  ???????????? |??????????????
|??????????????????? |????????????????? |
> >>>>>>>>>>  ???????????? |??????????????
|??????????????????? |????????????????? |
> >>>>>>>>>> +--------v-------+? +----v------+?
+----------v----------+? +----v-----+
> >>>>>>>>>> |? PCI EPF VHOST |? | NTB VHOST |?
|PLATFORM DEVICE VHOST|? |??? X???? |
> >>>>>>>>>> +----------------+? +-----------+?
+---------------------+? +----------+
So, the upper half is basically various functionality types, e.g. a net
device. What is the lower half, a hardware interface? Would it be
equivalent to e.g. a normal PCI device?
> >>>>>>>>>>
> >>>>>>>>>> This was initially proposed here
[1]
> >>>>>>>>>>
> >>>>>>>>>> [1] ->
> >>>>>>>>>>
https://lore.kernel.org/r/2cf00ec4-1ed6-f66e-6897-006d1a5b6390 at ti.com
> >>>>>>>>> I find this very interesting. A huge
patchset so will take a bit
> >>>>>>>>> to review, but I certainly plan to do
that. Thanks!
> >>>>>>>> Yes, it would be better if there's a
git branch for us to have a look.
> >>>>>>> I've pushed the branch
> >>>>>>> https://github.com/kishon/linux-wip.git
vhost_rpmsg_pci_ntb_rfc
> >>>>>> Thanks
> >>>>>>
> >>>>>>  
> >>>>>>>> Btw, I'm not sure I get the big
picture, but I vaguely feel some of the
> >>>>>>>> work is
> >>>>>>>> duplicated with vDPA (e.g the epf
transport or vhost bus).
> >>>>>>> This is about connecting two different HW
systems both running Linux and
> >>>>>>> doesn't necessarily involve
virtualization.
> >>>>>> Right, this is something similar to VOP
> >>>>>> (Documentation/misc-devices/mic/mic_overview.rst).
The different is the
> >>>>>> hardware I guess and VOP use userspace application
to implement the device.
> >>>>> I'd also like to point out, this series tries to
have communication between
> >>>>> two
> >>>>> SoCs in vendor agnostic way. Since this series solves
for 2 usecases (PCIe
> >>>>> RC<->EP and NTB), for the NTB case it directly
plugs into NTB framework and
> >>>>> any
> >>>>> of the HW in NTB below should be able to use a
virtio-vhost communication
> >>>>>
> >>>>> #ls drivers/ntb/hw/
> >>>>> amd? epf? idt? intel? mscc
> >>>>>
> >>>>> And similarly for the PCIe RC<->EP
communication, this adds a generic endpoint
> >>>>> function driver and hence any SoC that supports
configurable PCIe endpoint can
> >>>>> use virtio-vhost communication
> >>>>>
> >>>>> # ls drivers/pci/controller/dwc/*ep*
> >>>>> drivers/pci/controller/dwc/pcie-designware-ep.c
> >>>>> drivers/pci/controller/dwc/pcie-uniphier-ep.c
> >>>>> drivers/pci/controller/dwc/pci-layerscape-ep.c  
> >>>> Thanks for those backgrounds.
> >>>>
> >>>>  
> >>>>>>>  ??? So there is no guest or host as in
> >>>>>>> virtualization but two entirely different
systems connected via PCIe cable,
> >>>>>>> one
> >>>>>>> acting as guest and one as host. So one system
will provide virtio
> >>>>>>> functionality reserving memory for virtqueues
and the other provides vhost
> >>>>>>> functionality providing a way to access the
virtqueues in virtio memory.
> >>>>>>> One is
> >>>>>>> source and the other is sink and there is no
intermediate entity. (vhost was
> >>>>>>> probably intermediate entity in
virtualization?)
> >>>>>> (Not a native English speaker) but
"vhost" could introduce some confusion for
> >>>>>> me since it was use for implementing virtio
backend for userspace drivers. I
> >>>>>> guess "vringh" could be better.  
> >>>>> Initially I had named this vringh but later decided to
choose vhost instead of
> >>>>> vringh. vhost is still a virtio backend (not
necessarily userspace) though it
> >>>>> now resides in an entirely different system. Whatever
virtio is for a frontend
> >>>>> system, vhost can be that for a backend system. vring
can be for accessing
> >>>>> virtqueue and can be used either in frontend or
backend.
I guess that clears up at least some of my questions from above...
> >>>> Ok.
> >>>>
> >>>>  
> >>>>>>>> Have you considered to implement these
through vDPA?
> >>>>>>> IIUC vDPA only provides an interface to
userspace and an in-kernel rpmsg
> >>>>>>> driver
> >>>>>>> or vhost net driver is not provided.
> >>>>>>>
> >>>>>>> The HW connection looks something like
https://pasteboard.co/JfMVVHC.jpg
> >>>>>>> (usecase2 above),  
> >>>>>> I see.
> >>>>>>
> >>>>>>  
> >>>>>>>  ??? all the boards run Linux. The middle
board provides NTB
> >>>>>>> functionality and board on either side
provides virtio/vhost
> >>>>>>> functionality and
> >>>>>>> transfer data using rpmsg. 
This setup looks really interesting (sometimes, it's really hard to
imagine this in the abstract.)
 
> >>>>>> So I wonder whether it's worthwhile for a new
bus. Can we use
> >>>>>> the existed virtio-bus/drivers? It might work as,
except for
> >>>>>> the epf transport, we can introduce a epf
"vhost" transport
> >>>>>> driver.  
> >>>>> IMHO we'll need two buses one for frontend and
other for
> >>>>> backend because the two components can then
co-operate/interact
> >>>>> with each other to provide a functionality. Though
both will
> >>>>> seemingly provide similar callbacks, they are both
provide
> >>>>> symmetrical or complimentary funcitonality and need
not be same
> >>>>> or identical.
> >>>>>
> >>>>> Having the same bus can also create sequencing issues.
> >>>>>
> >>>>> If you look at virtio_dev_probe() of virtio_bus
> >>>>>
> >>>>> device_features =
dev->config->get_features(dev);
> >>>>>
> >>>>> Now if we use same bus for both front-end and
back-end, both
> >>>>> will try to get_features when there has been no
set_features.
> >>>>> Ideally vhost device should be initialized first with
the set
> >>>>> of features it supports. Vhost and virtio should use
"status"
> >>>>> and "features" complimentarily and not
identically.
> >>>> Yes, but there's no need for doing status/features
passthrough
> >>>> in epf vhost drivers.b
> >>>>
> >>>>  
> >>>>> virtio device (or frontend) cannot be initialized
before vhost
> >>>>> device (or backend) gets initialized with data such as
> >>>>> features. Similarly vhost (backend)
> >>>>> cannot access virqueues or buffers before virtio
(frontend) sets
> >>>>> VIRTIO_CONFIG_S_DRIVER_OK whereas that requirement is
not there
> >>>>> for virtio as the physical memory for virtqueues are
created by
> >>>>> virtio (frontend).  
> >>>> epf vhost drivers need to implement two devices:
vhost(vringh)
> >>>> device and virtio device (which is a mediated device). The
> >>>> vhost(vringh) device is doing feature negotiation with the
> >>>> virtio device via RC/EP or NTB. The virtio device is doing
> >>>> feature negotiation with local virtio drivers. If
there're
> >>>> feature mismatch, epf vhost drivers and do mediation
between
> >>>> them.  
> >>> Here epf vhost should be initialized with a set of features
for
> >>> it to negotiate either as vhost device or virtio device no?
Where
> >>> should the initial feature set for epf vhost come from?  
> >>
> >> I think it can work as:
> >>
> >> 1) Having an initial features (hard coded in the code) set X in
> >> epf vhost 2) Using this X for both virtio device and vhost(vringh)
> >> device 3) local virtio driver will negotiate with virtio device
> >> with feature set Y 4) remote virtio driver will negotiate with
> >> vringh device with feature set Z 5) mediate between feature Y and
> >> feature Z since both Y and Z are a subset of X
> >>
> >>  
> > okay. I'm also thinking if we could have configfs for configuring
> > this. Anyways we could find different approaches of configuring
> > this.  
> 
> 
> Yes, and I think some management API is needed even in the design of 
> your "Software Layering". In that figure, rpmsg vhost need some
> pre-set or hard-coded features.
When I saw the plumbers talk, my first idea was "this needs to be a new
transport". You have some hard-coded or pre-configured features, and
then features are negotiated via a transport-specific means in the
usual way. There's basically an extra/extended layer for this (and
status, and whatever).

Does that make any sense?
> 
> 
> >>>>>> It will have virtqueues but only used for the
communication
> >>>>>> between itself and
> >>>>>> uppter virtio driver. And it will have vringh
queues which
> >>>>>> will be probe by virtio epf transport drivers. And
it needs to
> >>>>>> do datacopy between virtqueue and
> >>>>>> vringh queues.
> >>>>>>
> >>>>>> It works like:
> >>>>>>
> >>>>>> virtio drivers <- virtqueue/virtio-bus ->
epf vhost drivers <-
> >>>>>> vringh queue/epf>  
> >>>>>>
> >>>>>> The advantages is that there's no need for
writing new buses
> >>>>>> and drivers.  
> >>>>> I think this will work however there is an addtional
copy
> >>>>> between vringh queue and virtqueue,  
> >>>> I think not? E.g in use case 1), if we stick to virtio
bus, we
> >>>> will have:
> >>>>
> >>>> virtio-rpmsg (EP) <- virtio ring(1) -> epf vhost
driver (EP) <-
> >>>> virtio ring(2) -> virtio pci (RC) <-> virtio
rpmsg (RC)
> >>> IIUC epf vhost driver (EP) will access virtio ring(2) using
> >>> vringh?  
> >>
> >> Yes.
> >>
> >>  
> >>> And virtio
> >>> ring(2) is created by virtio pci (RC).  
> >>
> >> Yes.
> >>
> >>  
> >>>> What epf vhost driver did is to read from virtio ring(1)
about
> >>>> the buffer len and addr and them DMA to Linux(RC)?  
> >>> okay, I made some optimization here where vhost-rpmsg using a
> >>> helper writes a buffer from rpmsg's upper layer directly
to
> >>> remote Linux (RC) as against here were it has to be first
written
> >>> to virtio ring (1).
> >>>
> >>> Thinking how this would look for NTB
> >>> virtio-rpmsg (HOST1) <- virtio ring(1) -> NTB(HOST1)
<->
> >>> NTB(HOST2)? <- virtio ring(2) -> virtio-rpmsg (HOST2)
> >>>
> >>> Here the NTB(HOST1) will access the virtio ring(2) using
vringh?
> >>
> >> Yes, I think so it needs to use vring to access virtio ring (1) as
> >> well.  
> > NTB(HOST1) and virtio ring(1) will be in the same system. So it
> > doesn't have to use vring. virtio ring(1) is by the virtio device
> > the NTB(HOST1) creates.  
> 
> 
> Right.
> 
> 
> >>  
> >>> Do you also think this will work seamlessly with virtio_net.c,
> >>> virtio_blk.c?  
> >>
> >> Yes.  
> > okay, I haven't looked at this but the backend of virtio_blk
should
> > access an actual storage device no?  
> 
> 
> Good point, for non-peer device like storage. There's probably no
> need for it to be registered on the virtio bus and it might be better
> to behave as you proposed.
I might be missing something; but if you expose something as a block
device, it should have something it can access with block reads/writes,
shouldn't it? Of course, that can be a variety of things.
> 
> Just to make sure I understand the design, how is VHOST SCSI expected
> to work in your proposal, does it have a device for file as a backend?
> 
> 
> >>  
> >>> I'd like to get clarity on two things in the approach you
> >>> suggested, one is features (since epf vhost should ideally be
> >>> transparent to any virtio driver)  
> >>
> >> We can have have an array of pre-defined features indexed by
> >> virtio device id in the code.
> >>
> >>  
> >>> and the other is how certain inputs to virtio device such as
> >>> number of buffers be determined.  
> >>
> >> We can start from hard coded the value like 256, or introduce some
> >> API for user to change the value.
> >>
> >>  
> >>> Thanks again for your suggestions!  
> >>
> >> You're welcome.
> >>
> >> Note that I just want to check whether or not we can reuse the
> >> virtio bus/driver. It's something similar to what you proposed
in
> >> Software Layering but we just replace "vhost core" with
"virtio
> >> bus" and move the vhost core below epf/ntb/platform
transport.
> > Got it. My initial design was based on my understanding of your
> > comments [1].  
> 
> 
> Yes, but that's just for a networking device. If we want something
> more generic, it may require more thought (bus etc).
I believe that we indeed need something bus-like to be able to support
a variety of devices.
> 
> 
> >
> > I'll try to create something based on your proposed design here.  
> 
> 
> Sure, but for coding, we'd better wait for other's opinion here.
Please tell me if my thoughts above make any sense... I have just
started looking at that, so I might be completely off.

