[PATCH 00/13] clk: add new flag CLK_ROUNDING_FW_MANAGED

[PATCH 00/13] clk: add new flag CLK_ROUNDING_FW_MANAGED

[Brian Masney]

There are some clocks where the rounding is managed by the hardware, and
the determine_rate() clk ops is just a noop that simply returns 0. Based
on discussions with Stephen at Linux Plumbers Conference, he suggested
adding a flag for this particular case. So let's add a new flag, and
update the clk core so that the determine_rate() clk op is not required
when this flag is set.
    
This series adds the flag, some kunit tests, and updates all of the
relevant drivers under drivers/clk to use the new flag.
    
Once this is merged, and in Linus's tree, I can update the few remaining
clk drivers that are outside of drivers/clk via those subsystems at a
later time.

Merge Strategy
--------------
All of this needs to be directly merged by Stephen as one series into
his tree. Subsystem maintainers: please leave a Reviewed-by or Acked-by.
To reduce the noise, I am only CCing people on their respective drivers.

Note this series depends on 3 previously-posted patches in this git pull
to Stephen for v7.1.
https://lore.kernel.org/linux-clk/aZuK4-QJCXUeSxtL@redhat.com/
Hopefully I set the depeendencies up correctly in b4.

Signed-off-by: Brian Masney <bmasney@redhat.com>
---
Brian Masney (13):
      clk: add new flag CLK_ROUNDING_FW_MANAGED
      clk: test: add test suite for CLK_ROUNDING_FW_MANAGED flag
      clk: rp1: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag
      clk: scpi: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag
      clk: hisilicon: hi3660-stub: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag
      clk: imx: scu: drop redundant init.ops variable assignment
      clk: imx: scu: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag
      clk: qcom: rpm: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag
      clk: qcom: rpmh: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag
      clk: qcom: smd-rpm: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag
      clk: renesas: rzg2l-cpg: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag
      clk: samsung: acpm: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag
      clk: sprd: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag

 drivers/clk/clk-rp1.c                   | 11 +----
 drivers/clk/clk-scpi.c                  | 14 +-----
 drivers/clk/clk.c                       | 24 ++++++++--
 drivers/clk/clk_test.c                  | 85 +++++++++++++++++++++++++++++++++
 drivers/clk/hisilicon/clk-hi3660-stub.c | 14 +-----
 drivers/clk/imx/clk-scu.c               | 23 +--------
 drivers/clk/qcom/clk-rpm.c              | 16 ++-----
 drivers/clk/qcom/clk-rpmh.c             |  8 +---
 drivers/clk/qcom/clk-smd-rpm.c          | 15 +-----
 drivers/clk/renesas/rzg2l-cpg.c         |  9 +---
 drivers/clk/samsung/clk-acpm.c          | 14 +-----
 drivers/clk/sprd/pll.c                  |  7 ---
 drivers/clk/sprd/pll.h                  |  2 +-
 include/linux/clk-provider.h            |  2 +
 14 files changed, 123 insertions(+), 121 deletions(-)
---
base-commit: 7d6661873f6b54c75195780a40d66bad3d482d8f
change-id: 20260226-clk-det-rate-fw-managed-4b8d061f31be
prerequisite-patch-id: 59198edc95aca82a29327137ad2af82ec13295b6
prerequisite-patch-id: 8932e170649711d7a80c57784033a37faadd519b
prerequisite-patch-id: 91c7b1851c5d77e504c49ce6bf14b3f8b84e826a

Best regards,
-- 
Brian Masney <bmasney@redhat.com>

[PATCH 01/13] clk: add new flag CLK_ROUNDING_FW_MANAGED

[Brian Masney]

There are some clocks where the rounding is managed by the hardware, and
the determine_rate() clk ops is just a noop that simply returns 0. Add a
new flag for these type of clocks, and update the clk core so that the
determine_rate() clk op is not required when this flag is set.

