mfd: omap-usb-host: fix OF populate on driver rebind

PATCH Applied Johan Hovold 3 participants 10 messages Dec 19, 2025 – Jan 20

Subsystems (3)

Stable branch Multifunction devices (MFD) OMAP2+ support

lore

Share & tools

lore

Johan Hovold

author

2025-12-19 11:08 UTC #0 lore raw

PATCH 1/1 mfd: omap-usb-host: fix OF populate on driver rebind

Since commit c6e126de43e7 ("of: Keep track of populated platform devices") child devices will not be created by of_platform_populate() if the devices had previously been deregistered individually so that the OF_POPULATED flag is still set in the corresponding OF nodes.

Switch to using of_platform_depopulate() instead of open coding so that the child devices are created if the driver is rebound.

Fixes c6e126de43e7 ("of: Keep track of populated platform devices")

Cc [email protected] # 3.16

Signed-off-by Johan Hovold <[email protected]>

---
 drivers/mfd/omap-usb-host.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

drivers/mfd/omap-usb-host.c +4 -2

index a77b6fc790f2..4d29a6e2ed87 100644
@@ -819,8 +819,10 @@ static void usbhs_omap_remove(struct platform_device *pdev)
 {
 	pm_runtime_disable(&pdev->dev);
 
-	/* remove children */
-	device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);
+	if (pdev->dev.of_node)
+		of_platform_depopulate(&pdev->dev);
+	else
+		device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);
 }
 
 static const struct dev_pm_ops usbhsomap_dev_pm_ops = {

Signature · 1 line

2.51.2

Andreas Kemnade

maintainer Reviewed-by on v1 1/1

2025-12-29 15:04 UTC #1 lore raw

Quoted 11 lines · On Fri, 19 Dec 2025 12:07:14 +0100 Johan Hovold <[email protected]> wrote:

Since commit c6e126de43e7 ("of: Keep track of populated platform devices") child devices will not be created by of_platform_populate() if the devices had previously been deregistered individually so that the OF_POPULATED flag is still set in the corresponding OF nodes.
Switch to using of_platform_depopulate() instead of open coding so that the child devices are created if the driver is rebound.
Fixes c6e126de43e7 ("of: Keep track of populated platform devices")
Cc [email protected] # 3.16
Signed-off-by Johan Hovold <[email protected]>

Reviewed-by Andreas Kemnade <[email protected]>

Lee Jones

maintainer

2026-01-09 16:37 UTC #2 lore raw

In reply to Johan Hovold

Quoted 20 lines · Johan Hovold

Since commit c6e126de43e7 ("of: Keep track of populated platform devices") child devices will not be created by of_platform_populate() if the devices had previously been deregistered individually so that the OF_POPULATED flag is still set in the corresponding OF nodes.
Switch to using of_platform_depopulate() instead of open coding so that the child devices are created if the driver is rebound.
Fixes c6e126de43e7 ("of: Keep track of populated platform devices")
Cc [email protected] # 3.16
Signed-off-by Johan Hovold <[email protected]>
---
 drivers/mfd/omap-usb-host.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/mfd/omap-usb-host.c +0 -0
index a77b6fc790f2..4d29a6e2ed87 100644

drivers/mfd/omap-usb-host.c static void usbhs_omap_remove(struct platform_device *pdev) · line 823

Quoted 7 lines

drivers/mfd/omap-usb-host.c +2 -2

@@ -819,8 +819,10 @@ static void usbhs_omap_remove(struct platform_device *pdev)
 {
 	pm_runtime_disable(&pdev->dev);
 
-	/* remove children */
-	device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);
+	if (pdev->dev.of_node)
+		of_platform_depopulate(&pdev->dev);

devm_of_platform_populate()?

drivers/mfd/omap-usb-host.c static void usbhs_omap_remove(struct platform_device *pdev) · line 825

Quoted 2 lines

drivers/mfd/omap-usb-host.c +2 -0

+	else
+		device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);

What on earth is this driver even doing in MFD?

