Re: [PATCH 2/6] platform/x86/intel/tpmi: Don't create devices for disabled features

[srinivas pandruvada <srinivas.pandruvada@linux.intel.com>]

On Thu, 2023-11-30 at 10:00 -0500, srinivas pandruvada wrote:
> On Thu, 2023-11-30 at 16:38 +0200, Andy Shevchenko wrote:
> > On Thu, Nov 30, 2023 at 04:33:00PM +0200, Ilpo Järvinen wrote:
> > > On Thu, 30 Nov 2023, srinivas pandruvada wrote:
> > > > On Thu, 2023-11-30 at 14:26 +0200, Ilpo Järvinen wrote:
> > > > > On Tue, 28 Nov 2023, Srinivas Pandruvada wrote:
> > 
> > ...
> > 
> > > > > > +       if (!feature_state.enabled)
> > > > > > +               return -EOPNOTSUPP;
> > > > > 
> > > > > -ENODEV sounds more appropriate.  
> > > > 
> > > > The -EOPNOTSUPP is returned matching the next return statement,
> > > > which
> > > > causes to continue to create devices which are supported and
> > > > not
> > > > disabled. Any other error is real device creation will causes
> > > > driver
> > > > modprobe to fail.
> > > 
> > > Oh, I see... I didn't look that deep into the code during my
> > > review
> > > (perhaps note that down into the commit message?).
> > 
> > Maybe we should even use -ENOTSUPP (Linux internal error code), so
> > it will be clear that it's _not_ going to user space?
> 
> That will be better. I will change and resubmit.
The checkpatch gives error with this.


WARNING: ENOTSUPP is not a SUSV4 error code, prefer EOPNOTSUPP
#25: FILE: drivers/platform/x86/intel/tpmi.c:613:
+		return -ENOTSUPP;

Thanks,
Srinivas
> 
> Thanks,
> Srinivas
> 
> > 
>