STM32MP157 USB OTG host mode with internal FS PHY issue

[hcodina]

Hi,

We use the USB OTG controller in host mode with its internal FS PHY.

Our DT node is the following:

&i2c4 {
    ...
    pmic: stpmic@33 {
        compatible = "st,stpmic1";
        ...
        regulators {
            compatible = "st,stpmic1-regulators";
            ...
            pwr_sw2-supply = <&vin>;
            ...
            vbus_sw: pwr_sw2 {
                regulator-name = "vbus_sw";
                regulator-active-discharge = <1>;
            };
        };
    };
};

&usbotg_hs {
    compatible = "st,stm32mp15-fsotg", "snps,dwc2";
    pinctrl-names = "default";
    pinctrl-0 = <&usbotg_hs_pins &usbotg_fs_dp_dm_pins>;
    dr_mode = "host";
    vbus-supply = <&vbus_sw>;
};

&pinctrl {
    usbotg_hs_pins: usbotg-hs-0 {
        pins {
            pinmux = <STM32_PINMUX('A', 10, ANALOG)>; /* OTG_ID */
        };
    };

    usbotg_fs_dp_dm_pins: usbotg-fs-dp-dm-0 {
        pins {
            pinmux = <STM32_PINMUX('A', 11, ANALOG)>, /* OTG_FS_DM */
                     <STM32_PINMUX('A', 12, ANALOG)>; /* OTG_FS_DP */
        };
    };
};

This configuration was perfecly working with ST linux v5.10-stm32mp-r2.1.

We bumped ST linux to v6.6-stm32mp-r1 and issues appeared with this controller.

The device enumeration failed when a device is plugged with the following kernel traces

[ 49.310271] usb 1-1: new low-speed USB device number 2 using dwc2
[ 49.530249] usb 1-1: device descriptor read/64, error -71
[ 49.870243] usb 1-1: device descriptor read/64, error -71
[ 50.210243] usb 1-1: new low-speed USB device number 3 using dwc2
[ 50.430255] usb 1-1: device descriptor read/64, error -71
[ 50.770226] usb 1-1: device descriptor read/64, error -71
[ 50.890323] usb usb1-port1: attempt power cycle
[ 51.460226] usb 1-1: new low-speed USB device number 4 using dwc2
[ 51.900229] usb 1-1: device not accepting address 4, error -71
[ 52.120291] usb 1-1: new low-speed USB device number 5 using dwc2
[ 52.560263] usb 1-1: device not accepting address 5, error -71
[ 52.564966] usb usb1-port1: unable to enumerate USB device

Also, the USB detection leading to the enumeration is done only on the first device USB connection. When the device is unplugged and re-plugged, no detection are performed by the kernel and so no enumeration started.

On suspend/resume (PM runtime) a WARN() is triggered on resume process:

# echo +5 > /sys/class/rtc/rtc0/wakealarm && echo mem > /sys/power/state
[ 63.470362] PM: suspend entry (deep)
[ 63.473574] Filesystems sync: 0.000 seconds
[ 63.478270] Freezing user space processes
[ 63.483305] Freezing user space processes completed (elapsed 0.001 seconds)
[ 63.490040] OOM killer disabled.
[ 63.493137] Freezing remaining freezable tasks
[ 63.498836] Freezing remaining freezable tasks completed (elapsed 0.001 seconds)
[ 63.506119] printk: Suspending console(s) (use no_console_suspend to debug)
NOTICE: CPU: STM32MP157CAC Rev.Z
NOTICE: Model: xxxxxxxx
NOTICE: BL2: v2.10-stm32mp1-r1.0(release):v2.10-stm32mp-r1-10-ga6bea6a20(a6bea6a2)
NOTICE: BL2: Built : 12:50:23, Apr 16 2025
NOTICE: get_v(b4) 1800
NOTICE: ldo3 bypass
NOTICE: pwr ddr done
NOTICE: BL2: Booting BL32
[ 63.515176] hy463x_ts 0-0038: Failed to enter suspend mode err:-6
[ 63.631633] ------------[ cut here ]------------
[ 63.631662] WARNING: CPU: 0 PID: 135 at drivers/usb/dwc2/core.c:523 dwc2_suspend+0x160/0x200
[ 63.631715] Modules linked in:
[ 63.631741] CPU: 0 PID: 135 Comm: sh Not tainted 6.6.48-gfa392e36fe98 #45
[ 63.631762] Hardware name: STM32 (Device Tree Support)
[ 63.631791] unwind_backtrace from show_stack+0x10/0x14
[ 63.631830] show_stack from dump_stack_lvl+0x40/0x4c
[ 63.631864] dump_stack_lvl from __warn+0xcc/0x15c
[ 63.631898] __warn from warn_slowpath_fmt+0x18c/0x1b0
[ 63.631925] warn_slowpath_fmt from dwc2_suspend+0x160/0x200
[ 63.631952] dwc2_suspend from dpm_run_callback+0x4c/0x178
[ 63.631979] dpm_run_callback from __device_suspend+0x11c/0x534
[ 63.632003] __device_suspend from dpm_suspend+0x114/0x324
[ 63.632029] dpm_suspend from dpm_suspend_start+0x64/0x6c
[ 63.632055] dpm_suspend_start from suspend_devices_and_enter+0xb4/0x834
[ 63.632092] suspend_devices_and_enter from pm_suspend+0x24c/0x2dc
[ 63.632127] pm_suspend from state_store+0x6c/0xc8
[ 63.632160] state_store from kernfs_fop_write_iter+0x118/0x1b4
[ 63.632199] kernfs_fop_write_iter from vfs_write+0x324/0x3f4
[ 63.632232] vfs_write from ksys_write+0xa0/0xd0
[ 63.632255] ksys_write from ret_fast_syscall+0x0/0x54
[ 63.632277] Exception stack(0xdeca5fa8 to 0xdeca5ff0)
[ 63.632296] 5fa0: 00000000 00000001 00000001 005a2260 00000004 b6f37070
[ 63.632314] 5fc0: 00000000 00000001 005a2260 00000004 0059f6bc b6f38cf0 beffb528 00000020
[ 63.632326] 5fe0: 00000000 beffb4f4 004ce500 b6e8db78
[ 63.632370] ---[ end trace 0000000000000000 ]---
[ 63.646249] Disabling non-boot CPUs ...
[ 63.647360] CPU1 killed.
[ 63.648102] Enabling non-boot CPUs ...
[ 63.649611] CPU1 is up
[ 63.913955] usb usb3: root hub lost power or was reset
[ 63.914063] usb usb2: root hub lost power or was reset
[ 64.282182] OOM killer enabled.
[ 64.285715] Restarting tasks ... done.
[ 64.292430] random: crng reseeded on system resumption
[ 64.298032] PM: suspend exit
# 

Those issues seems to be a kernel regression but I didn't succeed in identifying the root cause.

Did anyone encounter this issue? Any advice or clue will be appreciated.

Best regards,

Herve