Nightmare - quick dig, bury it!

Signature · 1 line

Lee Jones [李琼斯]

Lee Jones

maintainer

2026-01-09 16:39 UTC #3 lore raw

Quoted 32 lines · Lee Jones

Quoted 20 lines · Johan Hovold
Since commit c6e126de43e7 ("of: Keep track of populated platform devices") child devices will not be created by of_platform_populate() if the devices had previously been deregistered individually so that the OF_POPULATED flag is still set in the corresponding OF nodes.
Switch to using of_platform_depopulate() instead of open coding so that the child devices are created if the driver is rebound.
Fixes c6e126de43e7 ("of: Keep track of populated platform devices")
Cc [email protected] # 3.16
Signed-off-by Johan Hovold <[email protected]>
---
 drivers/mfd/omap-usb-host.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/mfd/omap-usb-host.c +0 -0
index a77b6fc790f2..4d29a6e2ed87 100644
drivers/mfd/omap-usb-host.c static void usbhs_omap_remove(struct platform_device *pdev) · line 823
Quoted 7 lines
drivers/mfd/omap-usb-host.c +2 -2
@@ -819,8 +819,10 @@ static void usbhs_omap_remove(struct platform_device *pdev)
 {
 	pm_runtime_disable(&pdev->dev);
 
-	/* remove children */
-	device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);
+	if (pdev->dev.of_node)
+		of_platform_depopulate(&pdev->dev);
devm_of_platform_populate()?

Quoted 6 lines

drivers/mfd/omap-usb-host.c static void usbhs_omap_remove(struct platform_device *pdev) · line 825
Quoted 2 lines
drivers/mfd/omap-usb-host.c +2 -0
+	else
+		device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);
What on earth is this driver even doing in MFD?
Nightmare - quick dig, bury it!

Is the old platform method even supported anymore?

$ git grep \"usbhs_omap\""
drivers/mfd/omap-usb-host.c:#define USBHS_DRIVER_NAME   "usbhs_omap"

Signature · 1 line

Lee Jones [李琼斯]

Andreas Kemnade

maintainer

2026-01-09 17:18 UTC #4 lore raw

Quoted 39 lines · On Fri, 9 Jan 2026 16:39:39 +0000 Lee Jones <[email protected]> wrote:

Quoted 37 lines · Lee Jones
Quoted 20 lines · Johan Hovold
Since commit c6e126de43e7 ("of: Keep track of populated platform devices") child devices will not be created by of_platform_populate() if the devices had previously been deregistered individually so that the OF_POPULATED flag is still set in the corresponding OF nodes.
Switch to using of_platform_depopulate() instead of open coding so that the child devices are created if the driver is rebound.
Fixes c6e126de43e7 ("of: Keep track of populated platform devices")
Cc [email protected] # 3.16
Signed-off-by Johan Hovold <[email protected]>
---
 drivers/mfd/omap-usb-host.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/mfd/omap-usb-host.c +0 -0
index a77b6fc790f2..4d29a6e2ed87 100644
drivers/mfd/omap-usb-host.c static void usbhs_omap_remove(struct platform_device *pdev) · line 823
Quoted 7 lines
drivers/mfd/omap-usb-host.c +2 -2
@@ -819,8 +819,10 @@ static void usbhs_omap_remove(struct platform_device *pdev)
 {
 	pm_runtime_disable(&pdev->dev);
 
-	/* remove children */
-	device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);
+	if (pdev->dev.of_node)
+		of_platform_depopulate(&pdev->dev);  
devm_of_platform_populate()?
drivers/mfd/omap-usb-host.c static void usbhs_omap_remove(struct platform_device *pdev) · line 825
Quoted 2 lines
drivers/mfd/omap-usb-host.c +2 -0
+	else
+		device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);  
What on earth is this driver even doing in MFD?

