linux-clk.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Emil Renner Berthing <kernel@esmil.dk>
To: linux-clk@vger.kernel.org, devicetree@vger.kernel.org
Cc: Emil Renner Berthing <kernel@esmil.dk>,
	Michael Turquette <mturquette@baylibre.com>,
	Stephen Boyd <sboyd@kernel.org>, Rob Herring <robh+dt@kernel.org>,
	Andy Shevchenko <andy.shevchenko@gmail.com>,
	Geert Uytterhoeven <geert@linux-m68k.org>,
	Arnd Bergmann <arnd@arndb.de>,
	Michael Zhu <michael.zhu@starfivetech.com>,
	Fu Wei <tekkamanninja@gmail.com>,
	linux-kernel@vger.kernel.org
Subject: [PATCH v1 1/7] clk: starfive: jh7100: Don't round divisor up twice
Date: Wed, 26 Jan 2022 18:39:47 +0100	[thread overview]
Message-ID: <20220126173953.1016706-2-kernel@esmil.dk> (raw)
In-Reply-To: <20220126173953.1016706-1-kernel@esmil.dk>

The problem is best illustrated by an example. Suppose a consumer wants
a 4MHz clock rate from a divider with a 10MHz parent. It would then
call

  clk_round_rate(clk, 4000000)

which would call into our determine_rate() callback that correctly
rounds up and finds that a divisor of 3 gives the highest possible
frequency below the requested 4MHz and returns 10000000 / 3 = 3333333Hz.

However the consumer would then call

  clk_set_rate(clk, 3333333)

but since 3333333 doesn't divide 10000000 evenly our set_rate() callback
would again round the divisor up and set it to 4 which results in an
unnecessarily low rate of 2.5MHz.

Fix it by using DIV_ROUND_CLOSEST in the set_rate() callback.

Fixes: 4210be668a09 ("clk: starfive: Add JH7100 clock generator driver")
Signed-off-by: Emil Renner Berthing <kernel@esmil.dk>
---
 drivers/clk/starfive/clk-starfive-jh7100.c | 14 +++-----------
 1 file changed, 3 insertions(+), 11 deletions(-)

diff --git a/drivers/clk/starfive/clk-starfive-jh7100.c b/drivers/clk/starfive/clk-starfive-jh7100.c
index 25d31afa0f87..db6a4dc203af 100644
--- a/drivers/clk/starfive/clk-starfive-jh7100.c
+++ b/drivers/clk/starfive/clk-starfive-jh7100.c
@@ -399,22 +399,13 @@ static unsigned long jh7100_clk_recalc_rate(struct clk_hw *hw,
 	return div ? parent_rate / div : 0;
 }
 
-static unsigned long jh7100_clk_bestdiv(struct jh7100_clk *clk,
-					unsigned long rate, unsigned long parent)
-{
-	unsigned long max = clk->max_div;
-	unsigned long div = DIV_ROUND_UP(parent, rate);
-
-	return min(div, max);
-}
-
 static int jh7100_clk_determine_rate(struct clk_hw *hw,
 				     struct clk_rate_request *req)
 {
 	struct jh7100_clk *clk = jh7100_clk_from(hw);
 	unsigned long parent = req->best_parent_rate;
 	unsigned long rate = clamp(req->rate, req->min_rate, req->max_rate);
-	unsigned long div = jh7100_clk_bestdiv(clk, rate, parent);
+	unsigned long div = min_t(unsigned long, DIV_ROUND_UP(parent, rate), clk->max_div);
 	unsigned long result = parent / div;
 
 	/*
@@ -442,7 +433,8 @@ static int jh7100_clk_set_rate(struct clk_hw *hw,
 			       unsigned long parent_rate)
 {
 	struct jh7100_clk *clk = jh7100_clk_from(hw);
-	unsigned long div = jh7100_clk_bestdiv(clk, rate, parent_rate);
+	unsigned long div = clamp(DIV_ROUND_CLOSEST(parent_rate, rate),
+				  1UL, (unsigned long)clk->max_div);
 
 	jh7100_clk_reg_rmw(clk, JH7100_CLK_DIV_MASK, div);
 	return 0;
-- 
2.34.1


  reply	other threads:[~2022-01-26 17:40 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-01-26 17:39 [PATCH v1 0/7] JH7100 Audio Clocks Emil Renner Berthing
2022-01-26 17:39 ` Emil Renner Berthing [this message]
2022-03-11  2:53   ` [PATCH v1 1/7] clk: starfive: jh7100: Don't round divisor up twice Stephen Boyd
2022-01-26 17:39 ` [PATCH v1 2/7] clk: starfive: jh7100: Handle audio_div clock properly Emil Renner Berthing
2022-03-11  2:53   ` Stephen Boyd
2022-01-26 17:39 ` [PATCH v1 3/7] dt-bindings: clock: Add JH7100 audio clock definitions Emil Renner Berthing
2022-02-09  3:32   ` Rob Herring
2022-03-11  2:53   ` Stephen Boyd
2022-01-26 17:39 ` [PATCH v1 4/7] dt-bindings: clock: Add starfive,jh7100-audclk bindings Emil Renner Berthing
2022-02-09  3:33   ` Rob Herring
2022-03-11  2:54   ` Stephen Boyd
2022-01-26 17:39 ` [PATCH v1 5/7] clk: starfive: jh7100: Make hw clock implementation reusable Emil Renner Berthing
2022-03-11  2:54   ` Stephen Boyd
2022-01-26 17:39 ` [PATCH v1 6/7] clk: starfive: jh7100: Support more clock types Emil Renner Berthing
2022-03-11  2:54   ` Stephen Boyd
2022-01-26 17:39 ` [PATCH v1 7/7] clk: starfive: Add JH7100 audio clock driver Emil Renner Berthing
2022-03-11  2:54   ` Stephen Boyd

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20220126173953.1016706-2-kernel@esmil.dk \
    --to=kernel@esmil.dk \
    --cc=andy.shevchenko@gmail.com \
    --cc=arnd@arndb.de \
    --cc=devicetree@vger.kernel.org \
    --cc=geert@linux-m68k.org \
    --cc=linux-clk@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=michael.zhu@starfivetech.com \
    --cc=mturquette@baylibre.com \
    --cc=robh+dt@kernel.org \
    --cc=sboyd@kernel.org \
    --cc=tekkamanninja@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).