diff mbox

[v6] clk: mediatek: Export CPU mux clocks for CPU frequency control

Message ID 1442230072-31593-1-git-send-email-pi-cheng.chen@linaro.org (mailing list archive)
State New, archived
Headers show

Commit Message

pi-cheng.chen Sept. 14, 2015, 11:27 a.m. UTC
From: "pi-cheng.chen" <pi-cheng.chen@linaro.org>

This patch adds CPU mux clocks which are used by Mediatek cpufreq driver
for intermediate clock source switching.

Signed-off-by: Pi-Cheng Chen <pi-cheng.chen@linaro.org>
---
Changes in v6:
- Relace __clk_get_num_parents with clk_hw_get_num_parents to reflect
  API change

Changes in v5:
- Replace __initdata with __initconst to fix compiling error

Changes in v4:
- Fix some minor issues for v3
- Rebase to the patch that adds 13mhz clock for MT8173[1]

Changes in v3:
- Rebase to 4.2-rc1
- Fix some issues of v2

Changes in v2:
- Remove use of .determine_rate callback
---
 drivers/clk/mediatek/Makefile          |   2 +-
 drivers/clk/mediatek/clk-cpumux.c      | 127 +++++++++++++++++++++++++++++++++
 drivers/clk/mediatek/clk-cpumux.h      |  22 ++++++
 drivers/clk/mediatek/clk-mt8173.c      |  23 ++++++
 include/dt-bindings/clock/mt8173-clk.h |   4 +-
 5 files changed, 176 insertions(+), 2 deletions(-)
 create mode 100644 drivers/clk/mediatek/clk-cpumux.c
 create mode 100644 drivers/clk/mediatek/clk-cpumux.h

Comments

Daniel Kurtz Nov. 30, 2015, 3:48 a.m. UTC | #1
Hi Pi-Cheng and Michael,

On Mon, Sep 14, 2015 at 7:27 PM, Pi-Cheng Chen <pi-cheng.chen@linaro.org> wrote:
> From: "pi-cheng.chen" <pi-cheng.chen@linaro.org>
>
> This patch adds CPU mux clocks which are used by Mediatek cpufreq driver
> for intermediate clock source switching.

It looks like this patch was never picked up.
It is required for MT8173 cpufreq.

What is the plan to get this one merged?

-Dan

>
> Signed-off-by: Pi-Cheng Chen <pi-cheng.chen@linaro.org>
> ---
> Changes in v6:
> - Relace __clk_get_num_parents with clk_hw_get_num_parents to reflect
>   API change
>
> Changes in v5:
> - Replace __initdata with __initconst to fix compiling error
>
> Changes in v4:
> - Fix some minor issues for v3
> - Rebase to the patch that adds 13mhz clock for MT8173[1]
>
> Changes in v3:
> - Rebase to 4.2-rc1
> - Fix some issues of v2
>
> Changes in v2:
> - Remove use of .determine_rate callback
> ---
>  drivers/clk/mediatek/Makefile          |   2 +-
>  drivers/clk/mediatek/clk-cpumux.c      | 127 +++++++++++++++++++++++++++++++++
>  drivers/clk/mediatek/clk-cpumux.h      |  22 ++++++
>  drivers/clk/mediatek/clk-mt8173.c      |  23 ++++++
>  include/dt-bindings/clock/mt8173-clk.h |   4 +-
>  5 files changed, 176 insertions(+), 2 deletions(-)
>  create mode 100644 drivers/clk/mediatek/clk-cpumux.c
>  create mode 100644 drivers/clk/mediatek/clk-cpumux.h

[snip]
pi-cheng.chen Dec. 1, 2015, 3:09 p.m. UTC | #2
On Mon, Nov 30, 2015 at 11:48 AM, Daniel Kurtz <djkurtz@chromium.org> wrote:
> Hi Pi-Cheng and Michael,
>
> On Mon, Sep 14, 2015 at 7:27 PM, Pi-Cheng Chen <pi-cheng.chen@linaro.org> wrote:
>> From: "pi-cheng.chen" <pi-cheng.chen@linaro.org>
>>
>> This patch adds CPU mux clocks which are used by Mediatek cpufreq driver
>> for intermediate clock source switching.
>
> It looks like this patch was never picked up.
> It is required for MT8173 cpufreq.
>
> What is the plan to get this one merged?