Quoted 7 lines

Quoted 1 line
Nightmare - quick dig, bury it!
Is the old platform method even supported anymore?
$ git grep \"usbhs_omap\""
drivers/mfd/omap-usb-host.c:#define USBHS_DRIVER_NAME   "usbhs_omap"

:~/linux/arch/arm/boot/dts/ti/omap$ grep ti,usbhs-host *.dtsi
omap3.dtsi:			compatible = "ti,usbhs-host";
omap4-l4.dtsi:				compatible = "ti,usbhs-host";
omap5-l4.dtsi:				compatible = "ti,usbhs-host";

So it is still in use.

Regards,
Andreas

Andreas Kemnade

maintainer

2026-01-09 17:37 UTC #5 lore raw

Quoted 49 lines · On Fri, 9 Jan 2026 18:18:39 +0100 Andreas Kemnade <[email protected]> wrote:

Quoted 39 lines · On Fri, 9 Jan 2026 16:39:39 +0000 Lee Jones <[email protected]> wrote:
Quoted 37 lines · Lee Jones
Quoted 20 lines · Johan Hovold
Since commit c6e126de43e7 ("of: Keep track of populated platform devices") child devices will not be created by of_platform_populate() if the devices had previously been deregistered individually so that the OF_POPULATED flag is still set in the corresponding OF nodes.
Switch to using of_platform_depopulate() instead of open coding so that the child devices are created if the driver is rebound.
Fixes c6e126de43e7 ("of: Keep track of populated platform devices")
Cc [email protected] # 3.16
Signed-off-by Johan Hovold <[email protected]>
---
 drivers/mfd/omap-usb-host.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/mfd/omap-usb-host.c +0 -0
index a77b6fc790f2..4d29a6e2ed87 100644
drivers/mfd/omap-usb-host.c static void usbhs_omap_remove(struct platform_device *pdev) · line 823
Quoted 7 lines
drivers/mfd/omap-usb-host.c +2 -2
@@ -819,8 +819,10 @@ static void usbhs_omap_remove(struct platform_device *pdev)
 {
 	pm_runtime_disable(&pdev->dev);
 
-	/* remove children */
-	device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);
+	if (pdev->dev.of_node)
+		of_platform_depopulate(&pdev->dev);    
devm_of_platform_populate()?
drivers/mfd/omap-usb-host.c static void usbhs_omap_remove(struct platform_device *pdev) · line 825
Quoted 2 lines
drivers/mfd/omap-usb-host.c +2 -0
+	else
+		device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);    
What on earth is this driver even doing in MFD?
Quoted 7 lines
Quoted 1 line
Nightmare - quick dig, bury it!
Is the old platform method even supported anymore?
$ git grep \"usbhs_omap\""
drivers/mfd/omap-usb-host.c:#define USBHS_DRIVER_NAME   "usbhs_omap"

Quoted 7 lines

:~/linux/arch/arm/boot/dts/ti/omap$ grep ti,usbhs-host *.dtsi
omap3.dtsi:			compatible = "ti,usbhs-host";
omap4-l4.dtsi:				compatible = "ti,usbhs-host";
omap5-l4.dtsi:				compatible = "ti,usbhs-host";

So it is still in use.

ok, that was just answering in panic mode to prevent removal of a needed driver. Reading it again: I think omap2/3/4/5 without devicetree will not work anyway, so we can remove anything done on !dev->of_node.