Signed-off-by: Brian Masney <bmasney@redhat.com>

---
To: Michael Turquette <mturquette@baylibre.com>
To: Stephen Boyd <sboyd@kernel.org>
To: Abel Vesa <abelvesa@kernel.org>
To: Andrea della Porta <andrea.porta@suse.com>
To: Baolin Wang <baolin.wang@linux.alibaba.com>
To: Bjorn Andersson <andersson@kernel.org>
To: Chanwoo Choi <cw00.choi@samsung.com>
To: Frank Li <Frank.Li@nxp.com>
To: Geert Uytterhoeven <geert+renesas@glider.be>
To: Krzysztof Kozlowski <krzk@kernel.org>
To: Orson Zhai <orsonzhai@gmail.com>
To: Sascha Hauer <s.hauer@pengutronix.de>
To: Sudeep Holla <sudeep.holla@kernel.org>
To: Sylwester Nawrocki <s.nawrocki@samsung.com>
To: Tudor Ambarus <tudor.ambarus@linaro.org>
Cc: linux-clk@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: Alim Akhtar <alim.akhtar@samsung.com>
Cc: arm-scmi@vger.kernel.org
Cc: Chunyan Zhang <zhang.lyra@gmail.com>
Cc: Cristian Marussi <cristian.marussi@arm.com>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: imx@lists.linux.dev
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-arm-msm@vger.kernel.org
Cc: linux-renesas-soc@vger.kernel.org
Cc: linux-samsung-soc@vger.kernel.org
Cc: Peng Fan <peng.fan@nxp.com>
Cc: Pengutronix Kernel Team <kernel@pengutronix.de>
---
 drivers/clk/clk.c            | 24 +++++++++++++++++++++---
 include/linux/clk-provider.h |  2 ++
 2 files changed, 23 insertions(+), 3 deletions(-)

diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c
index fd418dc988b1c60c49e3ac9c0c44aa132dd5da28..0a522a0817411c7f7c6e9cffd6f024e672a331a8 100644
--- a/drivers/clk/clk.c
+++ b/drivers/clk/clk.c
@@ -1557,6 +1557,20 @@ static int __init clk_disable_unused(void)
 }
 late_initcall_sync(clk_disable_unused);
 
+/**
+ * clk_is_rounding_fw_managed - Check to see if clk rounding is handled by the
+ * firmware.
+ * @core: the clk to check
+ *
+ * Clks that have this flag enabled do not need to have a determine_rate() op
+ * set, and will always return success for any rounding operation since the
+ * firmware will deal with the rounding.
+ */
+static inline bool clk_is_rounding_fw_managed(struct clk_core *core)
+{
+	return core->flags & CLK_ROUNDING_FW_MANAGED;
+}
+
 static int clk_core_determine_round_nolock(struct clk_core *core,
 					   struct clk_rate_request *req)
 {
@@ -1589,6 +1603,8 @@ static int clk_core_determine_round_nolock(struct clk_core *core,
 		req->rate = core->rate;
 	} else if (core->ops->determine_rate) {
 		return core->ops->determine_rate(core->hw, req);
+	} else if (clk_is_rounding_fw_managed(core)) {
+		return 0;
 	} else {
 		return -EINVAL;
 	}
@@ -1673,7 +1689,7 @@ EXPORT_SYMBOL_GPL(clk_hw_forward_rate_request);
 
 static bool clk_core_can_round(struct clk_core * const core)
 {
-	return core->ops->determine_rate;
+	return core->ops->determine_rate || clk_is_rounding_fw_managed(core);
 }
 
 static int clk_core_round_rate_nolock(struct clk_core *core,
@@ -3528,6 +3544,7 @@ static const struct {
 	ENTRY(CLK_IS_CRITICAL),
 	ENTRY(CLK_OPS_PARENT_ENABLE),
 	ENTRY(CLK_DUTY_CYCLE_PARENT),
+	ENTRY(CLK_ROUNDING_FW_MANAGED),
 #undef ENTRY
 };
 
@@ -3906,7 +3923,7 @@ static int __clk_core_init(struct clk_core *core)
 
 	/* check that clk_ops are sane.  See Documentation/driver-api/clk.rst */
 	if (core->ops->set_rate && !core->ops->determine_rate &&
-	      core->ops->recalc_rate) {
+	      core->ops->recalc_rate && !clk_is_rounding_fw_managed(core)) {
 		pr_err("%s: %s must implement .determine_rate in addition to .recalc_rate\n",
 		       __func__, core->name);
 		ret = -EINVAL;
@@ -3920,7 +3937,8 @@ static int __clk_core_init(struct clk_core *core)
 		goto out;
 	}
 
-	if (core->ops->set_parent && !core->ops->determine_rate) {
+	if (core->ops->set_parent && !core->ops->determine_rate &&
+	    !clk_is_rounding_fw_managed(core)) {
 		pr_err("%s: %s must implement .set_parent & .determine_rate\n",
 			__func__, core->name);
 		ret = -EINVAL;
diff --git a/include/linux/clk-provider.h b/include/linux/clk-provider.h
index 1cda2c78dffaff037f0f16b0f11106b63b3a746f..187f8248a9c840c701cbbba99bb7cdeef7b654ee 100644
--- a/include/linux/clk-provider.h
+++ b/include/linux/clk-provider.h
@@ -32,6 +32,8 @@
 #define CLK_OPS_PARENT_ENABLE	BIT(12)
 /* duty cycle call may be forwarded to the parent clock */
 #define CLK_DUTY_CYCLE_PARENT	BIT(13)
+/* clock rate rounding is managed by firmware, don't require determine_rate */
+#define CLK_ROUNDING_FW_MANAGED		BIT(14)
 
 struct clk;
 struct clk_hw;

-- 
2.53.0

[PATCH 08/13] clk: qcom: rpm: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag

[Brian Masney]

This clk driver has a noop determine_rate clk op. Drop this empty
function, and enable the CLK_ROUNDING_FW_MANAGED flag.

