rockchip: Fix devices suspend freeze on RK3568/RK3566
Subsystems (3)
v2 cover rockchip: Fix devices suspend freeze on RK3568/RK3566
This series fixes a system freeze during suspend in ohci_suspend() due to clk_usbphy1_480m not being enabled when EHCI/OHCI registers are accessed on e.g. a Raxa ROCK 3C board.
Following pm_test modes work on my ROCK 3C with the missing usbphy clk refs added:
echo N > /sys/module/printk/parameters/console_suspend
echo devices > /sys/power/pm_test
echo platform > /sys/power/pm_test
echo processors > /sys/power/pm_test
echo core > /sys/power/pm_test
echo mem > /sys/power/state
Changes in v2:
- Include rockchip,rk3588-ehci in the EHCI constraint
- Make clocks prop required for EHCI and OHCI
- Collect t-b tag
Jonas Karlman (2):
dt-bindings: usb: Add Rockchip RK3568 compatible for EHCI and OHCI
arm64: dts: rockchip: Fix devices suspend freeze on RK3568/RK3566
.../devicetree/bindings/usb/generic-ehci.yaml | 14 ++++++++++++++ .../devicetree/bindings/usb/generic-ohci.yaml | 7 ++++++- arch/arm64/boot/dts/rockchip/rk356x-base.dtsi | 16 ++++++++-------- 3 files changed, 28 insertions(+), 9 deletions(-)
Signature · 7 lines
2.54.0 _______________________________________________ Linux-rockchip mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-rockchip
PATCH v2 1/2 dt-bindings: usb: Add Rockchip RK3568 compatible for EHCI and OHCI
The Rockchip RK3568 EHCI/OHCI controller depends on clk_usbphy1_480m being enabled, or the system may freeze when registers are accessed.
Add Rockchip RK3568 EHCI and OHCI compatibles with a similar four-clock constraint as RK3588, also extend the EHCI constraint to include RK3588 to match similar requirements of RK3588.
---
Existing DTs for RK3568 use the plain generic-ehci/ohci compatible, next patch make use of these new compatibles and adds the missing clk_usbphy1_480m clock references.
Existing DTs for RK3588 have contained the required four clocks since the initial addition of the EHCI/OHCI nodes.
Changes in v2:
- Include rockchip,rk3588-ehci in the EHCI constraint
- Make clocks prop required for EHCI and OHCI
--- .../devicetree/bindings/usb/generic-ehci.yaml | 14 ++++++++++++++ .../devicetree/bindings/usb/generic-ohci.yaml | 7 ++++++- 2 files changed, 20 insertions(+), 1 deletion(-)
Documentation/devicetree/bindings/usb/generic-ehci.yaml +14 -0
@@ -52,6 +52,7 @@ properties: - ibm,476gtr-ehci - nxp,lpc1850-ehci - qca,ar7100-ehci + - rockchip,rk3568-ehci - rockchip,rk3588-ehci - snps,hsdk-v1.0-ehci - socionext,uniphier-ehci
@@ -186,6 +187,19 @@ allOf: required: - clocks - clock-names + - if: + properties: + compatible: + contains: + enum: + - rockchip,rk3568-ehci + - rockchip,rk3588-ehci + then: + properties: + clocks: + minItems: 4 + required: + - clocks unevaluatedProperties: false
Documentation/devicetree/bindings/usb/generic-ohci.yaml +6 -1
@@ -47,6 +47,7 @@ properties: - hpe,gxp-ohci - ibm,476gtr-ohci - ingenic,jz4740-ohci + - rockchip,rk3568-ohci - rockchip,rk3588-ohci - snps,hsdk-v1.0-ohci - const: generic-ohci
@@ -198,11 +199,15 @@ allOf: properties: compatible: contains: - const: rockchip,rk3588-ohci + enum: + - rockchip,rk3568-ohci + - rockchip,rk3588-ohci then: properties: clocks: minItems: 4 + required: + - clocks else: properties: clocks:
Signature · 1 line
2.54.0
Re: [PATCH v2 1/2] dt-bindings: usb: Add Rockchip RK3568 compatible for EHCI and OHCI
Quoted 43 lines · Jonas Karlman
The Rockchip RK3568 EHCI/OHCI controller depends on
clk_usbphy1_480mbeing enabled, or the system may freeze when registers are accessed.Add Rockchip RK3568 EHCI and OHCI compatibles with a similar four-clock constraint as RK3588, also extend the EHCI constraint to include RK3588 to match similar requirements of RK3588.