On 2020/9/1 ??1:24, Kishon Vijay Abraham I wrote:
> Hi,
>
> On 28/08/20 4:04 pm, Cornelia Huck wrote:
>> On Thu, 9 Jul 2020 14:26:53 +0800
>> Jason Wang <jasowang at redhat.com> wrote:
>>
>> [Let me note right at the beginning that I first noted this while
>> listening to Kishon's talk at LPC on Wednesday. I might be very
>> confused about the background here, so let me apologize beforehand for
>> any confusion I might spread.]
>>
>>> On 2020/7/8 ??9:13, Kishon Vijay Abraham I wrote:
>>>> Hi Jason,
>>>>
>>>> On 7/8/2020 4:52 PM, Jason Wang wrote:
>>>>> On 2020/7/7 ??10:45, Kishon Vijay Abraham I wrote:
>>>>>> Hi Jason,
>>>>>>
>>>>>> On 7/7/2020 3:17 PM, Jason Wang wrote:
>>>>>>> On 2020/7/6 ??5:32, Kishon Vijay Abraham I wrote:
>>>>>>>> Hi Jason,
>>>>>>>>
>>>>>>>> On 7/3/2020 12:46 PM, Jason Wang wrote:
>>>>>>>>> On 2020/7/2 ??9:35, Kishon Vijay Abraham I
wrote:
>>>>>>>>>> Hi Jason,
>>>>>>>>>>
>>>>>>>>>> On 7/2/2020 3:40 PM, Jason Wang wrote:
>>>>>>>>>>> On 2020/7/2 ??5:51, Michael S.
Tsirkin wrote:
>>>>>>>>>>>> On Thu, Jul 02, 2020 at
01:51:21PM +0530, Kishon Vijay
>>>>>>>>>>>> Abraham I wrote:
>>>>>>>>>>>>> This series enhances Linux
Vhost support to enable SoC-to-SoC
>>>>>>>>>>>>> communication over MMIO.
This series enables rpmsg
>>>>>>>>>>>>> communication between
>>>>>>>>>>>>> two SoCs using both PCIe
RC<->EP and HOST1-NTB-HOST2
>>>>>>>>>>>>>
>>>>>>>>>>>>> 1) Modify vhost to use
standard Linux driver model
>>>>>>>>>>>>> 2) Add support in vring to
access virtqueue over MMIO
>>>>>>>>>>>>> 3) Add vhost client driver
for rpmsg
>>>>>>>>>>>>> 4) Add PCIe RC driver (uses
virtio) and PCIe EP driver
>>>>>>>>>>>>> (uses vhost) for
>>>>>>>>>>>>> ? ?????? rpmsg
communication between two SoCs connected to
>>>>>>>>>>>>> each other
>>>>>>>>>>>>> 5) Add NTB Virtio driver
and NTB Vhost driver for rpmsg
>>>>>>>>>>>>> communication
>>>>>>>>>>>>> ? ?????? between two SoCs
connected via NTB
>>>>>>>>>>>>> 6) Add configfs to
configure the components
>>>>>>>>>>>>>
>>>>>>>>>>>>> UseCase1 :
>>>>>>>>>>>>>
>>>>>>>>>>>>> ? ???? VHOST RPMSG VIRTIO
RPMSG
>>>>>>>>>>>>>
+?????????????????????????????? +
>>>>>>>>>>>>>
|?????????????????????????????? |
>>>>>>>>>>>>>
|?????????????????????????????? |
>>>>>>>>>>>>>
|?????????????????????????????? |
>>>>>>>>>>>>>
|?????????????????????????????? |
>>>>>>>>>>>>> +-----v------+
+------v-------+
>>>>>>>>>>>>> |?? Linux???
|???????????????? | Linux??? |
>>>>>>>>>>>>> |? Endpoint?
|???????????????? | Root Complex |
>>>>>>>>>>>>> |
<----------------->????????????? |
>>>>>>>>>>>>> |??????????? |
|????????????? |
>>>>>>>>>>>>> |??? SOC1???
|???????????????? | SOC2???? |
>>>>>>>>>>>>> +------------+
+--------------+
>>>>>>>>>>>>>
>>>>>>>>>>>>> UseCase 2:
>>>>>>>>>>>>>
>>>>>>>>>>>>> ? ???????? VHOST RPMSG
VIRTIO RPMSG
>>>>>>>>>>>>> + +
>>>>>>>>>>>>> | |
>>>>>>>>>>>>> | |
>>>>>>>>>>>>> | |
>>>>>>>>>>>>> | |
>>>>>>>>>>>>> +------v------+
+------v------+
>>>>>>>>>>>>> ? ?????? | | |????????????
|
>>>>>>>>>>>>> ? ?????? |??? HOST1
|?????????????????????????????????? |
>>>>>>>>>>>>> HOST2??? |
>>>>>>>>>>>>> ? ?????? | | |????????????
|
>>>>>>>>>>>>> +------^------+
+------^------+
>>>>>>>>>>>>> | |
>>>>>>>>>>>>> | |
>>>>>>>>>>>>>
+---------------------------------------------------------------------+
>>>>>>>>>>>>>
>>>>>>>>>>>>> | +------v------+
+------v------+? |
>>>>>>>>>>>>> |? | | |???????????? |? |
>>>>>>>>>>>>> |? |???? EP
|?????????????????????????????????? | EP?????
>>>>>>>>>>>>> |? |
>>>>>>>>>>>>> |? | CONTROLLER1
|?????????????????????????????????? |
>>>>>>>>>>>>> CONTROLLER2 |? |
>>>>>>>>>>>>> |? |
<-----------------------------------> |? |
>>>>>>>>>>>>> |? | | |???????????? |? |
>>>>>>>>>>>>> |? | | |???????????? |? |
>>>>>>>>>>>>> |? |???????????? |? SoC
With Multiple EP Instances??
>>>>>>>>>>>>> |???????????? |? |
>>>>>>>>>>>>> |? |???????????? |?
(Configured using NTB Function)?
>>>>>>>>>>>>> |???????????? |? |
>>>>>>>>>>>>> | +-------------+
+-------------+? |
>>>>>>>>>>>>>
+---------------------------------------------------------------------+
>>>>>>>>>>>>>
>>
>> First of all, to clarify the terminology:
>> Is "vhost rpmsg" acting as what the virtio standard calls the
'device',
>> and "virtio rpmsg" as the 'driver'? Or is the
"vhost" part mostly just
>
> Right, vhost_rpmsg is 'device' and virtio_rpmsg is
'driver'.
>> virtqueues + the exiting vhost interfaces?
>
> It's implemented to provide the full 'device' functionality.
>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> Software Layering:
>>>>>>>>>>>>>
>>>>>>>>>>>>> The high-level SW layering
should look something like
>>>>>>>>>>>>> below. This series
>>>>>>>>>>>>> adds support only for RPMSG
VHOST, however something
>>>>>>>>>>>>> similar should be
>>>>>>>>>>>>> done for net and scsi. With
that any vhost device (PCI,
>>>>>>>>>>>>> NTB, Platform
>>>>>>>>>>>>> device, user) can use any
of the vhost client driver.
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> ? ???????
+----------------+ +-----------+? +------------+
>>>>>>>>>>>>> +----------+
>>>>>>>>>>>>> ? ??????? |? RPMSG VHOST??
|? | NET VHOST |? | SCSI VHOST
>>>>>>>>>>>>> |? |??? X???? |
>>>>>>>>>>>>> ? ???????
+-------^--------+ +-----^-----+? +-----^------+
>>>>>>>>>>>>> +----^-----+
>>>>>>>>>>>>> ? ??????????????? |
|????????????? |????????????? |
>>>>>>>>>>>>> ? ??????????????? |
|????????????? |????????????? |
>>>>>>>>>>>>> ? ??????????????? |
|????????????? |????????????? |
>>>>>>>>>>>>>
+-----------v-----------------v--------------v--------------v----------+
>>>>>>>>>>>>>
>>>>>>>>>>>>>
|??????????????????????????? VHOST
>>>>>>>>>>>>>
CORE??????????????????????????????? |
>>>>>>>>>>>>>
+--------^---------------^--------------------^------------------^-----+
>>>>>>>>>>>>>
>>>>>>>>>>>>> ? ???????????? |
|??????????????????? |????????????????? |
>>>>>>>>>>>>> ? ???????????? |
|??????????????????? |????????????????? |
>>>>>>>>>>>>> ? ???????????? |
|??????????????????? |????????????????? |
>>>>>>>>>>>>> +--------v-------+?
+----v------+ +----------v----------+?
>>>>>>>>>>>>> +----v-----+
>>>>>>>>>>>>> |? PCI EPF VHOST |? | NTB
VHOST | |PLATFORM DEVICE VHOST|?
>>>>>>>>>>>>> |??? X???? |
>>>>>>>>>>>>> +----------------+?
+-----------+ +---------------------+?
>>>>>>>>>>>>> +----------+
>>
>> So, the upper half is basically various functionality types, e.g. a net
>> device. What is the lower half, a hardware interface? Would it be
>> equivalent to e.g. a normal PCI device?
>
> Right, the upper half should provide the functionality.
> The bottom layer could be a HW interface (like PCIe device or NTB 
> device) or it could be a SW interface (for accessing virtio ring in 
> userspace) that could be used by Hypervisor.
>
> The top half should be transparent to what type of device is actually 
> using it.
>
>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> This was initially proposed
here [1]
>>>>>>>>>>>>>
>>>>>>>>>>>>> [1] ->
>>>>>>>>>>>>>
https://lore.kernel.org/r/2cf00ec4-1ed6-f66e-6897-006d1a5b6390 at ti.com
>>>>>>>>>>>>>
>>>>>>>>>>>> I find this very interesting. A
huge patchset so will take
>>>>>>>>>>>> a bit
>>>>>>>>>>>> to review, but I certainly plan
to do that. Thanks!
>>>>>>>>>>> Yes, it would be better if
there's a git branch for us to
>>>>>>>>>>> have a look.
>>>>>>>>>> I've pushed the branch
>>>>>>>>>> https://github.com/kishon/linux-wip.git
vhost_rpmsg_pci_ntb_rfc
>>>>>>>>> Thanks
>>>>>>>>>
>>>>>>>>>>> Btw, I'm not sure I get the big
picture, but I vaguely feel
>>>>>>>>>>> some of the
>>>>>>>>>>> work is
>>>>>>>>>>> duplicated with vDPA (e.g the epf
transport or vhost bus).
>>>>>>>>>> This is about connecting two different
HW systems both
>>>>>>>>>> running Linux and
>>>>>>>>>> doesn't necessarily involve
virtualization.
>>>>>>>>> Right, this is something similar to VOP
>>>>>>>>>
(Documentation/misc-devices/mic/mic_overview.rst). The
>>>>>>>>> different is the
>>>>>>>>> hardware I guess and VOP use userspace
application to
>>>>>>>>> implement the device.
>>>>>>>> I'd also like to point out, this series
tries to have
>>>>>>>> communication between
>>>>>>>> two
>>>>>>>> SoCs in vendor agnostic way. Since this series
solves for 2
>>>>>>>> usecases (PCIe
>>>>>>>> RC<->EP and NTB), for the NTB case it
directly plugs into NTB
>>>>>>>> framework and
>>>>>>>> any
>>>>>>>> of the HW in NTB below should be able to use a
virtio-vhost
>>>>>>>> communication
>>>>>>>>
>>>>>>>> #ls drivers/ntb/hw/
>>>>>>>> amd? epf? idt? intel? mscc
>>>>>>>>
>>>>>>>> And similarly for the PCIe RC<->EP
communication, this adds a
>>>>>>>> generic endpoint
>>>>>>>> function driver and hence any SoC that supports
configurable
>>>>>>>> PCIe endpoint can
>>>>>>>> use virtio-vhost communication
>>>>>>>>
>>>>>>>> # ls drivers/pci/controller/dwc/*ep*
>>>>>>>> drivers/pci/controller/dwc/pcie-designware-ep.c
>>>>>>>> drivers/pci/controller/dwc/pcie-uniphier-ep.c
>>>>>>>> drivers/pci/controller/dwc/pci-layerscape-ep.c
>>>>>>> Thanks for those backgrounds.
>>>>>>>
>>>>>>>>>> ? ??? So there is no guest or host as
in
>>>>>>>>>> virtualization but two entirely
different systems connected
>>>>>>>>>> via PCIe cable,
>>>>>>>>>> one
>>>>>>>>>> acting as guest and one as host. So one
system will provide
>>>>>>>>>> virtio
>>>>>>>>>> functionality reserving memory for
virtqueues and the other
>>>>>>>>>> provides vhost
>>>>>>>>>> functionality providing a way to access
the virtqueues in
>>>>>>>>>> virtio memory.
>>>>>>>>>> One is
>>>>>>>>>> source and the other is sink and there
is no intermediate
>>>>>>>>>> entity. (vhost was
>>>>>>>>>> probably intermediate entity in
virtualization?)
>>>>>>>>> (Not a native English speaker) but
"vhost" could introduce
>>>>>>>>> some confusion for
>>>>>>>>> me since it was use for implementing virtio
backend for
>>>>>>>>> userspace drivers. I
>>>>>>>>> guess "vringh" could be better.
>>>>>>>> Initially I had named this vringh but later
decided to choose
>>>>>>>> vhost instead of
>>>>>>>> vringh. vhost is still a virtio backend (not
necessarily
>>>>>>>> userspace) though it
>>>>>>>> now resides in an entirely different system.
Whatever virtio is
>>>>>>>> for a frontend
>>>>>>>> system, vhost can be that for a backend system.
vring can be
>>>>>>>> for accessing
>>>>>>>> virtqueue and can be used either in frontend or
backend.
>>
>> I guess that clears up at least some of my questions from above...
>>
>>>>>>> Ok.
>>>>>>>
>>>>>>>>>>> Have you considered to implement
these through vDPA?
>>>>>>>>>> IIUC vDPA only provides an interface to
userspace and an
>>>>>>>>>> in-kernel rpmsg
>>>>>>>>>> driver
>>>>>>>>>> or vhost net driver is not provided.
>>>>>>>>>>
>>>>>>>>>> The HW connection looks something like 
>>>>>>>>>> https://pasteboard.co/JfMVVHC.jpg
>>>>>>>>>> (usecase2 above),
>>>>>>>>> I see.
>>>>>>>>>
>>>>>>>>>> ? ??? all the boards run Linux. The
middle board provides NTB
>>>>>>>>>> functionality and board on either side
provides virtio/vhost
>>>>>>>>>> functionality and
>>>>>>>>>> transfer data using rpmsg.
>>
>> This setup looks really interesting (sometimes, it's really hard to
>> imagine this in the abstract.)
>>>>>>>>> So I wonder whether it's worthwhile for
a new bus. Can we use
>>>>>>>>> the existed virtio-bus/drivers? It might
work as, except for
>>>>>>>>> the epf transport, we can introduce a epf
"vhost" transport
>>>>>>>>> driver.
>>>>>>>> IMHO we'll need two buses one for frontend
and other for
>>>>>>>> backend because the two components can then
co-operate/interact
>>>>>>>> with each other to provide a functionality.
Though both will
>>>>>>>> seemingly provide similar callbacks, they are
both provide
>>>>>>>> symmetrical or complimentary funcitonality and
need not be same
>>>>>>>> or identical.
>>>>>>>>
>>>>>>>> Having the same bus can also create sequencing
issues.
>>>>>>>>
>>>>>>>> If you look at virtio_dev_probe() of virtio_bus
>>>>>>>>
>>>>>>>> device_features =
dev->config->get_features(dev);
>>>>>>>>
>>>>>>>> Now if we use same bus for both front-end and
back-end, both
>>>>>>>> will try to get_features when there has been no
set_features.
>>>>>>>> Ideally vhost device should be initialized
first with the set
>>>>>>>> of features it supports. Vhost and virtio
should use "status"
>>>>>>>> and "features" complimentarily and
not identically.
>>>>>>> Yes, but there's no need for doing
status/features passthrough
>>>>>>> in epf vhost drivers.b
>>>>>>>
>>>>>>>> virtio device (or frontend) cannot be
initialized before vhost
>>>>>>>> device (or backend) gets initialized with data
such as
>>>>>>>> features. Similarly vhost (backend)
>>>>>>>> cannot access virqueues or buffers before
virtio (frontend) sets
>>>>>>>> VIRTIO_CONFIG_S_DRIVER_OK whereas that
requirement is not there
>>>>>>>> for virtio as the physical memory for
virtqueues are created by
>>>>>>>> virtio (frontend).
>>>>>>> epf vhost drivers need to implement two devices:
vhost(vringh)
>>>>>>> device and virtio device (which is a mediated
device). The
>>>>>>> vhost(vringh) device is doing feature negotiation
with the
>>>>>>> virtio device via RC/EP or NTB. The virtio device
is doing
>>>>>>> feature negotiation with local virtio drivers. If
there're
>>>>>>> feature mismatch, epf vhost drivers and do
mediation between
>>>>>>> them.
>>>>>> Here epf vhost should be initialized with a set of
features for
>>>>>> it to negotiate either as vhost device or virtio device
no? Where
>>>>>> should the initial feature set for epf vhost come from?
>>>>>
>>>>> I think it can work as:
>>>>>
>>>>> 1) Having an initial features (hard coded in the code) set
X in
>>>>> epf vhost 2) Using this X for both virtio device and
vhost(vringh)
>>>>> device 3) local virtio driver will negotiate with virtio
device
>>>>> with feature set Y 4) remote virtio driver will negotiate
with
>>>>> vringh device with feature set Z 5) mediate between feature
Y and
>>>>> feature Z since both Y and Z are a subset of X
>>>>>
>>>> okay. I'm also thinking if we could have configfs for
configuring
>>>> this. Anyways we could find different approaches of configuring
>>>> this.
>>>
>>>
>>> Yes, and I think some management API is needed even in the design
of
>>> your "Software Layering". In that figure, rpmsg vhost
need some
>>> pre-set or hard-coded features.
>>
>> When I saw the plumbers talk, my first idea was "this needs to be
a new
>> transport". You have some hard-coded or pre-configured features,
and
>> then features are negotiated via a transport-specific means in the
>> usual way. There's basically an extra/extended layer for this (and
>> status, and whatever).
>
> I think for PCIe root complex to PCIe endpoint communication it's 
> still "Virtio Over PCI Bus", though existing layout cannot be
used in
> this context (find virtio capability will fail for modern interface 
> and loading queue status immediately after writing queue number is not 
> possible for root complex to endpoint communication; setup_vq() in 
> virtio_pci_legacy.c).

