Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Fri, 8 Apr 2022 17:53:59 +0300
From:   Andy Shevchenko <andriy.shevchenko@linux.intel.com>
To:     Sakari Ailus <sakari.ailus@linux.intel.com>
Cc:     Michael Walle <michael@walle.cc>, linux-acpi@vger.kernel.org,
        linux-kernel@vger.kernel.org, Daniel Scally <djrscally@gmail.com>,
        Heikki Krogerus <heikki.krogerus@linux.intel.com>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        "Rafael J. Wysocki" <rafael@kernel.org>,
        Len Brown <lenb@kernel.org>,
        Nuno =?iso-8859-1?Q?S=E1?= <nuno.sa@analog.com>
Subject: Re: [PATCH v5 1/4] device property: Allow error pointer to be passed
 to fwnode APIs
Message-ID: <YlBMh3A1UlsoDu1Q@smile.fi.intel.com>
References: <20220406130552.30930-1-andriy.shevchenko@linux.intel.com>
 <df3a78036864716fbeecf3cd94dbcbbe@walle.cc>
 <Yk66wHWlMg3QLy6u@smile.fi.intel.com>
 <YlAuEzW0fUuuXTN6@smile.fi.intel.com>
 <YlA4Pp11ZXG3eSX/@paasikivi.fi.intel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <YlA4Pp11ZXG3eSX/@paasikivi.fi.intel.com>
Organization: Intel Finland Oy - BIC 0357606-4 - Westendinkatu 7, 02160 Espoo
Precedence: bulk

On Fri, Apr 08, 2022 at 04:27:26PM +0300, Sakari Ailus wrote:
> On Fri, Apr 08, 2022 at 03:44:03PM +0300, Andy Shevchenko wrote:
> > On Thu, Apr 07, 2022 at 01:19:44PM +0300, Andy Shevchenko wrote:
> > > On Wed, Apr 06, 2022 at 08:05:23PM +0200, Michael Walle wrote:

...

> > > > > +	if (IS_ERR_OR_NULL(fwnode))
> > > > > +		return -ENOENT;
> > > > > +
> > > > >  	ret = fwnode_call_int_op(fwnode, get_reference_args, prop, nargs_prop,
> > > > >  				 nargs, index, args);
> > > > > +	if (ret == 0)
> > > > > +		return ret;
> > > > > 
> > > > > -	if (ret < 0 && !IS_ERR_OR_NULL(fwnode) &&
> > > > > -	    !IS_ERR_OR_NULL(fwnode->secondary))
> > > > > -		ret = fwnode_call_int_op(fwnode->secondary, get_reference_args,
> > > > > -					 prop, nargs_prop, nargs, index, args);
> > > > > +	if (IS_ERR_OR_NULL(fwnode->secondary))
> > > > > +		return -ENOENT;
> > > > 
> > > > Doesn't this mean you overwrite any return code != 0 with -ENOENT?
> > > > Is this intended?
> > > 
> > > Indeed, it would shadow the error code.
> > 
> > I was thinking more on this and am not sure about the best approach here.
> > On one hand in the original code this returns the actual error code from
> > the call against primary fwnode. But it can be at least -ENOENT or -EINVAL.
> > 
> > But when we check the secondary fwnode we want to have understanding that it's
> > secondary fwnode which has not been found, but this requires to have a good
> > distinguishing between error codes from the callback.
> > 
> > That said, the error codes convention of ->get_reference_args() simply
> > sucks. Sakari, do you have it on your TODO to fix this mess out, if it's
> > even feasible?
> 
> What would you expect to see compared to what it is now?
> 
> I guess the error code could be different for a missing property and a
> property with invalid data,

Yes, something like this.

> but I'm not sure any caller would be interested
> in that.

Yes, but it would be good for the consistency and working with fwnodes in
general. Esp. if we move at some point from primary-secondary to a full
linked list of fwnodes.

> > To be on safest side, I will change as suggested in previous mail (see below)
> > so it won't have impact on -EINVAL case.
> > 
> > > So, it should go with
> > > 
> > > 	if (IS_ERR_OR_NULL(fwnode->secondary))
> > > 		return ret;
> > > 
> > > then.

-- 
With Best Regards,
Andy Shevchenko