Signed-off-by Jonas Karlman <[email protected]>---
Existing DTs for RK3568 use the plain generic-ehci/ohci compatible, next patch make use of these new compatibles and adds the missingclk_usbphy1_480mclock references.Existing DTs for RK3588 have contained the required four clocks since the initial addition of the EHCI/OHCI nodes.
Changes in v2:
- Include rockchip,rk3588-ehci in the EHCI constraint
- Make clocks prop required for EHCI and OHCI--- .../devicetree/bindings/usb/generic-ehci.yaml | 14 ++++++++++++++ .../devicetree/bindings/usb/generic-ohci.yaml | 7 ++++++- 2 files changed, 20 insertions(+), 1 deletion(-)Documentation/devicetree/bindings/usb/generic-ehci.yaml +4
-0@@ -52,6 +52,7 @@ properties: - ibm,476gtr-ehci - nxp,lpc1850-ehci - qca,ar7100-ehci + - rockchip,rk3568-ehci - rockchip,rk3588-ehci - snps,hsdk-v1.0-ehci - socionext,uniphier-ehci@@ -186,6 +187,19 @@ allOf: required: - clocks - clock-names + - if: + properties: + compatible:
Documentation/devicetree/bindings/usb/generic-ehci.yaml allOf: · line 202
Quoted 10 lines
Documentation/devicetree/bindings/usb/generic-ehci.yaml +10
-0+ contains: + enum: + - rockchip,rk3568-ehci + - rockchip,rk3588-ehci + then: + properties: + clocks: + minItems: 4 + required: + - clocks
This is ABI break for RK3588, so your commit msg should also mention that RK3588 is not working for example. Otherwise you provided rationale only for breaking RK3568 ABI.
Same for OHCI part.
Best regards,
Krzysztof
PATCH v2 2/2 arm64: dts: rockchip: Fix devices suspend freeze on RK3568/RK3566
In reply to Jonas KarlmanThe EHCI/OHCI controller depends on clk_usbphy1_480m being enabled, or the system may freeze when registers are accessed, i.e. during suspend in ohci_suspend().
Add the missing clk_usbphy1_480m clocks reference to EHCI/OHCI controllers to ensure the clock is enabled when ECHI/OHCI registers are accessed to prevent a system freeze.