Then you need something that is functional equivalent to virtio PCI 
which is actually the concept of vDPA (e.g vDPA provides alternatives if 
the queue_sel is hard in the EP implementation).

>
> "Virtio Over NTB" should anyways be a new transport.
>>
>> Does that make any sense?
>
> yeah, in the approach I used the initial features are hard-coded in 
> vhost-rpmsg (inherent to the rpmsg) but when we have to use adapter 
> layer (vhost only for accessing virtio ring and use virtio drivers on 
> both front end and backend), based on the functionality (e.g, rpmsg), 
> the vhost should be configured with features (to be presented to the 
> virtio) and that's why additional layer or APIs will be required.

A question here, if we go with vhost bus approach, does it mean the 
virtio device can only be implemented in EP's userspace?

Thanks

>>
>>>
>>>
>>>>>>>>> It will have virtqueues but only used for
the communication
>>>>>>>>> between itself and
>>>>>>>>> uppter virtio driver. And it will have
vringh queues which
>>>>>>>>> will be probe by virtio epf transport
drivers. And it needs to
>>>>>>>>> do datacopy between virtqueue and
>>>>>>>>> vringh queues.
>>>>>>>>>
>>>>>>>>> It works like:
>>>>>>>>>
>>>>>>>>> virtio drivers <- virtqueue/virtio-bus
-> epf vhost drivers <-
>>>>>>>>> vringh queue/epf>
>>>>>>>>>
>>>>>>>>> The advantages is that there's no need
for writing new buses
>>>>>>>>> and drivers.
>>>>>>>> I think this will work however there is an
addtional copy
>>>>>>>> between vringh queue and virtqueue,
>>>>>>> I think not? E.g in use case 1), if we stick to
virtio bus, we
>>>>>>> will have:
>>>>>>>
>>>>>>> virtio-rpmsg (EP) <- virtio ring(1) -> epf
vhost driver (EP) <-
>>>>>>> virtio ring(2) -> virtio pci (RC) <->
virtio rpmsg (RC)
>>>>>> IIUC epf vhost driver (EP) will access virtio ring(2)
using
>>>>>> vringh?
>>>>>
>>>>> Yes.
>>>>>
>>>>>> And virtio
>>>>>> ring(2) is created by virtio pci (RC).
>>>>>
>>>>> Yes.
>>>>>
>>>>>>> What epf vhost driver did is to read from virtio
ring(1) about
>>>>>>> the buffer len and addr and them DMA to Linux(RC)?
>>>>>> okay, I made some optimization here where vhost-rpmsg
using a
>>>>>> helper writes a buffer from rpmsg's upper layer
directly to
>>>>>> remote Linux (RC) as against here were it has to be
first written
>>>>>> to virtio ring (1).
>>>>>>
>>>>>> Thinking how this would look for NTB
>>>>>> virtio-rpmsg (HOST1) <- virtio ring(1) ->
NTB(HOST1) <->
>>>>>> NTB(HOST2)? <- virtio ring(2) -> virtio-rpmsg
(HOST2)
>>>>>>
>>>>>> Here the NTB(HOST1) will access the virtio ring(2)
using vringh?
>>>>>
>>>>> Yes, I think so it needs to use vring to access virtio ring
(1) as
>>>>> well.
>>>> NTB(HOST1) and virtio ring(1) will be in the same system. So it
>>>> doesn't have to use vring. virtio ring(1) is by the virtio
device
>>>> the NTB(HOST1) creates.
>>>
>>>
>>> Right.
>>>
>>>
>>>>>> Do you also think this will work seamlessly with
virtio_net.c,
>>>>>> virtio_blk.c?
>>>>>
>>>>> Yes.
>>>> okay, I haven't looked at this but the backend of
virtio_blk should
>>>> access an actual storage device no?
>>>
>>>
>>> Good point, for non-peer device like storage. There's probably
no
>>> need for it to be registered on the virtio bus and it might be
better
>>> to behave as you proposed.
>>
>> I might be missing something; but if you expose something as a block
>> device, it should have something it can access with block reads/writes,
>> shouldn't it? Of course, that can be a variety of things.
>>
>>>
>>> Just to make sure I understand the design, how is VHOST SCSI
expected
>>> to work in your proposal, does it have a device for file as a
backend?
>>>
>>>
>>>>>> I'd like to get clarity on two things in the
approach you
>>>>>> suggested, one is features (since epf vhost should
ideally be
>>>>>> transparent to any virtio driver)
>>>>>
>>>>> We can have have an array of pre-defined features indexed
by
>>>>> virtio device id in the code.
>>>>>
>>>>>> and the other is how certain inputs to virtio device
such as
>>>>>> number of buffers be determined.
>>>>>
>>>>> We can start from hard coded the value like 256, or
introduce some
>>>>> API for user to change the value.
>>>>>
>>>>>> Thanks again for your suggestions!
>>>>>
>>>>> You're welcome.
>>>>>
>>>>> Note that I just want to check whether or not we can reuse
the
>>>>> virtio bus/driver. It's something similar to what you
proposed in
>>>>> Software Layering but we just replace "vhost
core" with "virtio
>>>>> bus" and move the vhost core below epf/ntb/platform
transport.
>>>> Got it. My initial design was based on my understanding of your
>>>> comments [1].
>>>
>>>
>>> Yes, but that's just for a networking device. If we want
something
>>> more generic, it may require more thought (bus etc).
>>
>> I believe that we indeed need something bus-like to be able to support
>> a variety of devices.
>
> I think we could still have adapter layers for different types of 
> devices ([1]) and use existing virtio bus for both front end and back 
> end. Using bus-like will however simplify adding support for new types 
> of devices and adding adapters for devices will be slightly more complex.
>
> [1] -> Page 13 in 
>
https://linuxplumbersconf.org/event/7/contributions/849/attachments/642/1175/Virtio_for_PCIe_RC_EP_NTB.pdf
>>
>>>
>>>
>>>>
>>>> I'll try to create something based on your proposed design
here.
>>>
>>>
>>> Sure, but for coding, we'd better wait for other's opinion
here.
>>
>> Please tell me if my thoughts above make any sense... I have just
>> started looking at that, so I might be completely off.
>
> I think your understanding is correct! Thanks for your inputs.
>
> Thanks
> Kishon