Regards,
Andreas

Lee Jones

maintainer

2026-01-09 17:49 UTC #6 lore raw

Quoted 57 lines · Andreas Kemnade

Quoted 54 lines · On Fri, 9 Jan 2026 18:18:39 +0100 Andreas Kemnade <[email protected]> wrote:
Quoted 39 lines · On Fri, 9 Jan 2026 16:39:39 +0000 Lee Jones <[email protected]> wrote:
Quoted 37 lines · Lee Jones
Quoted 20 lines · Johan Hovold
Since commit c6e126de43e7 ("of: Keep track of populated platform devices") child devices will not be created by of_platform_populate() if the devices had previously been deregistered individually so that the OF_POPULATED flag is still set in the corresponding OF nodes.
Switch to using of_platform_depopulate() instead of open coding so that the child devices are created if the driver is rebound.
Fixes c6e126de43e7 ("of: Keep track of populated platform devices")
Cc [email protected] # 3.16
Signed-off-by Johan Hovold <[email protected]>
---
 drivers/mfd/omap-usb-host.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/mfd/omap-usb-host.c +0 -0
index a77b6fc790f2..4d29a6e2ed87 100644
drivers/mfd/omap-usb-host.c static void usbhs_omap_remove(struct platform_device *pdev) · line 823
Quoted 7 lines
drivers/mfd/omap-usb-host.c +2 -2
@@ -819,8 +819,10 @@ static void usbhs_omap_remove(struct platform_device *pdev)
 {
 	pm_runtime_disable(&pdev->dev);
 
-	/* remove children */
-	device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);
+	if (pdev->dev.of_node)
+		of_platform_depopulate(&pdev->dev);    
devm_of_platform_populate()?
drivers/mfd/omap-usb-host.c static void usbhs_omap_remove(struct platform_device *pdev) · line 825
Quoted 2 lines
drivers/mfd/omap-usb-host.c +2 -0
+	else
+		device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);    
What on earth is this driver even doing in MFD?
Quoted 7 lines
Quoted 1 line
Nightmare - quick dig, bury it!
Is the old platform method even supported anymore?
$ git grep \"usbhs_omap\""
drivers/mfd/omap-usb-host.c:#define USBHS_DRIVER_NAME   "usbhs_omap"
:~/linux/arch/arm/boot/dts/ti/omap$ grep ti,usbhs-host *.dtsi
omap3.dtsi:			compatible = "ti,usbhs-host";
omap4-l4.dtsi:				compatible = "ti,usbhs-host";
omap5-l4.dtsi:				compatible = "ti,usbhs-host";

Quoted 5 lines

Quoted 2 lines
So it is still in use.
ok, that was just answering in panic mode to prevent removal of a needed driver. Reading it again: I think omap2/3/4/5 without devicetree will not work anyway, so we can remove anything done on !dev->of_node.

Okay, leave it with me.

Signature · 1 line

Lee Jones [李琼斯]

Lee Jones

maintainer

2026-01-09 17:43 UTC #7 lore raw

In reply to Andreas Kemnade

Quoted 46 lines · Andreas Kemnade

Quoted 43 lines · On Fri, 9 Jan 2026 16:39:39 +0000 Lee Jones <[email protected]> wrote:
Quoted 32 lines · Lee Jones
Quoted 20 lines · Johan Hovold
Since commit c6e126de43e7 ("of: Keep track of populated platform devices") child devices will not be created by of_platform_populate() if the devices had previously been deregistered individually so that the OF_POPULATED flag is still set in the corresponding OF nodes.
Switch to using of_platform_depopulate() instead of open coding so that the child devices are created if the driver is rebound.
Fixes c6e126de43e7 ("of: Keep track of populated platform devices")
Cc [email protected] # 3.16
Signed-off-by Johan Hovold <[email protected]>
---
 drivers/mfd/omap-usb-host.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/mfd/omap-usb-host.c +0 -0
index a77b6fc790f2..4d29a6e2ed87 100644
drivers/mfd/omap-usb-host.c static void usbhs_omap_remove(struct platform_device *pdev) · line 823
Quoted 7 lines
drivers/mfd/omap-usb-host.c +2 -2
@@ -819,8 +819,10 @@ static void usbhs_omap_remove(struct platform_device *pdev)
 {
 	pm_runtime_disable(&pdev->dev);
 
-	/* remove children */
-	device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);
+	if (pdev->dev.of_node)
+		of_platform_depopulate(&pdev->dev);  
devm_of_platform_populate()?
Quoted 6 lines
drivers/mfd/omap-usb-host.c static void usbhs_omap_remove(struct platform_device *pdev) · line 825
Quoted 2 lines
drivers/mfd/omap-usb-host.c +2 -0
+	else
+		device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);  
What on earth is this driver even doing in MFD?
Nightmare - quick dig, bury it!
Is the old platform method even supported anymore?