Hi Dan,
Mike is working on a new CCF feature called "coordinated clock rate changes"
which would help to migrate mt8173-cpufreq driver to generic cpufreq-dt
driver, and I am working with him to port MT8173 clock driver on the new
feature. So I think that's why this patch was not picked since it will
be migrated to new implementation once the new feature's ready.

@Mike:
Are you plannig to post those patches on the mailing list ?
Or is it fine for you to take this first and then I will refine it for
the new feature later?

Best Regards,
Pi-Cheng

>
> -Dan
>
>>
>> Signed-off-by: Pi-Cheng Chen <pi-cheng.chen@linaro.org>
>> ---
>> Changes in v6:
>> - Relace __clk_get_num_parents with clk_hw_get_num_parents to reflect
>>   API change
>>
>> Changes in v5:
>> - Replace __initdata with __initconst to fix compiling error
>>
>> Changes in v4:
>> - Fix some minor issues for v3
>> - Rebase to the patch that adds 13mhz clock for MT8173[1]
>>
>> Changes in v3:
>> - Rebase to 4.2-rc1
>> - Fix some issues of v2
>>
>> Changes in v2:
>> - Remove use of .determine_rate callback
>> ---
>>  drivers/clk/mediatek/Makefile          |   2 +-
>>  drivers/clk/mediatek/clk-cpumux.c      | 127 +++++++++++++++++++++++++++++++++
>>  drivers/clk/mediatek/clk-cpumux.h      |  22 ++++++
>>  drivers/clk/mediatek/clk-mt8173.c      |  23 ++++++
>>  include/dt-bindings/clock/mt8173-clk.h |   4 +-
>>  5 files changed, 176 insertions(+), 2 deletions(-)
>>  create mode 100644 drivers/clk/mediatek/clk-cpumux.c
>>  create mode 100644 drivers/clk/mediatek/clk-cpumux.h
>
> [snip]
Michael Turquette Dec. 17, 2015, 9:06 a.m. UTC | #3
On 12/01, Pi-Cheng Chen wrote:
> On Mon, Nov 30, 2015 at 11:48 AM, Daniel Kurtz <djkurtz@chromium.org> wrote:
> > Hi Pi-Cheng and Michael,
> >
> > On Mon, Sep 14, 2015 at 7:27 PM, Pi-Cheng Chen <pi-cheng.chen@linaro.org> wrote:
> >> From: "pi-cheng.chen" <pi-cheng.chen@linaro.org>
> >>
> >> This patch adds CPU mux clocks which are used by Mediatek cpufreq driver
> >> for intermediate clock source switching.
> >
> > It looks like this patch was never picked up.
> > It is required for MT8173 cpufreq.
> >
> > What is the plan to get this one merged?
> 
> Hi Dan,
> Mike is working on a new CCF feature called "coordinated clock rate changes"
> which would help to migrate mt8173-cpufreq driver to generic cpufreq-dt
> driver, and I am working with him to port MT8173 clock driver on the new
> feature. So I think that's why this patch was not picked since it will
> be migrated to new implementation once the new feature's ready.
> 
> @Mike:
> Are you plannig to post those patches on the mailing list ?
> Or is it fine for you to take this first and then I will refine it for
> the new feature later?

I plan to post those patches next week.