Fixes suspend pm_test issue with EHCI/OHCI devices due to the missing clk_usbphy1_480m reference and makes following pm_test modes work:
echo N > /sys/module/printk/parameters/console_suspend
echo devices > /sys/power/pm_test
echo platform > /sys/power/pm_test
echo processors > /sys/power/pm_test
echo core > /sys/power/pm_test
echo mem > /sys/power/state
---
v2: Collect t-b tag
--- arch/arm64/boot/dts/rockchip/rk356x-base.dtsi | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-)
arch/arm64/boot/dts/rockchip/rk356x-base.dtsi +8 -8
@@ -321,44 +321,44 @@ its: msi-controller@fd440000 { }; usb_host0_ehci: usb@fd800000 { - compatible = "generic-ehci"; + compatible = "rockchip,rk3568-ehci", "generic-ehci"; reg = <0x0 0xfd800000 0x0 0x40000>; interrupts = <GIC_SPI 130 IRQ_TYPE_LEVEL_HIGH>; clocks = <&cru HCLK_USB2HOST0>, <&cru HCLK_USB2HOST0_ARB>, - <&cru PCLK_USB>; + <&cru PCLK_USB>, <&usb2phy1>; phys = <&usb2phy1_otg>; phy-names = "usb"; status = "disabled"; }; usb_host0_ohci: usb@fd840000 { - compatible = "generic-ohci"; + compatible = "rockchip,rk3568-ohci", "generic-ohci"; reg = <0x0 0xfd840000 0x0 0x40000>; interrupts = <GIC_SPI 131 IRQ_TYPE_LEVEL_HIGH>; clocks = <&cru HCLK_USB2HOST0>, <&cru HCLK_USB2HOST0_ARB>, - <&cru PCLK_USB>; + <&cru PCLK_USB>, <&usb2phy1>; phys = <&usb2phy1_otg>; phy-names = "usb"; status = "disabled"; }; usb_host1_ehci: usb@fd880000 { - compatible = "generic-ehci"; + compatible = "rockchip,rk3568-ehci", "generic-ehci"; reg = <0x0 0xfd880000 0x0 0x40000>; interrupts = <GIC_SPI 133 IRQ_TYPE_LEVEL_HIGH>; clocks = <&cru HCLK_USB2HOST1>, <&cru HCLK_USB2HOST1_ARB>, - <&cru PCLK_USB>; + <&cru PCLK_USB>, <&usb2phy1>; phys = <&usb2phy1_host>; phy-names = "usb"; status = "disabled"; }; usb_host1_ohci: usb@fd8c0000 { - compatible = "generic-ohci"; + compatible = "rockchip,rk3568-ohci", "generic-ohci"; reg = <0x0 0xfd8c0000 0x0 0x40000>; interrupts = <GIC_SPI 134 IRQ_TYPE_LEVEL_HIGH>; clocks = <&cru HCLK_USB2HOST1>, <&cru HCLK_USB2HOST1_ARB>, - <&cru PCLK_USB>; + <&cru PCLK_USB>, <&usb2phy1>; phys = <&usb2phy1_host>; phy-names = "usb"; status = "disabled";
Signature · 1 line
2.54.0
Re: [PATCH v2 2/2] arm64: dts: rockchip: Fix devices suspend freeze on RK3568/RK3566
Hi Jonas,
Quoted 18 lines · Jonas Karlman
The EHCI/OHCI controller depends on
clk_usbphy1_480mbeing enabled, or the system may freeze when registers are accessed, i.e. during suspend inohci_suspend().Add the missing
clk_usbphy1_480mclocks reference to EHCI/OHCI controllers to ensure the clock is enabled when ECHI/OHCI registers are accessed to prevent a system freeze.Fixes suspend
pm_testissue with EHCI/OHCI devices due to the missingclk_usbphy1_480mreference and makes followingpm_testmodes work:echo N >
/sys/module/printk/parameters/console_suspendecho devices >
/sys/power/pm_test
echo platform >/sys/power/pm_test
echo processors >/sys/power/pm_test
echo core >/sys/power/pm_test
Quoted 6 lines
echo mem >
/sys/power/stateFixes 91c4c3e06a25 ("arm64: dts: rockchip: add usb2 nodes to rk3568 device tree")Fixes 78f7186095db ("arm64: dts: rockchip: rename and sort the rk356x usb2 phy handles")Tested-by Diederik de Haas <[email protected]>Signed-off-by Jonas Karlman <[email protected]>
Thanks and best regards,
Michael
Quoted 53 lines
---
