[PATCH v2 0/2] clk: ti: mux: resolve parent clocks by DT index, not by name

[PATCH v2 0/2] clk: ti: mux: resolve parent clocks by DT index, not by name

[Mathieu Dubois-Briand]

This commit aims to solve an issue I've been describing a few months ago
on AM335x SoC [1]. I believe using the parent_data field of
clk_init_data structure is now the preferred way to convey that data,
and it should be more reliable than string comparisons.

[1]: https://lore.kernel.org/all/DI4RUFQNSSNP.2QMSSQWJW9I2O@bootlin.com/

Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
---
Changes in v2:
- Also use clk_parent_data structure in composite.c.
- Rebase on v7.2-rc5.
- Link to v1: https://lore.kernel.org/r/20260715-mathieu-wdt-clock-theo-v1-1-da65bba1828b@bootlin.com

---
Mathieu Dubois-Briand (2):
      clk: ti: mux: resolve parent clocks by DT index, not by name
      clk: ti: composite: resolve parent clocks by DT index, not by name

 drivers/clk/ti/composite.c | 26 ++++++++++++++------------
 drivers/clk/ti/mux.c       | 20 +++++++++++---------
 2 files changed, 25 insertions(+), 21 deletions(-)
---
base-commit: f5098b6bae761e346ebcd9da7f95622c04733cff
change-id: 20260713-mathieu-wdt-clock-theo-f0c5ba58e258

Best regards,
-- 
Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>

[PATCH v2 1/2] clk: ti: mux: resolve parent clocks by DT index, not by name

[Mathieu Dubois-Briand]

Resolve parent clocks by their index into the device tree "clocks"
property rather than matching names as strings. Name-based matching is
fragile because a clock's "clock-output-names" value in its provider
node can differ from the name used to reference it in a consumer node,
and because names must be globally unique across all clock providers.

On AM335x, this caused broken clock trees where some clocks failed to
enable because their parents could not be found.

Replace of_clk_parent_fill() with a clk_parent_data array that sets
.index to the array position.

Fixes: ec7aa25fa483 ("ARM: dts: Use clock-output-names for am3")
Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
---
 drivers/clk/ti/mux.c | 20 +++++++++++---------
 1 file changed, 11 insertions(+), 9 deletions(-)

diff --git a/drivers/clk/ti/mux.c b/drivers/clk/ti/mux.c
index d6a0ccfd81db..ded4432f7528 100644
--- a/drivers/clk/ti/mux.c
+++ b/drivers/clk/ti/mux.c
@@ -119,7 +119,7 @@ const struct clk_ops ti_clk_mux_ops = {
 };
 
 static struct clk *_register_mux(struct device_node *node, const char *name,
-				 const char * const *parent_names,
+				 const struct clk_parent_data *parent_data,
 				 u8 num_parents, unsigned long flags,
 				 struct clk_omap_reg *reg, u8 shift, u32 mask,
 				 s8 latch, u8 clk_mux_flags, u32 *table)
@@ -136,7 +136,7 @@ static struct clk *_register_mux(struct device_node *node, const char *name,
 	init.name = name;
 	init.ops = &ti_clk_mux_ops;
 	init.flags = flags;
-	init.parent_names = parent_names;
+	init.parent_data = parent_data;
 	init.num_parents = num_parents;
 
 	/* struct clk_mux assignments */
@@ -167,24 +167,26 @@ static void of_mux_clk_setup(struct device_node *node)
 	struct clk *clk;
 	struct clk_omap_reg reg;
 	unsigned int num_parents;
-	const char **parent_names;
+	struct clk_parent_data *parent_data;
 	const char *name;
 	u8 clk_mux_flags = 0;
 	u32 mask = 0;
 	u32 shift = 0;
 	s32 latch = -EINVAL;
 	u32 flags = CLK_SET_RATE_NO_REPARENT;
+	int i;
 
 	num_parents = of_clk_get_parent_count(node);
 	if (num_parents < 2) {
 		pr_err("mux-clock %pOFn must have parents\n", node);
 		return;
 	}
-	parent_names = kcalloc(num_parents, sizeof(char *), GFP_KERNEL);
-	if (!parent_names)
-		goto cleanup;
+	parent_data = kcalloc(num_parents, sizeof(*parent_data), GFP_KERNEL);
+	if (!parent_data)
+		return;
 