On Tue, 1 Sep 2020 16:50:03 +0800
Jason Wang <jasowang at redhat.com> wrote:
> On 2020/9/1 ??1:24, Kishon Vijay Abraham I wrote:
> > Hi,
> >
> > On 28/08/20 4:04 pm, Cornelia Huck wrote:  
> >> On Thu, 9 Jul 2020 14:26:53 +0800
> >> Jason Wang <jasowang at redhat.com> wrote:
> >>
> >> [Let me note right at the beginning that I first noted this while
> >> listening to Kishon's talk at LPC on Wednesday. I might be
very
> >> confused about the background here, so let me apologize beforehand
for
> >> any confusion I might spread.]
> >>  
> >>> On 2020/7/8 ??9:13, Kishon Vijay Abraham I wrote:  
> >>>> Hi Jason,
> >>>>
> >>>> On 7/8/2020 4:52 PM, Jason Wang wrote:  
> >>>>> On 2020/7/7 ??10:45, Kishon Vijay Abraham I wrote:  
> >>>>>> Hi Jason,
> >>>>>>
> >>>>>> On 7/7/2020 3:17 PM, Jason Wang wrote:  
> >>>>>>> On 2020/7/6 ??5:32, Kishon Vijay Abraham I
wrote:
> >>>>>>>> Hi Jason,
> >>>>>>>>
> >>>>>>>> On 7/3/2020 12:46 PM, Jason Wang wrote:  
> >>>>>>>>> On 2020/7/2 ??9:35, Kishon Vijay
Abraham I wrote:
> >>>>>>>>>> Hi Jason,
> >>>>>>>>>>
> >>>>>>>>>> On 7/2/2020 3:40 PM, Jason Wang
wrote:
> >>>>>>>>>>> On 2020/7/2 ??5:51, Michael S.
Tsirkin wrote:
> >>>>>>>>>>>> On Thu, Jul 02, 2020 at
01:51:21PM +0530, Kishon Vijay
> >>>>>>>>>>>> Abraham I wrote:  
> >>>>>>>>>>>>> This series enhances
Linux Vhost support to enable SoC-to-SoC
> >>>>>>>>>>>>> communication over
MMIO. This series enables rpmsg
> >>>>>>>>>>>>> communication between
> >>>>>>>>>>>>> two SoCs using both
PCIe RC<->EP and HOST1-NTB-HOST2
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> 1) Modify vhost to use
standard Linux driver model
> >>>>>>>>>>>>> 2) Add support in
vring to access virtqueue over MMIO
> >>>>>>>>>>>>> 3) Add vhost client
driver for rpmsg
> >>>>>>>>>>>>> 4) Add PCIe RC driver
(uses virtio) and PCIe EP driver
> >>>>>>>>>>>>> (uses vhost) for
> >>>>>>>>>>>>> ? ?????? rpmsg
communication between two SoCs connected to
> >>>>>>>>>>>>> each other
> >>>>>>>>>>>>> 5) Add NTB Virtio
driver and NTB Vhost driver for rpmsg
> >>>>>>>>>>>>> communication
> >>>>>>>>>>>>> ? ?????? between two
SoCs connected via NTB
> >>>>>>>>>>>>> 6) Add configfs to
configure the components
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> UseCase1 :
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> ? ???? VHOST RPMSG
VIRTIO RPMSG
> >>>>>>>>>>>>>
+?????????????????????????????? +
> >>>>>>>>>>>>>
|?????????????????????????????? |
> >>>>>>>>>>>>>
|?????????????????????????????? |
> >>>>>>>>>>>>>
|?????????????????????????????? |
> >>>>>>>>>>>>>
|?????????????????????????????? |
> >>>>>>>>>>>>> +-----v------+
+------v-------+
> >>>>>>>>>>>>> |?? Linux???
|???????????????? | Linux??? |
> >>>>>>>>>>>>> |? Endpoint?
|???????????????? | Root Complex |
> >>>>>>>>>>>>> |
<----------------->????????????? |
> >>>>>>>>>>>>> |??????????? |
|????????????? |
> >>>>>>>>>>>>> |??? SOC1???
|???????????????? | SOC2???? |
> >>>>>>>>>>>>> +------------+
+--------------+
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> UseCase 2:
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> ? ???????? VHOST RPMSG
VIRTIO RPMSG
> >>>>>>>>>>>>> + +
> >>>>>>>>>>>>> | |
> >>>>>>>>>>>>> | |
> >>>>>>>>>>>>> | |
> >>>>>>>>>>>>> | |
> >>>>>>>>>>>>> +------v------+
+------v------+
> >>>>>>>>>>>>> ? ?????? | |
|???????????? |
> >>>>>>>>>>>>> ? ?????? |??? HOST1
|?????????????????????????????????? |
> >>>>>>>>>>>>> HOST2??? |
> >>>>>>>>>>>>> ? ?????? | |
|???????????? |
> >>>>>>>>>>>>> +------^------+
+------^------+
> >>>>>>>>>>>>> | |
> >>>>>>>>>>>>> | |
> >>>>>>>>>>>>>
+---------------------------------------------------------------------+
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> | +------v------+
+------v------+? |
> >>>>>>>>>>>>> |? | | |????????????
|? |
> >>>>>>>>>>>>> |? |???? EP
|?????????????????????????????????? | EP?????
> >>>>>>>>>>>>> |? |
> >>>>>>>>>>>>> |? | CONTROLLER1
|?????????????????????????????????? |
> >>>>>>>>>>>>> CONTROLLER2 |? |
> >>>>>>>>>>>>> |? |
<-----------------------------------> |? |
> >>>>>>>>>>>>> |? | | |????????????
|? |
> >>>>>>>>>>>>> |? | | |????????????
|? |
> >>>>>>>>>>>>> |? |???????????? |?
SoC With Multiple EP Instances??
> >>>>>>>>>>>>> |???????????? |? |
> >>>>>>>>>>>>> |? |???????????? |?
(Configured using NTB Function)?
> >>>>>>>>>>>>> |???????????? |? |
> >>>>>>>>>>>>> | +-------------+
+-------------+? |
> >>>>>>>>>>>>>
+---------------------------------------------------------------------+
> >>>>>>>>>>>>>  
> >>
> >> First of all, to clarify the terminology:
> >> Is "vhost rpmsg" acting as what the virtio standard
calls the 'device',
> >> and "virtio rpmsg" as the 'driver'? Or is the
"vhost" part mostly just
> >
> > Right, vhost_rpmsg is 'device' and virtio_rpmsg is
'driver'.
> >> virtqueues + the exiting vhost interfaces?  
> >
> > It's implemented to provide the full 'device'
functionality.
Ok.
> >>  
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> Software Layering:
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> The high-level SW
layering should look something like
> >>>>>>>>>>>>> below. This series
> >>>>>>>>>>>>> adds support only for
RPMSG VHOST, however something
> >>>>>>>>>>>>> similar should be
> >>>>>>>>>>>>> done for net and scsi.
With that any vhost device (PCI,
> >>>>>>>>>>>>> NTB, Platform
> >>>>>>>>>>>>> device, user) can use
any of the vhost client driver.
> >>>>>>>>>>>>>
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> ? ???????
+----------------+ +-----------+? +------------+
> >>>>>>>>>>>>> +----------+
> >>>>>>>>>>>>> ? ??????? |? RPMSG
VHOST?? |? | NET VHOST |? | SCSI VHOST
> >>>>>>>>>>>>> |? |??? X???? |
> >>>>>>>>>>>>> ? ???????
+-------^--------+ +-----^-----+? +-----^------+
> >>>>>>>>>>>>> +----^-----+
> >>>>>>>>>>>>> ? ??????????????? |
|????????????? |????????????? |
> >>>>>>>>>>>>> ? ??????????????? |
|????????????? |????????????? |
> >>>>>>>>>>>>> ? ??????????????? |
|????????????? |????????????? |
> >>>>>>>>>>>>>
+-----------v-----------------v--------------v--------------v----------+
> >>>>>>>>>>>>>
> >>>>>>>>>>>>>
|??????????????????????????? VHOST
> >>>>>>>>>>>>>
CORE??????????????????????????????? |
> >>>>>>>>>>>>>
+--------^---------------^--------------------^------------------^-----+
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> ? ???????????? |
|??????????????????? |????????????????? |
> >>>>>>>>>>>>> ? ???????????? |
|??????????????????? |????????????????? |
> >>>>>>>>>>>>> ? ???????????? |
|??????????????????? |????????????????? |
> >>>>>>>>>>>>> +--------v-------+?
+----v------+ +----------v----------+?
> >>>>>>>>>>>>> +----v-----+
> >>>>>>>>>>>>> |? PCI EPF VHOST |? |
NTB VHOST | |PLATFORM DEVICE VHOST|?
> >>>>>>>>>>>>> |??? X???? |
> >>>>>>>>>>>>> +----------------+?
+-----------+ +---------------------+?
> >>>>>>>>>>>>> +----------+  
> >>
> >> So, the upper half is basically various functionality types, e.g.
a net
> >> device. What is the lower half, a hardware interface? Would it be
> >> equivalent to e.g. a normal PCI device?  
> >
> > Right, the upper half should provide the functionality.
> > The bottom layer could be a HW interface (like PCIe device or NTB 
> > device) or it could be a SW interface (for accessing virtio ring in 
> > userspace) that could be used by Hypervisor.
Ok. In that respect, the SW interface is sufficiently device-like, I
guess.
> >
> > The top half should be transparent to what type of device is actually 
> > using it.
> >  
> >>  
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> This was initially
proposed here [1]
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> [1] ->
> >>>>>>>>>>>>>
https://lore.kernel.org/r/2cf00ec4-1ed6-f66e-6897-006d1a5b6390 at ti.com
> >>>>>>>>>>>>>  
> >>>>>>>>>>>> I find this very
interesting. A huge patchset so will take
> >>>>>>>>>>>> a bit
> >>>>>>>>>>>> to review, but I certainly
plan to do that. Thanks!
> >>>>>>>>>>> Yes, it would be better if
there's a git branch for us to
> >>>>>>>>>>> have a look.  
> >>>>>>>>>> I've pushed the branch
> >>>>>>>>>>
https://github.com/kishon/linux-wip.git vhost_rpmsg_pci_ntb_rfc
> >>>>>>>>> Thanks
> >>>>>>>>>  
> >>>>>>>>>>> Btw, I'm not sure I get
the big picture, but I vaguely feel
> >>>>>>>>>>> some of the
> >>>>>>>>>>> work is
> >>>>>>>>>>> duplicated with vDPA (e.g the
epf transport or vhost bus).
> >>>>>>>>>> This is about connecting two
different HW systems both
> >>>>>>>>>> running Linux and
> >>>>>>>>>> doesn't necessarily involve
virtualization.
> >>>>>>>>> Right, this is something similar to
VOP
> >>>>>>>>>
(Documentation/misc-devices/mic/mic_overview.rst). The
> >>>>>>>>> different is the
> >>>>>>>>> hardware I guess and VOP use userspace
application to
> >>>>>>>>> implement the device.  
> >>>>>>>> I'd also like to point out, this
series tries to have
> >>>>>>>> communication between
> >>>>>>>> two
> >>>>>>>> SoCs in vendor agnostic way. Since this
series solves for 2
> >>>>>>>> usecases (PCIe
> >>>>>>>> RC<->EP and NTB), for the NTB case
it directly plugs into NTB
> >>>>>>>> framework and
> >>>>>>>> any
> >>>>>>>> of the HW in NTB below should be able to
use a virtio-vhost
> >>>>>>>> communication
> >>>>>>>>
> >>>>>>>> #ls drivers/ntb/hw/
> >>>>>>>> amd? epf? idt? intel? mscc
> >>>>>>>>
> >>>>>>>> And similarly for the PCIe RC<->EP
communication, this adds a
> >>>>>>>> generic endpoint
> >>>>>>>> function driver and hence any SoC that
supports configurable
> >>>>>>>> PCIe endpoint can
> >>>>>>>> use virtio-vhost communication
> >>>>>>>>
> >>>>>>>> # ls drivers/pci/controller/dwc/*ep*
> >>>>>>>>
drivers/pci/controller/dwc/pcie-designware-ep.c
> >>>>>>>>
drivers/pci/controller/dwc/pcie-uniphier-ep.c
> >>>>>>>>
drivers/pci/controller/dwc/pci-layerscape-ep.c
> >>>>>>> Thanks for those backgrounds.
> >>>>>>>  
> >>>>>>>>>> ? ??? So there is no guest or host
as in
> >>>>>>>>>> virtualization but two entirely
different systems connected
> >>>>>>>>>> via PCIe cable,
> >>>>>>>>>> one
> >>>>>>>>>> acting as guest and one as host.
So one system will provide
> >>>>>>>>>> virtio
> >>>>>>>>>> functionality reserving memory for
virtqueues and the other
> >>>>>>>>>> provides vhost
> >>>>>>>>>> functionality providing a way to
access the virtqueues in
> >>>>>>>>>> virtio memory.
> >>>>>>>>>> One is
> >>>>>>>>>> source and the other is sink and
there is no intermediate
> >>>>>>>>>> entity. (vhost was
> >>>>>>>>>> probably intermediate entity in
virtualization?)
> >>>>>>>>> (Not a native English speaker) but
"vhost" could introduce
> >>>>>>>>> some confusion for
> >>>>>>>>> me since it was use for implementing
virtio backend for
> >>>>>>>>> userspace drivers. I
> >>>>>>>>> guess "vringh" could be
better.
> >>>>>>>> Initially I had named this vringh but
later decided to choose
> >>>>>>>> vhost instead of
> >>>>>>>> vringh. vhost is still a virtio backend
(not necessarily
> >>>>>>>> userspace) though it
> >>>>>>>> now resides in an entirely different
system. Whatever virtio is
> >>>>>>>> for a frontend
> >>>>>>>> system, vhost can be that for a backend