Thanks,
Mike

> 
> Best Regards,
> Pi-Cheng
> 
> >
> > -Dan
> >
> >>
> >> Signed-off-by: Pi-Cheng Chen <pi-cheng.chen@linaro.org>
> >> ---
> >> Changes in v6:
> >> - Relace __clk_get_num_parents with clk_hw_get_num_parents to reflect
> >>   API change
> >>
> >> Changes in v5:
> >> - Replace __initdata with __initconst to fix compiling error
> >>
> >> Changes in v4:
> >> - Fix some minor issues for v3
> >> - Rebase to the patch that adds 13mhz clock for MT8173[1]
> >>
> >> Changes in v3:
> >> - Rebase to 4.2-rc1
> >> - Fix some issues of v2
> >>
> >> Changes in v2:
> >> - Remove use of .determine_rate callback
> >> ---
> >>  drivers/clk/mediatek/Makefile          |   2 +-
> >>  drivers/clk/mediatek/clk-cpumux.c      | 127 +++++++++++++++++++++++++++++++++
> >>  drivers/clk/mediatek/clk-cpumux.h      |  22 ++++++
> >>  drivers/clk/mediatek/clk-mt8173.c      |  23 ++++++
> >>  include/dt-bindings/clock/mt8173-clk.h |   4 +-
> >>  5 files changed, 176 insertions(+), 2 deletions(-)
> >>  create mode 100644 drivers/clk/mediatek/clk-cpumux.c
> >>  create mode 100644 drivers/clk/mediatek/clk-cpumux.h
> >
> > [snip]
Daniel Kurtz Jan. 26, 2016, 10:23 a.m. UTC | #4
Hi Michael, Pi-Cheng,

On Thu, Dec 17, 2015 at 1:06 AM, Michael Turquette
<mturquette@baylibre.com> wrote:
>
> On 12/01, Pi-Cheng Chen wrote:
> > On Mon, Nov 30, 2015 at 11:48 AM, Daniel Kurtz <djkurtz@chromium.org> wrote:
> > > Hi Pi-Cheng and Michael,
> > >
> > > On Mon, Sep 14, 2015 at 7:27 PM, Pi-Cheng Chen <pi-cheng.chen@linaro.org> wrote:
> > >> From: "pi-cheng.chen" <pi-cheng.chen@linaro.org>
> > >>
> > >> This patch adds CPU mux clocks which are used by Mediatek cpufreq driver
> > >> for intermediate clock source switching.
> > >
> > > It looks like this patch was never picked up.
> > > It is required for MT8173 cpufreq.
> > >
> > > What is the plan to get this one merged?
> >
> > Hi Dan,
> > Mike is working on a new CCF feature called "coordinated clock rate changes"
> > which would help to migrate mt8173-cpufreq driver to generic cpufreq-dt
> > driver, and I am working with him to port MT8173 clock driver on the new
> > feature. So I think that's why this patch was not picked since it will
> > be migrated to new implementation once the new feature's ready.
> >
> > @Mike:
> > Are you plannig to post those patches on the mailing list ?
> > Or is it fine for you to take this first and then I will refine it for
> > the new feature later?
>
> I plan to post those patches next week.

Were these patches ever posted?
Or should we consider going ahead and trying to merge this 4-month old
MTK CPU mux clock patch into v4.5 before it is too late?

Thanks,
-Dan
Matthias Brugger Feb. 1, 2016, 11:46 a.m. UTC | #5
On 26/01/16 11:23, Daniel Kurtz wrote:
> Hi Michael, Pi-Cheng,
>
> On Thu, Dec 17, 2015 at 1:06 AM, Michael Turquette
> <mturquette@baylibre.com> wrote:
>>
>> On 12/01, Pi-Cheng Chen wrote:
>>> On Mon, Nov 30, 2015 at 11:48 AM, Daniel Kurtz <djkurtz@chromium.org> wrote:
>>>> Hi Pi-Cheng and Michael,
>>>>
>>>> On Mon, Sep 14, 2015 at 7:27 PM, Pi-Cheng Chen <pi-cheng.chen@linaro.org> wrote:
>>>>> From: "pi-cheng.chen" <pi-cheng.chen@linaro.org>
>>>>>
>>>>> This patch adds CPU mux clocks which are used by Mediatek cpufreq driver
>>>>> for intermediate clock source switching.
>>>>
>>>> It looks like this patch was never picked up.
>>>> It is required for MT8173 cpufreq.
>>>>
>>>> What is the plan to get this one merged?
>>>
>>> Hi Dan,
>>> Mike is working on a new CCF feature called "coordinated clock rate changes"
>>> which would help to migrate mt8173-cpufreq driver to generic cpufreq-dt
>>> driver, and I am working with him to port MT8173 clock driver on the new
>>> feature. So I think that's why this patch was not picked since it will
>>> be migrated to new implementation once the new feature's ready.
>>>
>>> @Mike:
>>> Are you plannig to post those patches on the mailing list ?
>>> Or is it fine for you to take this first and then I will refine it for
>>> the new feature later?
>>
>> I plan to post those patches next week.
>
> Were these patches ever posted?
> Or should we consider going ahead and trying to merge this 4-month old
> MTK CPU mux clock patch into v4.5 before it is too late?
>

