[PATCH 1/1] clk: qcom: ipq8074: set BRANCH_HALT_VOTED flag for usb master clocks

[PATCH 1/1] clk: qcom: ipq8074: set BRANCH_HALT_VOTED flag for usb master clocks

[Chukun Pan]

Currently, on some boards with older versions of U-Boot, attempting
to disable usb master clocks results in a "stuck at on" warning:

[    3.005088] gcc_usb1_master_clk status stuck at 'on'
[    3.187119] gcc_usb0_master_clk status stuck at 'on'

The USB master clocks of IPQ5332 and IPQ5424 have the BRANCH_HALT_VOTED
flag set, so let's do the same to fix the issue.

Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>
---
These boards run a vendor-specific version of U-Boot, and upgrading to
a newer version could brick the device. Therefore, I think this should
be fixed within the Linux kernel.
---
 drivers/clk/qcom/gcc-ipq8074.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/clk/qcom/gcc-ipq8074.c b/drivers/clk/qcom/gcc-ipq8074.c
index 1329ea28d703..10bd2f80010c 100644
--- a/drivers/clk/qcom/gcc-ipq8074.c
+++ b/drivers/clk/qcom/gcc-ipq8074.c
@@ -2612,6 +2612,7 @@ static struct clk_branch gcc_sys_noc_usb0_axi_clk = {
 
 static struct clk_branch gcc_usb0_master_clk = {
 	.halt_reg = 0x3e000,
+	.halt_check = BRANCH_HALT_VOTED,
 	.clkr = {
 		.enable_reg = 0x3e000,
 		.enable_mask = BIT(0),
@@ -2725,6 +2726,7 @@ static struct clk_branch gcc_sys_noc_usb1_axi_clk = {
 
 static struct clk_branch gcc_usb1_master_clk = {
 	.halt_reg = 0x3f000,
+	.halt_check = BRANCH_HALT_VOTED,
 	.clkr = {
 		.enable_reg = 0x3f000,
 		.enable_mask = BIT(0),
-- 
2.34.1

Re: [PATCH 1/1] clk: qcom: ipq8074: set BRANCH_HALT_VOTED flag for usb master clocks

[Taniya Das]

On 9/10/2026 3:30 PM, Chukun Pan wrote:
> Currently, on some boards with older versions of U-Boot, attempting
> to disable usb master clocks results in a "stuck at on" warning:
> 
> [    3.005088] gcc_usb1_master_clk status stuck at 'on'
> [    3.187119] gcc_usb0_master_clk status stuck at 'on'
> 
> The USB master clocks of IPQ5332 and IPQ5424 have the BRANCH_HALT_VOTED
> flag set, so let's do the same to fix the issue.
> 

Fixes tag?

> Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>
> ---
> These boards run a vendor-specific version of U-Boot, and upgrading to
> a newer version could brick the device. Therefore, I think this should
> be fixed within the Linux kernel.
> ---
>  drivers/clk/qcom/gcc-ipq8074.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/clk/qcom/gcc-ipq8074.c b/drivers/clk/qcom/gcc-ipq8074.c
> index 1329ea28d703..10bd2f80010c 100644
> --- a/drivers/clk/qcom/gcc-ipq8074.c
> +++ b/drivers/clk/qcom/gcc-ipq8074.c
> @@ -2612,6 +2612,7 @@ static struct clk_branch gcc_sys_noc_usb0_axi_clk = {
>  
>  static struct clk_branch gcc_usb0_master_clk = {
>  	.halt_reg = 0x3e000,
> +	.halt_check = BRANCH_HALT_VOTED,

BRANCH_HALT didn't work?

>  	.clkr = {
>  		.enable_reg = 0x3e000,
>  		.enable_mask = BIT(0),
> @@ -2725,6 +2726,7 @@ static struct clk_branch gcc_sys_noc_usb1_axi_clk = {
>  
>  static struct clk_branch gcc_usb1_master_clk = {
>  	.halt_reg = 0x3f000,
> +	.halt_check = BRANCH_HALT_VOTED,

BRANCH_HALT didn't work?

>  	.clkr = {
>  		.enable_reg = 0x3f000,
>  		.enable_mask = BIT(0),

-- 
Thanks,
Taniya Das

Re: [PATCH 1/1] clk: qcom: ipq8074: set BRANCH_HALT_VOTED flag for usb master clocks

[Chukun Pan]

Hi,

> Fixes tag?

This only occurs on a few devices, so I didn't add the Fixes tag.
Should I send a v2 patch with the Fixes tag?

> > +	.halt_check = BRANCH_HALT_VOTED,
>
> BRANCH_HALT didn't work?

I have tried these:

BRANCH_HALT - still has a status stuck at 'on' warning.
BRANCH_HALT_DELAY - No stuck warning, but EMI warning
may occur once: disabled by hub (EMI?), re-enabling...
BRANCH_HALT_VOTED - No warnings, USB working.

Thanks,
Chukun