Signed-off-by: Brian Masney <bmasney@redhat.com>

---
To: Bjorn Andersson <andersson@kernel.org>
To: Michael Turquette <mturquette@baylibre.com>
To: Stephen Boyd <sboyd@kernel.org>
Cc: linux-arm-msm@vger.kernel.org
Cc: linux-clk@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
---
 drivers/clk/qcom/clk-rpm.c | 16 +++-------------
 1 file changed, 3 insertions(+), 13 deletions(-)

diff --git a/drivers/clk/qcom/clk-rpm.c b/drivers/clk/qcom/clk-rpm.c
index be0145631197bea65438f3bed10344f18d6de802..7875cd1815f524572f630242e3b71ff0810cdeda 100644
--- a/drivers/clk/qcom/clk-rpm.c
+++ b/drivers/clk/qcom/clk-rpm.c
@@ -42,6 +42,7 @@ static const struct clk_parent_data gcc_cxo[] = {
 			.name = #_name "_clk",				      \
 			.parent_data = gcc_pxo,				      \
 			.num_parents = ARRAY_SIZE(gcc_pxo),		      \
+			.flags = CLK_ROUNDING_FW_MANAGED,		      \
 		},							      \
 	};								      \
 	static struct clk_rpm clk_rpm_##_name##_a_clk = {		      \
@@ -54,6 +55,7 @@ static const struct clk_parent_data gcc_cxo[] = {
 			.name = #_name "_a_clk",			      \
 			.parent_data = gcc_pxo,				      \
 			.num_parents = ARRAY_SIZE(gcc_pxo),		      \
+			.flags = CLK_ROUNDING_FW_MANAGED,		      \
 		},							      \
 	}
 
@@ -78,6 +80,7 @@ static const struct clk_parent_data gcc_cxo[] = {
 			.name = #_name "_clk",				      \
 			.parent_data = gcc_pxo,				      \
 			.num_parents = ARRAY_SIZE(gcc_pxo),		      \
+			.flags = CLK_ROUNDING_FW_MANAGED,		      \
 		},							      \
 	}
 
@@ -351,17 +354,6 @@ static int clk_rpm_set_rate(struct clk_hw *hw,
 	return 0;
 }
 
-static int clk_rpm_determine_rate(struct clk_hw *hw,
-				  struct clk_rate_request *req)
-{
-	/*
-	 * RPM handles rate rounding and we don't have a way to
-	 * know what the rate will be, so just return whatever
-	 * rate is requested.
-	 */
-	return 0;
-}
-
 static unsigned long clk_rpm_recalc_rate(struct clk_hw *hw,
 					 unsigned long parent_rate)
 {
@@ -383,7 +375,6 @@ static const struct clk_ops clk_rpm_xo_ops = {
 static const struct clk_ops clk_rpm_fixed_ops = {
 	.prepare	= clk_rpm_fixed_prepare,
 	.unprepare	= clk_rpm_fixed_unprepare,
-	.determine_rate = clk_rpm_determine_rate,
 	.recalc_rate	= clk_rpm_recalc_rate,
 };
 
@@ -391,7 +382,6 @@ static const struct clk_ops clk_rpm_ops = {
 	.prepare	= clk_rpm_prepare,
 	.unprepare	= clk_rpm_unprepare,
 	.set_rate	= clk_rpm_set_rate,
-	.determine_rate = clk_rpm_determine_rate,
 	.recalc_rate	= clk_rpm_recalc_rate,
 };
 

-- 
2.53.0

[PATCH 09/13] clk: qcom: rpmh: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag

[Brian Masney]

This clk driver has a noop determine_rate clk op. Drop this empty
function, and enable the CLK_ROUNDING_FW_MANAGED flag.

