[dpdk-dev] [PATCH v2] doc: add known uio_pci_generic issue for i40e

Ferruh Yigit ferruh.yigit at intel.com
Wed Jan 18 19:49:34 CET 2017

On 1/18/2017 2:48 AM, Guo, Jia wrote:
> On 1/18/2017 1:36 AM, Ferruh Yigit wrote:
>> On 1/17/2017 5:00 PM, Wu, Jingjing wrote:
>>>> -----Original Message-----
>>>> From: Yigit, Ferruh
>>>> Sent: Tuesday, January 17, 2017 4:31 AM
>>>> To: Guo, Jia <jia.guo at intel.com>; Zhang, Helin <helin.zhang at intel.com>; Wu,
>>>> Jingjing <jingjing.wu at intel.com>
>>>> Cc: dev at dpdk.org
>>>> Subject: Re: [dpdk-dev] [PATCH v2] doc: add known uio_pci_generic issue
>>>> for i40e
>>>> On 2/12/2016 4:39 PM, Jeff Guo wrote:
>>>>> From: "Guo, Jia" <jia.guo at intel.com>
>>>>> When bind with uio_pci_generic in X710/XL710/XXV710, the result is
>>>>> failed. uio_pci_generic is not supported by X710/XL710/XXV710.
>>>> Hi Jeff,
>>>> There is a Linux commit causing the issue, right. Commit done because of the
>>>> reasons mentioned in "Description". Do you think does it make sense to note
>>>> that Linux commit, and Linux kernel version that has been introduced?
>>> Hi, Ferruh
>>> At beginning, we found this issue in latest kernel version, such as kernel 4.7.9 on fedora24.
>>> But later, we observed that this issue also exist on REL7.3.
>> It won't work custom kernels, but upstream kernel version. If you think
>> that will be confusion, that is OK.
>> What do you think just adding Linux kernel commit information, can it be
>> useful?
> yes, the root cause of the issue is the HW' errata, but not the linux 
> kernel commit,  because even without the linux kernel commit and 
> uio_pci_generic can even be bind and use,  but still will cause other 
> interrupt status issue. fix this interrupt status issue that is the 
> purpose of the linux kernel commit.
> so i think add linux kernel commit information in dpdk can be somehow to 
> useful for track issue, 

That was my reasoning for asking this. It may help others to understand
the issue more.

But is seems both you and Jingjing think this cause more confusion than
benefit. So I am OK with your proposal, if nobody else have objection.

> but if let the kernel issue be manage by kernel 
> org must be better.
> ferruh and jingjing, Do you all agree with me. beg on your feedback.
>>> So, it may be a little difficult to us to figure out exactly which kernel version
>>> doesn't support. And the root cause is the HW's errata which makes uio_pci_generic
>>> cannot be used on XL710 series NICs.
>>> Thanks
>>> Jingjing

More information about the dev mailing list