system. vring can be
> >>>>>>>> for accessing
> >>>>>>>> virtqueue and can be used either in
frontend or backend.
> >>
> >> I guess that clears up at least some of my questions from above...
> >>  
> >>>>>>> Ok.
> >>>>>>>  
> >>>>>>>>>>> Have you considered to
implement these through vDPA?
> >>>>>>>>>> IIUC vDPA only provides an
interface to userspace and an
> >>>>>>>>>> in-kernel rpmsg
> >>>>>>>>>> driver
> >>>>>>>>>> or vhost net driver is not
provided.
> >>>>>>>>>>
> >>>>>>>>>> The HW connection looks something
like
> >>>>>>>>>> https://pasteboard.co/JfMVVHC.jpg
> >>>>>>>>>> (usecase2 above),  
> >>>>>>>>> I see.
> >>>>>>>>>  
> >>>>>>>>>> ? ??? all the boards run Linux.
The middle board provides NTB
> >>>>>>>>>> functionality and board on either
side provides virtio/vhost
> >>>>>>>>>> functionality and
> >>>>>>>>>> transfer data using rpmsg.  
> >>
> >> This setup looks really interesting (sometimes, it's really
hard to
> >> imagine this in the abstract.)  
> >>>>>>>>> So I wonder whether it's
worthwhile for a new bus. Can we use
> >>>>>>>>> the existed virtio-bus/drivers? It
might work as, except for
> >>>>>>>>> the epf transport, we can introduce a
epf "vhost" transport
> >>>>>>>>> driver.  
> >>>>>>>> IMHO we'll need two buses one for
frontend and other for
> >>>>>>>> backend because the two components can
then co-operate/interact
> >>>>>>>> with each other to provide a
functionality. Though both will
> >>>>>>>> seemingly provide similar callbacks, they
are both provide
> >>>>>>>> symmetrical or complimentary funcitonality
and need not be same
> >>>>>>>> or identical.
> >>>>>>>>
> >>>>>>>> Having the same bus can also create
sequencing issues.
> >>>>>>>>
> >>>>>>>> If you look at virtio_dev_probe() of
virtio_bus
> >>>>>>>>
> >>>>>>>> device_features =
dev->config->get_features(dev);
> >>>>>>>>
> >>>>>>>> Now if we use same bus for both front-end
and back-end, both
> >>>>>>>> will try to get_features when there has
been no set_features.
> >>>>>>>> Ideally vhost device should be initialized
first with the set
> >>>>>>>> of features it supports. Vhost and virtio
should use "status"
> >>>>>>>> and "features" complimentarily
and not identically.
> >>>>>>> Yes, but there's no need for doing
status/features passthrough
> >>>>>>> in epf vhost drivers.b
> >>>>>>>  
> >>>>>>>> virtio device (or frontend) cannot be
initialized before vhost
> >>>>>>>> device (or backend) gets initialized with
data such as
> >>>>>>>> features. Similarly vhost (backend)
> >>>>>>>> cannot access virqueues or buffers before
virtio (frontend) sets
> >>>>>>>> VIRTIO_CONFIG_S_DRIVER_OK whereas that
requirement is not there
> >>>>>>>> for virtio as the physical memory for
virtqueues are created by
> >>>>>>>> virtio (frontend).  
> >>>>>>> epf vhost drivers need to implement two
devices: vhost(vringh)
> >>>>>>> device and virtio device (which is a mediated
device). The
> >>>>>>> vhost(vringh) device is doing feature
negotiation with the
> >>>>>>> virtio device via RC/EP or NTB. The virtio
device is doing
> >>>>>>> feature negotiation with local virtio drivers.
If there're
> >>>>>>> feature mismatch, epf vhost drivers and do
mediation between
> >>>>>>> them.  
> >>>>>> Here epf vhost should be initialized with a set of
features for
> >>>>>> it to negotiate either as vhost device or virtio
device no? Where
> >>>>>> should the initial feature set for epf vhost come
from?
> >>>>>
> >>>>> I think it can work as:
> >>>>>
> >>>>> 1) Having an initial features (hard coded in the code)
set X in
> >>>>> epf vhost 2) Using this X for both virtio device and
vhost(vringh)
> >>>>> device 3) local virtio driver will negotiate with
virtio device
> >>>>> with feature set Y 4) remote virtio driver will
negotiate with
> >>>>> vringh device with feature set Z 5) mediate between
feature Y and
> >>>>> feature Z since both Y and Z are a subset of X
> >>>>>  
> >>>> okay. I'm also thinking if we could have configfs for
configuring
> >>>> this. Anyways we could find different approaches of
configuring
> >>>> this.  
> >>>
> >>>
> >>> Yes, and I think some management API is needed even in the
design of
> >>> your "Software Layering". In that figure, rpmsg
vhost need some
> >>> pre-set or hard-coded features.  
> >>
> >> When I saw the plumbers talk, my first idea was "this needs
to be a new
> >> transport". You have some hard-coded or pre-configured
features, and
> >> then features are negotiated via a transport-specific means in the
> >> usual way. There's basically an extra/extended layer for this
(and
> >> status, and whatever).  
> >
> > I think for PCIe root complex to PCIe endpoint communication it's 
> > still "Virtio Over PCI Bus", though existing layout cannot
be used in
> > this context (find virtio capability will fail for modern interface 
> > and loading queue status immediately after writing queue number is not
> > possible for root complex to endpoint communication; setup_vq() in 
> > virtio_pci_legacy.c).  
> 
> 
> Then you need something that is functional equivalent to virtio PCI 
> which is actually the concept of vDPA (e.g vDPA provides alternatives if 
> the queue_sel is hard in the EP implementation).
It seems I really need to read up on vDPA more... do you have a pointer
for diving into this alternatives aspect?
> 
> 
> >
> > "Virtio Over NTB" should anyways be a new transport.  
> >>
> >> Does that make any sense?  
> >
> > yeah, in the approach I used the initial features are hard-coded in 
> > vhost-rpmsg (inherent to the rpmsg) but when we have to use adapter 
> > layer (vhost only for accessing virtio ring and use virtio drivers on 
> > both front end and backend), based on the functionality (e.g, rpmsg), 
> > the vhost should be configured with features (to be presented to the 
> > virtio) and that's why additional layer or APIs will be required.
> 
> 
> A question here, if we go with vhost bus approach, does it mean the 
> virtio device can only be implemented in EP's userspace?
Can we maybe implement an alternative bus as well that would allow us
to support different virtio device implementations (in addition to the
vhost bus + userspace combination)?
> 
> Thanks
> 
> 
> >>  
> >>>
> >>>  
> >>>>>>>>> It will have virtqueues but only used
for the communication
> >>>>>>>>> between itself and
> >>>>>>>>> uppter virtio driver. And it will have
vringh queues which
> >>>>>>>>> will be probe by virtio epf transport
drivers. And it needs to
> >>>>>>>>> do datacopy between virtqueue and
> >>>>>>>>> vringh queues.
> >>>>>>>>>
> >>>>>>>>> It works like:
> >>>>>>>>>
> >>>>>>>>> virtio drivers <-
virtqueue/virtio-bus -> epf vhost drivers <-
> >>>>>>>>> vringh queue/epf>
> >>>>>>>>>
> >>>>>>>>> The advantages is that there's no
need for writing new buses
> >>>>>>>>> and drivers.  
> >>>>>>>> I think this will work however there is an
addtional copy
> >>>>>>>> between vringh queue and virtqueue,  
> >>>>>>> I think not? E.g in use case 1), if we stick
to virtio bus, we
> >>>>>>> will have:
> >>>>>>>
> >>>>>>> virtio-rpmsg (EP) <- virtio ring(1) ->
epf vhost driver (EP) <-
> >>>>>>> virtio ring(2) -> virtio pci (RC) <->
virtio rpmsg (RC)
> >>>>>> IIUC epf vhost driver (EP) will access virtio
ring(2) using
> >>>>>> vringh?  
> >>>>>
> >>>>> Yes.
> >>>>>  
> >>>>>> And virtio
> >>>>>> ring(2) is created by virtio pci (RC).  
> >>>>>
> >>>>> Yes.
> >>>>>  
> >>>>>>> What epf vhost driver did is to read from
virtio ring(1) about
> >>>>>>> the buffer len and addr and them DMA to
Linux(RC)?
> >>>>>> okay, I made some optimization here where
vhost-rpmsg using a
> >>>>>> helper writes a buffer from rpmsg's upper
layer directly to
> >>>>>> remote Linux (RC) as against here were it has to
be first written
> >>>>>> to virtio ring (1).
> >>>>>>
> >>>>>> Thinking how this would look for NTB
> >>>>>> virtio-rpmsg (HOST1) <- virtio ring(1) ->
NTB(HOST1) <->
> >>>>>> NTB(HOST2)? <- virtio ring(2) ->
virtio-rpmsg (HOST2)
> >>>>>>
> >>>>>> Here the NTB(HOST1) will access the virtio ring(2)
using vringh?
> >>>>>
> >>>>> Yes, I think so it needs to use vring to access virtio
ring (1) as
> >>>>> well.  
> >>>> NTB(HOST1) and virtio ring(1) will be in the same system.
So it
> >>>> doesn't have to use vring. virtio ring(1) is by the
virtio device
> >>>> the NTB(HOST1) creates.  
> >>>
> >>>
> >>> Right.
> >>>
> >>>  
> >>>>>> Do you also think this will work seamlessly with
virtio_net.c,
> >>>>>> virtio_blk.c?  
> >>>>>
> >>>>> Yes.  
> >>>> okay, I haven't looked at this but the backend of
virtio_blk should
> >>>> access an actual storage device no?  
> >>>
> >>>
> >>> Good point, for non-peer device like storage. There's
probably no
> >>> need for it to be registered on the virtio bus and it might be
better
> >>> to behave as you proposed.  
> >>
> >> I might be missing something; but if you expose something as a
block
> >> device, it should have something it can access with block
reads/writes,
> >> shouldn't it? Of course, that can be a variety of things.
> >>  
> >>>
> >>> Just to make sure I understand the design, how is VHOST SCSI
expected
> >>> to work in your proposal, does it have a device for file as a
backend?
> >>>
> >>>  
> >>>>>> I'd like to get clarity on two things in the
approach you
> >>>>>> suggested, one is features (since epf vhost should
ideally be
> >>>>>> transparent to any virtio driver)  
> >>>>>
> >>>>> We can have have an array of pre-defined features
indexed by
> >>>>> virtio device id in the code.
> >>>>>  
> >>>>>> and the other is how certain inputs to virtio
device such as
> >>>>>> number of buffers be determined.  
> >>>>>
> >>>>> We can start from hard coded the value like 256, or
introduce some
> >>>>> API for user to change the value.
> >>>>>  
> >>>>>> Thanks again for your suggestions!  
> >>>>>
> >>>>> You're welcome.
> >>>>>
> >>>>> Note that I just want to check whether or not we can
reuse the
> >>>>> virtio bus/driver. It's something similar to what
you proposed in
> >>>>> Software Layering but we just replace "vhost
core" with "virtio
> >>>>> bus" and move the vhost core below
epf/ntb/platform transport.
> >>>> Got it. My initial design was based on my understanding of
your
> >>>> comments [1].  
> >>>
> >>>
> >>> Yes, but that's just for a networking device. If we want
something
> >>> more generic, it may require more thought (bus etc).  
> >>
> >> I believe that we indeed need something bus-like to be able to
support
> >> a variety of devices.  
> >
> > I think we could still have adapter layers for different types of 
> > devices ([1]) and use existing virtio bus for both front end and back 
> > end. Using bus-like will however simplify adding support for new types
> > of devices and adding adapters for devices will be slightly more
complex.
> >
> > [1] -> Page 13 in 
> >
https://linuxplumbersconf.org/event/7/contributions/849/attachments/642/1175/Virtio_for_PCIe_RC_EP_NTB.pdf
So, I guess it's a trade-off: do we expect a variety of device types,
or more of a variety of devices needing different adapters?

