If the dwc2 driver fails to probe after having enabled the regulators,
it ends up being unregistered with regulators enabled, something the
core regulator code is legitimately upset about:
dwc2
ff400000.usb: supply vusb_d not found, using dummy regulator
dwc2
ff400000.usb: supply vusb_a not found, using dummy regulator
dwc2
ff400000.usb: dwc2_core_reset: HANG! AHB Idle timeout GRSTCTL GRSTCTL_AHBIDLE
WARNING: CPU: 2 PID: 112 at drivers/regulator/core.c:2074 _regulator_put.part.0+0x16c/0x174
Modules linked in: dwc2(E+) dwc3(E) udc_core(E) rtc_hym8563(E) dwmac_generic(E) ulpi(E) usbcore(E) dwc3_meson_g12a(E) roles(E) meson_gx_mmc(E+) i2c_meson(E) mdio_mux_meson_g12a(E) mdio_mux(E) dwmac_meson8b(E) stmmac_platform(E) stmmac(E) mdio_xpcs(E) phylink(E) of_mdio(E) fixed_phy(E) libphy(E) pwm_regulator(E) fixed(E)
CPU: 2 PID: 112 Comm: systemd-udevd Tainted: G E 5.9.0-rc4-00102-g423583bc8cf9 #1840
Hardware name: amlogic w400/w400, BIOS 2020.04 05/22/2020
pstate:
80400009 (Nzcv daif +PAN -UAO BTYPE=--)
pc : _regulator_put.part.0+0x16c/0x174
lr : regulator_bulk_free+0x6c/0x9c
sp :
ffffffc012353820
x29:
ffffffc012353820 x28:
ffffff805a4b7000
x27:
ffffff8059c2eac0 x26:
ffffff8059c2e810
x25:
ffffff805a4b7d00 x24:
ffffffc008cf3028
x23:
ffffffc011729ef8 x22:
ffffff807e2761d8
x21:
ffffffc01171df78 x20:
ffffff805a4b7700
x19:
ffffff805a4b7700 x18:
0000000000000030
x17:
0000000000000000 x16:
0000000000000000
x15:
ffffff807ea8d178 x14:
3935312820435455
x13:
2038323a36313a37 x12:
ffffffffffffffff
x11:
0000000000000040 x10:
0000000000000007
x9 :
ffffffc0106f77d0 x8 :
ffffffffffffffe0
x7 :
ffffffffffffffff x6 :
0000000000017702
x5 :
ffffff805a4b7400 x4 :
0000000000000000
x3 :
ffffffc01171df78 x2 :
ffffff807ea8cc40
x1 :
0000000000000000 x0 :
0000000000000001
Call trace:
_regulator_put.part.0+0x16c/0x174
regulator_bulk_free+0x6c/0x9c
devm_regulator_bulk_release+0x28/0x3c
release_nodes+0x1c8/0x2c0
devres_release_all+0x44/0x6c
really_probe+0x1ec/0x504
driver_probe_device+0x100/0x170
device_driver_attach+0xcc/0xd4
__driver_attach+0xb0/0x17c
bus_for_each_dev+0x7c/0xd4
driver_attach+0x30/0x3c
bus_add_driver+0x154/0x250
driver_register+0x84/0x140
__platform_driver_register+0x54/0x60
dwc2_platform_driver_init+0x2c/0x1000 [dwc2]
do_one_initcall+0x54/0x2d0
do_init_module+0x68/0x29c
In order to fix this, tie the regulator disabling to the teardown
process by registering a devm action callback. This makes sure that
the regulators are disabled at the right time (just before they are
released).
Cc: Minas Harutyunyan <hminas@synopsys.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Felipe Balbi <balbi@kernel.org>
return 0;
}
+static void __dwc2_disable_regulators(void *data)
+{
+ struct dwc2_hsotg *hsotg = data;
+
+ regulator_bulk_disable(ARRAY_SIZE(hsotg->supplies), hsotg->supplies);
+}
+
static int __dwc2_lowlevel_hw_enable(struct dwc2_hsotg *hsotg)
{
struct platform_device *pdev = to_platform_device(hsotg->dev);
if (ret)
return ret;
+ ret = devm_add_action_or_reset(&pdev->dev,
+ __dwc2_disable_regulators, hsotg);
+ if (ret)
+ return ret;
+
if (hsotg->clk) {
ret = clk_prepare_enable(hsotg->clk);
if (ret)
if (hsotg->clk)
clk_disable_unprepare(hsotg->clk);
- ret = regulator_bulk_disable(ARRAY_SIZE(hsotg->supplies),
- hsotg->supplies);
-
- return ret;
+ return 0;
}
/**