Re: [PATCH v5 03/28] clk: Introduce clk_get_rate_range()

From: Stephen Boyd <sboyd@kernel.org>
To: Maxime Ripard <maxime@cerno.tech>,
	Mike Turquette <mturquette@baylibre.com>,
	linux-clk@vger.kernel.org
Cc: Yassine Oudjana <y.oudjana@protonmail.com>,
	Naresh Kamboju <naresh.kamboju@linaro.org>,
	Dmitry Baryshkov <dmitry.baryshkov@linaro.org>,
	Tony Lindgren <tony@atomide.com>,
	Neil Armstrong <narmstrong@baylibre.com>,
	Alexander Stein <alexander.stein@ew.tq-group.com>,
	Marek Szyprowski <m.szyprowski@samsung.com>,
	Jerome Brunet <jbrunet@baylibre.com>,
	Maxime Ripard <maxime@cerno.tech>
Subject: Re: [PATCH v5 03/28] clk: Introduce clk_get_rate_range()
Date: Mon, 27 Jun 2022 16:23:59 -0700	[thread overview]
Message-ID: <20220627232401.81710C34115@smtp.kernel.org> (raw)
In-Reply-To: <20220516132527.328190-4-maxime@cerno.tech>

Quoting Maxime Ripard (2022-05-16 06:25:02)
> With the recent introduction of clock drivers that will force their
> clock rate to either the minimum or maximum boundaries, it becomes
> harder for clock users to discover either boundary of their clock.
> 
> Indeed, the best way to do that previously was to call clk_round_rate()
> on either 0 or ULONG_MAX and count on the driver to clamp the rate to
> the current boundary, but that won't work anymore.
> 
> Since any other alternative (calling clk_set_rate_range() and looking at
> the returned value, calling clk_round_rate() still, or just doing
> nothing) depends on how the driver will behaves, we actually are
> punching a hole through the abstraction provided by the clock framework.
> 
> In order to avoid any abstraction violation, let's create a bunch of
> accessors that will return the current minimum and maximum for a given
> clock.

Why does a clk consumer need to know the rate range? I don't see any
information in the commit text about that.