Signed-off-by: Brian Masney <bmasney@redhat.com>

---
To: Bjorn Andersson <andersson@kernel.org>
To: Michael Turquette <mturquette@baylibre.com>
To: Stephen Boyd <sboyd@kernel.org>
Cc: linux-arm-msm@vger.kernel.org
Cc: linux-clk@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
---
 drivers/clk/qcom/clk-rpmh.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)

diff --git a/drivers/clk/qcom/clk-rpmh.c b/drivers/clk/qcom/clk-rpmh.c
index 547729b1a8ee01cf28c11ee8c4bd2f36d7536e6d..18fc94e6b98713e6aaf20a6d6144234f435d07a4 100644
--- a/drivers/clk/qcom/clk-rpmh.c
+++ b/drivers/clk/qcom/clk-rpmh.c
@@ -129,6 +129,7 @@ static DEFINE_MUTEX(rpmh_clk_lock);
 		.hw.init = &(struct clk_init_data){			\
 			.ops = &clk_rpmh_bcm_ops,			\
 			.name = #_name,					\
+			.flags = CLK_ROUNDING_FW_MANAGED,		\
 		},							\
 	}
 
@@ -321,12 +322,6 @@ static int clk_rpmh_bcm_set_rate(struct clk_hw *hw, unsigned long rate,
 	return 0;
 }
 
-static int clk_rpmh_determine_rate(struct clk_hw *hw,
-				   struct clk_rate_request *req)
-{
-	return 0;
-}
-
 static unsigned long clk_rpmh_bcm_recalc_rate(struct clk_hw *hw,
 					unsigned long prate)
 {
@@ -339,7 +334,6 @@ static const struct clk_ops clk_rpmh_bcm_ops = {
 	.prepare	= clk_rpmh_bcm_prepare,
 	.unprepare	= clk_rpmh_bcm_unprepare,
 	.set_rate	= clk_rpmh_bcm_set_rate,
-	.determine_rate = clk_rpmh_determine_rate,
 	.recalc_rate	= clk_rpmh_bcm_recalc_rate,
 };
 

-- 
2.53.0

[PATCH 10/13] clk: qcom: smd-rpm: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag

[Brian Masney]

This clk driver has a noop determine_rate clk op. Drop this empty
function, and enable the CLK_ROUNDING_FW_MANAGED flag.

Signed-off-by: Brian Masney <bmasney@redhat.com>

---
To: Bjorn Andersson <andersson@kernel.org>
To: Michael Turquette <mturquette@baylibre.com>
To: Stephen Boyd <sboyd@kernel.org>
Cc: linux-arm-msm@vger.kernel.org
Cc: linux-clk@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
---
 drivers/clk/qcom/clk-smd-rpm.c | 15 ++-------------
 1 file changed, 2 insertions(+), 13 deletions(-)

diff --git a/drivers/clk/qcom/clk-smd-rpm.c b/drivers/clk/qcom/clk-smd-rpm.c
index 103db984a40b950bd33fba668a292be46af6326e..96af781195a23d8b56e2b977811a3304a452e1cb 100644
--- a/drivers/clk/qcom/clk-smd-rpm.c
+++ b/drivers/clk/qcom/clk-smd-rpm.c
@@ -35,6 +35,7 @@
 					.name = "xo_board",		      \
 			},						      \
 			.num_parents = 1,				      \
+			.flags = CLK_ROUNDING_FW_MANAGED,		      \
 		},							      \
 	};								      \
 	static struct clk_smd_rpm clk_smd_rpm_##_prefix##_active = {	      \
@@ -52,7 +53,7 @@
 					.name = "xo_board",		      \
 			},						      \
 			.num_parents = 1,				      \
-			.flags = (ao_flags),				      \
+			.flags = (CLK_ROUNDING_FW_MANAGED | (ao_flags)),      \
 		},							      \
 	}
 
@@ -370,17 +371,6 @@ static int clk_smd_rpm_set_rate(struct clk_hw *hw, unsigned long rate,
 	return 0;
 }
 
-static int clk_smd_rpm_determine_rate(struct clk_hw *hw,
-				      struct clk_rate_request *req)
-{
-	/*
-	 * RPM handles rate rounding and we don't have a way to
-	 * know what the rate will be, so just return whatever
-	 * rate is requested.
-	 */
-	return 0;
-}
-
 static unsigned long clk_smd_rpm_recalc_rate(struct clk_hw *hw,
 					     unsigned long parent_rate)
 {
@@ -427,7 +417,6 @@ static const struct clk_ops clk_smd_rpm_ops = {
 	.prepare	= clk_smd_rpm_prepare,
 	.unprepare	= clk_smd_rpm_unprepare,
 	.set_rate	= clk_smd_rpm_set_rate,
-	.determine_rate = clk_smd_rpm_determine_rate,
 	.recalc_rate	= clk_smd_rpm_recalc_rate,
 };
 

-- 
2.53.0

Re: [PATCH 08/13] clk: qcom: rpm: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag

[Dmitry Baryshkov]

On Thu, Feb 26, 2026 at 01:16:52PM -0500, Brian Masney wrote:
> This clk driver has a noop determine_rate clk op. Drop this empty
> function, and enable the CLK_ROUNDING_FW_MANAGED flag.
> 
> Signed-off-by: Brian Masney <bmasney@redhat.com>
> 
> ---
> To: Bjorn Andersson <andersson@kernel.org>
> To: Michael Turquette <mturquette@baylibre.com>
> To: Stephen Boyd <sboyd@kernel.org>
> Cc: linux-arm-msm@vger.kernel.org
> Cc: linux-clk@vger.kernel.org
> Cc: linux-kernel@vger.kernel.org
> ---
>  drivers/clk/qcom/clk-rpm.c | 16 +++-------------
>  1 file changed, 3 insertions(+), 13 deletions(-)
> 

Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>


-- 
With best wishes
Dmitry

Re: [PATCH 09/13] clk: qcom: rpmh: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag

[Dmitry Baryshkov]

On Thu, Feb 26, 2026 at 01:16:53PM -0500, Brian Masney wrote:
> This clk driver has a noop determine_rate clk op. Drop this empty
> function, and enable the CLK_ROUNDING_FW_MANAGED flag.
> 
> Signed-off-by: Brian Masney <bmasney@redhat.com>
> 
> ---
> To: Bjorn Andersson <andersson@kernel.org>
> To: Michael Turquette <mturquette@baylibre.com>
> To: Stephen Boyd <sboyd@kernel.org>
> Cc: linux-arm-msm@vger.kernel.org
> Cc: linux-clk@vger.kernel.org
> Cc: linux-kernel@vger.kernel.org
> ---
>  drivers/clk/qcom/clk-rpmh.c | 8 +-------
>  1 file changed, 1 insertion(+), 7 deletions(-)
> 

Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>

-- 
With best wishes
Dmitry

Re: [PATCH 10/13] clk: qcom: smd-rpm: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag

[Dmitry Baryshkov]

On Thu, Feb 26, 2026 at 01:16:54PM -0500, Brian Masney wrote:
> This clk driver has a noop determine_rate clk op. Drop this empty
> function, and enable the CLK_ROUNDING_FW_MANAGED flag.
> 
> Signed-off-by: Brian Masney <bmasney@redhat.com>
> 
> ---
> To: Bjorn Andersson <andersson@kernel.org>
> To: Michael Turquette <mturquette@baylibre.com>
> To: Stephen Boyd <sboyd@kernel.org>
> Cc: linux-arm-msm@vger.kernel.org
> Cc: linux-clk@vger.kernel.org
> Cc: linux-kernel@vger.kernel.org
> ---
>  drivers/clk/qcom/clk-smd-rpm.c | 15 ++-------------
>  1 file changed, 2 insertions(+), 13 deletions(-)
> 

Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>


-- 
With best wishes
Dmitry

Re: [PATCH 01/13] clk: add new flag CLK_ROUNDING_FW_MANAGED

[Geert Uytterhoeven]

Hi Brian,

Thanks for your patch!

On Thu, 26 Feb 2026 at 19:17, Brian Masney <bmasney@redhat.com> wrote:
> There are some clocks where the rounding is managed by the hardware, and

s/hardware/firmware/

You got me totally confused, also/especially in the cover letter! ;-)

> the determine_rate() clk ops is just a noop that simply returns 0. Add a
> new flag for these type of clocks, and update the clk core so that the
> determine_rate() clk op is not required when this flag is set.
>
> Signed-off-by: Brian Masney <bmasney@redhat.com>

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

RE: [PATCH 01/13] clk: add new flag CLK_ROUNDING_FW_MANAGED

[Biju Das]

Thanks, I will test and provide the feedback.

Cheers,
Biju

> -----Original Message-----
> From: linux-arm-kernel <linux-arm-kernel-bounces@lists.infradead.org> On Behalf Of Geert Uytterhoeven
> Sent: 27 February 2026 08:17
> Subject: Re: [PATCH 01/13] clk: add new flag CLK_ROUNDING_FW_MANAGED
> 
> Hi Brian,
> 
> Thanks for your patch!
> 
> On Thu, 26 Feb 2026 at 19:17, Brian Masney <bmasney@redhat.com> wrote:
> > There are some clocks where the rounding is managed by the hardware,
> > and
> 
> s/hardware/firmware/
> 
> You got me totally confused, also/especially in the cover letter! ;-)
> 
> > the determine_rate() clk ops is just a noop that simply returns 0. Add
> > a new flag for these type of clocks, and update the clk core so that
> > the
> > determine_rate() clk op is not required when this flag is set.
> >
> > Signed-off-by: Brian Masney <bmasney@redhat.com>
> 
> Gr{oetje,eeting}s,
> 
>                         Geert
> 
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
> 
> In personal conversations with technical people, I call myself a hacker. But when I'm talking to
> journalists I just say "programmer" or something like that.
>                                 -- Linus Torvalds