It's already too late for v4.5. The -rc are mostly bug fixes and new 
features get merged after the official v4.5 release. They will show up 
in v4.6-rc1
Daniel Kurtz Feb. 1, 2016, 11:48 a.m. UTC | #6
On Mon, Feb 1, 2016 at 7:46 PM, Matthias Brugger <matthias.bgg@gmail.com> wrote:
>
>
> On 26/01/16 11:23, Daniel Kurtz wrote:
>>
>> Hi Michael, Pi-Cheng,
>>
>> On Thu, Dec 17, 2015 at 1:06 AM, Michael Turquette
>> <mturquette@baylibre.com> wrote:
>>>
>>>
>>> On 12/01, Pi-Cheng Chen wrote:
>>>>
>>>> On Mon, Nov 30, 2015 at 11:48 AM, Daniel Kurtz <djkurtz@chromium.org>
>>>> wrote:
>>>>>
>>>>> Hi Pi-Cheng and Michael,
>>>>>
>>>>> On Mon, Sep 14, 2015 at 7:27 PM, Pi-Cheng Chen
>>>>> <pi-cheng.chen@linaro.org> wrote:
>>>>>>
>>>>>> From: "pi-cheng.chen" <pi-cheng.chen@linaro.org>
>>>>>>
>>>>>> This patch adds CPU mux clocks which are used by Mediatek cpufreq
>>>>>> driver
>>>>>> for intermediate clock source switching.
>>>>>
>>>>>
>>>>> It looks like this patch was never picked up.
>>>>> It is required for MT8173 cpufreq.
>>>>>
>>>>> What is the plan to get this one merged?
>>>>
>>>>
>>>> Hi Dan,
>>>> Mike is working on a new CCF feature called "coordinated clock rate
>>>> changes"
>>>> which would help to migrate mt8173-cpufreq driver to generic cpufreq-dt
>>>> driver, and I am working with him to port MT8173 clock driver on the new
>>>> feature. So I think that's why this patch was not picked since it will
>>>> be migrated to new implementation once the new feature's ready.
>>>>
>>>> @Mike:
>>>> Are you plannig to post those patches on the mailing list ?
>>>> Or is it fine for you to take this first and then I will refine it for
>>>> the new feature later?
>>>
>>>
>>> I plan to post those patches next week.
>>
>>
>> Were these patches ever posted?
>> Or should we consider going ahead and trying to merge this 4-month old
>> MTK CPU mux clock patch into v4.5 before it is too late?
>>
>
> It's already too late for v4.5. The -rc are mostly bug fixes and new
> features get merged after the official v4.5 release. They will show up in
> v4.6-rc1

Which will show up in v4.6-rc1?  MediaTek's CPU mux clocks, or Mike's
"coordinated clock rate changes"?
Daniel Kurtz Feb. 15, 2016, 3:28 a.m. UTC | #7
Hi Mike, Pi-Cheng,