Hi Kishon:

On 2020/9/14 ??3:23, Kishon Vijay Abraham I wrote:
>> Then you need something that is functional equivalent to virtio PCI
>> which is actually the concept of vDPA (e.g vDPA provides alternatives
if
>> the queue_sel is hard in the EP implementation).
> Okay, I just tried to compare the 'struct vdpa_config_ops' and
'struct
> vhost_config_ops' ( introduced in [RFC PATCH 03/22] vhost: Add ops for
> the VHOST driver to configure VHOST device).
>
> struct vdpa_config_ops {
> 	/* Virtqueue ops */
> 	int (*set_vq_address)(struct vdpa_device *vdev,
> 			      u16 idx, u64 desc_area, u64 driver_area,
> 			      u64 device_area);
> 	void (*set_vq_num)(struct vdpa_device *vdev, u16 idx, u32 num);
> 	void (*kick_vq)(struct vdpa_device *vdev, u16 idx);
> 	void (*set_vq_cb)(struct vdpa_device *vdev, u16 idx,
> 			  struct vdpa_callback *cb);
> 	void (*set_vq_ready)(struct vdpa_device *vdev, u16 idx, bool ready);
> 	bool (*get_vq_ready)(struct vdpa_device *vdev, u16 idx);
> 	int (*set_vq_state)(struct vdpa_device *vdev, u16 idx,
> 			    const struct vdpa_vq_state *state);
> 	int (*get_vq_state)(struct vdpa_device *vdev, u16 idx,
> 			    struct vdpa_vq_state *state);
> 	struct vdpa_notification_area
> 	(*get_vq_notification)(struct vdpa_device *vdev, u16 idx);
> 	/* vq irq is not expected to be changed once DRIVER_OK is set */
> 	int (*get_vq_irq)(struct vdpa_device *vdv, u16 idx);
>
> 	/* Device ops */
> 	u32 (*get_vq_align)(struct vdpa_device *vdev);
> 	u64 (*get_features)(struct vdpa_device *vdev);
> 	int (*set_features)(struct vdpa_device *vdev, u64 features);
> 	void (*set_config_cb)(struct vdpa_device *vdev,
> 			      struct vdpa_callback *cb);
> 	u16 (*get_vq_num_max)(struct vdpa_device *vdev);
> 	u32 (*get_device_id)(struct vdpa_device *vdev);
> 	u32 (*get_vendor_id)(struct vdpa_device *vdev);
> 	u8 (*get_status)(struct vdpa_device *vdev);
> 	void (*set_status)(struct vdpa_device *vdev, u8 status);
> 	void (*get_config)(struct vdpa_device *vdev, unsigned int offset,
> 			   void *buf, unsigned int len);
> 	void (*set_config)(struct vdpa_device *vdev, unsigned int offset,
> 			   const void *buf, unsigned int len);
> 	u32 (*get_generation)(struct vdpa_device *vdev);
>
> 	/* DMA ops */
> 	int (*set_map)(struct vdpa_device *vdev, struct vhost_iotlb *iotlb);
> 	int (*dma_map)(struct vdpa_device *vdev, u64 iova, u64 size,
> 		       u64 pa, u32 perm);
> 	int (*dma_unmap)(struct vdpa_device *vdev, u64 iova, u64 size);
>
> 	/* Free device resources */
> 	void (*free)(struct vdpa_device *vdev);
> };
>
> +struct vhost_config_ops {
> +	int (*create_vqs)(struct vhost_dev *vdev, unsigned int nvqs,
> +			  unsigned int num_bufs, struct vhost_virtqueue *vqs[],
> +			  vhost_vq_callback_t *callbacks[],
> +			  const char * const names[]);
> +	void (*del_vqs)(struct vhost_dev *vdev);
> +	int (*write)(struct vhost_dev *vdev, u64 vhost_dst, void *src, int len);
> +	int (*read)(struct vhost_dev *vdev, void *dst, u64 vhost_src, int len);
> +	int (*set_features)(struct vhost_dev *vdev, u64 device_features);
> +	int (*set_status)(struct vhost_dev *vdev, u8 status);
> +	u8 (*get_status)(struct vhost_dev *vdev);
> +};
> +
> struct virtio_config_ops
> I think there's some overlap here and some of the ops tries to do the
> same thing.
>
> I think it differs in (*set_vq_address)() and (*create_vqs)().
> [create_vqs() introduced in struct vhost_config_ops provides
> complimentary functionality to (*find_vqs)() in struct
> virtio_config_ops. It seemingly encapsulates the functionality of
> (*set_vq_address)(), (*set_vq_num)(), (*set_vq_cb)(),..].
>
> Back to the difference between (*set_vq_address)() and (*create_vqs)(),
> set_vq_address() directly provides the virtqueue address to the vdpa
> device but create_vqs() only provides the parameters of the virtqueue
> (like the number of virtqueues, number of buffers) but does not directly
> provide the address. IMO the backend client drivers (like net or vhost)
> shouldn't/cannot by itself know how to access the vring created on
> virtio front-end. The vdpa device/vhost device should have logic for
> that. That will help the client drivers to work with different types of
> vdpa device/vhost device and can access the vring created by virtio
> irrespective of whether the vring can be accessed via mmio or kernel
> space or user space.
>
> I think vdpa always works with client drivers in userspace and providing
> userspace address for vring.

Sorry for being unclear. What I meant is not replacing vDPA with the 
vhost(bus) you proposed but the possibility of replacing virtio-pci-epf 
with vDPA in:

My question is basically for the part of virtio_pci_epf_send_command(), 
so it looks to me you have a vendor specific API to replace the 
virtio-pci layout of the BAR:


+static int virtio_pci_epf_send_command(struct virtio_pci_device *vp_dev,
+??? ??? ??? ??? ?????? u32 command)
+{
+??? struct virtio_pci_epf *pci_epf;
+??? void __iomem *ioaddr;
+??? ktime_t timeout;
+??? bool timedout;
+??? int ret = 0;
+??? u8 status;
+
+??? pci_epf = to_virtio_pci_epf(vp_dev);
+??? ioaddr = vp_dev->ioaddr;
+
+??? mutex_lock(&pci_epf->lock);
+??? writeb(command, ioaddr + HOST_CMD);
+??? timeout = ktime_add_ms(ktime_get(), COMMAND_TIMEOUT);
+??? while (1) {
+??? ??? timedout = ktime_after(ktime_get(), timeout);
+??? ??? status = readb(ioaddr + HOST_CMD_STATUS);
+

Several questions:

- It's not clear to me how the synchronization is done between the RC 
and EP. E.g how and when the value of HOST_CMD_STATUS can be changed.? 
If you still want to introduce a new transport, a virtio spec patch 
would be helpful for us to understand the device API.
- You have you vendor specific layout (according to 
virtio_pci_epb_table()), so I guess you it's better to have a vendor 
specific vDPA driver instead
- The advantage of vendor specific vDPA driver is that it can 1) have 
less codes 2) support userspace drivers through vhost-vDPA (instead of 
inventing new APIs since we can't use vfio-pci here).