RE: [PATCH 01/13] clk: add new flag CLK_ROUNDING_FW_MANAGED

[Biju Das]

Hi Brain,

Thanks for the patch

> -----Original Message-----
> From: linux-arm-kernel <linux-arm-kernel-bounces@lists.infradead.org> On Behalf Of Brian Masney
> Sent: 26 February 2026 18:17
> Subject: [PATCH 01/13] clk: add new flag CLK_ROUNDING_FW_MANAGED
> 
> There are some clocks where the rounding is managed by the hardware, and the determine_rate() clk ops
> is just a noop that simply returns 0. Add a new flag for these type of clocks, and update the clk core
> so that the
> determine_rate() clk op is not required when this flag is set.
> 
> Signed-off-by: Brian Masney <bmasney@redhat.com>
> 
> ---
> To: Michael Turquette <mturquette@baylibre.com>
> To: Stephen Boyd <sboyd@kernel.org>
> To: Abel Vesa <abelvesa@kernel.org>
> To: Andrea della Porta <andrea.porta@suse.com>
> To: Baolin Wang <baolin.wang@linux.alibaba.com>
> To: Bjorn Andersson <andersson@kernel.org>
> To: Chanwoo Choi <cw00.choi@samsung.com>
> To: Frank Li <Frank.Li@nxp.com>
> To: Geert Uytterhoeven <geert+renesas@glider.be>
> To: Krzysztof Kozlowski <krzk@kernel.org>
> To: Orson Zhai <orsonzhai@gmail.com>
> To: Sascha Hauer <s.hauer@pengutronix.de>
> To: Sudeep Holla <sudeep.holla@kernel.org>
> To: Sylwester Nawrocki <s.nawrocki@samsung.com>
> To: Tudor Ambarus <tudor.ambarus@linaro.org>
> Cc: linux-clk@vger.kernel.org
> Cc: linux-kernel@vger.kernel.org
> Cc: Alim Akhtar <alim.akhtar@samsung.com>
> Cc: arm-scmi@vger.kernel.org
> Cc: Chunyan Zhang <zhang.lyra@gmail.com>
> Cc: Cristian Marussi <cristian.marussi@arm.com>
> Cc: Fabio Estevam <festevam@gmail.com>
> Cc: imx@lists.linux.dev
> Cc: linux-arm-kernel@lists.infradead.org
> Cc: linux-arm-msm@vger.kernel.org
> Cc: linux-renesas-soc@vger.kernel.org
> Cc: linux-samsung-soc@vger.kernel.org
> Cc: Peng Fan <peng.fan@nxp.com>
> Cc: Pengutronix Kernel Team <kernel@pengutronix.de>
> ---
>  drivers/clk/clk.c            | 24 +++++++++++++++++++++---
>  include/linux/clk-provider.h |  2 ++
>  2 files changed, 23 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c index
> fd418dc988b1c60c49e3ac9c0c44aa132dd5da28..0a522a0817411c7f7c6e9cffd6f024e672a331a8 100644
> --- a/drivers/clk/clk.c
> +++ b/drivers/clk/clk.c
> @@ -1557,6 +1557,20 @@ static int __init clk_disable_unused(void)  }
> late_initcall_sync(clk_disable_unused);
> 
> +/**
> + * clk_is_rounding_fw_managed - Check to see if clk rounding is handled
> +by the
> + * firmware.
> + * @core: the clk to check
> + *
> + * Clks that have this flag enabled do not need to have a
> +determine_rate() op
> + * set, and will always return success for any rounding operation since
> +the
> + * firmware will deal with the rounding.
> + */
> +static inline bool clk_is_rounding_fw_managed(struct clk_core *core) {
> +	return core->flags & CLK_ROUNDING_FW_MANAGED; }
> +
>  static int clk_core_determine_round_nolock(struct clk_core *core,
>  					   struct clk_rate_request *req)
>  {
> @@ -1589,6 +1603,8 @@ static int clk_core_determine_round_nolock(struct clk_core *core,
>  		req->rate = core->rate;
>  	} else if (core->ops->determine_rate) {
>  		return core->ops->determine_rate(core->hw, req);
> +	} else if (clk_is_rounding_fw_managed(core)) {
> +		return 0;
>  	} else {
>  		return -EINVAL;
>  	}
> @@ -1673,7 +1689,7 @@ EXPORT_SYMBOL_GPL(clk_hw_forward_rate_request);
> 
>  static bool clk_core_can_round(struct clk_core * const core)  {
> -	return core->ops->determine_rate;
> +	return core->ops->determine_rate || clk_is_rounding_fw_managed(core);
>  }
> 
>  static int clk_core_round_rate_nolock(struct clk_core *core, @@ -3528,6 +3544,7 @@ static const
> struct {
>  	ENTRY(CLK_IS_CRITICAL),
>  	ENTRY(CLK_OPS_PARENT_ENABLE),
>  	ENTRY(CLK_DUTY_CYCLE_PARENT),
> +	ENTRY(CLK_ROUNDING_FW_MANAGED),
>  #undef ENTRY
>  };
> 
> @@ -3906,7 +3923,7 @@ static int __clk_core_init(struct clk_core *core)
> 
>  	/* check that clk_ops are sane.  See Documentation/driver-api/clk.rst */
>  	if (core->ops->set_rate && !core->ops->determine_rate &&
> -	      core->ops->recalc_rate) {
> +	      core->ops->recalc_rate && !clk_is_rounding_fw_managed(core)) {
>  		pr_err("%s: %s must implement .determine_rate in addition to .recalc_rate\n",
>  		       __func__, core->name);
>  		ret = -EINVAL;
> @@ -3920,7 +3937,8 @@ static int __clk_core_init(struct clk_core *core)
>  		goto out;
>  	}
> 
> -	if (core->ops->set_parent && !core->ops->determine_rate) {
> +	if (core->ops->set_parent && !core->ops->determine_rate &&
> +	    !clk_is_rounding_fw_managed(core)) {

