Re: [RFC 3/4] wwan: add interface creation support

From: Johannes Berg <johannes@sipsolutions.net>
To: Loic Poulain <loic.poulain@linaro.org>
Cc: linux-wireless@vger.kernel.org,
	Network Development <netdev@vger.kernel.org>,
	M Chetan Kumar <m.chetan.kumar@intel.com>
Subject: Re: [RFC 3/4] wwan: add interface creation support
Date: Tue, 01 Jun 2021 12:35:09 +0200	[thread overview]
Message-ID: <0555025c6d7a88f4f3dcdd6704612ed8ba33b175.camel@sipsolutions.net> (raw)
In-Reply-To: <CAMZdPi-ZaH8WWKfhfKzy0OKpUtNAiCUfekh9R1de5awFP-ed=A@mail.gmail.com> (sfid-20210601_112832_203691_AAB7DA4E)

Hi,

> > +int wwan_register_ops(struct device *parent, const struct wwan_ops *ops,
> > +                     void *ctxt)
> > +{
> > +       struct wwan_dev_reg *reg;
> > +       int ret;
> > +
> > +       if (WARN_ON(!parent || !ops))
> > +               return -EINVAL;
> > +
> > +       mutex_lock(&wwan_mtx);
> > +       list_for_each_entry(reg, &wwan_devs, list) {
> > +               if (WARN_ON(reg->dev == parent)) {
> > +                       ret = -EBUSY;
> > +                       goto out;
> > +               }
> > +       }
> 
> Thanks for this, overall it looks good to me, but just checking why
> you're not using the wwan_dev internally to create-or-pick wwan_dev
> (wwan_dev_create) and register ops to it, instead of having a global
> new wwan_devs list.

Uh, no good reason. I just missed that all that infrastructure is
already there, oops.

johannes