-	of_clk_parent_fill(node, parent_names, num_parents);
+	for (i = 0; i < num_parents; i++)
+		parent_data[i].index = i;
 
 	if (ti_clk_get_reg_addr(node, 0, &reg))
 		goto cleanup;
@@ -207,7 +209,7 @@ static void of_mux_clk_setup(struct device_node *node)
 	mask = (1 << fls(mask)) - 1;
 
 	name = ti_dt_clk_name(node);
-	clk = _register_mux(node, name, parent_names, num_parents,
+	clk = _register_mux(node, name, parent_data, num_parents,
 			    flags, &reg, shift, mask, latch, clk_mux_flags,
 			    NULL);
 
@@ -215,7 +217,7 @@ static void of_mux_clk_setup(struct device_node *node)
 		of_clk_add_provider(node, of_clk_src_simple_get, clk);
 
 cleanup:
-	kfree(parent_names);
+	kfree(parent_data);
 }
 CLK_OF_DECLARE(mux_clk, "ti,mux-clock", of_mux_clk_setup);
 

-- 
2.47.3

Re: [PATCH v2 1/2] clk: ti: mux: resolve parent clocks by DT index, not by name

[Brian Masney]

On Mon, Jul 27, 2026 at 09:41:40AM +0200, Mathieu Dubois-Briand wrote:
> Resolve parent clocks by their index into the device tree "clocks"
> property rather than matching names as strings. Name-based matching is
> fragile because a clock's "clock-output-names" value in its provider
> node can differ from the name used to reference it in a consumer node,
> and because names must be globally unique across all clock providers.
> 
> On AM335x, this caused broken clock trees where some clocks failed to
> enable because their parents could not be found.
> 
> Replace of_clk_parent_fill() with a clk_parent_data array that sets
> .index to the array position.
> 
> Fixes: ec7aa25fa483 ("ARM: dts: Use clock-output-names for am3")
> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>

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

Re: [PATCH v2 1/2] clk: ti: mux: resolve parent clocks by DT index, not by name

[Stephen Boyd]

Quoting Mathieu Dubois-Briand (2026-07-27 00:41:40)
> Resolve parent clocks by their index into the device tree "clocks"
> property rather than matching names as strings. Name-based matching is
> fragile because a clock's "clock-output-names" value in its provider
> node can differ from the name used to reference it in a consumer node,
> and because names must be globally unique across all clock providers.
> 
> On AM335x, this caused broken clock trees where some clocks failed to
> enable because their parents could not be found.
> 
> Replace of_clk_parent_fill() with a clk_parent_data array that sets
> .index to the array position.
> 
> Fixes: ec7aa25fa483 ("ARM: dts: Use clock-output-names for am3")
> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
> ---

Applied to clk-next

Re: [PATCH v2 1/2] clk: ti: mux: resolve parent clocks by DT index, not by name

[Geert Uytterhoeven]

Hi Mathieu,

On Mon, 27 Jul 2026 at 09:44, Mathieu Dubois-Briand
<mathieu.dubois-briand@bootlin.com> wrote:
> Resolve parent clocks by their index into the device tree "clocks"
> property rather than matching names as strings. Name-based matching is
> fragile because a clock's "clock-output-names" value in its provider
> node can differ from the name used to reference it in a consumer node,
> and because names must be globally unique across all clock providers.
>
> On AM335x, this caused broken clock trees where some clocks failed to
> enable because their parents could not be found.
>
> Replace of_clk_parent_fill() with a clk_parent_data array that sets
> .index to the array position.
>
> Fixes: ec7aa25fa483 ("ARM: dts: Use clock-output-names for am3")
> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>

Thanks for your patch, which is now commit 667f420c09f1417c ("clk: ti:
mux: resolve parent clocks by DT index, not by name") in clk/clk-next.