>>> "Virtio Over NTB" should anyways be a new transport.
>>>> Does that make any sense?
>>> yeah, in the approach I used the initial features are hard-coded in
>>> vhost-rpmsg (inherent to the rpmsg) but when we have to use adapter
>>> layer (vhost only for accessing virtio ring and use virtio drivers
on
>>> both front end and backend), based on the functionality (e.g,
rpmsg),
>>> the vhost should be configured with features (to be presented to
the
>>> virtio) and that's why additional layer or APIs will be
required.
>> A question here, if we go with vhost bus approach, does it mean the
>> virtio device can only be implemented in EP's userspace?
> The vhost bus approach doesn't provide any restriction in where the
> virto backend device should be created. This series creates two types of
> virtio backend device (one for PCIe endpoint and the other for NTB) and
> both these devices are created in kernel.

Ok.

Thanks

>
> Thanks
> Kishon
>

On 2020/9/15 ??11:47, Kishon Vijay Abraham I wrote:
> Hi Jason,
>
> On 15/09/20 1:48 pm, Jason Wang wrote:
>> Hi Kishon:
>>
>> On 2020/9/14 ??3:23, Kishon Vijay Abraham I wrote:
>>>> Then you need something that is functional equivalent to virtio
PCI
>>>> which is actually the concept of vDPA (e.g vDPA provides
alternatives if
>>>> the queue_sel is hard in the EP implementation).
>>> Okay, I just tried to compare the 'struct vdpa_config_ops'
and 'struct
>>> vhost_config_ops' ( introduced in [RFC PATCH 03/22] vhost: Add
ops for
>>> the VHOST driver to configure VHOST device).
>>>
>>> struct vdpa_config_ops {
>>>  ????/* Virtqueue ops */
>>>  ????int (*set_vq_address)(struct vdpa_device *vdev,
>>>  ????????????????? u16 idx, u64 desc_area, u64 driver_area,
>>>  ????????????????? u64 device_area);
>>>  ????void (*set_vq_num)(struct vdpa_device *vdev, u16 idx, u32
num);
>>>  ????void (*kick_vq)(struct vdpa_device *vdev, u16 idx);
>>>  ????void (*set_vq_cb)(struct vdpa_device *vdev, u16 idx,
>>>  ????????????? struct vdpa_callback *cb);
>>>  ????void (*set_vq_ready)(struct vdpa_device *vdev, u16 idx, bool
ready);
>>>  ????bool (*get_vq_ready)(struct vdpa_device *vdev, u16 idx);
>>>  ????int (*set_vq_state)(struct vdpa_device *vdev, u16 idx,
>>>  ??????????????? const struct vdpa_vq_state *state);
>>>  ????int (*get_vq_state)(struct vdpa_device *vdev, u16 idx,
>>>  ??????????????? struct vdpa_vq_state *state);
>>>  ????struct vdpa_notification_area
>>>  ????(*get_vq_notification)(struct vdpa_device *vdev, u16 idx);
>>>  ????/* vq irq is not expected to be changed once DRIVER_OK is set
*/
>>>  ????int (*get_vq_irq)(struct vdpa_device *vdv, u16 idx);
>>>
>>>  ????/* Device ops */
>>>  ????u32 (*get_vq_align)(struct vdpa_device *vdev);
>>>  ????u64 (*get_features)(struct vdpa_device *vdev);
>>>  ????int (*set_features)(struct vdpa_device *vdev, u64 features);
>>>  ????void (*set_config_cb)(struct vdpa_device *vdev,
>>>  ????????????????? struct vdpa_callback *cb);
>>>  ????u16 (*get_vq_num_max)(struct vdpa_device *vdev);
>>>  ????u32 (*get_device_id)(struct vdpa_device *vdev);
>>>  ????u32 (*get_vendor_id)(struct vdpa_device *vdev);
>>>  ????u8 (*get_status)(struct vdpa_device *vdev);
>>>  ????void (*set_status)(struct vdpa_device *vdev, u8 status);
>>>  ????void (*get_config)(struct vdpa_device *vdev, unsigned int
offset,
>>>  ?????????????? void *buf, unsigned int len);
>>>  ????void (*set_config)(struct vdpa_device *vdev, unsigned int
offset,
>>>  ?????????????? const void *buf, unsigned int len);
>>>  ????u32 (*get_generation)(struct vdpa_device *vdev);
>>>
>>>  ????/* DMA ops */
>>>  ????int (*set_map)(struct vdpa_device *vdev, struct vhost_iotlb
*iotlb);
>>>  ????int (*dma_map)(struct vdpa_device *vdev, u64 iova, u64 size,
>>>  ?????????????? u64 pa, u32 perm);
>>>  ????int (*dma_unmap)(struct vdpa_device *vdev, u64 iova, u64
size);
>>>
>>>  ????/* Free device resources */
>>>  ????void (*free)(struct vdpa_device *vdev);
>>> };
>>>
>>> +struct vhost_config_ops {
>>> +??? int (*create_vqs)(struct vhost_dev *vdev, unsigned int nvqs,
>>> +????????????? unsigned int num_bufs, struct vhost_virtqueue
*vqs[],
>>> +????????????? vhost_vq_callback_t *callbacks[],
>>> +????????????? const char * const names[]);
>>> +??? void (*del_vqs)(struct vhost_dev *vdev);
>>> +??? int (*write)(struct vhost_dev *vdev, u64 vhost_dst, void *src,
>>> int len);
>>> +??? int (*read)(struct vhost_dev *vdev, void *dst, u64 vhost_src,
int
>>> len);
>>> +??? int (*set_features)(struct vhost_dev *vdev, u64
device_features);
>>> +??? int (*set_status)(struct vhost_dev *vdev, u8 status);
>>> +??? u8 (*get_status)(struct vhost_dev *vdev);
>>> +};
>>> +
>>> struct virtio_config_ops
>>> I think there's some overlap here and some of the ops tries to
do the
>>> same thing.
>>>
>>> I think it differs in (*set_vq_address)() and (*create_vqs)().
>>> [create_vqs() introduced in struct vhost_config_ops provides
>>> complimentary functionality to (*find_vqs)() in struct
>>> virtio_config_ops. It seemingly encapsulates the functionality of
>>> (*set_vq_address)(), (*set_vq_num)(), (*set_vq_cb)(),..].
>>>
>>> Back to the difference between (*set_vq_address)() and
(*create_vqs)(),
>>> set_vq_address() directly provides the virtqueue address to the
vdpa
>>> device but create_vqs() only provides the parameters of the
virtqueue
>>> (like the number of virtqueues, number of buffers) but does not
directly
>>> provide the address. IMO the backend client drivers (like net or
vhost)
>>> shouldn't/cannot by itself know how to access the vring created
on
>>> virtio front-end. The vdpa device/vhost device should have logic
for
>>> that. That will help the client drivers to work with different
types of
>>> vdpa device/vhost device and can access the vring created by virtio
>>> irrespective of whether the vring can be accessed via mmio or
kernel
>>> space or user space.
>>>
>>> I think vdpa always works with client drivers in userspace and
providing
>>> userspace address for vring.
>>
>> Sorry for being unclear. What I meant is not replacing vDPA with the
>> vhost(bus) you proposed but the possibility of replacing virtio-pci-epf
>> with vDPA in:
> Okay, so the virtio back-end still use vhost and front end should use
> vDPA. I see. So the host side PCI driver for EPF should populate
> vdpa_config_ops and invoke vdpa_register_device().

Yes.

>> My question is basically for the part of virtio_pci_epf_send_command(),
>> so it looks to me you have a vendor specific API to replace the
>> virtio-pci layout of the BAR:
> Even when we use vDPA, we have to use some sort of
> virtio_pci_epf_send_command() to communicate with virtio backend right?

Right.

>
> Right, the layout is slightly different from the standard layout.
>
> This is the layout
> struct epf_vhost_reg_queue {
>          u8 cmd;
>          u8 cmd_status;
>          u16 status;
>          u16 num_buffers;
>          u16 msix_vector;
>          u64 queue_addr;

What's the meaning of queue_addr here?

Does not mean the device expects a contiguous memory for avail/desc/used 
ring?

> } __packed;
>
> struct epf_vhost_reg {
>          u64 host_features;
>          u64 guest_features;
>          u16 msix_config;
>          u16 num_queues;
>          u8 device_status;
>          u8 config_generation;
>          u32 isr;
>          u8 cmd;
>          u8 cmd_status;
>          struct epf_vhost_reg_queue vq[MAX_VQS];
> } __packed;
>>
>> +static int virtio_pci_epf_send_command(struct virtio_pci_device
*vp_dev,
>> +??? ??? ??? ??? ?????? u32 command)
>> +{
>> +??? struct virtio_pci_epf *pci_epf;
>> +??? void __iomem *ioaddr;
>> +??? ktime_t timeout;
>> +??? bool timedout;
>> +??? int ret = 0;
>> +??? u8 status;
>> +
>> +??? pci_epf = to_virtio_pci_epf(vp_dev);
>> +??? ioaddr = vp_dev->ioaddr;
>> +
>> +??? mutex_lock(&pci_epf->lock);
>> +??? writeb(command, ioaddr + HOST_CMD);
>> +??? timeout = ktime_add_ms(ktime_get(), COMMAND_TIMEOUT);
>> +??? while (1) {
>> +??? ??? timedout = ktime_after(ktime_get(), timeout);
>> +??? ??? status = readb(ioaddr + HOST_CMD_STATUS);
>> +
>>
>> Several questions:
>>
>> - It's not clear to me how the synchronization is done between the
RC
>> and EP. E.g how and when the value of HOST_CMD_STATUS can be changed.
> The HOST_CMD (commands sent to the EP) is serialized by using mutex.
> Once the EP reads the command, it resets the value in HOST_CMD. So
> HOST_CMD is less likely an issue.

Here's my understanding of the protocol:

1) RC write to HOST_CMD
2) RC wait for HOST_CMD_STATUS to be HOST_CMD_STATUS_OKAY

It looks to me what EP should do is

1) EP reset HOST_CMD after reading new command

And it looks to me EP should also reset HOST_CMD_STATUS here?

(I thought there should be patch to handle stuffs like this but I didn't 
find it in this series)

>
> A sufficiently large time is given for the EP to complete it's
operation
> (1 Sec) where the EP provides the status in HOST_CMD_STATUS. After it
> expires, HOST_CMD_STATUS_NONE is written to HOST_CMD_STATUS. There could
> be case where EP updates HOST_CMD_STATUS after RC writes
> HOST_CMD_STATUS_NONE, but by then HOST has already detected this as
> failure and error-ed out.
>   
>> If you still want to introduce a new transport, a virtio spec patch
>> would be helpful for us to understand the device API.
> Okay, that should be on https://github.com/oasis-tcs/virtio-spec.git?

Yes.

>> - You have you vendor specific layout (according to
>> virtio_pci_epb_table()), so I guess you it's better to have a
vendor
>> specific vDPA driver instead
> Okay, with vDPA, we are free to define our own layouts.

Right, but vDPA have other requirements. E.g it requires the device have 
the ability to save/restore the state (e.g the last_avail_idx).

So it actually depends on what you want. If you don't care about 
userspace drivers and want to have a standard transport, you can still 
go virtio.

>> - The advantage of vendor specific vDPA driver is that it can 1) have
>> less codes 2) support userspace drivers through vhost-vDPA (instead of
>> inventing new APIs since we can't use vfio-pci here).
> I see there's an additional level of indirection from virtio to vDPA
and
> probably no need for spec update but don't exactly see how it'll
reduce
> code.

AFAIK you don't need to implement your own setup_vq and del_vq.

>
> For 2, Isn't vhost-vdpa supposed to run on virtio backend?

Not currently, vDPA is a superset of virtio (e.g it support virtqueue 
state save/restore). This it should be possible in the future probably.

>
>  From a high level, I think I should be able to use vDPA for
> virtio_pci_epf.c. Would you also suggest using vDPA for ntb_virtio.c?
> ([RFC PATCH 20/22] NTB: Add a new NTB client driver to implement VIRTIO
> functionality).