On Thu, Dec 17, 2015 at 5:06 PM, Michael Turquette
<mturquette@baylibre.com> wrote:
> On 12/01, Pi-Cheng Chen wrote:
>> On Mon, Nov 30, 2015 at 11:48 AM, Daniel Kurtz <djkurtz@chromium.org> wrote:
>> > Hi Pi-Cheng and Michael,
>> >
>> > On Mon, Sep 14, 2015 at 7:27 PM, Pi-Cheng Chen <pi-cheng.chen@linaro.org> wrote:
>> >> From: "pi-cheng.chen" <pi-cheng.chen@linaro.org>
>> >>
>> >> This patch adds CPU mux clocks which are used by Mediatek cpufreq driver
>> >> for intermediate clock source switching.
>> >
>> > It looks like this patch was never picked up.
>> > It is required for MT8173 cpufreq.
>> >
>> > What is the plan to get this one merged?
>>
>> Hi Dan,
>> Mike is working on a new CCF feature called "coordinated clock rate changes"
>> which would help to migrate mt8173-cpufreq driver to generic cpufreq-dt
>> driver, and I am working with him to port MT8173 clock driver on the new
>> feature. So I think that's why this patch was not picked since it will
>> be migrated to new implementation once the new feature's ready.
>>
>> @Mike:
>> Are you plannig to post those patches on the mailing list ?
>> Or is it fine for you to take this first and then I will refine it for
>> the new feature later?
>
> I plan to post those patches next week.

Were these patches ever posted?
I'm not really sure what patch set I am looking for.

Thanks,
-Dan
Daniel Kurtz Feb. 12, 2017, 3:31 a.m. UTC | #8
Hi Mike, Pi-Cheng, Matthias,

On Mon, Feb 15, 2016 at 11:28 AM, Daniel Kurtz <djkurtz@chromium.org> wrote:
> Hi Mike, Pi-Cheng,
>
> On Thu, Dec 17, 2015 at 5:06 PM, Michael Turquette
> <mturquette@baylibre.com> wrote:
>> On 12/01, Pi-Cheng Chen wrote:
>>> On Mon, Nov 30, 2015 at 11:48 AM, Daniel Kurtz <djkurtz@chromium.org> wrote:
>>> > Hi Pi-Cheng and Michael,
>>> >
>>> > On Mon, Sep 14, 2015 at 7:27 PM, Pi-Cheng Chen <pi-cheng.chen@linaro.org> wrote:
>>> >> From: "pi-cheng.chen" <pi-cheng.chen@linaro.org>
>>> >>
>>> >> This patch adds CPU mux clocks which are used by Mediatek cpufreq driver
>>> >> for intermediate clock source switching.
>>> >
>>> > It looks like this patch was never picked up.
>>> > It is required for MT8173 cpufreq.
>>> >
>>> > What is the plan to get this one merged?
>>>
>>> Hi Dan,
>>> Mike is working on a new CCF feature called "coordinated clock rate changes"
>>> which would help to migrate mt8173-cpufreq driver to generic cpufreq-dt
>>> driver, and I am working with him to port MT8173 clock driver on the new
>>> feature. So I think that's why this patch was not picked since it will
>>> be migrated to new implementation once the new feature's ready.
>>>
>>> @Mike:
>>> Are you plannig to post those patches on the mailing list ?
>>> Or is it fine for you to take this first and then I will refine it for
>>> the new feature later?
>>
>> I plan to post those patches next week.
>
> Were these patches ever posted?
> I'm not really sure what patch set I am looking for.

Another friendly (+1 year) ping on the status of the "coordinated
clock rate changes" patches that were a prerequisite to finally finish
merging the complete Mediatek MT8173 cpufreq implementation.