v2: Collect t-b tag--- arch/arm64/boot/dts/rockchip/rk356x-base.dtsi | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-)arch/arm64/boot/dts/rockchip/rk356x-base.dtsi +6
-6@@ -321,44 +321,44 @@ its: msi-controller@fd440000 { }; usb_host0_ehci: usb@fd800000 { - compatible = "generic-ehci"; + compatible = "rockchip,rk3568-ehci", "generic-ehci"; reg = <0x0 0xfd800000 0x0 0x40000>; interrupts = <GIC_SPI 130 IRQ_TYPE_LEVEL_HIGH>; clocks = <&cru HCLK_USB2HOST0>, <&cru HCLK_USB2HOST0_ARB>, - <&cru PCLK_USB>; + <&cru PCLK_USB>, <&usb2phy1>; phys = <&usb2phy1_otg>; phy-names = "usb"; status = "disabled"; }; usb_host0_ohci: usb@fd840000 { - compatible = "generic-ohci"; + compatible = "rockchip,rk3568-ohci", "generic-ohci"; reg = <0x0 0xfd840000 0x0 0x40000>; interrupts = <GIC_SPI 131 IRQ_TYPE_LEVEL_HIGH>; clocks = <&cru HCLK_USB2HOST0>, <&cru HCLK_USB2HOST0_ARB>, - <&cru PCLK_USB>; + <&cru PCLK_USB>, <&usb2phy1>; phys = <&usb2phy1_otg>; phy-names = "usb"; status = "disabled"; }; usb_host1_ehci: usb@fd880000 { - compatible = "generic-ehci"; + compatible = "rockchip,rk3568-ehci", "generic-ehci"; reg = <0x0 0xfd880000 0x0 0x40000>; interrupts = <GIC_SPI 133 IRQ_TYPE_LEVEL_HIGH>; clocks = <&cru HCLK_USB2HOST1>, <&cru HCLK_USB2HOST1_ARB>, - <&cru PCLK_USB>; + <&cru PCLK_USB>, <&usb2phy1>; phys = <&usb2phy1_host>; phy-names = "usb"; status = "disabled"; }; usb_host1_ohci: usb@fd8c0000 {
arch/arm64/boot/dts/rockchip/rk356x-base.dtsi its: msi-controller@fd440000 { · line 364
Quoted 10 lines
arch/arm64/boot/dts/rockchip/rk356x-base.dtsi +2
-2- compatible = "generic-ohci"; + compatible = "rockchip,rk3568-ohci", "generic-ohci"; reg = <0x0 0xfd8c0000 0x0 0x40000>; interrupts = <GIC_SPI 134 IRQ_TYPE_LEVEL_HIGH>; clocks = <&cru HCLK_USB2HOST1>, <&cru HCLK_USB2HOST1_ARB>, - <&cru PCLK_USB>; + <&cru PCLK_USB>, <&usb2phy1>; phys = <&usb2phy1_host>; phy-names = "usb"; status = "disabled";
_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip
Re: [PATCH v2 2/2] arm64: dts: rockchip: Fix devices suspend freeze on RK3568/RK3566
In reply to Jonas KarlmanTested on ODROID-M1 (RK3568) with 7.3-rc7: without this patch pm_test=devices hung while suspending the first OHCI controller 4 of 14 times, with it all 31 runs passed.
_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip
Heiko Stuebner
maintainerHi Jonas,
Am Mittwoch, 24. Juni 2026, 21:27:23 Mitteleuropäische Sommerzeit schrieb Jonas Karlman:
Quoted 16 lines
This series fixes a system freeze during suspend in
ohci_suspend()due toclk_usbphy1_480mnot being enabled when EHCI/OHCI registers are accessed on e.g. a Raxa ROCK 3C board.Following
pm_testmodes work on my ROCK 3C with the missing usbphy clk refs added:echo N >
/sys/module/printk/parameters/console_suspendecho devices >
/sys/power/pm_test
echo platform >/sys/power/pm_test
echo processors >/sys/power/pm_test
echo core >/sys/power/pm_testecho mem >
/sys/power/state
Quoted 7 lines
Changes in v2:
- Include rockchip,rk3588-ehci in the EHCI constraint
- Make clocks prop required for EHCI and OHCI
- Collect t-b tagJonas Karlman (2):
dt-bindings: usb: Add Rockchip RK3568 compatible for EHCI and OHCI
it seems I wasn't in Cc for the first patch, so can't know if and when that gets/got applied.
Also I think I remember Greg preferring not having to split apart patch series', so I think the binding patch might be better off standing alone.
Heiko