I think it's your call. If you want

1) a well-defined standard virtio transport
2) willing to finalize d and maintain the spec
3) doesn't care about userspace drivers

You can go with virtio, otherwise vDPA.

Thanks

>
> Thanks
> Kishon
>

On 2020/9/16 ??7:47, Kishon Vijay Abraham I wrote:
> Hi Jason,
>
> On 16/09/20 8:40 am, Jason Wang wrote:
>> On 2020/9/15 ??11:47, Kishon Vijay Abraham I wrote:
>>> Hi Jason,
>>>
>>> On 15/09/20 1:48 pm, Jason Wang wrote:
>>>> Hi Kishon:
>>>>
>>>> On 2020/9/14 ??3:23, Kishon Vijay Abraham I wrote:
>>>>>> Then you need something that is functional equivalent
to virtio PCI
>>>>>> which is actually the concept of vDPA (e.g vDPA
provides
>>>>>> alternatives if
>>>>>> the queue_sel is hard in the EP implementation).
>>>>> Okay, I just tried to compare the 'struct
vdpa_config_ops' and 'struct
>>>>> vhost_config_ops' ( introduced in [RFC PATCH 03/22]
vhost: Add ops for
>>>>> the VHOST driver to configure VHOST device).
>>>>>
>>>>> struct vdpa_config_ops {
>>>>>  ?????/* Virtqueue ops */
>>>>>  ?????int (*set_vq_address)(struct vdpa_device *vdev,
>>>>>  ?????????????????? u16 idx, u64 desc_area, u64
driver_area,
>>>>>  ?????????????????? u64 device_area);
>>>>>  ?????void (*set_vq_num)(struct vdpa_device *vdev, u16 idx,
u32 num);
>>>>>  ?????void (*kick_vq)(struct vdpa_device *vdev, u16 idx);
>>>>>  ?????void (*set_vq_cb)(struct vdpa_device *vdev, u16 idx,
>>>>>  ?????????????? struct vdpa_callback *cb);
>>>>>  ?????void (*set_vq_ready)(struct vdpa_device *vdev, u16
idx, bool
>>>>> ready);
>>>>>  ?????bool (*get_vq_ready)(struct vdpa_device *vdev, u16
idx);
>>>>>  ?????int (*set_vq_state)(struct vdpa_device *vdev, u16
idx,
>>>>>  ???????????????? const struct vdpa_vq_state *state);
>>>>>  ?????int (*get_vq_state)(struct vdpa_device *vdev, u16
idx,
>>>>>  ???????????????? struct vdpa_vq_state *state);
>>>>>  ?????struct vdpa_notification_area
>>>>>  ?????(*get_vq_notification)(struct vdpa_device *vdev, u16
idx);
>>>>>  ?????/* vq irq is not expected to be changed once
DRIVER_OK is set */
>>>>>  ?????int (*get_vq_irq)(struct vdpa_device *vdv, u16 idx);
>>>>>
>>>>>  ?????/* Device ops */
>>>>>  ?????u32 (*get_vq_align)(struct vdpa_device *vdev);
>>>>>  ?????u64 (*get_features)(struct vdpa_device *vdev);
>>>>>  ?????int (*set_features)(struct vdpa_device *vdev, u64
features);
>>>>>  ?????void (*set_config_cb)(struct vdpa_device *vdev,
>>>>>  ?????????????????? struct vdpa_callback *cb);
>>>>>  ?????u16 (*get_vq_num_max)(struct vdpa_device *vdev);
>>>>>  ?????u32 (*get_device_id)(struct vdpa_device *vdev);
>>>>>  ?????u32 (*get_vendor_id)(struct vdpa_device *vdev);
>>>>>  ?????u8 (*get_status)(struct vdpa_device *vdev);
>>>>>  ?????void (*set_status)(struct vdpa_device *vdev, u8
status);
>>>>>  ?????void (*get_config)(struct vdpa_device *vdev, unsigned
int offset,
>>>>>  ??????????????? void *buf, unsigned int len);
>>>>>  ?????void (*set_config)(struct vdpa_device *vdev, unsigned
int offset,
>>>>>  ??????????????? const void *buf, unsigned int len);
>>>>>  ?????u32 (*get_generation)(struct vdpa_device *vdev);
>>>>>
>>>>>  ?????/* DMA ops */
>>>>>  ?????int (*set_map)(struct vdpa_device *vdev, struct
vhost_iotlb
>>>>> *iotlb);
>>>>>  ?????int (*dma_map)(struct vdpa_device *vdev, u64 iova,
u64 size,
>>>>>  ??????????????? u64 pa, u32 perm);
>>>>>  ?????int (*dma_unmap)(struct vdpa_device *vdev, u64 iova,
u64 size);
>>>>>
>>>>>  ?????/* Free device resources */
>>>>>  ?????void (*free)(struct vdpa_device *vdev);
>>>>> };
>>>>>
>>>>> +struct vhost_config_ops {
>>>>> +??? int (*create_vqs)(struct vhost_dev *vdev, unsigned int
nvqs,
>>>>> +????????????? unsigned int num_bufs, struct
vhost_virtqueue *vqs[],
>>>>> +????????????? vhost_vq_callback_t *callbacks[],
>>>>> +????????????? const char * const names[]);
>>>>> +??? void (*del_vqs)(struct vhost_dev *vdev);
>>>>> +??? int (*write)(struct vhost_dev *vdev, u64 vhost_dst,
void *src,
>>>>> int len);
>>>>> +??? int (*read)(struct vhost_dev *vdev, void *dst, u64
vhost_src, int
>>>>> len);
>>>>> +??? int (*set_features)(struct vhost_dev *vdev, u64
device_features);
>>>>> +??? int (*set_status)(struct vhost_dev *vdev, u8 status);
>>>>> +??? u8 (*get_status)(struct vhost_dev *vdev);
>>>>> +};
>>>>> +
>>>>> struct virtio_config_ops
>>>>> I think there's some overlap here and some of the ops
tries to do the
>>>>> same thing.
>>>>>
>>>>> I think it differs in (*set_vq_address)() and
(*create_vqs)().
>>>>> [create_vqs() introduced in struct vhost_config_ops
provides
>>>>> complimentary functionality to (*find_vqs)() in struct
>>>>> virtio_config_ops. It seemingly encapsulates the
functionality of
>>>>> (*set_vq_address)(), (*set_vq_num)(), (*set_vq_cb)(),..].
>>>>>
>>>>> Back to the difference between (*set_vq_address)() and
(*create_vqs)(),
>>>>> set_vq_address() directly provides the virtqueue address to
the vdpa
>>>>> device but create_vqs() only provides the parameters of the
virtqueue
>>>>> (like the number of virtqueues, number of buffers) but does
not
>>>>> directly
>>>>> provide the address. IMO the backend client drivers (like
net or vhost)
>>>>> shouldn't/cannot by itself know how to access the vring
created on
>>>>> virtio front-end. The vdpa device/vhost device should have
logic for
>>>>> that. That will help the client drivers to work with
different types of
>>>>> vdpa device/vhost device and can access the vring created
by virtio
>>>>> irrespective of whether the vring can be accessed via mmio
or kernel
>>>>> space or user space.
>>>>>
>>>>> I think vdpa always works with client drivers in userspace
and
>>>>> providing
>>>>> userspace address for vring.
>>>> Sorry for being unclear. What I meant is not replacing vDPA
with the
>>>> vhost(bus) you proposed but the possibility of replacing
virtio-pci-epf
>>>> with vDPA in:
>>> Okay, so the virtio back-end still use vhost and front end should
use
>>> vDPA. I see. So the host side PCI driver for EPF should populate
>>> vdpa_config_ops and invoke vdpa_register_device().
>>
>> Yes.
>>
>>
>>>> My question is basically for the part of
virtio_pci_epf_send_command(),
>>>> so it looks to me you have a vendor specific API to replace the
>>>> virtio-pci layout of the BAR:
>>> Even when we use vDPA, we have to use some sort of
>>> virtio_pci_epf_send_command() to communicate with virtio backend
right?
>>
>> Right.
>>
>>
>>> Right, the layout is slightly different from the standard layout.
>>>
>>> This is the layout
>>> struct epf_vhost_reg_queue {
>>>  ???????? u8 cmd;
>>>  ???????? u8 cmd_status;
>>>  ???????? u16 status;
>>>  ???????? u16 num_buffers;
>>>  ???????? u16 msix_vector;
>>>  ???????? u64 queue_addr;
>>
>> What's the meaning of queue_addr here?
> Using queue_addr, the virtio front-end communicates the address of the
> allocated memory for virtqueue to the virtio back-end.
>> Does not mean the device expects a contiguous memory for
avail/desc/used
>> ring?
> It's contiguous memory. Isn't this similar to other virtio
transport
> (both PCI legacy and modern interface)?.

That's only for legacy device, for modern device we don't have such 
restriction.

>>
>>> } __packed;
>>>
>>> struct epf_vhost_reg {
>>>  ???????? u64 host_features;
>>>  ???????? u64 guest_features;
>>>  ???????? u16 msix_config;
>>>  ???????? u16 num_queues;
>>>  ???????? u8 device_status;
>>>  ???????? u8 config_generation;
>>>  ???????? u32 isr;
>>>  ???????? u8 cmd;
>>>  ???????? u8 cmd_status;
>>>  ???????? struct epf_vhost_reg_queue vq[MAX_VQS];
>>> } __packed;
>>>> +static int virtio_pci_epf_send_command(struct
virtio_pci_device
>>>> *vp_dev,
>>>> +??? ??? ??? ??? ?????? u32 command)
>>>> +{
>>>> +??? struct virtio_pci_epf *pci_epf;
>>>> +??? void __iomem *ioaddr;
>>>> +??? ktime_t timeout;
>>>> +??? bool timedout;
>>>> +??? int ret = 0;
>>>> +??? u8 status;
>>>> +
>>>> +??? pci_epf = to_virtio_pci_epf(vp_dev);
>>>> +??? ioaddr = vp_dev->ioaddr;
>>>> +
>>>> +??? mutex_lock(&pci_epf->lock);
>>>> +??? writeb(command, ioaddr + HOST_CMD);
>>>> +??? timeout = ktime_add_ms(ktime_get(), COMMAND_TIMEOUT);
>>>> +??? while (1) {
>>>> +??? ??? timedout = ktime_after(ktime_get(), timeout);
>>>> +??? ??? status = readb(ioaddr + HOST_CMD_STATUS);
>>>> +
>>>>
>>>> Several questions:
>>>>
>>>> - It's not clear to me how the synchronization is done
between the RC
>>>> and EP. E.g how and when the value of HOST_CMD_STATUS can be
changed.
>>> The HOST_CMD (commands sent to the EP) is serialized by using
mutex.
>>> Once the EP reads the command, it resets the value in HOST_CMD. So
>>> HOST_CMD is less likely an issue.
>>
>> Here's my understanding of the protocol:
>>
>> 1) RC write to HOST_CMD
>> 2) RC wait for HOST_CMD_STATUS to be HOST_CMD_STATUS_OKAY
> That's right!
>> It looks to me what EP should do is
>>
>> 1) EP reset HOST_CMD after reading new command
> That's right! It does.
>> And it looks to me EP should also reset HOST_CMD_STATUS here?
> yeah, that would require RC to send another command to reset the status.
> Didn't see it required in the normal scenario but good to add this.
>> (I thought there should be patch to handle stuffs like this but I
didn't
>> find it in this series)
> This is added in [RFC PATCH 19/22] PCI: endpoint: Add EP function driver
> to provide VHOST interface
>
> pci_epf_vhost_cmd_handler() gets commands from RC using
"reg->cmd;". On
> the EP side, it is local memory access (mapped to BAR memory exposed to
> the host) and hence accessed using structure member access.

Thanks for the pointer, will have a look at and I think this part need 
to be carefully designed and the key to the success of the epf transport.