>  		pr_err("%s: %s must implement .set_parent & .determine_rate\n",
>  			__func__, core->name);
>  		ret = -EINVAL;


After applying patch#11, I get a message as you removed .determine_rate, Also it breaks display.

[    0.096414] __clk_core_init: .pll5_foutpostdiv must implement .round_rate or .determine_rate in addition to .recalc_rate


Cheers,
Biju

Re: [PATCH 01/13] clk: add new flag CLK_ROUNDING_FW_MANAGED

[Brian Masney]

On Fri, Feb 27, 2026 at 12:00:55PM +0000, Biju Das wrote:
> > -----Original Message-----
> > From: linux-arm-kernel <linux-arm-kernel-bounces@lists.infradead.org> On Behalf Of Brian Masney
> > Sent: 26 February 2026 18:17
> > Subject: [PATCH 01/13] clk: add new flag CLK_ROUNDING_FW_MANAGED
> > 
> > There are some clocks where the rounding is managed by the hardware, and the determine_rate() clk ops
> > is just a noop that simply returns 0. Add a new flag for these type of clocks, and update the clk core
> > so that the
> > determine_rate() clk op is not required when this flag is set.
> > 
> > Signed-off-by: Brian Masney <bmasney@redhat.com>
> > 
[snip]
> >  drivers/clk/clk.c            | 24 +++++++++++++++++++++---
> >  include/linux/clk-provider.h |  2 ++
> >  2 files changed, 23 insertions(+), 3 deletions(-)
> > 
> > diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c index
> > fd418dc988b1c60c49e3ac9c0c44aa132dd5da28..0a522a0817411c7f7c6e9cffd6f024e672a331a8 100644
> > --- a/drivers/clk/clk.c
> > +++ b/drivers/clk/clk.c
> > @@ -1557,6 +1557,20 @@ static int __init clk_disable_unused(void)  }
> > late_initcall_sync(clk_disable_unused);
> > 
> > +/**
> > + * clk_is_rounding_fw_managed - Check to see if clk rounding is handled
> > +by the
> > + * firmware.
> > + * @core: the clk to check
> > + *
> > + * Clks that have this flag enabled do not need to have a
> > +determine_rate() op
> > + * set, and will always return success for any rounding operation since
> > +the
> > + * firmware will deal with the rounding.
> > + */
> > +static inline bool clk_is_rounding_fw_managed(struct clk_core *core) {
> > +	return core->flags & CLK_ROUNDING_FW_MANAGED; }
> > +
> >  static int clk_core_determine_round_nolock(struct clk_core *core,
> >  					   struct clk_rate_request *req)
> >  {
> > @@ -1589,6 +1603,8 @@ static int clk_core_determine_round_nolock(struct clk_core *core,
> >  		req->rate = core->rate;
> >  	} else if (core->ops->determine_rate) {
> >  		return core->ops->determine_rate(core->hw, req);
> > +	} else if (clk_is_rounding_fw_managed(core)) {
> > +		return 0;
> >  	} else {
> >  		return -EINVAL;
> >  	}
> > @@ -1673,7 +1689,7 @@ EXPORT_SYMBOL_GPL(clk_hw_forward_rate_request);
> > 
> >  static bool clk_core_can_round(struct clk_core * const core)  {
> > -	return core->ops->determine_rate;
> > +	return core->ops->determine_rate || clk_is_rounding_fw_managed(core);
> >  }
> > 
> >  static int clk_core_round_rate_nolock(struct clk_core *core, @@ -3528,6 +3544,7 @@ static const
> > struct {
> >  	ENTRY(CLK_IS_CRITICAL),
> >  	ENTRY(CLK_OPS_PARENT_ENABLE),
> >  	ENTRY(CLK_DUTY_CYCLE_PARENT),
> > +	ENTRY(CLK_ROUNDING_FW_MANAGED),
> >  #undef ENTRY
> >  };
> > 
> > @@ -3906,7 +3923,7 @@ static int __clk_core_init(struct clk_core *core)
> > 
> >  	/* check that clk_ops are sane.  See Documentation/driver-api/clk.rst */
> >  	if (core->ops->set_rate && !core->ops->determine_rate &&
> > -	      core->ops->recalc_rate) {
> > +	      core->ops->recalc_rate && !clk_is_rounding_fw_managed(core)) {
> >  		pr_err("%s: %s must implement .determine_rate in addition to .recalc_rate\n",
> >  		       __func__, core->name);
> >  		ret = -EINVAL;
> > @@ -3920,7 +3937,8 @@ static int __clk_core_init(struct clk_core *core)
> >  		goto out;
> >  	}
> > 
> > -	if (core->ops->set_parent && !core->ops->determine_rate) {
> > +	if (core->ops->set_parent && !core->ops->determine_rate &&
> > +	    !clk_is_rounding_fw_managed(core)) {
> 
> >  		pr_err("%s: %s must implement .set_parent & .determine_rate\n",
> >  			__func__, core->name);
> >  		ret = -EINVAL;
> 
> 
> After applying patch#11, I get a message as you removed .determine_rate, Also it breaks display.
> 
> [    0.096414] __clk_core_init: .pll5_foutpostdiv must implement .round_rate or .determine_rate in addition to .recalc_rate