This breaks booting BeagleBone Black for me: I don't get any output
on the serial console, even with "earlycon keep_bootcon".  When the
watchdog kicks in (or on kernel panic?), the system reboots.

Reverting this commit fixes the issue.
After reverting, I added debug code to print all parents, and compared
them to the DTB, but didn't see any discrepancies.

Your similar change to drivers/clk/ti/composite.c does not cause any
issues for me.

Do you have a clue?
Thanks!

> --- a/drivers/clk/ti/mux.c
> +++ b/drivers/clk/ti/mux.c
> @@ -119,7 +119,7 @@ const struct clk_ops ti_clk_mux_ops = {
>  };
>
>  static struct clk *_register_mux(struct device_node *node, const char *name,
> -                                const char * const *parent_names,
> +                                const struct clk_parent_data *parent_data,
>                                  u8 num_parents, unsigned long flags,
>                                  struct clk_omap_reg *reg, u8 shift, u32 mask,
>                                  s8 latch, u8 clk_mux_flags, u32 *table)
> @@ -136,7 +136,7 @@ static struct clk *_register_mux(struct device_node *node, const char *name,
>         init.name = name;
>         init.ops = &ti_clk_mux_ops;
>         init.flags = flags;
> -       init.parent_names = parent_names;
> +       init.parent_data = parent_data;
>         init.num_parents = num_parents;
>
>         /* struct clk_mux assignments */
> @@ -167,24 +167,26 @@ static void of_mux_clk_setup(struct device_node *node)
>         struct clk *clk;
>         struct clk_omap_reg reg;
>         unsigned int num_parents;
> -       const char **parent_names;
> +       struct clk_parent_data *parent_data;
>         const char *name;
>         u8 clk_mux_flags = 0;
>         u32 mask = 0;
>         u32 shift = 0;
>         s32 latch = -EINVAL;
>         u32 flags = CLK_SET_RATE_NO_REPARENT;
> +       int i;
>
>         num_parents = of_clk_get_parent_count(node);
>         if (num_parents < 2) {
>                 pr_err("mux-clock %pOFn must have parents\n", node);
>                 return;
>         }
> -       parent_names = kcalloc(num_parents, sizeof(char *), GFP_KERNEL);
> -       if (!parent_names)
> -               goto cleanup;
> +       parent_data = kcalloc(num_parents, sizeof(*parent_data), GFP_KERNEL);
> +       if (!parent_data)
> +               return;
>
> -       of_clk_parent_fill(node, parent_names, num_parents);
> +       for (i = 0; i < num_parents; i++)
> +               parent_data[i].index = i;
>
>         if (ti_clk_get_reg_addr(node, 0, &reg))
>                 goto cleanup;
> @@ -207,7 +209,7 @@ static void of_mux_clk_setup(struct device_node *node)
>         mask = (1 << fls(mask)) - 1;
>
>         name = ti_dt_clk_name(node);
> -       clk = _register_mux(node, name, parent_names, num_parents,
> +       clk = _register_mux(node, name, parent_data, num_parents,
>                             flags, &reg, shift, mask, latch, clk_mux_flags,
>                             NULL);
>
> @@ -215,7 +217,7 @@ static void of_mux_clk_setup(struct device_node *node)
>                 of_clk_add_provider(node, of_clk_src_simple_get, clk);
>
>  cleanup:
> -       kfree(parent_names);
> +       kfree(parent_data);
>  }
>  CLK_OF_DECLARE(mux_clk, "ti,mux-clock", of_mux_clk_setup);

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 v2 1/2] clk: ti: mux: resolve parent clocks by DT index, not by name

[Brian Masney]

