Re: [PATCH v2 1/3] fpga: manager: change status api prototype, don't use older

[Xu Yilun]

On 2022-08-17 at 11:16:14 +0000, Manne, Nava kishore wrote:
> Hi Yilun,
> 
> 	Please find my response inline.
> 
> > -----Original Message-----
> > From: Xu Yilun <yilun.xu@xxxxxxxxx>
> > Sent: Tuesday, June 28, 2022 2:02 PM
> > To: Nava kishore Manne <nava.manne@xxxxxxxxxx>
> > Cc: michal.simek@xxxxxxxxxx; hao.wu@xxxxxxxxx; trix@xxxxxxxxxx;
> > mdf@xxxxxxxxxx; gregkh@xxxxxxxxxxxxxxxxxxx; ronak.jain@xxxxxxxxxx;
> > rajan.vaja@xxxxxxxxxx; abhyuday.godhasara@xxxxxxxxxx;
> > piyush.mehta@xxxxxxxxxx; harsha.harsha@xxxxxxxxxx;
> > lakshmi.sai.krishna.potthuri@xxxxxxxxxx; linux-arm-
> > kernel@xxxxxxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; linux-
> > fpga@xxxxxxxxxxxxxxx; git@xxxxxxxxxx
> > Subject: Re: [PATCH v2 1/3] fpga: manager: change status api prototype,
> > don't use older
> > 
> > On Tue, Jun 21, 2022 at 02:58:31PM +0530, Nava kishore Manne wrote:
> > > Different vendors have different error sets defined by Hardware.
> > > If we always define the new bits when we cannot find an exact 1:1
> > > mapping in the core the 64 bits would soon be used out. Also, it's
> > > hard to understand the mixture of different error sets.
> > >
> > > To address these issues updated the status interface to handle the
> > > vendor-specific messages in a generic way. With the updated status
> > > interface the vendor-specific driver files can independently handle
> > > the error messages.
> > 
> > I think we don't have to provide the vendor specific HW errors in a generic
> > way, maybe the vendor specific drivers could handle them by its own device
> > attributes.
> > 
> > Since the output value set of the interface is specific to each driver, users
> > should still interpret them in specific manners. So doesn't see much value for
> > a class interface.
> >
> 
> Agree,  vendor specific drivers could handle them by its own device attributes.
> If it is the case, can we remove the existing status interface relevant changes from the core?

We don't have to. Some *user* interfaces may become hard to use as time
goes, but we still try to make them stable as there are exsiting users
working on them once they are released.

That also means we should be more careful when introducing new user
interfaces.

Thanks,
Yilun

> 
> Regards,
> Navakishore