Thanks for testing. This happens because rzg2l_cpg_pll_clk_register()
doesn't have the new flag set. I'll fix this, and go through all of the
others again just to make sure I don't miss any others.

Brian

Re: [PATCH 01/13] clk: add new flag CLK_ROUNDING_FW_MANAGED

[Brian Masney]

On Thu, Feb 26, 2026 at 01:16:45PM -0500, Brian Masney wrote:
> There are some clocks where the rounding is managed by the hardware, and
> the determine_rate() clk ops is just a noop that simply returns 0. Add a
> new flag for these type of clocks, and update the clk core so that the
> determine_rate() clk op is not required when this flag is set.
> 
> Signed-off-by: Brian Masney <bmasney@redhat.com>
> 
...
> diff --git a/include/linux/clk-provider.h b/include/linux/clk-provider.h
> index 1cda2c78dffaff037f0f16b0f11106b63b3a746f..187f8248a9c840c701cbbba99bb7cdeef7b654ee 100644
> --- a/include/linux/clk-provider.h
> +++ b/include/linux/clk-provider.h
> @@ -32,6 +32,8 @@
>  #define CLK_OPS_PARENT_ENABLE	BIT(12)
>  /* duty cycle call may be forwarded to the parent clock */
>  #define CLK_DUTY_CYCLE_PARENT	BIT(13)
> +/* clock rate rounding is managed by firmware, don't require determine_rate */
> +#define CLK_ROUNDING_FW_MANAGED		BIT(14)

Based on a conversation in the Renesas driver change, I think we should
rename this flag to CLK_ROUNDING_NOOP. Let me know if there are any
objections, or suggestions for a better name. I'll let let this sit out
there for about a week or so before sending a v2.

Brian

Re: [PATCH 01/13] clk: add new flag CLK_ROUNDING_FW_MANAGED

[Sudeep Holla]

On Thu, Feb 26, 2026 at 01:16:45PM -0500, Brian Masney wrote:
> There are some clocks where the rounding is managed by the hardware, and
> the determine_rate() clk ops is just a noop that simply returns 0. Add a
> new flag for these type of clocks, and update the clk core so that the
> determine_rate() clk op is not required when this flag is set.
> 
> Signed-off-by: Brian Masney <bmasney@redhat.com>
> 

Acked-by: Sudeep Holla <sudeep.holla@kernel.org>

-- 
Regards,
Sudeep