Re: [dpdk-dev] [PATCH v2] kernel: remove igb_uio

["谢华伟(此时此刻）" <huawei.xhw@alibaba-inc.com>]

On 2020/10/5 17:11, Maxime Coquelin wrote:
>
> On 10/5/20 10:57 AM, Thomas Monjalon wrote:
>> 24/09/2020 07:41, Stephen Hemminger:
>>> On Fri, 11 Sep 2020 17:54:48 +0200
>>> Thomas Monjalon <thomas@monjalon.net> wrote:
>>>
>>>> As decided in the Technical Board in November 2019,
>>>> the kernel module igb_uio is moved to the dpdk-kmods repository
>>>> in the /linux/igb_uio/ directory.
>>>>
>>>> Minutes of Technical Board meeting:
>>>> https://mails.dpdk.org/archives/dev/2019-November/151763.html
>>>>
>>>> Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
>>>> ---
>>>> v2: update few docs (including release notes)
>>> Good so far:
>>> Acked-by: Stephen Hemminger <stephen@networkplumber.org>
>>>
>>> You may want to address all the references to igb_uio in guides/nics
>>>
>>> ark.rst
>>> axgbe.rst
>>> bnx2x.rst
>>> bnxt.rst
>>> build_and_test.rst
>>> ena.rst
>>> enic.rst
>>> features.rst
>>> hns3.rst
>>> i40e.rst
>>> intel_vf.rst
>>> ixgbe.rst
>>> liquidio.rst
>>> mlx4.rst
>>> mlx5.rst
>>> nfp.rst
>>> qede.rst
>>> virtio.rst
>> igb_uio is still available.
>> A next step in deprecation might be to remove igb_uio references.
>>
>>> What about drivers like ark which don't mention vfio?
>> They should be updated by their maintainer.
>>
>>> Does virtio still require igb_uio? or x86 I/O port for doorbell?
>>> Or is this just stale language.
>> Maxime, any update on the use of igb_uio with virtio?
> For sure Virtio don't require igb_uio, I always use vfio myself.
> It seems the doc needs an update, I'll try to look at it later in this
> release.
>
> Regards,
> Maxime

PIO/MMIO write(notify backend) needs to go through vfio ioctl call,

which impacts performance.  I fix this in another patch. PIO/MMIO

read/write will be executed directly in user space whatever driver is being

used.


/huawei