Hi Stephen / Geert,

On Tue, Aug 18, 2026 at 06:02:56PM +0200, Geert Uytterhoeven wrote:
> On Mon, 27 Jul 2026 at 09:44, Mathieu Dubois-Briand
> <mathieu.dubois-briand@bootlin.com> wrote:
> > Resolve parent clocks by their index into the device tree "clocks"
> > property rather than matching names as strings. Name-based matching is
> > fragile because a clock's "clock-output-names" value in its provider
> > node can differ from the name used to reference it in a consumer node,
> > and because names must be globally unique across all clock providers.
> >
> > On AM335x, this caused broken clock trees where some clocks failed to
> > enable because their parents could not be found.
> >
> > Replace of_clk_parent_fill() with a clk_parent_data array that sets
> > .index to the array position.
> >
> > Fixes: ec7aa25fa483 ("ARM: dts: Use clock-output-names for am3")
> > Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
> 
> Thanks for your patch, which is now commit 667f420c09f1417c ("clk: ti:
> mux: resolve parent clocks by DT index, not by name") in clk/clk-next.
> 
> This breaks booting BeagleBone Black for me: I don't get any output
> on the serial console, even with "earlycon keep_bootcon".  When the
> watchdog kicks in (or on kernel panic?), the system reboots.
> 
> Reverting this commit fixes the issue.
> After reverting, I added debug code to print all parents, and compared
> them to the DTB, but didn't see any discrepancies.
> 
> Your similar change to drivers/clk/ti/composite.c does not cause any
> issues for me.
> 
> Do you have a clue?

Stephen: I think the safest thing is to just drop / revert the two
patches in this series in the pull to Linus next week. I'd do it now,
however I'm still waiting for my kernel.org account to get created,
and don't have write access to the clk tree yet.

Mathieu: Feel free to post an updated version, and we can get it merged
earlier in the development cycle so it'll get more testing in next.
Hopefully Geert will have time to test an updated version.

Brian

Re: [PATCH v2 1/2] clk: ti: mux: resolve parent clocks by DT index, not by name

[Mathieu Dubois-Briand]

Hi Geert,

On Tue Aug 18, 2026 at 6:02 PM CEST, Geert Uytterhoeven wrote:
> Hi Mathieu,
>
> On Mon, 27 Jul 2026 at 09:44, Mathieu Dubois-Briand
> <mathieu.dubois-briand@bootlin.com> wrote:
>> Resolve parent clocks by their index into the device tree "clocks"
>> property rather than matching names as strings. Name-based matching is
>> fragile because a clock's "clock-output-names" value in its provider
>> node can differ from the name used to reference it in a consumer node,
>> and because names must be globally unique across all clock providers.
>>
>> On AM335x, this caused broken clock trees where some clocks failed to
>> enable because their parents could not be found.
>>
>> Replace of_clk_parent_fill() with a clk_parent_data array that sets
>> .index to the array position.
>>
>> Fixes: ec7aa25fa483 ("ARM: dts: Use clock-output-names for am3")
>> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
>
> Thanks for your patch, which is now commit 667f420c09f1417c ("clk: ti:
> mux: resolve parent clocks by DT index, not by name") in clk/clk-next.
>
> This breaks booting BeagleBone Black for me: I don't get any output
> on the serial console, even with "earlycon keep_bootcon".  When the
> watchdog kicks in (or on kernel panic?), the system reboots.
>

Sorry for that!

It's a bit unexpected, as the beaglebone black was one of my test
platforms and I didn't see any issue.

> Reverting this commit fixes the issue.
> After reverting, I added debug code to print all parents, and compared
> them to the DTB, but didn't see any discrepancies.
>
> Your similar change to drivers/clk/ti/composite.c does not cause any
> issues for me.
>
> Do you have a clue?
> Thanks!

Can you describe what bootloader and device tree you are using for your
tests? Any specific configuration?

On my side, I have:
- BeagleBone Black Wireless A5
- am335x-boneblack-wireless.dtb
- U-boot 2025.10-ti-g6825d60bea17. Maybe I should try with u-boot master
  branch.