Thanks,
-Dan
diff mbox

Patch

diff --git a/drivers/clk/mediatek/Makefile b/drivers/clk/mediatek/Makefile
index 8e4b2a4..299917a 100644
--- a/drivers/clk/mediatek/Makefile
+++ b/drivers/clk/mediatek/Makefile
@@ -1,4 +1,4 @@ 
-obj-y += clk-mtk.o clk-pll.o clk-gate.o
+obj-y += clk-mtk.o clk-pll.o clk-gate.o clk-cpumux.o
 obj-$(CONFIG_RESET_CONTROLLER) += reset.o
 obj-y += clk-mt8135.o
 obj-y += clk-mt8173.o
diff --git a/drivers/clk/mediatek/clk-cpumux.c b/drivers/clk/mediatek/clk-cpumux.c
new file mode 100644
index 0000000..91b5238
--- /dev/null
+++ b/drivers/clk/mediatek/clk-cpumux.c
@@ -0,0 +1,127 @@ 
+/*
+ * Copyright (c) 2015 Linaro Ltd.
+ * Author: Pi-Cheng Chen <pi-cheng.chen@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/clk-provider.h>
+#include <linux/mfd/syscon.h>
+#include <linux/slab.h>
+
+#include "clk-mtk.h"
+#include "clk-cpumux.h"
+
+struct mtk_clk_cpumux {
+	struct clk_hw	hw;
+	struct regmap	*regmap;
+	u32		reg;
+	u32		mask;
+	u8		shift;
+};
+
+static inline struct mtk_clk_cpumux *to_clk_mux(struct clk_hw *_hw)
+{
+	return container_of(_hw, struct mtk_clk_cpumux, hw);
+}
+
+static u8 clk_cpumux_get_parent(struct clk_hw *hw)
+{
+	struct mtk_clk_cpumux *mux = to_clk_mux(hw);
+	int num_parents = clk_hw_get_num_parents(hw);
+	unsigned int val;
+
+	regmap_read(mux->regmap, mux->reg, &val);
+
+	val >>= mux->shift;
+	val &= mux->mask;
+
+	if (val >= num_parents)
+		return -EINVAL;
+
+	return val;
+}
+
+static int clk_cpumux_set_parent(struct clk_hw *hw, u8 index)
+{
+	struct mtk_clk_cpumux *mux = to_clk_mux(hw);
+	u32 mask, val;
+
+	val = index << mux->shift;
+	mask = mux->mask << mux->shift;
+
+	return regmap_update_bits(mux->regmap, mux->reg, mask, val);
+}
+
+static const struct clk_ops clk_cpumux_ops = {
+	.get_parent = clk_cpumux_get_parent,
+	.set_parent = clk_cpumux_set_parent,
+};
+
+static struct clk __init *mtk_clk_register_cpumux(const struct mtk_composite *mux,
+					   struct regmap *regmap)
+{
+	struct mtk_clk_cpumux *cpumux;
+	struct clk *clk;
+	struct clk_init_data init;
+
+	cpumux = kzalloc(sizeof(*cpumux), GFP_KERNEL);
+	if (!cpumux)
+		return ERR_PTR(-ENOMEM);
+
+	init.name = mux->name;
+	init.ops = &clk_cpumux_ops;
+	init.parent_names = mux->parent_names;
+	init.num_parents = mux->num_parents;
+	init.flags = mux->flags;
+
+	cpumux->reg = mux->mux_reg;
+	cpumux->shift = mux->mux_shift;
+	cpumux->mask = BIT(mux->mux_width) - 1;
+	cpumux->regmap = regmap;
+	cpumux->hw.init = &init;
+
+	clk = clk_register(NULL, &cpumux->hw);
+	if (IS_ERR(clk))
+		kfree(cpumux);
+
+	return clk;
+}
+
+int __init mtk_clk_register_cpumuxes(struct device_node *node,
+			      const struct mtk_composite *clks, int num,
+			      struct clk_onecell_data *clk_data)
+{
+	int i;
+	struct clk *clk;
+	struct regmap *regmap;
+
+	regmap = syscon_node_to_regmap(node);
+	if (IS_ERR(regmap)) {
+		pr_err("Cannot find regmap for %s: %ld\n", node->full_name,
+		       PTR_ERR(regmap));
+		return PTR_ERR(regmap);
+	}
+
+	for (i = 0; i < num; i++) {
+		const struct mtk_composite *mux = &clks[i];
+
+		clk = mtk_clk_register_cpumux(mux, regmap);
+		if (IS_ERR(clk)) {
+			pr_err("Failed to register clk %s: %ld\n",
+			       mux->name, PTR_ERR(clk));
+			continue;
+		}
+
+		clk_data->clks[mux->id] = clk;
+	}
+
+	return 0;
+}
diff --git a/drivers/clk/mediatek/clk-cpumux.h b/drivers/clk/mediatek/clk-cpumux.h
new file mode 100644
index 0000000..52c769f
--- /dev/null
+++ b/drivers/clk/mediatek/clk-cpumux.h
@@ -0,0 +1,22 @@ 
+/*
+ * Copyright (c) 2015 Linaro Ltd.
+ * Author: Pi-Cheng Chen <pi-cheng.chen@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef __DRV_CLK_CPUMUX_H
+#define __DRV_CLK_CPUMUX_H
+
+int mtk_clk_register_cpumuxes(struct device_node *node,
+			      const struct mtk_composite *clks, int num,
+			      struct clk_onecell_data *clk_data);
+
+#endif /* __DRV_CLK_CPUMUX_H */
diff --git a/drivers/clk/mediatek/clk-mt8173.c b/drivers/clk/mediatek/clk-mt8173.c
index 9ea6aa1..b02d310 100644
--- a/drivers/clk/mediatek/clk-mt8173.c
+++ b/drivers/clk/mediatek/clk-mt8173.c
@@ -20,6 +20,7 @@ 
 
