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;
MIME-Version: 1.0
References: <20220703183059.4133659-1-daniel.lezcano@linexp.org>
 <20220703183059.4133659-6-daniel.lezcano@linexp.org> <ad8dd950-9260-16b3-2134-72984e2ba261@arm.com>
 <e2d81210-a969-fe9d-a4fb-1826c991fbce@linexp.org> <f1cd08de-e8cb-18c2-9f92-01436810bc20@arm.com>
In-Reply-To: <f1cd08de-e8cb-18c2-9f92-01436810bc20@arm.com>
From:   "Rafael J. Wysocki" <rafael@kernel.org>
Date:   Tue, 5 Jul 2022 16:20:32 +0200
Message-ID: <CAJZ5v0iKcLM--L4ecoPDWpQ=8zgfeMxCku4zJt7hzVxSszgSNQ@mail.gmail.com>
Subject: Re: [PATCH v3 05/12] thermal/core: Remove unneeded EXPORT_SYMBOLS
To:     Lukasz Luba <lukasz.luba@arm.com>
Cc:     Daniel Lezcano <daniel.lezcano@linaro.org>,
        Linux PM <linux-pm@vger.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Kevin Hilman <khilman@baylibre.com>,
        Alexandre Bailon <abailon@baylibre.com>,
        Amit Kucheria <amitk@kernel.org>,
        Zhang Rui <rui.zhang@intel.com>, Todd Kjos <tkjos@google.com>,
        Wei Wang <wvw@google.com>,
        "Rafael J. Wysocki" <rafael@kernel.org>,
        Daniel Lezcano <daniel.lezcano@linexp.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Tue, Jul 5, 2022 at 9:30 AM Lukasz Luba <lukasz.luba@arm.com> wrote:
>
>
>
> On 7/4/22 22:14, Daniel Lezcano wrote:
> > On 04/07/2022 09:35, Lukasz Luba wrote:
> >> Hi Daniel,
> >>
> >> (+Todd and Wei on CC)
> >>
> >>
> >> On 7/3/22 19:30, Daniel Lezcano wrote:
> >
> > [ ... ]
> >
> >>>   }
> >>> -EXPORT_SYMBOL(get_tz_trend);
> >
> > [ ... ]
> >
> >>>   }
> >>> -EXPORT_SYMBOL(thermal_cdev_update);
> >>
> >> I wouldn't remove that export. I can see in my Pixel6 modules dir, that
> >> it's called in 7 places.
> >>
> >> I assume that in Android world this is common use.
> >
> > It is not possible to do changes taking into consideration out of tree
> > code. Moreover there is logically no good reason to use the
> > thermal_cdev_update() function from outside of the thermal core code.
> >
>
> I see your point which is 'upstream'. On the other hand the mostly
> deployed kernel is in Android devices and that brings a lot to the
> community.
>
> This symbol might also be used by other distros which might have
> modules for some accelerators, which also support tricky cooling.
>
> I would keep it as is...

I think that the long-term goal is to reduce differences between the
mainline kernel and Android.  From this angle, it would be good if
Android was aware that the mainline did stuff especially for them and
making them carry an extra patch would go a long way towards that
purpose.