Thanks,
Mathieu

-- 
Mathieu Dubois-Briand, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

Re: [PATCH v2 1/2] clk: ti: mux: resolve parent clocks by DT index, not by name

[Geert Uytterhoeven]

Hi Mathieu,

On Wed, 19 Aug 2026 at 07:39, Mathieu Dubois-Briand
<mathieu.dubois-briand@bootlin.com> wrote:
> On Tue Aug 18, 2026 at 6:02 PM CEST, Geert Uytterhoeven wrote:
> > On Mon, 27 Jul 2026 at 09:44, Mathieu Dubois-Briand
> > <mathieu.dubois-briand@bootlin.com> wrote:
> >> Resolve parent clocks by their index into the device tree "clocks"
> >> property rather than matching names as strings. Name-based matching is
> >> fragile because a clock's "clock-output-names" value in its provider
> >> node can differ from the name used to reference it in a consumer node,
> >> and because names must be globally unique across all clock providers.
> >>
> >> On AM335x, this caused broken clock trees where some clocks failed to
> >> enable because their parents could not be found.
> >>
> >> Replace of_clk_parent_fill() with a clk_parent_data array that sets
> >> .index to the array position.
> >>
> >> Fixes: ec7aa25fa483 ("ARM: dts: Use clock-output-names for am3")
> >> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
> >
> > Thanks for your patch, which is now commit 667f420c09f1417c ("clk: ti:
> > mux: resolve parent clocks by DT index, not by name") in clk/clk-next.
> >
> > This breaks booting BeagleBone Black for me: I don't get any output
> > on the serial console, even with "earlycon keep_bootcon".  When the
> > watchdog kicks in (or on kernel panic?), the system reboots.
>
> Sorry for that!
>
> It's a bit unexpected, as the beaglebone black was one of my test
> platforms and I didn't see any issue.
>
> > Reverting this commit fixes the issue.
> > After reverting, I added debug code to print all parents, and compared
> > them to the DTB, but didn't see any discrepancies.
> >
> > Your similar change to drivers/clk/ti/composite.c does not cause any
> > issues for me.
> >
> > Do you have a clue?
> > Thanks!
>
> Can you describe what bootloader and device tree you are using for your
> tests? Any specific configuration?
>
> On my side, I have:
> - BeagleBone Black Wireless A5
> - am335x-boneblack-wireless.dtb
> - U-boot 2025.10-ti-g6825d60bea17. Maybe I should try with u-boot master
>   branch.

BeagleBone Black Rev C3 (by Seeed Studio)
am335x-boneblack.dtb
Stock U-Boot SPL 2019.04-00002-g31a8ae0206 (May 13 2020 - 09:26:17 -0500)

I am using my own config.

With omap2plus_defconfig and multi_v7_defconfig, it does boot.  However,
this commit does have an impact on /sys/kernel/debug/clk/clk_summary:

--- summary-7.2.0-rc2-00106-g0d4d262c1664       2026-08-19
10:56:51.622275674 +0200
+++ summary-7.2.0-rc2-00107-g667f420c09f1       2026-08-19
10:58:13.929823890 +0200
@@ -77,10 +77,14 @@
 clk_summary:

deviceless                      no_connection_id
 clk_summary:                clkdiv32k_ck         0       1        0
     32786       0          0     50000      Y
deviceless                      of_clk_get_from_provider
 clk_summary:

deviceless                      no_connection_id
-clk_summary:                   clk-24mhz-clkctrl:0000:0 0       4
   0        32786       0          0     50000      N
   deviceless                      clkdiv32k_ick
+clk_summary:                   clk-24mhz-clkctrl:0000:0 0       5
   0        32786       0          0     50000      N
   deviceless                      clkdiv32k_ick
 clk_summary:

deviceless                      of_clk_get_from_provider
 clk_summary:

deviceless                      of_clk_get_from_provider
 clk_summary:

deviceless                      no_connection_id
+clk_summary:                      wdt1_fck       0       1        0
     32786       0          0     50000      Y
deviceless                      wdt1_fck
+clk_summary:

deviceless                      no_connection_id
+clk_summary:                         l4-wkup-clkctrl:00d4:0 0       1
       0        32786       0          0     50000      N
             44e35000.target-module          fck
+clk_summary:

  deviceless                      no_connection_id
 clk_summary:                      l4-rtc-clkctrl:0000:0 0       1
   0        32786       0          0     50000      N
      44e3e074.target-module          fck
 clk_summary:

deviceless                      no_connection_id
 clk_summary:                      l4ls-clkctrl:0074:18 0       2
  0        32786       0          0     50000      N
     4804c000.gpio                   dbclk
@@ -297,17 +301,13 @@
 clk_summary:       l4-cefuse-clkctrl:0020:0      0       0        0
     24000000    0          0     50000      N         deviceless
                no_connection_id
 clk_summary: virt_19200000_ck                    0       0        0
     19200000    0          0     50000      Y   deviceless
          of_clk_get_from_provider
 clk_summary:
                                                 deviceless
          no_connection_id
-clk_summary: clk_rc32k_ck                        0       2        0
     32000       0          0     50000      Y   deviceless
          of_clk_get_from_provider
+clk_summary: clk_rc32k_ck                        0       1        0
     32000       0          0     50000      Y   deviceless
          of_clk_get_from_provider
 clk_summary:
                                                 deviceless
          no_connection_id
 clk_summary:    gpio0_dbclk_mux_ck               0       1        0
     32000       0          0     50000      Y      deviceless
             no_connection_id
 clk_summary:       l4-wkup-clkctrl:0008:18       0       2        0
     32000       0          0     50000      N         44e07000.gpio
                dbclk
 clk_summary:

44e07000.target-module          dbclk
 clk_summary:
                                                       deviceless
                of_clk_get_from_provider
 clk_summary:
                                                       deviceless
                no_connection_id
-clk_summary:    wdt1_fck                         0       1        0
     32000       0          0     50000      Y      deviceless
             wdt1_fck
-clk_summary:
                                                    deviceless
             no_connection_id
-clk_summary:       l4-wkup-clkctrl:00d4:0        0       1        0
     32000       0          0     50000      N
44e35000.target-module          fck
-clk_summary:
                                                       deviceless
                no_connection_id
 clk_summary: clk_32768_ck                        1       1        0
     32768       0          0     50000      Y   deviceless
          of_clk_get_from_provider
 clk_summary:
                                                 deviceless
          no_connection_id
 clk_summary:    sysclkout_pre_ck                 1       1        0
     32768       0          0     50000      Y      deviceless
             no_connection_id

I.e. wdt1_fck changed its parent from clk_rc32k_ck to
clk-24mhz-clkctrl:0000:0, which is probably not intended?

However, that is not the cause of the crash.  My .config has
CONFIG_INIT_STACK_ALL_PATTERN enabled, so any uninitialized local
variables contain garbage.

> @@ -136,7 +136,7 @@ static struct clk *_register_mux(struct device_node *node, const char *name,
>         init.name = name;
>         init.ops = &ti_clk_mux_ops;
>         init.flags = flags;
> -       init.parent_names = parent_names;
> +       init.parent_data = parent_data;
>         init.num_parents = num_parents;
>
>         /* struct clk_mux assignments

Init is not fully initialized, hence init.parent_names contains garbage
(before, .parent_data contained garbage).
As clk_core_populate_parent_map() prioritizes .parent_names over
.parent_data, the uninitialized .parent_data didn't cause any harm before.

I will send a patch to fix this.  This seems to be a recurring issue:
lots of clk_init_data structures are not fully initialized...

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

[PATCH v2 2/2] clk: ti: composite: resolve parent clocks by DT index, not by name

[Mathieu Dubois-Briand]

Resolve parent clocks by their index into the device tree "clocks"
property rather than matching names as strings. This makes it consistent
with other parts of the same driver.

Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
---
 drivers/clk/ti/composite.c | 26 ++++++++++++++------------
 1 file changed, 14 insertions(+), 12 deletions(-)

diff --git a/drivers/clk/ti/composite.c b/drivers/clk/ti/composite.c
index c379bbdae25a..01eae8995254 100644
--- a/drivers/clk/ti/composite.c
+++ b/drivers/clk/ti/composite.c
@@ -52,7 +52,7 @@ static const struct clk_ops ti_composite_gate_ops = {
 
 struct component_clk {
 	int num_parents;
-	const char **parent_names;
+	struct clk_parent_data *parent_data;
 	struct device_node *node;
 	int type;
 	struct clk_hw *hw;
@@ -116,7 +116,7 @@ static void __init _register_composite(void *user,
 	struct clk_hw_omap_comp *cclk = to_clk_hw_comp(hw);
 	struct component_clk *comp;
 	int num_parents = 0;
-	const char **parent_names = NULL;
+	struct clk_parent_data *parent_data = NULL;
 	const char *name;
 	int i;
 	int ret;
@@ -155,7 +155,7 @@ static void __init _register_composite(void *user,
 			continue;
 		if (comp->num_parents) {
 			num_parents = comp->num_parents;
-			parent_names = comp->parent_names;
+			parent_data = comp->parent_data;
 			break;
 		}
 	}
@@ -166,8 +166,8 @@ static void __init _register_composite(void *user,
 	}
 
 	name = ti_dt_clk_name(node);
-	clk = clk_register_composite(NULL, name,
-				     parent_names, num_parents,
+	clk = clk_register_composite_pdata(NULL, name,
+				     parent_data, num_parents,
 				     _get_hw(cclk, CLK_COMPONENT_TYPE_MUX),
 				     &ti_clk_mux_ops,
 				     _get_hw(cclk, CLK_COMPONENT_TYPE_DIVIDER),
@@ -190,7 +190,7 @@ static void __init _register_composite(void *user,
 		if (!cclk->comp_clks[i])
 			continue;
 		list_del(&cclk->comp_clks[i]->link);
-		kfree(cclk->comp_clks[i]->parent_names);
+		kfree(cclk->comp_clks[i]->parent_data);
 		kfree(cclk->comp_clks[i]);
 	}
 
@@ -237,8 +237,9 @@ int __init ti_clk_add_component(struct device_node *node, struct clk_hw *hw,
 				int type)
 {
 	unsigned int num_parents;
-	const char **parent_names;
+	struct clk_parent_data *parent_data;
 	struct component_clk *clk;
+	unsigned int i;
 
 	num_parents = of_clk_get_parent_count(node);
 
@@ -247,20 +248,21 @@ int __init ti_clk_add_component(struct device_node *node, struct clk_hw *hw,
 		return -EINVAL;
 	}
 
-	parent_names = kcalloc(num_parents, sizeof(char *), GFP_KERNEL);
-	if (!parent_names)
+	parent_data = kcalloc(num_parents, sizeof(*parent_data), GFP_KERNEL);
+	if (!parent_data)
 		return -ENOMEM;
 
-	of_clk_parent_fill(node, parent_names, num_parents);
+	for (i = 0; i < num_parents; i++)
+		parent_data[i].index = i;
 
 	clk = kzalloc_obj(*clk);
 	if (!clk) {
-		kfree(parent_names);
+		kfree(parent_data);
 		return -ENOMEM;
 	}
 
 	clk->num_parents = num_parents;
-	clk->parent_names = parent_names;
+	clk->parent_data = parent_data;
 	clk->hw = hw;
 	clk->node = node;
 	clk->type = type;

-- 
2.47.3

Re: [PATCH v2 2/2] clk: ti: composite: resolve parent clocks by DT index, not by name

[Brian Masney]

On Mon, Jul 27, 2026 at 09:41:41AM +0200, Mathieu Dubois-Briand wrote:
> Resolve parent clocks by their index into the device tree "clocks"
> property rather than matching names as strings. This makes it consistent
> with other parts of the same driver.
> 
> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>

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

Re: [PATCH v2 2/2] clk: ti: composite: resolve parent clocks by DT index, not by name

[Stephen Boyd]

Quoting Mathieu Dubois-Briand (2026-07-27 00:41:41)
> Resolve parent clocks by their index into the device tree "clocks"
> property rather than matching names as strings. This makes it consistent
> with other parts of the same driver.
> 
> Signed-off-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
> ---

Applied to clk-next