 #include "clk-mtk.h"
 #include "clk-gate.h"
+#include "clk-cpumux.h"
 
 #include <dt-bindings/clock/mt8173-clk.h>
 
@@ -518,6 +519,25 @@  static const char * const i2s3_b_ck_parents[] __initconst = {
 	"apll2_div5"
 };
 
+static const char * const ca53_parents[] __initconst = {
+	"clk26m",
+	"armca7pll",
+	"mainpll",
+	"univpll"
+};
+
+static const char * const ca57_parents[] __initconst = {
+	"clk26m",
+	"armca15pll",
+	"mainpll",
+	"univpll"
+};
+
+static const struct mtk_composite cpu_muxes[] __initconst = {
+	MUX(CLK_INFRA_CA53SEL, "infra_ca53_sel", ca53_parents, 0x0000, 0, 2),
+	MUX(CLK_INFRA_CA57SEL, "infra_ca57_sel", ca57_parents, 0x0000, 2, 2),
+};
+
 static const struct mtk_composite top_muxes[] __initconst = {
 	/* CLK_CFG_0 */
 	MUX(CLK_TOP_AXI_SEL, "axi_sel", axi_parents, 0x0040, 0, 3),
@@ -760,6 +780,9 @@  static void __init mtk_infrasys_init(struct device_node *node)
 						clk_data);
 	mtk_clk_register_factors(infra_divs, ARRAY_SIZE(infra_divs), clk_data);
 
+	mtk_clk_register_cpumuxes(node, cpu_muxes, ARRAY_SIZE(cpu_muxes),
+						clk_data);
+
 	r = of_clk_add_provider(node, of_clk_src_onecell_get, clk_data);
 	if (r)
 		pr_err("%s(): could not register clock provider: %d\n",
diff --git a/include/dt-bindings/clock/mt8173-clk.h b/include/dt-bindings/clock/mt8173-clk.h
index fa2a2bb..a17cbaa 100644
--- a/include/dt-bindings/clock/mt8173-clk.h
+++ b/include/dt-bindings/clock/mt8173-clk.h
@@ -188,7 +188,9 @@ 
 #define CLK_INFRA_PMICSPI		10
 #define CLK_INFRA_PMICWRAP		11
 #define CLK_INFRA_CLK_13M		12
-#define CLK_INFRA_NR_CLK		13
+#define CLK_INFRA_CA53SEL		13
+#define CLK_INFRA_CA57SEL		14
+#define CLK_INFRA_NR_CLK		15
 
 /* PERI_SYS */