Quoted 7 lines

Quoted 3 lines

$ git grep \"usbhs_omap\""
drivers/mfd/omap-usb-host.c:#define USBHS_DRIVER_NAME   "usbhs_omap"

:~/linux/arch/arm/boot/dts/ti/omap$ grep ti,usbhs-host *.dtsi
omap3.dtsi:			compatible = "ti,usbhs-host";
omap4-l4.dtsi:				compatible = "ti,usbhs-host";
omap5-l4.dtsi:				compatible = "ti,usbhs-host";

That's the new "DT" method.

What about the old "usbhs_omap" one?

Looks to me like support was removed by Tony in 2016:

9080b8dc761a ("ARM: OMAP2+: Remove legacy usb-host.c platform init code")

Signature · 1 line

Lee Jones [李琼斯]

Johan Hovold

author

2026-01-13 09:37 UTC #8 lore raw

In reply to Lee Jones

Quoted 18 lines · Lee Jones

Quoted 16 lines · Johan Hovold
Since commit c6e126de43e7 ("of: Keep track of populated platform devices") child devices will not be created by of_platform_populate() if the devices had previously been deregistered individually so that the OF_POPULATED flag is still set in the corresponding OF nodes.
Switch to using of_platform_depopulate() instead of open coding so that the child devices are created if the driver is rebound.
Fixes c6e126de43e7 ("of: Keep track of populated platform devices")
Cc [email protected] # 3.16
Signed-off-by Johan Hovold <[email protected]>
drivers/mfd/omap-usb-host.c +0 -0
@@ -819,8 +819,10 @@ static void usbhs_omap_remove(struct platform_device *pdev)
 {
 	pm_runtime_disable(&pdev->dev);
 

Quoted 6 lines

drivers/mfd/omap-usb-host.c static void usbhs_omap_remove(struct platform_device *pdev) · line 823
Quoted 4 lines
drivers/mfd/omap-usb-host.c +2 -2
-	/* remove children */
-	device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);
+	if (pdev->dev.of_node)
+		of_platform_depopulate(&pdev->dev);
devm_of_platform_populate()?

Looks like that would work here, but I prefer using explicit removal for both the OF and non-OF for symmetry and readability reasons.

If you think that the non-OF support can be removed then a devres conversion can be done as part of that change.

Quoted 2 lines

drivers/mfd/omap-usb-host.c static void usbhs_omap_remove(struct platform_device *pdev) · line 825
Quoted 2 lines
drivers/mfd/omap-usb-host.c +2 -0
+	else
+		device_for_each_child(&pdev->dev, NULL, usbhs_omap_remove_child);

Johan

Lee Jones

maintainer Applied

2026-01-20 15:10 UTC #9 lore raw

Re: (subset) [PATCH] mfd: omap-usb-host: fix OF populate on driver rebind

In reply to Johan Hovold

Quoted 9 lines · Johan Hovold

Since commit c6e126de43e7 ("of: Keep track of populated platform devices") child devices will not be created by of_platform_populate() if the devices had previously been deregistered individually so that the OF_POPULATED flag is still set in the corresponding OF nodes.
Switch to using of_platform_depopulate() instead of open coding so that the child devices are created if the driver is rebound.
[...]

Applied, thanks!

[1/1] mfd: omap-usb-host: fix OF populate on driver rebind

      commit: 74283e1c49dd2d56e83bca070c163c56375c057d

--
Lee Jones [李琼斯]